An Evaluation of the Role of Microbiological Criteria for Foods and Food Ingredients (1985)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

9 Application of Microbiological Criteria to Foods and Food Ingredlients INTRODUCTION In preceding chapters, conditions necessary for establishing meaningful microbiological criteria were presented. In this chapter recommendations are given regarding the need or lack thereof for microbiological criteria for each of 22 food products or groups of products. The subcommittee elected not to give specific recommendations relative to microbiological limits but chose instead to emphasize that any criteria that are developed should be realistic and should be based on relevant background infor- mation. Although the organization of the individual sections of this chapter may vary for each of the foods or groups of foods, the subcommittee has attempted to address the following basic issues in each section: (1) the sensitivity of the food products relative to safety and quality, (2) the needs for a microbiological standards and/or guidelinets), (3) assessment of information necessary for establishment of a criterion if one seems to be indicated, and (4) where the criterion should be applied. The following foods and food groups are included in this chapter in the order in which they are listed below: A. Dairy Products B. Raw Meats C. Processed Meats D. Raw (Eviscerated, Ready To-Cook) Poultry E. Processed Poultry Products 184 F. Eggs and Egg Products G. Fish, Molluscs, and Crustaceans H. Fruits and Vegetables I. Fruit Beverages J. Low-Acid Canned Foods

APPLICATION TO FOODS AND FOOD INGREDIENTS K. Acid Canned Foods L. Water Activity-Controlled Canned Foods M. Cereals and Cereal Products N. Fats and Oils O. Sugar, Cocoa, Chocolate, and Confectioneries P. Spices A. DAIRY PRODUCTS Introduction 185 Q. Yeasts R. Formulated Foods S. Nuts T. Miscellaneous Additives U. Bottled Water, Processing Water, and Ice V. Pet Foods Microbial growth in the more perishable dairy products, i.e., pasteurized milks, condensed milks, ice cream mixes, creams, cottage cheese, and fermented milks, often results in development of objectionable flavors and textural changes. Even under conditions of good production, pro- cessing, distribution, and storage (including care in the home) such changes are inevitable and may be expected to occur within two to three weeks or less. However, the high acidity of cottage cheese and fermented milks and the high heat treatment given to ultrapasteurized milk permits some- what longer shelf-life. Recognition of the perishability of these products has led to the common practice of "sell by date" labelling as a means of alerting distributors and consumers to the products' limited shelf-life. On the other hand, the relatively stable dairy products, i.e., dried milks, evaporated milk, sterilized milk, ice cream, ripened cheese, butter, and sweetened condensed milk, may remain free of microbiologically induced deterioration for several months or years. In the early part of this century, health of dairy animals and production, processing, and distribution practices were often poor. At that time, un- pasteurized milk was a major vehicle for transmission to humans of dis- eases such as typhoid, diphtheria, septic sore throat, tuberculosis, and brucellosis (Bryan, 19831. Recognition of these problems by government and industry led to a series of recommendations embodied in the Milk Ordinance of 1924 and an interpretion of these recommendations in the Code in 1927. This model milk ordinance, now titled the "Grade A Milk Ordinance" (see below), is an example of the application of the HACCP  system to a major food industry. Maintenance of the quality and safety of dairy products, which includes optimum shelf-life, is now a well-accepted industry responsibility and is a necessity for economic survival in this highly competitive industry. Furthermore, it has become traditional for the public to expect, if not

186 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRlTERlA demand, high-quality products that are safe and esthetically acceptable. Therein lies the basis for current safety and quality assurance programs of regulatory agencies and of industry. As a component of such programs, microbiological criteria play an important role. Sensitivity of Products Relative to Quality Currently, most state and local regulatory agencies utilize almost ex- clusively the Grade A Pasteurized Milk Ordinance (USPHS/FDA, 1978) and USDA Standards for Grades of Dairy Products (USDA, 1975) as the bases for their regulatory programs for dairy products. As an integral part of these two documents, microbiological criteria are specified for most products (see Chapter 8, Table 8-41. Furthermore, the testing of dried milk products for Salmonella is provided for in accordance with a Mem- orandum of Understanding (FDA, 1975) (see Chapter 81. There can be little doubt that application of microbiological criteria has contributed significantly to the provision of high-quality, safe dairy prod- ucts. With the exception noted below and as needs are uncovered by future investigations and research, there appears to be no basis for imposing more severe standards or additional criteria. Industry imposes on itself criteria far more stringent than those that must be met to avoid the like- lihood of noncompliance. This has the salutary effect of providing rea- sonable assurance that products are esthetically acceptable and that aerobic plate count levels are maintained well below those likely to cause dete- riorative changes within a reasonable shelf-life period. One of the exceptions referred to above is the bacterial count limit for Grade 2 raw milk for manufacturing purposes as specified in the USDA "Standards for Grades" (USDA, 19751. Recent research has revealed the potential for heat-resistant enzymes of microbial origin to be involved in the deterioration of processed dairy products held for prolonged storage periods. Furthermore, these enzymes have been implicated in lowered cheese yields. Psychrotrophs are among the principal organisms that pro- duce these enzymes and because of modern milk-handling practices, they can comprise a large proportion of the microflora of raw milk. Thus, bacterial count levels as high as the 3 million per ml permitted in Grade 2 milk would appear to be excessive. Consideration might well be given to modifying this standard. Certainly, lower count levels are easily attained through application of modern milk-handling practices. Sensitivity of Products Relative to Safety Currently the microbiological safety of dairy products can be assured only through application of three preventive measures. These are:

APPLICATION TO FOODS AND FOOD INGREDIENTS 187 (1) pasteurization or more severe heat treatments; (2) prevention of post- heat treatment contamination; and, (3) for certain products, end-product testing for microorganisms and toxins for certain products. Microbiol- ogical criteria are useful in the application of the preventive measures listed above. Current standards for coliforms as specified in the Grade A Pasteurized Milk Ordinance and the USDA Standards for Grades are useful in detecting post-heat treatment contamination. However, failure to find these organ- isms in finished products or at critical control points does not necessarily indicate the absence of post-heat treatment contaminants. Dried Milk There is ample justification for continued finished-product testing and surveillance of dried milk products for the presence of Salmonella. These products are susceptible to Salmonella contamination and are often used without further heat treatment for fluid consumption as recombined milk or as ingredients in formulated foods. Furthermore, these recombined or formulated products are often consumed by high-risk populations. Monthly reports of the USDA's Salmonella surveillance program (USDA, 1980) administered in accordance with the USDA/FDA Memorandum of Un- derstanding (FDA, 1975) reveal a continuing low level of Salmonella- positive environmental samples and finished products from dried milk plants. Concurrent with the USDA/FDA programs, industry conducts ex- tensive testing. Subsequent routine follow-up procedures undoubtedly have prevented contaminated product from reaching the market. The above-mentioned program as well as FDA surveillance of products offered for import should be continued, strengthened when indicated, and reviewed periodically to ascertain that sampling plans, including methods used, are consistent with the hazards presented and in accord with current developments in methodology and with appropriate statistical concepts. Reference is here made to the USDA document entitled "Salmonella Surveillance Program" (USDA, 19801. This document refers to the Na- tional Academy of Sciences publication An Evaluation of the Salmonella Problem (NRC, 1969) and states that the report classifies dried milk and dried milk product in Food Category II and proposes acceptance of a lot on the basis of all negative results on twenty-nine 25-g samples (n = 291. The USDA document states "Instead of analyzing 25-g samples this In- struction provides a procedure whereby each test shall comprise a com- posite of four 100-g samples." It further states that, "based on low incidence of contamination, this procedure provides comparable sensitivity and permits greater coverage at a reduced cost." In this case n = 16. The sampling plan applies to finished products analyzed quarterly in ac

188 EVALUATION OF TlIE ROLE OF MICROBIOLOGICAL CRITERIA cordance with the surveillance program. This subcommittee does not agree that a sampling plan of n = 16 provides "comparable sensitivity" to the recommended plan where n = 29, though obviously it must agree that the cost is reduced. In essence the USDA Quarterly Surveillance Program should require the analysis of Weather than 1 composite sample. In this case the sampling plan would be n = 32, which would be only slightly more stringent than the plan recommended by the Salmonella committee. In practice, on a quarterly basis, the USDA collects samples from 3 days' production (preferably consecutive days). Four samples of product from each day are drawn, for a total of 12 samples. The four samples for each day are composited to yield a 400-g analytical unit that is analyzed for Salmonella. Thus, the quarterly finished-product surveillance consists of the analysis of three 400-g samples for a given plant. If the plant produces products other than a nonfat dried milk, i.e., buttermilk and whey, a separate set of 12 samples is taken from the dryerts) that is (are) used for each product. If the plant has several dry-milk dryers, each with its own bagging head, the product from only one dryer is sampled. The "Instruc- tion" directs that alternately a different dryer be sampled on each suc- cessive quarterly survey. This subcommittee believes that samples should be drawn from each dryer since each is an integral unit of production equipment. At the time of the quarterly survey three environmental samples are also collected, these being waste material from the vacuum cleaner, air filters, and tailings. When the surveillance on a plant's production shows a positive test, a letter is sent to the plant manager informing him of the single positive test and the three available options by which the positive product can be handled. These options appear to be adequate to provide reasonable assurance that contaminated product will not reach the market. They provide that the day's production having the positive Sal- monella test shall be either: 1. segregated, reprocessed and the reprocessed product tested for Sal- monella; 2. segregated and disposed of in a manner that poses no health problem to humans or animals, e.g., USDA certification that the product was discarded in a sanitary landfill; or 3. retested (verification test) at the rate of twelve 100-g samples. For test purposes the laboratory will composite 4 samples for a total of 3 tests in = 481. If none of the composites shows a positive test, the results are interpreted as meaning that the incidence of Salmonella is insignificant, and the day's production may be used for human purposes. If one or more of the composites shows a positive test, the day's production represented by the test is then handled as in

APPLlCATlON TO FOODS ID FOOD INGREDIENTS 189 (1) or (2) above, i.e., segregated, reprocessed, and retested or dis- posed of. In addition, eight 100-g samples are collected from product manufactured on each of two days immediately preceding the day(s) in which the positive product was noted and this rate of sampling and testing is also performed on each day's production made sub- sequent to the day(s) having a positive product test until the plant effects a complete cleanup of its drying facilities. This procedure would represent a sampling plan of n = 32, a plan comparable in stringency to that recommended by the Salmonella committee. If, in connection with a quarterly survey, more than one of the three finished product samples are positive, then verification testing (as outlined above) is not permitted. Positive lots are reconstituted and repasteurized or disposed of in a manner that poses no threat to human or animal health. The plant ~nanager is requested to furnish a list of production back to at least two days prior to the positive lot and up to the time of special cleanup. The list should show the date of manufacture, lot number, number of containers in each lot, and the present location of the product. "Because of possible serious contamination, the product should be recalled from distribution channels and held for sampling and testing for Salmonella." In this connection eight samples are drawn from each day's production held from distribution. If all eight test results on a day's production are negative, that product may be released for use or distribution. If a positive test is obtained on product for any day, all the product for that day shall be disposed of in such a manner so as to pose no health problem to humans or animals. This subcommittee believes that the actions taken subsequent to the detection of positive lots at time of quarterly surveillance sampling are adequate for the purposes intended. The effectiveness of cleanup (required as part of one of the three options indicated above) is determined on the basis of tests made on product manufactured subsequent to cleanup. Sampling involves the collection of eight 100-g samples on each of three production days immediately fol- lowing the cleanup (n = 321. In this instance, if a lot is declared positive, there is no option for verification testing; management must dispose of the product "in such a manner as to pose no health problem to humans or animals" or reprocess the lot and again test it for Salmonella. Fur- thermore, three environmental samples are taken presumably on each post- cleanup production day. The environmental sampling program is an integral part of the USDA Salmonella Surveillance Program. Concomitant collection of environ- mental samples at the time quarterly finished-product samples are taken materially strengthens the program. Experience has taught that if a Sal

190 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA monella problem exists within a milk-dr~ving plant, one is far more apt to detect this problem through the analysis of environmental samples than through finished product analysis. Thus, given an adequate finished prod- uct sampling plan (n = 29-30) combined with the testing of environmental samples, there is reasonable assurance that if all samples are negative, a serious Salmonella problem did not exist at the time of the quarterly inspection. Furthermore, the sampling programs followed in verification tests have a stringency comparable to that recommended by the Salmonella committee (NRC, 19691. The overall weakness of the program lies in the fact that (1) samples are collected only on a quarterly basis; (2) sampling plans used at time of quarterly sampling are not sufficiently stringent, i.e., consistent with recommendations of the NAS/NRC Salmonella report (NRC, 19691; and (3) in plants having more than one dryer for a given product, the product from only one of the dryers is sampled. Thus, the USDA surveillance program cannot substitute for in-house surveillance by the processor. As with the USDA program on eggs and egg products, the control of the Salmonella hazard in dry milk requires continuous testing of finished product and environmental samples by the processor. The present USDA program is to be commended and should be strengthened as indicated above and continued, with the realization that it is not a substitute for microbiological control by the manufacturer. Accordingly, the dry milk industry should be encouraged to test finished products reg- ularly and in accordance with sampling plans recommended in the NAS/ NRC Salmonella report. Cheese Cheese is the second dairy product for which finished product testing for presence of a pathogen or its toxins may be indicated. The organism of primary concern is Staphylococcus aureus; although pathogenic Esch- erichia cold have caused some concern. Also, the recent series of outbreaks of brucellosis due to unripened raw goat milk queso blanco cheese sold primarily from roadside vendors in the Houston, Texas area has empha- sized the hazard of cheese made from unpasteurized milk (Perkins et al., 19834. As in the case of the hazard of raw milk consumption, pasteurization of milk used in the manufacture of this cheese is the only rational means of control. Routine microbiological testing would not serve as an effective control measure. Certain cheese varieties have served as the vehicle involved in outbreaks of staphylococcal food poisoning. Others, although not involved in out- breaks, have been shown to permit the buildup of potentially hazardous levels of S. aureus under certain conditions of manufacture. Although few

APPLICATION TO FOODS AND FOOD INGREDIENTS 191 outbreaks have been reported in the United States in recent years, the problem does persist. The nature and control of the problem, including application of microbiological criteria, has been reviewed recently (ICMSF, 1980, 19851. The following is a brief summary of these considerations. The hazard of staphylococcal food poisoning presented by cheese is limited largely to hard varieties, i.e., Cheddar and similar types and Swiss or Emmenthaler. Although S. aureus has been demonstrated to grow in certain other varieties, i.e., Gouda, Brick, Roquefort, Blue, and Moz- zarella, their involvement in outbreaks has been rare or unreported (ICMSF, 19801. At some point during manufacture and subsequent handling, the first group of cheeses undergoes a sufficiently long period at moderate temperatures, which permits growth of lactic starter cultures. Impairment of such growth during these periods permits relatively unrestricted growth of various other organisms that may be present, including S. aureus. If the initial population and period of time of favorable growth conditions are sufficient, the number of S. aureus may reach several million per gram of product, at which point hazardous levels of enterotoxin may be present. Adequate heat treatment of cheese milk, good sanitary practices to avoid post-heat treatment contamination, and unimpaired starter culture activity are essential elements of good manufacturing practices for control of S. aureus in cheese. Appropriate analytical methods are available for the testing of cheese and the monitoring of critical control points in cheese manufacture. Ad- equate and relatively simple methods for detection and enumeration of S. aureus in cheese are available. S. aureus enterotoxins may be detected with specificity, although the procedures are somewhat complicated. Fur- thermore, a rapid test for staphylococcal thermonuclease is sufficiently reliable for determining whether a particular lot of cheese may contain enterotoxin or whether it may be safely released for distribution (see Chapter 15 of ICMSF, 19851. At present the incidence of staphylococcal food poisoning by domest- ically produced cheese does not justify routine testing by regulatory agen- cies. However, industry should be encouraged to (l) routinely monitor critical control points for presence or indication of staphylo.~occal growth and to (2) test all cheese for S. aureus and/or thermonuclease if abnormal lactic culture activity occurred during manufacture or if other conditions that might lead to extensive staphylococcal growth were encountered. FDA should routinely test all susceptible cheese varieties offered for import for presence of thermonuclease. This routine testing is advisable as regulatory agencies generally do not have knowledge of the cheese production con- ditions. Lots positive for thermonuclease should be further tested for presence of enterotoxins.

192 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Outbreaks of foodborne illness due to certain pathogenic strains of E. cold in imported Camembert cheese occurred in the United States in 1971. These were the first documented foodborne outbreaks due to E. cold to be reported in the United States. Nevertheless, they caused con- siderable concern about cheese as a vehicle for transmission of pathogenic E. cold to humans. No further outbreaks of cheeseborne illness due to E. cold were reported until 1983. In 1983 several outbreaks occurred again, resulting from consuming imported soft cheese (Brie and Camembert) of French origin. Investigations were not completed at time of this writing, but they seem to indicate that a certain strain of E. cold (027: H20) pro- ducing a heat-stable toxin was the causative organism (Francis and Davis, 19841. . Following the first episode in 1971 certain control measures were in- troduced by the French government and the industry. Apparently these measures were effective in view of the 12-year interval between the two series of outbreaks. However, now it would appear prudent for FDA to initiate appropriate research relative to the E. cold problem in soft cheese as well as the routine testing of soft cheese offered at import, i.e., Ca- membert, Brie, and similar varieties, for the presence and quantitative level of E. colt. Such studies would serve to further delineate the problem and assist in development of an appropriate control program. Fluidt Milk . Milkborne disease outbreaks caused by consuming legally purchased contaminated raw milk as well as raw milk cheese (see preceding dis- cussion of queso blanco cheese) continue with regularity. Even certified raw milk, which is produced under the most exacting sanitary conditions, continues to cause outbreaks (Werner et al., 19841. Recently, Bryan (1983) reviewed the epidemiology of milkborne dis- eases and concluded, as many others have done previously j that pasteur- ization is an essential process in providing milk that is free of disease- producing microorganisms. Application of microbiological criteria, al- though useful, cannot assure that contaminated raw milk will be detected. The problems of preventing the sale and consumption of raw milk have been emphasized in a recent editorial by Chin (1982) (see Appendix F). The sale of raw milk is still legal in some 20 states in spite of the fact that infectious disease professionals consider the scientific case against raw milk to be irrefutable. Nevertheless, the legal aspects involved in preventing the sale of raw milk are complex and provide impediments toward that end. Unfortunately, the small segment of the dairy industry that engages in public sale of raw milk for fluid consumption, deliberately or through ignorance of the consequences, continues to make available to

APPLICATION TO FOODS AND FOOD INGREDIENTS 193 consumers a product that sometimes is hazardous. The subcommittee agrees with Dr. Chin's conclusion that, "It is the responsibility of all health professionals to see that the public and the policymakers are ade- quately informed about the scientific findings so that public policy on raw milk may be compatible with scientific knowledge and protective of the public's health." References Bryan, F. L. 1983 Epidemiology of milk-borne diseases. J. Food Prot. 46:637-649. Chin, J. 1982 Raw milk: A continuing vehicle for the transmission of infectious disease agents in the United States. J. Infect. Dis. 46: 440-441. FDA (Food and Drug Administration) 1975 Memorandum of Understanding USDA/FDA on Salmonella Inspection of Dry Milk Plants. No. FDA 225-75-4002. Washington, D.C.: U.S. Department of Agriculture. Francis, B. J., and J. P. Davis 1984 Update: gastrointestinal illness associated with imported semi-soft cheese. Morb. Mort. Weekly Rpt. 33: 16, 22. ICMSF (International Commission on Microbiological Specifications for Foods) 1980 Milk and milk products. Pp. 470-520 in Microbial Ecology of Foods. Vol. 2. Food Commodities. New York: Academic Press. 1985 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. 2nd Ed. In preparation. NRC (National Research Council) 1969 An Evaluation of the Salmonella Problem. Committee on Salmonella. Washington, D.C.: National Academy of Sciences. Perkins, P., A. Rogers, M. Key, V. Pappas, R. Wende, J. Epstein, M. Thapar, F. Jensen, T. L. Gustafson, and E. Young 1983 Brucellosis Texas. Morb. Mort. Weekly Rpt. 32:548-553. USDA (U.S. Department of Agriculture) 1975 General specifications for approved dairy plants and standards for grades of dairy products. Federal Register 40(198):47910-47940. 1980 Salmonella Surveillance Program. DA Instruction No. 918-72. Washington, D.C.: U.S. Department of Agriculture. USPHS/FDA (U.S. Public Health Service/Food and Drug Administration) 1978 Grade A Pasteurized Milk Ordinance. 1978 Recommendations. PHS/FDA Publ. No. 229. Washington, D.C.: U.S. Government Printing Office. Werner, S. B., F. R. Morrison, G. L. Humphrey, R. A. Murray, and J. Chin 1984 Salmonella dublin and raw milk consumption California. Morb. Mort. Weekly Rpt. 33: 196-198. B. RAW MEATS  Sensitivity of Products Relative to Safety and Quality The microbiological condition of retail cuts of red meat (beef, pork, and lamb) is the result of a series of conditions and events including:

194 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA 1. the health and condition of the live animal; 2. slaughtering-dressing practices; 3. conditions of chilling of the carcass such as rate of cooling, tem- perature, and humidity; 4. sanitary conditions and practices during fabrication of a carcass into primal, subprimal, and retail cuts; 5. packaging conditions such as air versus vacuum-packaging; 6. conditions of distribution and storage (time-temperature profiles); 7. handling of cuts in food service establishments and in the home (proper refrigerated storage, adequate heat treatment, avoiding cross-con- tamination). Following is a brief summary of these conditions and events as they relate to shelf-life and wholesomeness of meat and the potential need for microbiological criteria. For more detailed information, the reader is re- ferred to the following reports (APHA, 1984; Ayres, 1955, 1960; ICMSF, 1980; Ingram and Roberts, 1976; and Roberts, 19741. Conditions prior to slaughter can have an impact on the microbiological condition of meat. Muscle tissue from carcasses of animals that have undergone prolonged muscular activity or long-term stress (lack of feed, temperature changes) before slaughter is often dark, firm, and dry (DFD meat), contains little or no glucose and has a higher pH ~-6.0) than that of unstressed animals (approximately 5.51. Under aerobic storage con- ditions, normal meat spoils when glucose is exhausted and amino acids are attacked. In DFD meat, however, amino acids are attacked without delay. The high pH of vacuum-packaged DFD meat allows the devel- opment of Serratia liquefaciens and Alteromonas putrefaciens, which pro- duce off-odors. For these reasons DFD meat spoils more rapidly than normal meat (Gill and Newton, 19811. Stress also may increase the prev- alence of Salmonella in pigs as they are transported from production units to slaughtering facilities (Ingram, 1972; Williams and Newell, 19701. Microorganisms associated with the live animal are located primarily on the surface of the animal (hide, hair, hooves) and in the gastrointestinal tract. The number of microorganisms in the muscle tissue (intrinsic bac- teria) of healthy animals is small (Gill, 1979~. Carcasses of normal, healthy animals appear to have considerable residual ability to maintain tissue sterility. It is often reported that muscle tissue from stressed animals is more likely to contain "intrinsic" bacteria. It is possible that certain forms of stress depress the immune defense mechanisms and therefore allow the survival of bacteria that otherwise would have been destroyed. Sources of microbial contamination of a carcass include: the anima (surface and gastro-intestinal tract), workers, clothing of workers, utensils

APPLICATION TO FOODS AND FOOD INGREDIENTS 195 equipment, air, and water. Hence, the level of microbial contamination of a carcass at this stage depends upon the degree of sanitation practiced during the slaughter-dressing procedures. Aerobic plate counts (APCs) of freshly dressed beef, pork, and lamb carcasses in the United States usually range from 102-104 per cm2, most of which are not psychrotrophic. Mi- crobial types reflect the various sources of contamination and include Micrococcus, Staphylococcus, Bacillus, Enterobacteriaceae, Pseudo- monas, yeasts and molds, and others. Because of location and handling practices certain areas of a carcass are more likely to be contaminated or to remain contaminated than are others. For these reasons, microorganisms are not uniformly distributed over the surface. Under conditions of rapid chilling, low storage temperature (-1 to + 1°C/ 30.2 to 33.8°F), proper air movement, and relative low humidity, the number of microorganisms on a carcass may increase little or in some cases actually decrease during the first 24 to 48 hours. During further refrigerated storage, psychrotrophic bacteria will become a more promi- nent part of the microbial flora. Fabrication of carcasses into primal, subprimal, and retail cuts involves extensive handling of the meat, markedly increases the surface area, and increases the aw as newly cut surfaces are made. Because these operations are carried out at refrigeration temperature, contamination at this stage is likely to include a high proportion of psychrotrophic bacteria. In recent years, approximately 60-70% of the beef leaving meat-packing plants in the United States is vacuum-packaged. Storage of cuts in vacuum packages results in a predominance of lactic acid bacteria, whereas gram- negative, aerobic, psychrotrophic rods such as Pseudomonas spp. pre- dominate on aerobically stored cuts. The latter are more active spoilage producers than the lactic acid bacteria. Thus, vacuum-packaged cuts with 106-107 bacteria per cm2 may be organoleptically acceptable whereas com- parable cuts stored in air-permeable films with similar numbers of gram- negative, aerobic psychrotrophic rods may exhibit off-odors. In vacuum- packaged red meat, high levels of lactic acid bacteria (107-108 per cm2) frequently contribute to odors characterized as sour, acid, buttermilk-like, cheesy and less frequently "sulphur-like" or "H2S-like" (Henna et al., 1983~. When vacuum-packaged cuts are fabricated into steaks or chops and then repackaged in oxygen-permeable films, gram-negative aerobic rods will soon become predominant on the cuts in the display cases. Conditions of refrigerated storage in the distribution chain (plant, whole- sale, retail, food service establishment, home) can have a profound effect on the microbial condition and therefore shelf-life of the product. Cuts with initial low numbers of psychrotrophic bacteria held at proper refrig- eration temperature (-1 to + 1°C/30.2 to 33.8°F) will have a greater shelf

196 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA life than will meats produced under less sanitary conditions. Even meats produced under good sanitary conditions will have reduced shelf-life when stored at marginal refrigeration temperatures. Improper cooking and handling of raw meats in homes and food service establishments is one of the main reasons for foodborne illness caused by consumption of "cooked" meats. It is not yet commercially feasible to produce red meat free from pathogenic bacteria. Small numbers of a variety of pathogenic organisms such as Salmonella, Clostridium perfringens, Yersinia enterocolitica, Campylobacter jejuni, and Staphylococcus aureus can be present on raw red meats. (See Chapter 4 for more detailed in- formation on these bacteria.) In pork Trichinella spiralis can be present. Adequate heat treatment of the meat will destroy most of these pathogens. Spores of C. perfringens, however, may survive in cooked meats. If such meats are held at temperatures between 20 to 50°C/68 to 122°F, growth of C. perfringens may occur, during either serving operations or subse- quent storage, thus creating a health hazard. To produce red meats with optimum shelf-life and safety requires the monitoring of a series of critical control points pertinent to each of the conditions or events listed above. Most of these critical control points can be monitored by careful inspection of the animals, evaluating the proce- dures and practices used, and by checking temperatures during chilling operations. Microbiological guidelines can be applied to monitor some of the critical control points. For example, several relatively simple procedures (APHA, 1984) are available to check the sanitary condition of equipment and utensils. They are useful to evaluate the effectiveness of cleaning and sanitizing procedures. Microbiological count data on freshly dressed car- casses need to be interpreted cautiously (Ingram and Roberts, 1976) be- cause (1) microorganisms are not distributed evenly over the carcass, (2) there is considerable variation in contamination between carcasses, (3) there are differences in contamination due to processing on different days in the same plant and (4) there are differences between plants. In addition, the tests that can be applied to examine carcasses under com- mercial conditions recover only a fraction (often a variable fraction) of the organisms present, usually from a rather small surface area of the total carcass. Nevertheless, APCs on freshly dressed carcasses may provide data that indicate a trend in the handling of animals during the slaughter- dressing operations. Several other methods have been tried to monitor microbial activity in red meats such as pH, dye reduction, titratable acidity, volatile nitrogen compounds, volatile acids, extract-release-volume, and others (see Chapter 51. None of these has proved to be a sensitive and reproducible indicator of incipient spoilage applicable to various raw meats.

APPLICATION TO FOODS AND FOOD INGREDIENTS 197 Critical control points associated with handling in retail outlets, food service establishments, and in the home are beyond the control of the meat processor. To obtain optimum shelf-life and wholesomeness, the HACCP concept has to be applied along the entire production, processing, distribution, and food preparation chain. Need for Microbiological Criteria Microorganisms of public health significance such as Salmonella, C. perfringens, S. aureus, Y. enterocolitica, and C. jejuni often are pres- ent in small numbers as part of the natural microbial flora of live animals. Even the best production and processing practices do not eliminate these organisms from raw meats. Therefore, limits for pathogenic microorgan- isms in microbiological criteria for raw meats are impractical. The same is true for indicator organisms such as the coliform group (coliforms, fecal coliforms, and Escherichia coli) because there is no direct relationship between the presence of these types and the presence or absence of path- ogens. The courts have held that Salmonella is an inherent defect in raw meats (APHA v. Butz, 19741. This decision was based upon the fact that even with the use of the best manufacturing practices consistent with present technology, Salmonella cannot be eliminated from raw meat. The APC of refrigerated red meats in distribution channels reflects microorganisms acquired from a series of events: slaughter-dressing pro- cedures, chilling, fabrication into primals, subprimals, and retail samples, and growth during refrigerated storage. At points remote from the pro- cessing line the APC does not distinguish between microorganisms from the carcass, those acquired during processing, and those resulting from growth during normal refrigerated storage. Although the APCs may be similar, vacuum-packaged cuts on which lactic acid bacteria frequently predominate can be organoleptically acceptable, whereas comparable cuts under aerobic storage, with gram-negative aerobic rods predominating, may be unacceptable. The APC may be of some value to a meat processor to evaluate processing conditions and perhaps shelf-life of a product under well-defined conditions of refrigerated storage. However, the APC of perishable red meats in distribution channels is of little value in the eval- uation of the microbial quality of the carcass and processing practices and has no relevance to health. The history of the Oregon Meat Standard is directly related to the above discussion. In 1973, the state of Oregon adopted a microbiological standard for fresh or frozen red meat at the retail level with limits of APC 5x106 and 50 E. cold per gram. The standard was revoked in 1977 (State of Oregon, 1977) for the following reasons:

198 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA There was no evidence that application of the standard improved sanitary conditions in the retail market. 2. There was no evidence of a significant change in the numbers of microorganisms in ground meat and probably no change in quality char- acteristics of the meat. 3. There was no evidence of a reduction of foodborne illness. 4. The promulgation of the standard may have created the impression on the part of the public that ground meat produced under the standard would have lower microbial counts, improved quality, and fewer health hazards. There was no clear evidence that the expectations were mater- ialized. 5. The additional costs of the program were not justified because the expected benefits, namely lower microbial counts, improved organoleptic quality, and reduced risk to public health, were not clearly demonstrated. The working group of the Codex Committee on Food Hygiene (FAO/ WHO, 1979) also concluded that no benefits would result for either public health or quality through the application of microbiological criteria for raw meats. In summary: 1. Carcasses, primal, subprimal, and retail cuts of meat from normal, healthy animals contain a variety of microorganisms including low levels of some pathogens. 2. Refrigerated raw meats will spoil eventually even if they are pro- duced from the carcasses of normal, healthy animals, fabricated under good manufacturing conditions, and properly refrigerated. 3. If red meats are not properly cooked, held, cooled, and stored, they can cause foodborne illness. 4. Microbiological standards for raw meats will prevent neither spoilage nor foodborne illness and thus do not appear warranted. Instead, appli- cation of the HACCP system to the entire processing and distribution chain including the meat-packing plant, retail units, food service establishment, and home should be used to produce a product with satisfactory shelf-life and public health safety. 5. Microbiological guidelines are applicable to monitor certain critical control points in the processing of raw meats such as the sanitary condition of equipment and utensils and the condition of freshly dressed carcasses. References APHA v. Butz 1974 American Public Health Association et al., Appellants vs. Earl Butz, Secretary of Agriculture et al. D.C. Civil Court. Suit to enjoin the Secretary of Agriculture against

APPLICATION TO FOODS AND FOOD INGREDIENTS 199 alleged violations of the Wholesome Meat Act. Pp. 331-338. 511 F. 2d 331 (D.C. Civ. 1974). APHA (American Public Health Association) 1984 Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed., M. L. Speck, ed. Washington, D.C.: APHA. Ayres, J. C. 1955 Microbiological implications in handling, slaughtering, and dressing meat animals. Adv. Food Res. 6:109-161. 1960 Temperature relationships and some other characteristics of the microbial flora de- veloping on refrigerated beef. Food Res. 25:1-18. FAD/WHO (Food and Agriculture Organization/World Health Organization) 1979 Report of a FAD/WHO working group on microbiological criteria for foods. Geneva: FAD/WHO. Gill, C. O. 1979 A review. Intrinsic bacteria in meat. J. Appl. Bacteriol. 47:367-378. Gill, C. O., and K. G. Newton 1981 Microbiology of DFD beef. Pp. 305-327 in The Problem of Dark-cutting in Beef, D. E. Hood and P. V. Tarrant, eds. The Hague: Martines Nijhoff. Hanna, M. O., J. W. Savell, G. C. Smith, D. E. Purser, F. A. Gardner, and C. Vanderzant 1983 Effect of growth of individual meat bacteria on pH, color and odor of aseptically prepared vacuum-packaged round steaks. J. Food Prot. 46:216-221, 225. ICMSF (International Commission on Microbiological Specifications for Foods) 1980 Microbial Ecology of Foods. 2. Food Commodities. New York: Academic Press. Ingram, M. 1972 Meat processing past, present and future. R. Soc. Health J. 92:121-131. Ingram, M., and T. A. Roberts 1976 The microbiology of the red meat carcass and the slaughterhouse. R. Soc. Health J. 96:270-276. Roberts, T. A. 1974 Hygiene in the production of meat. Meat Research Institute Memoir No. 696. Lang- ford, Bristol. State of Oregon 1977 Report of the meat bacterial standards review committee. Salem, Oregon: Department of Agriculture. Williams, L. P., and K. W. Newell 1970 Salmonella excretion in joy-r~ding pigs. Am. J. Publ. Health 60:926-929. C. PROCESSED MEATS Introduction After the animal carcass is chilled, it is generally broken down into primal, subprimal, and ultimately retail cuts. Unlike poultry, where the  various separated parts of the carcass are classified as "processed," the various cuts derived from carcass meat are still classified as "red meat," i.e., unprocessed.

200 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Sensitivity of Products Relative to Safety and Quality Raw Ground Beef Ground beef (hamburger) is an important by-product of the retail butcher shop. Trimmings from in-store cutting operations, as well as various "held-over" retail cuts, are ground at the store level. The product may be sold in bulk or may be packaged in oxygen-permeable films. A substantial quantity of raw ground beef is now being processed in central locations for distribution to retail outlets. The raw materials consist of trimmings from cutting operations, but in addition, substantial quantities of various cuts such as chucks and rounds are also utilized. Such ground meat is generally packaged in an oxygen-impermeable film. In some cases the product is frozen before distribution to retail outlets; in other cases it is shipped under refrigeration. Generally, the shelf-life of in-store produced ground beef is relatively short, due in part to high microbial numbers in the raw materials, but also due to lax practices with respect to equipment sanitation and temperature control. The initial flora reflects that of the raw materials and thus is comprised chiefly of gram-negative aerobic psychrotrophic bacteria. Dur- ing storage these organisms are responsible for surface spoilage; in the interior of the meat microaerophilic bacteria (Lactobacillaceae and Bro- chothrix thermosphacta) develop (Gardner, 19811. Packaged ground beef produced at central locations generally has sub- stantially greater shelf-life than in-store produced product. The oxygen- impermeable packaging material creates an environment that inhibits the growth of gram-negative psychrotrophic bacteria. Ultimately spoilage is brought about by the growth of lactic acid bacteria and B. thermosphacta. Successful producers of the product must exercise considerable control over the microbiological condition of raw materials, equipment sanitation and temperature control. Though the sensory changes caused by the de- veloping flora in this product are not nearly as disagreeable as those brought about by the gram-negative psychrotrophs, gas production gen- erally brought about by heterofermentative lactic acid bacteria is often a serious problem. Raw ground beef, as such, has not constituted a serious foodborne disease problem. Obviously, those who consume the product without cooking (as in steak tartar) may contract salmonellosis and, at least in theory, this practice could lead to infection with Yersinia and Campylo- bacter. The consumption of raw ground meat packaged at a single meat- processing plant, was the cause of a multistate outbreak of Salmonella newport disease (Fontaine et al., 1978~. Raw ground beef can constitute

APPLICATION TO FOODS AND FOOD INGREDIENTS 201 a trichinellosis hazard if the meat is ground in equipment that has pre- viously been used to grind pork. This is far more apt to happen with raw ground beef produced at the store level than with that manufactured in centralized locations where regulatory control is exercised. Pathogenic E. cold (0157:H7) in hamburger has been responsible for outbreaks of hemorrhagic colitis (CDC, 19821. Adequate cooking of the meat should minimize or eliminate the risk of contracting this disease from eating hamburgers. Cooked ground beef as an ingredient in a variety of foods can be associated with foodborne illness (Bryan, 19801. The critical control points in the production of raw ground beef are the raw meats, equipment sanitation, and temperature control. At centralized locations, the raw materials may be entirely generated "in-house." More often, at least a portion of the raw meats are purchased from outside sources. Outside purchased trimmings are generally inspected upon re- ceipt, and if sensory evaluation indicates off-condition the trimmings should be rejected. The prudent manufacturer will conduct microbiological anal- yses on the trimmings from various suppliers and eliminate those suppliers shipping raw meats of poor microbiological quality. Proper cleanup of equipment used to convey raw meats, grinders, and packaging machinery is essential to the production of a product with satisfactory shelf-life. Temperature control is of paramount importance. Most manufacturers attempt to grind and package product as near to the freezing point as possible. This is often accomplished through the use of dry ice. If proper steps are not taken, however, the product may be packaged before the ice has sublimed. If this occurs, the packaged product may become "gassy" due to evolution of gaseous CO2 after packaging. Microbiological mon- itoring of the finished product is only of retrospective value, because results become available after the product has entered merchandising chan- nels. The prudent manufacturer will regularly monitor critical control points to determine whether he has been successful in controlling the microbiological quality of raw meat, the sanitation of processing equip- ment, and storage temperature. Perishable Raw Salted and Salted Cured Meats Examples of these products are fresh pork sausage, Polish and Italian sausage, and uncooked ham, bacon (see also Cooked Cured Meats), and corned beef. Fresh pork sausage, Polish sausage, and Italian sausage are essentially ground meat to which spices and salt are added. The addition of salt may be somewhat inhibitory to the gram-negative psychrotrophic aerobic bac

202 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA teria that normally spoil red meats. However, these are highly perishable products that when packaged in oxygen-permeable films support the growth of a heterogeneous microbial flora similar to that of ground beef, though the rate of spoilage is somewhat slower due to the addition of salt. When packaged in oxygen-impermeable films, the spoilage flora is comprised of lactic acid bacteria and B. thermosphacta (Gardner, 1981; Johnston and Tompkin, 19841. These products may be merchandised refrigerated. Alternatively, they may be frozen at the processing level and merchandised in the frozen state, or frozen at the processing level and merchandised as refrigerated products. The addition of salt makes these products susceptible to oxidative rancidity. If merchandised in the frozen state, oxidative ran- cidity is the usual cause of spoilage. If merchandised as refrigerated prod- ucts, microbial spoilage precedes oxidative rancidity. As with ground beef, the critical control points are the raw materials, temperature and equipment sanitation. Again, finished product analysis is useful but of retrospective value. If properly cooked and promptly served, these products do not constitute an important foodborne disease problem. If not adequately cooked, products containing pork may pose a hazard of trichinello~is (see Chapter 4), e.g., the incorporation of fresh pork sausage in turkey dressing that is not adequately heat treated during roasting. Uncooked ham, bacon, and corned beef are raw salted cured meats. Uncooked ham is no longer an important commodity in the United States; bacon and corned beef are of considerable importance. In the production of bacon and corned beef, pork bellies and beef briskets respectively are injected with a "pickle" containing salt and nitrite and/or nitrate. The pork bellies are then subjected to a mild heat treatment (53.3 to 54.4°C/ 128 to 130°F), a temperature sufficient to "fix" the cured meat color but not sufficient to destroy trichinellae. A substantial volume of corned beef is distributed as a raw cured product to which no heat treatment is applied. Currently, most bacon is merchandised as a sliced product packaged in oxygen-impermeable films. Spoilage is due primarily to the growth of lactic acid bacteria. Likewise, uncooked corned beef is merchandised in oxygen-impermeable films and the spoilage flora is also comprised pri- marily of lactic acid bacteria. If bacon is packaged in oxygen-permeable films, its shelf-life is significantly shorter and the spoilage flora consists primarily of yeasts and micrococci. The critical control points in the production of these products include the microbiological status of the raw materials (beef briskets and pork bellies), the microbiological status of the pickle, and the environment of the packaging area. Generally, the pickle is recirculated and, if adequate control is not exercised, this recir- culated pickle can be a significant source of spoilage organisms, including those that cause spoilage due to oxidation of the cured meat pigment. In

APPLICATION TO FOODS AND FOOD INGREDIENTS 203 the past, when it was common to merchandise bacon in oxygen-permeable films, the air in the packaging area was a critical control point as mold growth commonly terminated the shelf-life of sliced bacon. Currently, most bacon is merchandised in oxygen-impermeable films, and mold development is no longer a significant problem. The perishable raw salted and salted cured meat products do not con- stitute a significant foodborne illness problem, although, in common with other meat products, misuse by the ultimate consumer can lead to food- borne illness. On occasion, bacon may be consumed by certain ethnic groups without cooking, and outbreaks of trichinellosis have resulted. Cooked Polish and Italian sausages, as well as heat-processed corned beef, also are merchandised. The spoilage and foodborne disease problems of these products are analogous to those presented by other cooked cured meats (see below). Shelf-Stable Raw Salted and Salted Cured Meats Included in this category are salt pork, dry cured bacon, and dry cured ham (country cured ham). These products were extremely important in the early history of the United States and continued to be so until home refrigeration procedures became common. Currently, salt pork, dry cured bacon, and country hams continue to be produced in agricultural regions. Salted pork, bacon, and hams are prepared by coating the meat with dry salt and storing it at low temperatures (below 10°C/50°F). At intervals, the meat is recoated with dry salt. The product contains high levels of salt and is racked and held at ambient temperatures until the surface of the meat dries. It may then be rubbed with a thin coating of salt and spices, netted, and sold. With bacon and some ham, nitrite and/or nitrate is added to the dry salt used for curing. The product may be hung to dry at ambient temperature for 35 to 140 days before offering it for sale as a shelf-stable meat product. During the initial phases of processing, salt and/or curing agents penetrate and equilibrate in the tissues. Non-spore- forming bacteria of public health concern are subjected to stress or are rendered nonviable. During the subsequent drying period, salt-tolerant micrococci and enterococci grow and appear to render the products even more retractile to the microorganisms of public health concern. The nitrite- and/or nitrate-cured products are required to be held in the dry room for a specified period to destroy trichinellae. Although most U.S. consumers cook these products before eating, dry cured hams have historically been eaten raw in Europe (Johnston and Tompkin, 19841. Although these products are produced by major meat packers as dry cured hams, a substantial volume is produced by small processors, i.e.,

204 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA as a cottage industry (country cured hams). These operators generally initiate production during the late fall and winter months; the curing pro- cess is continued into the early spring of the year. Processors depend upon cool weather to protect the product from spoilage during the curing period. They provide no mechanical refrigeration and, if climatic conditions are unfavorable, internal spoilage of the product may occur before salt and nitrite penetrate by diffusion to protect the product from spoilage. If unseasonably cold weather occurs, freezing of the meat may prevent pen- etration of salt and nitrite, this leading to spoilage when warm weather appears. Bone taint deep tissue spoilage near the bone occurs when improper salt equilibration occurs (Mundt and Kitchen, 1951~. These products have an excellent public health history in the United States. They are merchandised at ambient temperatures, and if properly processed, microorganisms of public health concern constitute no problem. Spoilage problems, as indicated above, relate to improper control of tem- perature during the curing process. Molds may develop on the finished product during merchandising but to date have constituted no known foodborne disease hazard. The critical control points in the manufacture of these products are the control of temperature during the curing period and, closely related to this, the proper penetration and equilibration of the curing ingredients. Improper control of humidity during the drying process can lead to the development of surface spoilage microorganisms. Microbiological criteria for either quality control or regulatory purposes are not applicable. . Cooked Uncured bleats i. Products in this category include cooked versions of the perishable raw salted items discussed above (pork sausage, and Polish and Italian sau- sage). In addition, cooked uncured meats are components in a variety of  frozen entrees such as "TV dinners," potpies, and sliced cooked meat with gravy. Still other cooked uncured meats are produced for bulk dis- tribution to food service establishments, e.g., cooked ground beef with chili, thinly sliced cooked beef for use in Mexican or other ethnic dishes. These products may be frozen, but more often are distributed under re- frigeration. A substantial volume of cooked "roast beef" is distributed to food service establishments. These products are not necessarily prepared from primal cuts. Sometimes they are fabricated from meat trimmings in such a manner that when sliced they have the appearance of cooked roast beef. Generally, the heat treatment used in the preparation of these products s sufficient to destroy most vegetative microorganisms. A notable ex

APPLICATION TO FOODS AND FOOD INGREDIENTS 205 ception is precooked roast beef. Post-cooking handling by the manufac- turer, the food service worker, and the consumer frequently results in contamination of these products with microorganisms, including those capable of producing spoilage and those of public health concern. If temperature abuse follows, the hazards of spoilage and foodborne disease increase. In recent years precooked "roast beef" has been a significant vehicle of human salmonellosis in the United States (CDC, 1976a,b; 1977a,b; 1981; and Bryan, 19801. Initially, the multistate outbreaks traced to this product were thought to be due to consumer desire for roast beef presenting a rare appearance and the necessity, therefore, for mild heat processing. Accordingly, research was conducted to determine time-temperature pa- rameters necessary to the destruction of salmonellae in "roast beef" (Goodfellow and Brown, 19781. Despite the promulgation of regulations consistent with the parameters of these studies, outbreaks of salmonellosis traced to cooked beef persist (CDC, 19781. Outbreaks can be attributed either to failure of the processor to comply with the established regulations or to post-heat treatment contamination. In the outbreak cited above, the USDA determined that the processor had not complied with USDA reg- ulations. However, cross-contamination between raw and adequately cooked products is the more usual problem (R. W. Johnston, USDA. 1983. Per- sonal communication). The last outbreak occurred in 1980 (R. W. John- ston, 1984, personal communication). Through a combination of regulation, education of processors, and microbiological monitoring by both industry and the USDA the problem of salmonellae in precooked roast beef appears to have been brought under control. The sampling of products that are rebagged after cooking is more intense than that of products cooked as rolls and not rebagged. However, the USDA requires industry monitoring of both types of product (R. W. Johnston, 1984, personal communication). The hazard potential for precooked uncured meats produced in com- mercial establishments is high (NRC, 19641. Yet the incidence of out- breaks traced to cooked meat produced in meat-processing plants has been low, except for the aforementioned important role of precooked roast beef in human salmonellosis. Four cases of botulism have been traced to frozen potpies (CDC, 1960; State of California, 1975; State of California, 1976; and CDC, 19831; in each instance, gross mishandling in homes was the direct cause of illness. Thus, despite the high potential of precooked uncured meats processed in meat-processing plants as a foodborne disease hazard, few outbreaks have been uncovered. However, cooked meat, mishandled in food service establishments and homes, has been an im- portant vehicle of foodborne disease (Bryan, 19801. The critical control points in the manufacture of these products include,

206 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA aside from the microbiological quality of raw materials, the heat process and post-heat process handling, proper refrigeration, and adequate re- heating. Assuming adequate heat processing, proper chilling of the heated product is a critical control point. In recognition of this, the USDA (1983) requires that the maximum chill time after cooking of cooked beef, roast beef and cooked corned beef be restricted to no more than six hours during which the product may be in the optimum growth zone for pathogens, defined as 48.9-12.8°C (120-55°F). It further stipulates the chill of the product must continue until the product reaches 4.4°C (40°F) and that the product not be packed until this occurs. Other post-heat treatment handling procedures constitute critical control points, e.g., the slicing of cooked products, boning, packaging, etc. As indicated previously, these proce- dures may add microorganisms of spoilage and/or foodborne illness sig- nificance. Subsequently, improper handling by persons in food service establishments and in homes may lead to spoilage or the potential for foodborne disease. In the commercial production of these products the processor has assumed many of the responsibilities met by the consumer who prepares such foods from raw meat and other ingredients. When the expanding market for these products first developed (in the 1950s and 1960s) there was concern over their high foodborne disease hazard po- tential (NRC, 1964), but the low incidence of foodborne disease outbreaks traced to them suggests that the processor is more able to circumvent the hazards attendant to their production than is the individual consumer. Cooked Cured Meats Cooked cured meats include domestic sausage, (e.g., frankfurters, bo- logna), luncheon meats (e.g., pepper and pimiento loaf), ham, bacon, and corned beef. The common shared property is that the products are cured, i.e., treated with nitrite and salt, and are subject to heat treatment. Some are ready to eat (bologna, luncheon meats, and fully cooked hams); others are cooked before eating to improve texture and flavor (smoked ham and frankfurters); still others must be cooked before eating to ensure safety, e.g., bacon that does not receive a heat process sufficiently severe to destroy trichinellae. The physical nature of the raw materials varies from primal cuts of red meats (ham, pork bellies, and beef briskets) to red meat trimmings that are the basic raw materials for domestic sausage and luncheon meats. This necessitates different methods for introducing the curing ingredients (ni- trite and salt, with or without nitrate). With ham, bacon, and beef briskets, the curing ingredients are introduced in a solution (pickle). Historically, the primal cuts were submerged in the pickle, which, over a period of

APPLICATION TO FOODS AND FOOD INGREDIENTS 207 time, diffused into the meat. Pork bellies were rubbed with dry salt con- taining nitrite, with or without nitrate; the curing ingredients dissolved in the moisture phase and, as with ham and beef briskets, penetrated by diffusion. At the present time the curing ingredients are introduced by "pumping" pickle into the primal cuts of meat. This has greatly changed the microbiological problems presented by these products. In former times internal microbial spoilage occurred before complete diffusion of the cur- ing ingredients (see section above on raw salted and salted cured meats). The modern process of injecting the pickle results in rapid distribution of the curing salts within the meats and largely obviates the problem of bone taint. On the other hand, microorganisms contained in the pumping pickle may be instantaneously distributed throughout the meat mass. At times this can result in interior spoilage problems. With domestic sausage and luncheon meats, the trimmings are ground and the curing ingredients in the form of a dry powder are mixed with the trimmings, generally with the concomitant introduction of various spices, sweeteners, and a curing accelerator. After the spices and curirig ingredients are mixed with ground trim- mings, the emulsion is "stuffed" into casings. In the past, natural casings, which were sections of animal intestines, were utilized. Natural casings are still used with some products, but the bulk of domestic sausage products are now processed in synthetic casings. Luncheon meat emulsions are generally introduced into metal containers that are fitted with watertight lids. Whether the product being manufactured is a primal cut or an emulsion prepared from ground trimmings, development of the stable cured meat color requires heat treatment. For this purpose hams and similar products are wrapped in nets. The netted hams are hung in smokehouses; bacon bellies are attached to metal hangers and hung on "trees." The trees are placed in smokehouses. Domestic sausage products, such as frankfurters and bologna, are generally hung in smokehouses, whereas luncheon meats are usually cooked by submersion in water. Historically, hams, bacon, and briskets were "held in cure" under refrigeration for many days, even weeks, before heat processing. This process has gradually died out in the meat industry. In most plants these products are heat treated soon after the cure is introduced, though some packers may hold the pumped product overnight before heat processing. Domestic sausage and meat emulsions are generally heat processed very soon after the emulsion has been pumped into casings or introduced into metal forms. The heat process applied to cooked cured meats depends upon the product (for temperatures required in meat processing, see Bailey, 19741. For example, bacon is cooked to a final internal temperature of 48.9 to

208 EVALUATION OF THE ROLE OF MlCROBlOLOGlCAL CRITERIA 54.4°C (120 to 130°F). The heat is applied for the sole purpose of "fixing" the cured color. This requires denaturation of the myoglobin, which takes place at temperatures between 48.9 and 54.4°C (120 and 130°F). Smoked hams, on the other hand, must achieve a temperature of 58.3°C (137°F). This minimum internal temperature is required for the destruction of tri- chinellae. Fully cooked hams must achieve a minimum internal temper- ature of 70°C (158°F). Domestic sausage and luncheon meats containing pork must achieve a minimum internal temperature of 58.3°C (137°F) to achieve freedom from trichinellae. From a practical point of view, how- ever, luncheon meats and domestic sausage are processed to much higher temperatures in order to ensure satisfactory shelf-life. The temperatures employed vary with the processor, but generally they range from approx- imately 68.3 to 76.7°C (155 to 170°F). The maximum temperature, as with the minimum, is self-limiting, since the higher the final temperature the greater the shrinkage loss during heat processing. The processes nor- mally applied are sufficient to destroy non-sporeforming bacteria of public health concern, namely, staphylococci and salmonellae, though the spores of Clostridium perfringens, Clostridium botulinum, and Bacillus cereus survive. Thermoduric non-sporeforming bacteria, such as Group D strep- tococci and some lactobacilli, may survive the heat process. The number of survivors is directly related to the level of microorganisms in the un- heated product as well as to the degree of heat treatment applied. There is some evidence that surviving non-sporeforming bacteria may be heat injured and thus initiate growth only slowly in properly refrigerated fin- ished product, the degree of heat injury being related to the severity of the process (Greenberg and Silliker, 19611. Various natural spices, e.g., black pepper, may contribute large numbers of aerobic sporeforming bac- teria to the sausage or luncheon meat emulsion. These organisms survive the heat process and at times are the predominant surviving microorgan- isms. They constitute neither a spoilage nor a public health hazard in properly refrigerated cooked cured meats; however, occasionally they may cause product spoilage during the heat process, particularly with large pieces of sausage such as bologna, which must be subjected to a long heat process. After heat processing, domestic sausage and luncheon meats are gen- erally showered with cold water and placed under refrigeration. The chilled meats are then further treated. The products cooked in molds are removed from the molds as loaves. The luncheon meats are then sliced and pack- aged. Similarly, domestic sausage products such as bologna are sliced and packaged. With products such as frankfurters that have been heat processed in artificial casings, it is the practice to mechanically remove (peel) the casings. Following this, they are packaged. Natural casings are generally not removed.

APPLICATION TO FOODS AND FOOD INGREDIENTS 209 It is in the handling of cooked sausage products prior to packaging that the greatest opportunities for contamination with microorganisms occur. For example, a contaminated slicer blade can inoculate the surface of every piece of meat it touches. The equipment used to peel frankfurters can similarly be a source of contamination to every product unit passing through it. Conveyors that transfer sliced or peeled products to packaging equipment can likewise contaminate finished products with spoilage or- ganisms. Most packaging areas are maintained at temperatures that select for the growth of psychrotrophic organisms, 1.1 to 4.4°C (34 to 40°F). This only serves to magnify post-heat treatment handling problems. Cooked cured meats are perishable commodities. In general their shelf-life under refrigerated storage conditions is directly related to the number of micro- organisms present upon them at the time of packaging. Most important in this regard is the contamination that occurs after heat processing. These microorganisms are specifically adapted to growth at low temperatures, and if the environment of the processing area is not properly controlled, the organisms are metabolically active. These microorganisms are more capable of initiating growth on the packaged product than are the survivors of the heat process, which may suffer from the effect of heat injury. Hams may be packaged as whole hams or may be cut to yield half hams or slices. Due to the smoking process, the surfaces of hams tend to be somewhat dry, and some of the components of smoke that are deposited on these surfaces have mild antimicrobial properties. Thus, hams receive less post-heat treatment handling and their surfaces are somewhat more resistant to microbial attack than is the case with domestic sausage and luncheon meats. Bacon, though occasionally sold unsliced, is generally sliced and packaged. As with domestic sausage, if the slicing equipment is not satisfactorily maintained, each sliced piece may become contami- nated with spoilage organisms. The spoilage flora of cooked cured meats depends upon the packaging system employed. In the past, packaging was in oxygen-permeable films, and the spoilage flora was comprised primarily of micrococci and yeasts. In recent years, oxygen-impermeable films have been used and the spoilage flora has consisted primarily of lactic acid bacteria. Formerly, these prod- ucts had a relatively short shelf-life. Packaging in oxygen-impermeable films has significantly increased shelf-life since the lactic acid bacteria normally developing during refrigerated storage may reach very high levels without significant organoleptic changes. Characteristically, the growth of these lactic acid bacteria is manifested by a milky appearance in the free moisture within the package. This is due to the production of lactic acid, which precipitates protein in the free moisture surrounding the meat. The public health history of sliced cured meats, domestic sausage, and luncheon meats in this country has been good even though these products

210 EVALUATION OF THE ROLE OF MlCROBlOLOGICAL CRITERIA are commonly subjected to temperature abuse in the hands of the con- sumer. Although salmonellae are quite capable of growing on cooked, cured meats, they have rarely been associated with foodborne illness from this source. Though staphylococci are capable of growth on sliced lunch- eon meats and domestic sausage, it appears likely that the relative freedom from foodborne disease may be attributed to competing microorganisms. The situation with ham is quite different. This cooked cured meat is perhaps the most important source of staphylococcal food poisoning. In general the food poisoning is not related to problems created by the food processor. Rather, most outbreaks of staphylococcal food poisoning are caused by contamination of the ham by food service personnel and persons in their homes after cooking. Such contamination, followed by temperature abuse, is the major cause of foodborne disease traced to this product. Many critical control points are associated with the production of cooked cured meats. Among these are the raw materials themselves. If, for ex- ample, raw materials of poor microbiological quality are utilized, the shelf- life of the finished product will be correspondingly compromised. It is common practice in the meat industry to use "rework" product as a raw ingredient in cooked products. Such product may be generated "in plant." On the other hand, the USDA permits the use of over-age product returned from retail stores if it is not organoleptically spoiled. At least in part, rework consists of over-age product returned from retail sources. Such raw materials may have extremely high loads of microorganisms. A spe- cific problem may relate to microorganisms that cause cured meat dis- coloration. Such organisms are generally lactic acid bacteria, which under aerobic conditions produce hydrogen peroxide. Some of these are ex- tremely heat-resistant (Niven et al., 1954~. A processor experiencing a problem with short shelf-life in his products will have an increasing amount of them retuned from the trade and thus increasing amounts that he is tempted, for econamip reasons, to rework. If heat-resistant, peroxide- producing lactic acid bacteria are present in the returned product, the use of rework may cause an increasing problem of spoilage that can only be solved through terminating rework as a raw material. As mentioned previously, spices may on occasion constitute a critical raw material if they contain excessive numbers of aerobic sporeforming bacteria that survive normal heat processing. The heat process itself is a critical control point for these products. The number of survivors is directly related to the severity of the heat process except of course with respect to heat-resistant spores. These spores introduced by spices may lead to spoilage during heat processing of certain products such as large pieces of bologna. The number of survivors of a given heat process is also related to the initial number of organisms in the emulsion or other product being treated.

APPLICATION TO FOODS AND FOOD INGREDIENTS 211 The sanitation of equipment for handling the cooked product (slicers, peelers, and conveyors) is also a critical control point. If such equipment, through insanitary practices, becomes a site of microbial growth, each product unit may become contaminated with spoilage organisms that re- duce shelf-life. The processor must be certain that the processing envi- ronment is so controlled that this does not happen. Control is accomplished by microbiological analyses of equipment during the course of the oper- ating day. Cleaning schedules are established upon the basis of such studies to break the cycle of microbiological development. Indeed, the USDA requires a processor to establish that a mid-shift cleanup is unnecessary in the plant. Proof is based upon data indicating that the processing equip- ment in the plant does not become a site of active microbial growth during the course of an operating shift. Critical control points in the manufacture of cooked cured meats include proper temperature control at all stages of processing, including the handling of raw material, the proper chilling of cooked product, the environment of the packaging operations, and the storage of packaged products. Prevention of cross-contamination and proper cold storage practices are also essential in retail stores, food service es- tablishments, and homes. Fermented Sausages The initial steps in the production of domestic sausage are followed in the production of fermented sausage: the raw materials (meat trimmings that are ground, appropriate spices, sugar, and curing agents) are mixed and the emulsion is "stuffed" into natural or artificial casings. At this point, the processes diverge. Fermented sausages are held at a temperature of about 21 to 32°C (70 to 90°F) for 24 to 96 hours. During this period, under proper conditions, lactic acid bacteria develop and their growth results in fermentation of the sugar in the initial emulsion, the development of lactic acid, and a drop in pH to 5.3 or lower. Processors may depend upon the selective nature of the cured meat environment for lactic acid bacteria (natural fermentation) or may inoculate the product with com- mercially available cultures of lactic acid bacteria to ensure that fermen- tation occurs. Alternatively, glucono-delta-lactone may be added to the meat emulsion, the hydrolysis of which yields gluconic acid and a rapid decrease in pH. As an alternative to commercial cultures, many processors practice "back slopping," the addition of naturally fermented meat from a previous batch to the new emulsion to provide fermentative organisms. In all procedures the 24- to 96-hour hold in the "green room" at tem- peratures around 21 to 32°C (70 to 90°F) should, under proper conditions, produce a fermented meat emulsion with a pH of 5.3 or lower. If the product is then subjected to a mild heat process, usually including smoke,

212 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA a semidry product results. If the emulsion contains pork, the product must be processed to a minimum of 58.3°C (137°F). If it does not contain pork, the degree of processing is then selected by the manufacturer. The finished semidry sausage, e.g., Lebanon bologna, Thuringer, Summer sausage, has a relatively long shelf-life due to the increased acidity produced by fermentation. Experience teaches that these products can be safely mer- chandised at ambient temperatures, though most of them are merchandised under refrigerated conditions as either sliced or unsliced products. If, after fermentation in the "green room" the product is held in a "dry room," a room maintained at low humidity with proper air flow, a dry sausage is produced. The product is held in the dry room for an extended period of time varying from several weeks to months, depending upon the volume of the sausage pieces. During this period, the low humidity and air movement in the room result in loss of moisture from the product. Secondary fermentations occur and nuances of flavor develop. Moisture loss results in concentration of salt in the moisture phase of the product. If the product contains pork, the length of time in the dry room is closely regulated by the USDA in order to assure that any trichinellae present in pork ingredients will be destroyed by low water activity. The resulting dry sausage, e.g., dry salami, Genoa, is stable at ambient temperatures and may be merchandised without refrigeration. Both semidry and dry sausage products are relatively resistant to mi- crobial spoilage. The exteriors of unsliced products may support the growth of mold, but such mold growth has generally been accepted by the con- sumer as a characteristic of the product. Its occurrence has not been associated with foodborne illness. The health hazard presented by fermented sausages relates primarily to improper fermentation, in which case the growth of staphylococci with attendant enterotoxin formation may pose a health hazard (Barber and Deibel, 19721. The American Meat Institute has recently formulated Good Manufacturing Procedures in connection with the production of fermented sausages (AMI, 19821. The key factor is rapid acid production, which prevents the development of enterotoxigenic staphylococci. The survival and/or growth of salmonellae in these products is likewise a potential problem. Salmonellae are relatively salt-tolerant. As indicated above, fermented sausages are resistant to microbial spoil- age, except for the surface growth of molds. The foodborne disease hazard relates primarily to the growth of enterotoxin-producing staphylococci where prompt fermentation and pH reduction do not occur. The critical control point is the fermentation itself, and the American Meat Institute has provided leadership by indicating methods for monitoring. Proper control of humidity and air movement in the drying room is likewise a

APPLICATION TO FOODS AND FOOD INGREDIENTS 213 critical control point, as this relates to proper drying, which in turn is directly related to control of the hazard of trichinellosis. This is under control of the USDA, and trichinellosis traced to these products is not a serious problem in the United States. Canned Uncured Meat Products in this category include low-acid canned foods (roast beef and gravy, beef stew, chili con came, and tamales) and acid canned foods (Sloppy Joe and spaghetti with meat sauce). The classification of a given canned meat depends upon the equilibrated pH at the time of retorting. If the pH is 4.6 or below, the food is classified as an acid canned product and need only be processed to an extent assuring stability. If the final pH of the product is above 4.6, it is classified as a low-acid canned food and must be heat processed to assure freedom from C. botulinum (see parts J and K in this chapter). In this regard, canned meat products are no different from other canned foods and require no additional discussion in this sec- tion. Shelf-Stable Canned Cured Meats Shelf-stable canned cured meats include (1) canned viennas, corned beef, frankfurters, and meat spreads all of which receive a botulinal cook or greater; (2) canned luncheon meat and canned hams that are given less than a botulinal cook; (3) canned sausages with hot oil in the final container and a water activity of <0.92, and sliced dried beef in vacuum- sealed jars and canned prefried bacon that rely upon a water activity of <0.86 for stability; and (4) vinegar pickled meats such as sausages and pig's feet (Johnston and Tompkin, 19841. Products in category ( 1 ) are treated as low-acid canned foods (see canned foods sections of this chapter). Those in category (2) are limited to 3 lbs or less. They receive significantly less than a botulinal cook. Empirically a heat process equivalent to Fo = 0.1 to 0.7 has been found to produce a shelf-stable product. This stability is dependent upon the presence of nitrite, salt, a low indigenous level of C. botulinum in meat, and a thermal process that injures surviving spores such that they are incapable of growth in the cured meat environment. Canned sausages, sliced dried beef, and prefried bacon (category 3) do not undergo spoilage unless their water activity is higher than prescribed or the vacuum seal is broken. Theoretically, it may be possible for S. aureus to grow if the water activity exceeds 0.86, but this has not been observed in commercial practice. The inability of S. aureus to compete

214 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA with other organisms developing in this environment is probably the major factor contributing to the absence of staphylococcal food poisoning. Pickled pig's feet, pickled sausages, and similar items in category (4) are immersed in vinegar brine. Their stability may be traced to low pH combined with the presence of undissociated acetic acid, little or no fer- mentable sugar in the tissue, and/or an airtight package (Niven, 19561. None of these products has been a significant source of foodborne disease or microbial spoilage. For spoilage problems relating to products in category (1) or those common to other low-acid canned foods receiving a botulinal cook, see Part J of this chapter. With respect to category (2), thermophilic spoilage can occur in products receiving a cook equivalent to Fo = 0.1-0.7, if the products are stored at an abnormally high temperature for a sufficient length of time. Furthermore, if inordinately high numbers of mesophilic sporeforming bacteria are present, mesophilic spoilage can occur. Faulty curing resulting in insufficient levels of salt or nitrite can also result in an unstable product. As with other low-acid canned foods, poor manu- facturing practices can lead to product spoilage. As previously indicated, products in category (3) do not undergo microbial spoilage unless their water activity is higher than recommended. Products in category (4) may contain moderate numbers of lactic acid bacteria and viable spores. Large numbers of lactic acid bacteria may develop and cause the brine to become cloudy (Niven, 19561. However, foodborne disease organisms do not survive. Hermetically sealed jars of pickled bone-in meat may develop gas and even explode from the action of the vinegar on bone. This spoilage is of nonmicrobial origin but may prove puzzling to microbiologists investigating this cause of spoilage. Critical control points in the production of these products include: for category (1), those points essential to the production of low-acid canned foods; for category (2), the proper curing of the meat, the control of the level of aerobic and anaerobic sporeforming bacteria, and proper heat processing; for category (3), the proper control of water activity and of container integrity; for category (4), control of brine and acid content, temperature control during storage for prevention of growth of acid-tol- erant bacteria with attendant gas production, and proper container integrity to prevent the growth of mold. Perishable Canned Cured Meats These products receive a heat process far less than that necessary for commercial sterility. They are merchandised under refrigeration with a label indicating that they are perishable. The major product in this category is perishable canned ham, which must achieve a minimum temperature

APPLICATION TO FOODS AND FOOD INGREl)IENTS 215 of 65.6°C (150°F) during processing. A substantial quantity of canned perishable luncheon meat is also manufactured, primarily for distribution to food service establishments. Thermoduric lactic acid bacteria and aerobic and anaerobic spores are the major survivors of the heat process. Lactic acid bacteria may grow extensively during extended refrigerated storage with or without evidence of spoilage. Sporeforming bacteria, which generally develop only with temperature abuse, present the chief health hazard. Canned cured perish able meats have enjoyed an excellent public health history. Hams distributed as retail products are subject to the hazards of mis- handling in food service establishments and homes. Contamination of the ham with S. aureus by persons during slicing or other improper handling followed by prolonged storage at ambient temperatures or storage of large masses in the refrigerator results in staphylococcal growth and enterotoxin formation. A substantial quantity of these hams is merchandised as sliced and packaged items. In such cases the hams must be removed from the con- tainers, sliced, and packaged. In this process, the cans are opened and the hams are unloaded onto tables that "feed" the slicing machine. Ham juices and gelatin are attached to the meat and unloaded onto the table. If the sanitation is not carefully controlled, substantial numbers of micro- organisms, including those capable of spoilage, may develop on the "un- loading table" feeding the slicer. Likewise, the slicing machine, if not maintained in a sanitary condition, is a key source of contamination to the sliced meat. As indicated, perishable canned cured meats mishandled after cooking may be an important source of foodborne disease, particularly staphylo- coccal food poisoning. If these products are sliced and packaged at the packing plant, spoilage problems may result from inadequate sanitation of plant equipment. Critical control points in the manufacturing process are analogous to those in the production of cooked cured meats. The table upon which perishable cured meats are unloaded prior to slicing constitutes a unique critical control point in the production of sliced items derived from per- ishable canned cured meats. Dried Meats Preservation of meat by drying predates recorded history. The combi- nation of this application with salting likely evolved from the prehistoric observation that the meat could be protected from spoilage during drying if it were salted. The following discussion is concerned with meat that is dried without the addition of salt, i.e., where microbial growth is inhibited

216 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA by reduced water activity achieved simply by the removal of water (ICMSF, 1980). Substantial quantities of meat are dried commercially, primarily for use in formulated foods, e.g., soup mixes and dried pasta with meat. Two approaches to drying are employed. In the first, meat is defatted, and the lean meat is cooked and minced, spread in thin layers, and dried under controlled ventilation in hot air tunnels. The air-dried product consists of granules of 5 to 10 mm in diameter and has a water content of less than 15%. The meat pieces must be small and must be separated to ensure thorough drying in one to two hours. In the second process, cooked or uncooked meat is freeze-dried. Pieces as large as chops or steaks can be dried, and hence this process has a wider range of uses than has the air- dried, cooked, mince process. During freeze-drying, microbiological prob- lems are minimized by subzero temperatures that are maintained until the water content is reduced to a level at which microbial growth is impossible. The key to stability and safety of dried meat products lies in water relations. If dried meat is stored at high humidity, water uptake may occur and this may petit the growth of xerophilic molds. Growth of these molds alters the appearance of the meat and may cause musty odors and/ or off-flavors. The following pathogenic microorganisms may be important for dried meats: (1) Clostridia, Salmonella, and other Enterobacteriaceae may be introduced by contamination associated with meat; and (2) staphylococci and B. cereus may be introduced by contamination during preparation and drying. The most dangerous circumstances arise when water is added to rehydrate dried meat. The rehydrated product is an excellent substrate for microbial growth if it is held in a temperature range permitting such development. The critical control points in the production of dried meat include the use of meat of suitable microbiological quality, control of contamination during preparation and transport to the dryer, strict control of time/tem- perature relations during drying, the avoidance of wet spots in the drying mass, drying to a sufficiently low moisture content, protection by suitable packaging of the dried product from reabsorption of moisture, and proper time-temperature control after rehydration to minimize possibilities of microbial multiplication. Need for Microbiological Criteria Raw Ground Beef Microbiological criteria can be usefully applied to assess the micro- biological quality of the raw materials used, the effectiveness of equipment

APPLICATION TO FOODS AND FOOD INGREDIENTS 217 sanitation, and the microbiological quality of finished product. The results obtained through the use of such criteria at the processing level are ret- rospective but serve as a useful guide to the processor. The application of microbiological criteria to the product after it has entered trade channels is without value, even though standards and guidelines have been pro- mulgated by a number of states and municipalities. ~ . Perishable Raw Salted and Salted Cured Meats The considerations discussed in the preceding paragraph apply equally to the perishable meat products in this category. Shelf-Stable Raw Salted and Salted Cured Meats Microbiological criteria have little value in connection with the pro- duction of these products, except perhaps in the evaluation of the micro- biological condition of raw materials. The application of microbiological criteria to the finished product would be without value. Microbiological control during processing is best exercised over critical control points identified previously, particularly the control of temperature during the curing process and the control of humidity during the postcuring, drying period. Cooked Uncured Meats As previously indicated, one product in this category (roast beef) has been a significant vehicle of human salmonellosis in recent years.-It ap- pears that this problem has now been brought under control. Regulation of processing as well as education of processors have played key roles. In addition, microbiological monitoring by processors and the microbiol- ogical surveillance program of the USDA have proven to be essential to assuring the safety of these products before their release for sale. The continued application of microbiological criteria to precooked roast beef and similar products is indicated. Cooked Cured Meats Microbiological criteria can appropriately be used to evaluate the mi- crobiological quality of raw materials, though the results are of retro- spective value with respect to the lot produced from them. Similarly, microbiological criteria can usefully be applied in the evaluation of equip- ment sanitation, particularly at critical control points such as slicers, con- veyors, and casing peelers. Microbiological criteria can effectively be

218 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA applied in the determination of the microbiological quality of finished products. The results of such tests should confirm the effectiveness of the control over critical control points. For example, a sliced product with a significantly higher APC than the unsliced product is an indication of failure to adequately clean and sanitize equipment and/or the use of pro- cessing equipment for too long a period before cleanup. High counts on the surface of such products as frankfurters have similar implications, since surface counts should be negligible. The application of microbio- logical criteria, including purchase specifications to nonmeat ingredients (in particular, spices), may be appropriate for some products. Cooked cured meats have not proven to be a Salmonella hazard. The application of microbiological criteria to these products after they have entered retail channels is inappropriate, despite the fact that many states and munici- palities have promulgated standards and guidelines. Fermented Sausages Microbiological criteria may be useful in the qualitative and/or quan- titative evaluation of starter culture activity (see Chapter 5~. Such a cri- terion might be a part of a purchase specification. The primary microbiological hazard is growth and enterotoxin formation by S. aureus (Barber and Deibel, 19721. To monitor this hazard, the outer 3-mm layer of the individual sausage may be sampled for viable S. aureus at the end of the fermentation cycle before the product is heated and/or dried. Since S. aureus death may have occurred in products ready for consumption, it may be necessary to test the casing or outer 4-mm of meat for thermo- nuclease and/or enterotoxin (Johnston and Tompkin, 19841. These mi- crobiological criteria are best applied at the processing plant. Presence of Staphylococcus enterotoxin should not be tolerated. The product at any point in the distribution should be recalled if there is evidence of the presence of enterotoxin. Other microbiological criteria are not warranted. Canned Uncured Meats Both low-acid and acid products in this category are discussed in parts J and K of this chapter. Shelf-Stable Canned Cured Meats Included in this category are products that receive a "botulinal cook," namely, canned viennas, corned beef, etc. Such products are discussed in the section on canned foods as are luncheon meats and small canned

APPLICATION TO FOODS AND FOOD INGREDIENTS 219 hams that receive less than a "botulinal cook." Products whose stability is related to reduced water activity are also included, i.e., canned sausage and sliced beef (aw < 0.92) and canned prefried bacon (aw ~ 0.861. Mi- crobiological control is best accomplished through monitoring water ac- tivity rather than through the use of microbiological criteria. Finally, the routine application of microbiological criteria to vinegar pickled meats seems unnecessary, though spoilage problems, both microbial and non- microbial, may occasionally occur. Perishable Canned, Cured Meats Microbiological criteria are not generally applicable to these products. However, if there is reason to believe that a consignment may have been subjected to temperature abuse, investigative sampling may be indicated (see ICMSF, 19741. Dried Meats Microbiological criteria may appropriately be applied to the finished product as a means of assessing the adequacy of moisture control during the drying process. The APC with proper baseline data may be effective in this monitoring. Tests for pathogens such as C. per~fringens, B. cereus, S. aureus, and Salmonella may be appropriately applied, depending upon the ultimate use of the product. Such monitoring is best carried out at the processing level, though regulatory authorities may apply criteria at any point in merchandising channels. A Salmonella standard is applicable to these products. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated These considerations have been largely addressed in sections above. With respect to criteria utilized to assess the microbiological status of various raw materials, e.g., meat trimmings or the adequacy of equipment cleanup, baseline information must be established by the processor. Only on the basis of such information can useful limits be established. Where Criteria Should Be Applied See sections on sensitivity of products relative to safety and quality and on need for microbiological criteria.

220 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA References AMI (American Meat Institute) 1982 Good Manufacturing Practices. I. Voluntary Guidelines for the Production of Dry Fermented Sausage; II. Voluntary Guidelines for the Production of Semi-dry Fer- mented Sausage. Washington, D.C.: American Meat Institute. Bailey, J. W. 1974 Encyclopaedia of labeling meat and poultry products. St. Louis: Meat Plant Magazine. Barber, J. E., and R. H. Deibel 1972 Effect of pH and oxygen tension on staphylococcal growth and enterotoxin formation in fermented sausage. Appl. Microbiol. 24:891-898. Bryan, F. L. 1980 Foodborne diseases in the United States associated with meat and poultry. J. Food Prot. 43:140-150. CDC (Centers for Disease Control) 1960 Botulism. Morb. Mort. Weekly Rpt. 9:2. 1976a Salmonella Saint-Paul in precooked roasts of beef New Jersey. Morb. Mort. Weekly Rpt. 25(5):34, 39. 1976b Salmonella bovis-morbificans in precooked roasts of beef. Morb. Mort. Weekly Rpt. 25(42):333-334. 1977a Multistate outbreak of Salmonella newport transmitted by precooked roasts of beef. Morb. Mort. Weekly Rpt. 26(34)277-278. 1977b Follow-up on Salmonella organisms in precooked roast beef. Morb. Mort. Weekly Rpt. 26(38):310. 1978 Salmonellae in precooked roasts of beef-New York. Morb. Mort. Weekly Rpt. 27(24):315. 1981a Multistate outbreak of salmonellosis caused by precooked roast beef. Morb. Mort. Weekly Rpt. 30:391-392. 1981b Multiple outbreaks of salmonellosis associated with precooked roast beef-Pennsyl vania, New York, Vermont. Morb. Mort. Weekly Rpt. 30:569-570. 1982 Isolation of E. cold 0157:H7 from sporadic cases of hemorrhagic colitis United States. Morb. Mort. Weekly Rpt. 31:580, 585. 1983 Botulism and commercial pot pie California. Morb. Mort. Weekly Rpt. 32:39-40, 45. Fontaine, R. E., S. Arnon, W. T. Martin, T. M. Vernon, E. J. Gangarosa, J. J. Farmer, A. B. Moran, J. H. Silliker, and D. L. Decker 1978 Raw hamburger: An interstate common source of human salmonellosis. Am. J. Ep- idemiol. 107:36-45. Gardner, G. A. 1981 Brochothr~c thermosphacta (Microbacterium thermosphactum) in the spoilage of meats. A review. Pp. 139-173 in Psychrotrophic Microorganisms in Spoilage and Patho- genicity, T. A. Roberts, G. Hobbs, J.H.B. Christian, and N. Skovgaard, eds. New York: Academic Press. Goodfellow, S. J., and W. L. Brown 1978 Fate of Salmonella inoculated into beef for cooking. J. Food Prot. 41:598-605. Greenberg, R. A., and J. H. Silliker  1961 Evidence for thermal injury in enterococci. J. Food Sci. 26:622-625. ICMSF (International Commission on Microbiological Specifications for Foods) 1974 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. Toronto: University of Toronto Press. Pp. 143-146.

APPLICATION TO FOODS AND FOOD INGREDIENTS 221 1980 Microbial Ecology of Foods. Volume 2. Food Commodities. New York: Academic Press. Pp. 378-383. Johnston, R. W., and R. B. Tompkin 1984 Meat and meat products. In Compendium of Methods for the Microbiological Ex- amination of Foods. 2nd Ed. M. L. Speck, ed. Washington, D.C.: American Public Health Association. Mundt, J. O., and H. M. Kitchen 1951 Taint in southern country-style hams. Food Res. 16:233-238. Niven, C. F. 1956 Vinegar pickled meats. A discussion of bacterial and curing problems encountered in processing. Bulletin No. 27. Chicago: American Meat Institute Foundation. Niven, C. F., L. G. Bultner, and J. B. Evans 1954 Thermal tolerance studies on the heterofermentative lactobacilli that cause greening of cured meat products. Appl. Microbiol. 2:26. NRC (National Research Council) 1964 An Evaluation of Public Health Hazards from Microbiological Contamination of Foods. Publication No. 1195. Washington, D.C.: National Academy of Sciences. State of California (Department of Health Services) 1975 Botulism home canned figs and chicken pot pie. Calif. Morbidity No. 46. 1976 Type A botulism associated with commercial pot pie. Calif. Morbidity No. 51. USDA (U.S. Department of Agriculture) 1983 Production requirements for cooked beef, roast beef and cooked corned beef. Federal Register 48 (106):24314-24318. D. RAW (EVISCERATED, READY-TO-COOK) POULTRY Sensitivity of Products Relative to Safety and Quality The poultry industry is one of the most integrated food industries in the United States. A single organization may control an entire operation including breeder farm, hatchery, grow-out farm, processing, and retail operations. Fertile eggs from breeder farms are delivered to hatcheries where the eggs are incubated and then placed in hatchers. After hatching, chicks or poults are delivered to grow-out farms, reared until ready for slaughter, and then transported to processing plants. The microbiological condition of eviscerated, ready-to-cook poultry (chickens, turkeys, ducks, and quail) is the result of a series of conditions and events (ICMSF, 1980) including: 1. conditions at the breeding farm and hatchery 2. health of the live animal 3. feed supply 4. environmental conditions under which the flock was raised (cross- contamination by man, rodents, wild birds, litter, drinking water) 5. transportation 6. scalding

222 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA 7. picking (defeathering) 8. washing 9. evisceration (opening of body cavity, viscera pull, inspection, vis- cera removal) 10. chilling 11. packaging 12. sanitary practices and conditions in the processing plant (involving workers, equipment, utensils) 13. storage (time-temperature profile) 14. thawing, storage, and handling practices in retail stores and kitchens Microbial contamination of the egg can occur in the ovaries during the development of the egg or later as a result of penetration of the egg shell. Fumigation at the breeder farm is generally practiced to control microbial penetration of eggs. Breeding flocks are tested for the presence of Sal- monella by blood agglutination tests and subsequent culturing of internal organs. The National Poultry Improvement Plan (USDA, 1982) is the most recognized plan charged with the elimination of Salmonella pullorum and Salmonella gallinarum from breeder flocks. Chicks or poults on hatch- ing contain an extensive microbial flora. Microorganisms associated with live poultry are located primarily on the surface of the bird (skin, feet, feathers) and in the gastro-intestinal tract. The numbers and types depend largely on the environmental con- ditions under which the flock was raised. This population nearly always includes a large variety of microbial types such as Pseudomonas, Mor- axella, Acinetobacter, Micrococcus, Enterobacteriaceae, Staphylococcus, Bacillus, Clostridium, Flavobacterium, molds, and yeasts. Soil, litter, feed, and drinking water are primary sources of these microbes. Other reservoirs include insects, rodents, wild birds, and farm workers. These sources contribute not only saprophytic species but may infect or contam- inate birds with pathogens such as Salmonella, Yersinia, and Campylo- bacter. During transportation to processing plants, the incidence of ~ 1_ _ contamination Increases Decause of further distribution of microorganisms from bird to bird primarily through contact with fecal material. During the overall processing operation in modern broiler or turkey processing plants a significant reduction (90-95%) in total bacterial num- bers can be achieved by the scalding, washing, and chilling operations (Gardner and Golan, 19764. In the first, microbial destruction is effected by heat; in the latter two operations, reduction is achieved by the rinsing effect (mechanical removal) of the spray or wash. Bacterial numbers on the carcass may increase during evisceration because of extensive manual handling. This is particularly so for turkey carcasses because of handling

APPLICATION TO FOODS AND FOOD INGREDIENTS 223 associated with carcass trussing, draining, and packaging. Post-chill op- erations in broiler processing, however, are largely automated. Microbial contamination of carcasses during processing should be kept to a minimum by employing sanitary handling procedures, proper cleaning and sanitation of equipment and utensils that come in contact with car- casses, and use of adequate quantities of rinse and chill waters. Microorganisms on freshly processed carcasses are located primarily on the surface, normally at a level of 103 to 104 per cm2. They constitute a variety of species including psychrotrophic bacteria and originate from various sources such as the incoming bird (feet, feathers, intestinal tract), water, ice, and air, and are spread from carcass to carcass by processing equipment, utensils, and line workers. The shelf-life of freshly processed carcasses depends on the number and types of microorganisms present, the time and temperature of storage, and the method of packaging. Shelf-life will be short when initial numbers of psychrotrophic spoilage bacteria are high, particularly when the car- casses are stored at marginal (above 1.7°C/35°F) refrigeration tempera- tures. When counts of psychrotrophic gram-negative bacteria such as Pseudomonas, Moraxella-Acinetobacter, and Flavobacterium on chilled poultry reach 107-108 per cm2, off-odors often followed by slime formation occur. When carcasses are stored refrigerated in vacuum-packages or in modified gaseous atmospheres containing CO2, lactic acid bacteria usually become predominant. Both of these packaging procedures significantly increase the shelf-life of raw poultry compared to that observed when oxygen-permeable films are used. In vacuum packages and in modified atmospheres containing elevated levels of CO2, gram-negative aerobic psychrotrophic bacteria are inhibited. Although large numbers of psy- chrotrophic lactic acid bacteria develop during refrigerated storage, sen- sory degradation of the product is not as rapid as is observed when gram- negative aerobic psychrotrophic bacteria predominate. Subsequent handling of carcasses in food service establishments or in homes can spread microorganisms, including pathogens, that may be as- sociated with the carcass to the cooked product or to other foods through contact with knives, tables, cutting boards, and cleaning cloths. Thawing frozen turkeys and then leaving them under conditions that allow microbial growth may lead to increases in bacterial count on the carcass. Proper thawing should be done under refrigeration or in cold running water. Eviscerated, ready-to-cook poultry carcasses often contain small num- bers of pathogens such as Salmonella, Staphylococcus aureus, Clostridium perfringens, Campylobacter fetus subsp. jejuni, and Yersinia enterocoli- tica.These organisms enter processing plants with live birds and are spread to carcass surfaces during processing. At the present time there are no

224 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA commercially applicable methods employed in the United States to elim- inate pathogens from carcasses, although ionizing radiation could accom- plish this. Good sanitary processing practices, however, can reduce the prevalence and extent of spread of pathogens in a processing plant. The presence of pathogens on ready-to-cook carcasses can lead to health haz- ards if the product is mishandled in a plant, food service establishment, or in the home (Bryan, 1980; Bryan and McKinley, 19741. These faulty practices include: 1. inadequate cooking of poultry resulting in survival of pathogens such as Salmonella, C. fetus subsp. jejuni, Y. enterocolitica, and S. aureus (spores of C. perfringens may survive adequate cooking); 2. transfer of pathogens from the raw carcass via hands, cleaning cloths, equipment, and utensils to cooked poultry or to foods that will not receive further heat treatment; 3. temperature abuse of adequately cooked poultry containing surviving spores of C. perfringens or of inadequately cooked poultry contain- ing Salmonella or other pathogens (this could involve improper cool- ing, hot-holding, and reheating); 4. time-temperature abuse of cooked poultry subsequently sliced or chopped and recontaminated with S. aureus or of salads in which this recontaminated poultry is used as an ingredient. Poultry products so abused have been identified as vehicles in outbreaks of foodborne disease. From 1968-1977, poultry was responsible for 14% of foodborne disease outbreaks in which a vehicle was ascertained (Bryan, 19801. Salmonellosis accounted for 19% of these outbreaks, staphylo- coccal intoxication for 16%, C. pe~fringens enteritis for 10%, other food- borne diseases of known etiology for 2%, and diseases of unknown etiology for 53%. In summary, cooked poultry products may become a hazard when raw or cooked products are mishandled. Need for Microbiological Criteria Salmonella The proportion of Salmonella-contaminated carcasses from a processing plant is determined mainly by the incidence of infected or contaminated live birds and by the extent of subsequent spread during processing (Bryan et al., 1968a; ICMSF, 19801. Unfortunately, even with the best manu- facturing practices, cross-contamination occurs during slaughtering, dress- ing, and further processing (Bryan et al., 1968a,b). Recommendations to reduce or eradicate Salmonella in animals including poultry have been discussed in various reports (NRC, 1969; Silliker, 1982~. These include:

APPLICATION TO FOODS AND FOOD INGREDIENTS 225 1. use of Salmonella-free feed 2. control of Salmonella in breeder flocks, hatchery, and production operations 3. application of the Nurmi concept 4. irradiation of packaged raw poultry 5. education of farm workers, processing plant workers, food service personnel, and homemakers Salmonella-contaminated feeds remain a major source of Salmonella in poultry (NRC, 1969; GAO, 19741. Infected breeding stock can spread Salmonella to other farms. The spread of Salmonella agona and Salmonella hadar in the United States and in England is part of the epidemiological evidence for the involvement of these sources (Clark et al., 1973; Rowe, 1980~. Until feeds are decontaminated by heating, irradiation, or chemical means they will continue to be a major source of Salmonella. The Nurmi concept (Pivnick et al., 1981; Stersky et al., 1981) involves oral administration of the gastrointestinal flora from adult birds into newly hatched chicks and poults. Newly hatched birds may be infected by a single Salmonella; however, immediately after introduction of gastroin- testinal flora from adult birds, chicks become resistant to between 1,000 and 1,000,000 infectious doses of Salmonella. The term "competitive exclusion" has been used to describe this phenomenon. Colonization of the intestinal tract with normal gut flora apparently discourages coloni- zation with Salmonella.Two important factors may be involved: (1) pro- duction of volatile fatty acids in the caecum and (2) occupation of sites on the mucosa that Salmonella normally invade. Recently, inocula derived from pure cultures isolated from the feces of adult birds have shown the protective effect achieved through the use of fecal material or mixed cultures derived from feces (H. Pivnick, 1983. Personal communication). The use of defined pure cultures, instead of fecal material or mixed cul- tures, may lead to a more widespread application of the Nurmi concept. Irradiation of poultry with a dosage of up to 0.7 Mrad is effective in eliminating pathogens, especially Salmonella (Kampelmacher, 19831. The World Health Organization (WHO) has cleared irradiation of foods with up to 0.7 Mrad and declared these foods unconditionally safe for human consumption (FAD/IAEA/WHO, 1977J. In the Netherlands, Canada, and the USSR, clearance has been given for test marketing of chilled and frozen poultry irradiated with dosages of 0.3, 0.75, and 0.6 Mrad, re- spectively (Vas, 19771. Government, industry, and educational institutions have made sporadic attempts to educate the public about foodborne illness including salmo- nellosis. A Gallup study (Anonymous, 1979) has shown that these efforts have been ineffective. Educational efforts such as those in Denmark and

226 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Canada that are directed to the young homemaker perhaps may be more successful. Until changes are made to minimize the infection and contamination of birds on the farm and/or until a method of decontamination is routinely applied to packaged carcasses, it is questionable whether practical micro- biological criteria can be set for Salmonella on raw poultry without risk of eliminating poultry as food. A working group of the Codex Committee on Food Hygiene (FAD/WHO, 1979) concluded that application of mi- crobiological criteria for raw poultry would not improve safety. Never- theless, concern must be maintained about Salmonella being brought into home kitchens, hospitals, and food service establishments on raw poultry, and being spread via hands, equipment utensils, and cleaning cloths back to either the cooked poultry or other foods. Extensive and continuous efforts should be made to educate the food service industry and public about the contamination potential associated with the handling of raw poultry, to inform them of the need to cook poultry to temperatures that kill Salmonella, to hold cooked poultry at temperatures that preclude multiplication of these organisms, and to reheat leftovers thoroughly. Campylobacter fetus subsp. jejuni and Yersinia enterocolitica Much of what has been said about Salmonella on raw poultry is ap- plicable to C. fetus subsp. jejuni and Y. enterocolitica because these organisms also are frequently present on raw poultry products. Poultry- associated outbreaks of Campylobacter infection have been reported in recent years (Cunningham, 19821. Prevention depends on thorough cook- ing and proper storage of cooked products rather than on microbiological criteria for raw poultry products. C/tostridFium perfringens C. perfringens is a part of the intestinal flora of fowl and is shed in their feces. Furthermore, it is found in soil, dust, or feces that get on their skin, feet, and feathers. Thus, it is impractical at this time to eliminate C. perfringens from live fowl. Even strict adherence to Good Manufac- turing Practices cannot prevent some contamination of poultry carcasses with C. perfringens. Microbiological criteria for C. perfringens for raw poultry would not accomplish anything because it only becomes a problem when cooked poultry (in which the spores have become heat activated and the redox potential is reduced) is held at sufficiently high temperatures for spores to germinate and the resulting cells to multiply to large numbers.

APPLICATION TO FOODS AND FOOD INGREDIENTS 227 Staphylococcus aureus S. aureus is part of the nasopharynx and skin flora of poultry and is sometimes associated with arthritic and bruised tissue. Staphylococci from these sources are readily spread during defeathering. Therefore, some contamination of raw poultry carcasses with S. aureus can be expected. These organisms neither compete well with the microbial flora on raw carcasses nor multiply on chilled carcasses. A microbiological criterion for S. aureus on raw poultry meat is impractical because the foodborne disease problem arises when cooked poultry (frequently leftovers) is con- taminated with S. aureus by persons who handle it and then subjected to room-temperature storage or improper refrigeration. Other Bacteria Microbiological criteria involving aerobic plate counts (APCs) and in- dicator organisms have limited application for eviscerated ready-to-cook poultry carcasses. Microbiological guidelines for psychrotrophic aerobic bacteria and indicator bacteria such as coliforms on carcasses immediately post-processing may be useful to indicate that a sanitation problem exists somewhere along the processing line. Counts in excess of those normally found should alert the processor to inspect critical control points more closely to locate and remedy the problem. With proper sanitary practices along the processing line, control of fecal contamination of carcasses and washing of carcasses, coliform contamination on the freshly processed carcasses can be limited (Gardner and Golan, 19761. These guidelines, however, are not applicable to poultry in distribution channels or at the retail level as suggested by some state standards or guidelines. This is so because aerobic, psychrotrophic bacteria responsible for quality loss and ultimate spoilage of refrigerated raw poultry (packaged in oxygen-perme- able film) continue to multiply on carcasses even at recommended cold storage temperatures. Microbiological guidelines are, however, applicable to monitor the sanitary condition of equipment and utensils. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Adequate data base is available (ICMSF, 1974, 1980, 1985; APHA, 1976, 1984) to guide processors in establishing and interpreting micro- biological guidelines that they may wish to apply.

228 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Where Criteria Should Be Applied A HACCP program should be applied to the entire poultry production, processing, distnbution, and food preparation chain. Critical control points at the farm include sanitary condition of feed, drinking water, equipment, and surroundings in which the birds are raised, and control of microbial contamination by farm workers, insects, rodents, and wild birds. Although each of these control measures can affect the general microbial population, the principal focus should be on reducing the incidence of pathogens, primarily Salmonella and Campylobacter. Critical control points to be monitored at the processing plant (evis- cerated ready-to-cook carcasses only) should include (1) carcass washing, cooling, and storage procedures, which include amount of water, degree of chlorination and temperature of water, and temperature of storage; (2) cleaning and sanitation of equipment; and (3) employee sanitary practices during processing. Microbiological guidelines should be applicable to: (1) periodic evaluation of equipment surfaces to check cleaning and sanitizing procedures and to check on potential buildup of microbial contaminants, and (2) examination fAPC at 20-25°C (68-77°F) for 2-3 days and coli- forms] of freshly processed carcasses. Excessive counts should alert the processor to check the product at various stages of processing to pinpoint the problem. Reduction of the number of birds infected with Salmonella perhaps will be possible sometime in the future by application of several measures such as use of Salmonella-free feed, application of the Nurmi concept, and irradiation of packaged raw poultry. References Anonymous 1979 Report on the Scandinavian Salmonella control programs in poultry with added ob- servations from Finland, Germany and Switzerland. Health/Agrieulture/Industry Com- mittee on Salmonella, Ottawa, Canada. APHA (American Public Health Association) 1976 Compendium of Methods for the Microbiological Examination of Foods. M. L. Speck, ed. Washington, D.C.: APHA. 1984 Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed. M. L. Speck, ed. Washington, D.C.: APHA. Bryan, F. L. 1980 Foodborne diseases in the United States associated with meat and poultry. J. Food Prot. 43:140-150. Bryan, F. L., and T. W. McKinley 1974 Prevention of foodborne illness by time-temperature control of thawing, cooking, chilling and reheating turkeys in school lunch kitchens. J. Milk Food Teehnol. 37:420- 429.

APPLICATION TO FOODS AND FOOD INGREDIENTS 229 Bryan, F. L., J. C. Ayres, and A. A. Kraft 1968a Contributory sources of salmonellae on turkey products. Am. J. Epidemiol. 87:578- 591. 1968b Salmonellae associated with further-processed turkey products. Appl. Microbiol. 16:1- 9. Clark, G. M., A. F. Kaufman, E. J. Gangarosa, and M. A. Thompson 1973 Epidemiology of an international outbreak of Salmonella agona. Lancet ii:490-493. Cunningham, F. E. 1982 Microbiological aspects of poultry and poultry products An update. J. Food Prot. 45:1149-1164. FAD/IAEA/WHO 1977 Wholesomeness of irradiated food. Report of a joint FAD/IAEA/WHO Expert Com- mittee. WHO Techn. Rep. Ser. 604. FAD/WHO (Food and Agriculture Organization/World Health Organization) 1979 Report of an FAD/WHO working group on Microbiological Criteria for Foods. Ge- neva: FAD/WHO. GAO (U.S. General Accounting Office) 1974 Report to the Congress, Salmonella in raw meat and poultry: An assessment of the problem [B-164031 (2), July 22, 1974]. Washington, D.C.: U.S. General Accounting Office. Gardner, F. A., and F. A. Golan 1976 Water usage in poultry processing An effective mechanism for bacterial reduction. Pp. 338-355 in Proc. 7th Natl. Symposium on Food Processing Wastes, Atlanta, GA. Cincinnati: U.S. Envir. Prot. Agency. ICMSF (International Commission on Microbiological Specifications for Foods) 1974 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. Toronto: University of Toronto Press. 1980 Microbial Ecology of Foods. Vol. 2. Food Commodities. New York: Academic Press. 1985 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications, 2nd Ed. In preparation. Kampelmacher E. H. 9 1983 Irradiation for control of Salmonella and other pathogens in poultry and fresh meats. Food Technol. 37(4): 117-119, 169. NRC (National Research Council) 1969 An evaluation of the Salmonella problem. Committee on Salmonella. Washington, D.C.: National Academy of Sciences. Pivnick, H., B. Blanchfield, and J.-Y. D'Aoust 1981 Prevention of Salmonella infections in chicks by treatment with fecal cultures from mature chickens (Nurmi Cultures). J. Food Prot. 44:909-916. Rowe, B. 1980 Salmonella hadar England and Wales. Morb. Mort. Weekly Rpt. 29:506-508, 513. Silliker, J. H. 1982 The Salmonella problem: Current status and future direction. J. Food Prot. 45:661-  666. Stersky, A., B. Blanchfield, C. Thacker, and H. Pivnick 1981 Reduction of Salmonella excretion into drinking water following treatment of chicks with Nurmi culture. J. Food Prot. 44:917-920. USDA (U.S. Department of Agriculture) 1982 National Poultry Improvement Plan and Auxiliary Provisions. APHIS-Veterinary Ser- vices, APHIS 91-40.

230 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Vas, K. 1977 General survey of irradiated food products cleared for human consumption in different countries. Joint FAD/IAEA/ WHO advisory group on international acceptance of irradiated food. GA- 143/INF/2-IAEA, Vienna. E. PROCESSED POULTRY PRODUCTS Sensitivity of Products Relative to Safety and Quality The term "further-processed poultry product" refers to any product beyond the ready-to-cook carcass, frozen or nonfrozen. Further-processed poultry products comprise a significant part of poultry marketed in the United States. In 1982, about 60% of the 16.8 billion pounds of broilers produced in the United States were in the form of cut-up parts and other further-processed products, and of the 2.5 billion pounds of turkey con- samed, about 90% was in the form of further-processed products (USDA, 19831. Examples of these further-processed products are: Broilers: Noncooked-tray pack (chilled or frozen) Cooked wieners, bologna Turkeys: Noncooked tray-pack, tenderloin, roasts, further-pro- cessed whole birds (injected,basted) Cooked boneless breast, cured smoked breast, hams, rolls, roasts, bologna, salami, pastrami, wieners The microbiological condition of further-processed products and hence shelf-life and safety depends upon a series of conditions and events in- cluding: 1. Microbiological condition of the carcass entering further processing concern: numbers and types of microorganisms, particularly psychrotrophic bacteria and Salmonella 2. Processing procedures c. a. parts removal-removal of drum, wings, tail, and neck concern: sanitary handling b. hand deponing removal of breast tissue and thigh tissue from carcass concern: sanitary handling mechanical deponing-grinding of racks and other bones fol- lowed by deboning concerns: sanitary condition of equipment, extensive increase in surface area of tissue, temperature increase of tissue at separation, handling of mechanically de- boned poultry meat (MDPM): rate of cooling, time- temperature profile during storage

APPLICATION TO FOODS AND FOOD INGREDIENTS 231 d. trim operation trimming of major muscles, breast, and thigh concerns: sanitary handling, continuity of product flow e. blending mixing of different tissues (dark, light, trim meat), seasonings and additives (salt, sugar, phosphates, erythorbate, sodium nitrite) concerns: sanitary condition of equipment, microbial condition of additives, time-temperature profile of product, use schedule of equipment (continuity of product flow) f. emulsification emulsification of tissue fractions concerns: sanitary condition of equipment, time-temperature profile of product (heat generation), increase in tis- sue surface area, use schedule of equipment (con- tinuity of product flow) g. studding-placing product in casing such as for wieners, bo- logna, salami, pastrami, rolls, roasts concerns: sanitary condition of equipment, time-temperature profile of product (time between stuffing and cook- ing) h. product formation (forming'-to provide shape to a product (for example, boneless turkey breast) concerns: sanitary condition of equipment, sanitary handling, time-temperature profile of product i. massaging (tumbling'-to extract protein to surface to bind product tissues together when cooked (examples turkey breast, rolls) concerns: sanitary condition of equipment, time-temperature profile of product j. curing to add or inject curing ingredients to tissue for flavor and preservative action concerns: microbiological condition of curing ingredients or solution, sanitary condition of equipment k. heat treatment (cooking, smoking, canningJ to provide desir- able body and texture characteristics and for destruc- tion of microorganisms concern: time-temperature profile of product 1. chilling concern: rapid chilling of cooked product m. packaging concerns: sanitary condition of equipment (slicers for ample), sanitary practices (particularly in repacking

232 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA and weighing of cooked product), type of pack- aging film, container integrity, gaseous atmosphere n. storage offurther-processed products concerns: time-temperature profile of product, sanitary con- dition of coolers and freezers, inventory control and rotation (Note: not all processing steps are applicable to each further- processed product.) 3. Handling in food service operations and in the home a. storage conditions offrozen and refrigerated products concern: time-temperature profile of product b. thawing of frozen product concern: time-temperature profile of product c. cross-contam~nation concern: contamination from raw to cooked product via cut- ting boards, handling, etc., and from raw product to food that receives no further heat treatment d. cooking of product concern: time-temperature profile of product e. cooling of cooked product concerns: rate of cooling, time-temperature profile of product I. storage of cooked product concern: time-temperature profile of product g. reheating of product concern: time-temperature profile of product Quantitatively, there will be differences in the degree of "concern" depending upon the type of further-processed product. The significance of individual concerns associated with various processing, storage, and preparation practices for individual further-processed products are dealt with on the following pages. No attempt will be made to evaluate the sensitivity of all of the above-listed types of further-processed poultry products relative to safety, shelf-life, and microbiological criteria. Instead, these issues and the need for microbiological criteria will be discussed for a few typical further-processed poultry products as examples of the various factors that should be considered: 1. tray-pack product (chilled or frozen): cut-up broilers (breast, wish- bone, wings, drum, thigh, back, neck); turkeys (drum, tail, wings, neck) 2. mechanically deboned poultry meat product: turkey wieners 3. trim meat product: sliced turkey ham 4. muscle product: boneless turkey breast

APPLICATION TO FOODS AND FOOD INGREDIENTS 5. canned product: canned chicken 6. poultry pot pies 233 To evaluate shelf-life and safety of other types of products, it would be necessary to conduct a hazard analysis and establish and monitor critical control points for these products. Tray-Pack Product The shelf-life of a tray-pack product stored at refrigeration temperatures depends upon the extent of microbial activities, primarily those of psy- chrotrophic spoilage bacteria such as Pseudomonas spp. Critical control points include the microbiological condition of the carcasses from which the parts were derived, sanitary handling of the product during the cut- up operation, sanitary condition of equipment, and the time-temperature profile of the product during processing and storage. Chill-packs should be held at the plant at-2.8°C (27°F), a temperature not conducive for rapid growth of psychrotrophic spoilage bacteria. At the retail level some- what higher storage temperatures (O to 4.5°C/ 32 to 40°F) can be expected, resulting in more rapid growth of psychrotrophic bacteria. As the aerobic plate count (APC) of raw poultry parts approaches 107 to 108 viable cells per cm2, off-odors often become noticeable with subsequent slime for- mation at somewhat higher APC. Safety of tray-pack products as well as that of ready-to-cook carcasses (see Raw Poultry section) can best be assured by adhering to the following practices in food service operations and in the home: 1. adequate cooking 2. proper hot and cold storage of cooked product 3. avoidance of cross-contamination of cooked food from raw food and from contaminated equipment and utensils 4. sanitary handling of raw and cooked product 5. proper cooling of cooked product 6. adequate reheating of cooked chilled product Bryan (1980) and Bryan and McKinley (1974) have shown that failure to observe these practices commonly leads to outbreaks of meat- and poultry- borne diseases. Clostridium perfringens enteritis, staphylococcal intoxi- cation, and salmonellosis are the more common foodborne diseases as- sociated with poultry. Low levels of C. perfringens are common on raw poultry. When poultry is cooked, some C. perfringens spores survive. As the temperature of the hot cooked product reaches 50°C (122°F) during holding at room temperature or on a warming device or during storage in

234 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA a refrigerator (when product is stored in thick layers), spore germination and multiplication of vegetative cells begin. Multiplication is favored by the absence of competitive bacteria and the low Eh of the cooked product. If there is enough time during warm holding or slow cooling, cells may reach numbers that can cause illness. If such foods are consumed without adequate reheating to kill vegetative cells, illness may occur. Contamination of cooked poultry with Staphylococcus aureus usually occurs through handling by humans. If such product is left without re- frigeration for several hours or cools slowly in refrigerators (when stored in containers in thick layers) growth of S. aureus and enterotoxin formation may occur. Growth of S. aureus in a cooked product is favored by lack of competitive bacteria, which are destroyed by heat. Small numbers of Salmonella may often be present on raw poultry. Salmonella on cooked poultry result either from inadequate cooking or from recontamination of adequately cooked poultry by contact with con- taminated hands, equipment, and utensils. Perishable Cooked Further-Processed Products: Wieners, Ham, Boneless Breast A flow diagram of basic processing procedures utilized in the production of turkey wieners, sliced turkey ham, and turkey boneless breast is given below: Wieners: (turkey, water, salt, sugar, phosphates, erythorbate, nitrite and flavoring) Mechanically deboned turkey meat > packaged frozen flaking emulsification > blending emulsification stuffing cooking chilling peeling packaging storage. Sliced turkey ham: (turkey, water, salt, sugar, phosphates, erythor- bate, nitrite, flavoring) Hand-deboned thigh meats trimming- > blending stuffing cook- ing~ chillings peeling slicing packaging storage. Boneless breast: (turkey, salt, sugar, phosphates) Hand-deboned breast skin and breast muscles > trimming > tum- bling~ forming cooking chilling ~ repackaging storage. Factors that influence the microbiological condition of further-processed poultry products listed in the beginning of this section apply to turkey wieners, sliced ham, and boneless breast. The importance of: using car- casses with low numbers of spoilage bacteria for further-processed prod- ucts is obvious. Unless sanitary procedures are used, hand-boning and trimming can contaminate products. Mechanical deboning systems offer

APPLICATION TO FOODS AND FOOD INGREDIENTS 235 opportunities for increases in bacterial counts. This is possibly an effect of increases in product temperature and increased nutrient availability due to increased surface area of tissue. Many pieces of equipment are used in the manufacture of further- processed products. To keep increases in bacterial numbers to a minimum, it is advisable to design a cleaning and sanitizing program to fit the production operations of a plant. Equipment and utensils should be pe- riodically cleaned and sanitized thoroughly. Product accumulation on equipment coupled with infrequent cleanup operations should be avoided because they result in substantial increases in bacterial count of the prod- uct. Extensive interruptions in product flow increase opportunities for microbial growth in the product. Proper cooking destroys most vegetative cells, including pathogens such as Salmonella, Yersinia, Campylobacter, and S. aureus. Cooked poultry must reach an internal temperature of at least 71°C (160°F); cured and smoked poultry at least 68°C (155°F) (Bailey, 19741. Surviving bacteria are primarily sporeformers. APCs of these products immediately post- cook usually are low, but recontamination may occur through contact with hands (packaging) and contaminated equipment such as in peeling, slicing, and packaging operations. Presence of pathogens such as Salmonella, Campylobacter, Yersinia, and S. aureus on a fully cooked perishable product indicates a lack of sanitary processing practices in post-cook operations. The hazard associated with eating a product recontaminated with Sal- monella depends upon the method of food preparation. Cooked poultry products such as rolls and roasts are not always reheated before eating. Salmonellosis has been traced to recontaminated precooked turkey roasts and rolls (Bryan, 19801. Wieners, sliced ham, and luncheon meats such as bologna, salami, and pastrami prepared from poultry meat are perishable cooked cured meats and are subject to shelf-life problems similar to those described for prod- ucts produced from beef and pork (see Chapter 9, part C). Cooked cured poultry products such as wieners, ham, bologna, and a wide variety of luncheon meats have a relatively long shelf-life [about 30 days at 1.7 to 4.4°C (35 to 40°F)~. Cooking [internal temperature of at least 68°C (155°F)] destroys most microorganisms except for some thermoduric bacteria and spores. Salt and nitrite have an inhibitory effect on some of the survivors and contaminants. During prolonged refrigerated storage in oxygen-perme- able films, micrococci, enterococci, lactic acid bacteria, and yeasts may develop. When packaged in oxygen-impermeable films (vacuum-pack- aged) lactic acid bacteria become dominant and eventually cause shelf- life problems.

236 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Canned Poultry Products For a discussion on the sensitivity of canned poultry products relative to shelf-life and safety see the sections in this chapter on canned foods. Poultry Potpies Commercially prepared and homemade potpies contain poultry meat, vegetables, gravy, and seasoning materials. They are cooked during pro- cessing and the contents are protected from subsequent recontamination by the crust or shell. Proper cooking destroys most vegetative cells, in- cluding Salmonella, S. aureus, Campylobacter, and Yersinia, but some spores can survive. If these products are subjected to temperature abuse, spoilage usually occurs. More serious, however, is that under such con- ditions spores of C. botulinum, which can survive cooking, germinate, and multiply. Mishandled poultry potpies have been implicated as a vehicle in family outbreaks of botulism (CDC, 19761. Temperature abuse of these products also presents a potential risk for C. perfringens enteritis. Need for Microbiological Criteria Most critical control points related to further-processed poultry products can be monitored by: (1) checks of time-temperature profile of product during processing (checking temperature during cooking, cooling, and storage), (2) evaluation of sanitary condition of processing equipment and facilities, and (3) evaluation of sanitary practices of employees. Microbiological criteria for pathogens in raw poultry would accomplish little because (1) low numbers of pathogens such as Salmonella, Cam- pylobacter, Yersinia, C. perfringens, and S. aureus are often present on raw poultry and for the most part are unavoidable under present conditions of handling, and (2) foodborne disease problems arise when cooked prod- ucts are mishandled in processing plants, food service operations, or in the home. A working group of the Codex Committee on Food Hygiene (FAD/WHO, 1979) concluded that application of microbiological criteria for raw poultry would not improve safety (see Part D, above). Microbiological guidelines (APC) are useful to check the condition of incoming carcasses to be used for further processing, particularly if they were obtained from other sources. Relatively simple microbiological pro- cedures (rinse, swab, direct agar contact) are available to evaluate the sanitary condition of equipment and utensils after each cleaning and san- itizing schedule and to monitor buildup of microorganisms on equipment during the processing day (APHA, 1984~. The results of counts on in ---c,

APPLICATION TO FOODS AND FOOD INGREDIENTS 237 coming carcasses as well as those on equipment and utensils are of ret- rospective value. They are useful because counts on carcasses can identify suppliers that provide products with excessively high counts. High counts on equipment and utensils would alert the processor to intensify cleanup operations or review cleaning procedures. Microbiological guidelines for products after a critical control point or for finished products can be useful (1) to check conditions along the processing line and (2) to meet "guar- anteed shelf life" of product or purchaser specifications. In a cooked product the guideline should include APC (to evaluate general condition along the processing line), S. aureus (to identify lack of hygienic practices and potential temperature abuse), Salmonella (post-heat cross-contami- nation), and coliforms (post-heat contamination). For raw finished prod- ucts, microbiological guidelines (APC) may be helpful to evaluate process control. If counts exceed guidelines that can be met when operating under good processing practices, then the processor will be alerted to check critical control points to locate and remedy the problem. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Although several studies have reported on the microbiological condition of further-processed poultry products (Bryan et al., 1968a; Mercuri et al., 1970; Zottola and Busta, 1971; Robach et al., 1980), few published reports (Bryan et al., 1968b; Denton and Gardner, 1982) are available that char- acterize the effect of individual processing practices on the microbial flora of these products. Information in reports by ICMSF (1974, 1980, 1985) and APHA (1976, 1984) should provide useful guidelines in evaluating the microbiological condition of further-processed poultry. Where Criteria Should Be Applied A hazard analysis critical control point program should be applied for these products at the plant level. Various procedures including micro- biological guidelines to monitor these control points have been discussed earlier in this section. Mishandling of poultry occurs in food service operations and in homes. Control of this problem should be through mon- itoring of critical control points, which include thawing, cooking, hot- holding, handling after cooking, chilling, and reheating. The HACCP concept described for food service operations should also be applied to the handling of foods in the home. This approach requires education of all persons in food handling procedures with particular emphasis at the grade and high school levels.

238 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA References APHA (American Public Health Association) 1976 Compendium of Methods for the Microbiological Examination of Foods. M. L. Speck, ed. Washington, D.C.: APHA. 1984 Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed., M. L. Speck, ed. Washington, D.C.: APHA. Bryan, F. 1974 Bailey, J. W. 1974 Encyclopedia of Labeling Meat and Poultry Products. 2nd Ed. St. Louis: Meat Plant Magazine. Bryan, F. L. 1980 Foodborne diseases in the United States associated with meat and poultry. J. Food Prot. 43:140-150. L., and T. W. McKinley Prevention of foodborne illness by time-temperature control of thawing, cooking, chilling and reheating turkeys in school lunch kitchens. J. Milk Food Technol. 37:420- 429. Bryan, F. L., J. C. Ayres, and A. A. Kraft 1968a Destruction of salmonellae and indicator organisms during thermal processing of turkey rolls. Poultry Sci., 47: 1966- 1978. 1968b Salmonellae associated with further processed turkey products. Appl. Microbiol. 16: 1-9. CDC (Center for Disease Control) 1976 Foodborne and Waterborne Disease Outbreaks. Annual Summary 1975. Atlanta: Cen- ter For Disease Control. Denton, J. H., and F. A. Gardner 1982 Effect of further processing systems on selected microbiological attributes of turkey meat products. J. Food Sci. 47:214-217. ICMSF (International Commission on Microbiological Specifications for Foods) 1974 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. Toronto: University of Toronto Press. 1980 Microbial Ecology of Foods. 2. Food Commodities. New York: Academic Press. 1985 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. In preparation. Mercuri, A. J., G. J. Banwart, J. A. Kinner, and A. R. Sessoms  1970 Bacteriological examination of commercial precooked Eastern-type turkey rolls. Appl. Microbiol. 19:768-771. Robach, M. C., E. C. To, S. Meydav, and C. F. Cook 1980 Effect of sorbates on microbiological growth in cooked turkey products. J. Food Sci. 45:638-640. USDA (U.S. Department of Agriculture) 1983 Poultry-Production, Disposition & Income 1981-1982. Statistical Reporting Service. Pou-2-3, April 83. Zottola, E. A., and F. F. Busta 1971 Microbiological quality of further-processed turkey products. J. Food Sci. 36:1001- 1004.

APPLICATION TO FOODS AND FOOD INGREDIENTS F. EGGS AND EGG PRODUCTS Sensitivity of Products Relative to Safety and Quality 239 Eggs and egg products are regulated by the USDA. the legislative base being the Egg Products Inspection Act of 1970 (U.S. Congress, 19701. The term "egg" means the shell egg of domesticated chicken, turkey, duck, goose, or guinea. The term "egg product" means any dried, frozen, or liquid eggs with or without added ingredients, excepting products that contain eggs only in a relatively small proportion. Processors of egg products are subject to continual USDA inspection. Though shell eggs are covered under the Egg Products Inspection Act, their processing is not under continual inspection. Historically, consumption of eggs and egg products has been an im- portant cause of human salmonellosis. Between 1963 and 1975 there were 651 outbreaks of human salmonellosis reported to the Center for Disease Control, and the vehicle of transmission was identified in 463 (71%) of these occurrences. Of these, poultry accounted for 99 (21%), meat for 69 (15%), and eggs for 53 (11%~. These data, reviewed by Cohen and Blake (1977), show that poultry, meat, and eggs were vehicles in ap- proximately half of the outbreaks of human salmonellosis during this 13- year period. The percentage of outbreaks caused by eggs decreased dra- matically after 1966 (see Figure 9-11. Of 10 egg-related outbreaks oc- curring between 1968 and 1975, 9 were associated with shell eggs and only 1 withOegg products. The vastly improved safety position of eggs and egg products has been maintained since 1975 through to the present time. This may be traced to consumer education and to improvements made by egg producers and processors. Of great importance was enforce- ment of the Egg Products Inspection Act of 1970, which placed egg products, as well as shell eggs, under supervision by USDA. The Egg Products Inspection Act, as of July 1972, eliminated human consumption of high-risk shell eggs, namely, checks, dirties, incubator rejects, and leakers. Furthermore, the Egg Products Inspection Act required that all egg products be pasteurized to render them free of salmonellae. SheR Eggs The interior contents of sound shell eggs are virtually free of micro- organisms, e.g., < 10/g (Bergquist et al., 1984~. However, shell eggs are subject to spoilage if microbial invasion occurs. This is apt to occur with "leakers," eggs of which the shell and shell membranes are broken.

240 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA 40r 30 25 Z20 15 10 5 o A Poultry . . . . . -.1? 1. 1. I · - I ·e I ·e I I _ /'~' / ; 1969 Meat ~ i\ V · ~ , ·. Egg \ 1971 1973 1975  1963 1965 1967 YEAR FIGURE 9-1 Percentage of salmonellosis outbreaks (outbreaks caused by uniden- tified vehicle are eliminated) caused by poultry, meat, or eggs. United States, 1963- 1975. SOURCE: Cohen and Blake, 1977. The egg contents may become contaminated with microorganisms if im- proper washing or storage conditions are used. For example, if the wash solution temperature is lower than the temperature of the egg, the wash water may be drawn into the egg through shell pores when the contents contract. Despite the perishability of shell eggs, microbiological criteria are of little value. Before marketing, shell eggs go through a mechanized process of sorting, washing, drying, candling, weighing, and packaging. The procedures, particularly candling, are depended upon to eliminate shell eggs of poor quality from entering commercial channels. At the present time, egg-associated foodborne disease outbreaks almost always involve consumer misuse of shell eggs, viz. their use as ingredients in foods under conditions where the shell egg contents are not subjected to pasteurizing temperatures as components of eclairs, meringues, and ice cream. In such circumstances, contamination by bacteria residing on the surface of the shell is carried to the "broken out egg" and thereafter into the food that is prepared without cooking. Compounding the problem is the frequent exposure of broken out shell egg contents to temperatures permitting microbial growth prior to their use as ingredients. To the degree that eggs

APPLICATION TO FOODS AND FOOD INGREDIENTS 241 continue to be a source of human salmonellosis, this is the most significant route of human infection. Clearly, microbiological criteria can play no significant role in the control of shell egg quality. Both safety and general microbiological qual- ity are dependent upon sorting (elimination of defective eggs from com- merce) and candling, as well as proper washing and storage conditions. Egg Products ~ ,. _ _ ~ Egg products present quite different microbiological problems from those of shell eggs. The contents of shell eggs are subject to contamination during the breaking operation and subsequently during further handling of the broken out egg contents. A single highly contaminated egg can c.ont~minnte ~ large volume of fluid egg with spoilage and/or pathogenic microorganisms. If the liquid egg is to be dried, then it must first be desugared. Microbial growth usually occurs during the desugaring process, even when this is accomplished by enzyme treatment. Whether the liquid eggs are frozen or dried, they must be pasteurized. Egg albumen cannot withstand as high temperatures as can whole eggs and yolks. As a con- sequence, spray-dried albumen must be held at a temperature of not less than 54.4°C (130°F) for not less than 7 days and until it is negative for Salmonella. Pan-dried albumen must be held at a temperature not less than 51.7°C (125°F) continuously for not less than 5 days and until it is negative for Salmonella (USDA, 19801. The risk of postprocessing contamination exists with both liquid and dried egg products. Finally, frozen eggs must be thawed before use, and if this is not done properly, microbial growth may occur. Similarly, dried eggs must be reconstituted. If the thawed frozen eggs or reconstituted dried eggs are not held under proper conditions (time and temperature), microbial growth may lead to spoilage and/or the growth of Salmonella. If egg products are contaminated with salmonellae, there is the added hazard of cross-contamination to other products in the food-processing or kitchen environments. The only pathogen of concern is Salmonella. Federal law requires that to assure that adequate pasteurization has occurred, pasteurized egg prod- ucts and heat-treated dried egg whites shall be sampled and analyzed for the presence of Salmonella (USDA, 1980~. Thus, a standard requiring that egg products be free of Salmonella exists. Other than a direct microscopic standard, no standards relating to spoil- age or indicator organisms exist. Microbiological criteria may be profitably used by the processor and user of egg products. APCs may be used to  assess the microbiological quality of raw materials and processing con

242 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA ditions. Pasteurized products may be tested for coliforms to detect post- pasteurization contamination. Many processors of egg products use microbiological criteria to monitor critical control points, and purchase specifications commonly include criteria with limits for aerobic plate count and coliforms, as well as for salmonellae. The Salmonella Standard As indicated above, federal law has established a standard requiring that egg products, except unpasteurized salted egg products used in acidic dressings, be free of Salmonella. The USDA is responsible for adminis- tration of this standard. Details of this activity are contained in AMS-PY instruction number 910, Egg Products-1 (USDA, 1975~. This document recognizes three different types of sampling and testing, namely certifi- cation, surveillance, and confirmation. Certification means certifying to the quality of a specific lot. Samples are drawn by the USDA inspector and submitted to a USDA laboratory. An attribute sampling plan is prescribed. Samples are randomly drawn from the final shipping containers. The number of containers sampled is determined by the number of containers in the lot. For example, with liquid and frozen egg products, the sampling plan below is used. No. of shipping No. of shipping containers in containers to select Quantity to be the lot for sampling analyzed 1,200 or less 4 1 x 100 g 1,201-3,200 8 2x 100 g Over3,200 16 3x100g For dried egg products the following sampling plan is prescribed. No. of shipping containers in the lot No. of shipping containers to select for sampling Quantity to be analyzed 50 or less 4 1 x 100 g 51-150 8 2x100g 151-500 12 3 x 100 g 501-1,500 16 4x 100 g Overl,500 20 Sx100g  In applying this plan, the units randomly selected for sampling are segregated into groups of four. Two to three ounces of product are with- drawn from each unit in the group, and these samples are combined in a single container, mixed, and a 100-g composite sample is taken and ana- lyzed for Salmonella. With reference to dried products, then, a single 100-g sample is analyzed for lots comprised of 50 or fewer containers;

APPLICATION TO FOODS AND FOOD INGREDIENTS 243 whereas 5 x 100-g composites are analyzed for lots comprised of over 1,500 containers. Thus, the stringency of the sampling plan is related to the number of containers within a lot. If one considers a lot to be comprised of an infinite number of 25-g units, then for dried egg products, the plan applied varies from n = 4 to n = 20. For liquid and frozen egg products, the plan varies from n = 4 to n = 16. As indicated in Chapter 6, lot size has little effect on the probability of acceptance of large lots. Selecting the number of sample units as a percentage of lot size really serves no purpose. In the above case, however, lot size not only influences the number of units sampled but also the amount of sample analyzed. Thus, the USDA approach to certification samples encompasses variation in the stringency of the sampling plan with lot size, and furthermore the plan utilized may vary from one that is extremely loose to one that has a moderate degree of stringency. Surveillance as applied to the Egg Products Inspection Act means "the sampling of pasteurized products on a statistical basis and analyzing for the presence of Salmonella by any laboratory using the USDA method of analysis" (AMS-PY instruction no. 910) (USDA, 19754. Plants without an established history of producing products free of Salmonella are re- quired to start sampling 100% of the lots in the order produced. As indicated in Figure 9-2, if 83 consecutive lots are found negative for Salmonella, the plan permits inspection of every other lot. If 83 inspected lots at this level are found negative for Salmonella then the frequency is shifted to one lot in four (level 21. If 83 inspected lots are found negative, then the sampling rate becomes one in eight lots (level 31. As further indicated in Figure 9-2, the finding of a positive lot "triggers" an in- creased frequency of sampling. The sampling plan initially applies to each day's production, each category of product, or each product. When a product is found to be positive, that product or category of product has a separate sampling plan until a satisfactory history of analyses is estab- lished. Any reduction or tightening of frequency of sampling applies separately to each product or category and not to cumulative results. It is of interest to note that the USDA instruction (AMS-PY instruction no. 910) (USDA, 1975) states that no rigid or set sampling pattern is to be followed. It suggests alternation between frozen and liquid product and between first and last product packed. For surveillance sampling a single 100-g frozen or liquid product is subjected to analysis. For dried samples three separate 25-g samples are analyzed (H. Maguire, 1982. Personal communication). Thus, again considering a lot to be composed of an infinite number of 25-g units, for liquid products n = 4 and for dried products n = 3. Confirmation is "sampling and analyses, at government expense and direction, to verify the accuracy of the company's surveillance program

244 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA 100% Level Inspect 100% of the lots in the order produced If 83 consecutive lots are found to be free of Sal monel la 1 ~ ~ Sampling Rate Level No. 1 inspect at rate f = 1/2 If 83 inspected lots are found to be free of Salmonella State 1 Resume inspec- tion at rate f= 1/2 If 79 inspected lots are found to be free of Salmonella Sampling Rate Level No. 2 Inspect at rate f2= 1/4 If 83 inspected lots are found to be free of Salmonella State 2 Resume inspec tion at rate f2= 1/4 I f 79 i nspected I ots , ~are found to be free of Sal monel la I Sampl i ng Rate Level No. 3 e inspect at  rate f 3 = 1 /8 State 3 Resume inspec l tion at rate f3 = 1/8 If 79 inspected lots are found to be free of Salmonella If a lot is found with Salmonella while sampling every other one If the 4 inspected lots are f ound to be free of Salmonella If a lot with Salmonella is found after resuming inspection at rate f = 1/2 If a lot is found with Salmonella while sampling every fourth lot If the 4 inspected lots are found to be free of Sal monel la If a lot is found with Salmonella after resuming inspection at rate f2= 1/4 If a lot is found with Salmonella while sampling every eighth lot If the 4 inspected lots are found to be free of Salmonella If a lot with Salmonella is found among the 4 lots inspected, shift to 100% inspection State 1 R Inspect next 4 lots following ~ . the one found with Salmonella T  If a lot with Salmonella is found among the 4 lots inspected State 2R Inspect next 4 lots following the one found with Salmonella If a lot with Salmonella is found among the 4 lots inspected State 3R Inspect next 4 lots following the one found with Salmonella If a lot is found with Salmonella after resuming inspection at rate f 3 = 1 /8 FIGURE 9-2 Flow process chart for multilevel continuous sampling plans (i = 83, f= '/2, 6= t/4, [3= tail). SOURCE: USDA, 1975 (From AMS-PY Instruction No. 910 [Egg Products] 1; Exhibit 8.) and analyses" (USDA, 19751. The USDA inspector draws confirmation samples from pasteurized liquid, frozen, and dried egg products in the final type of package in which the product is to be shipped. The inspector does not draw confirmation samples when the plant submits all their surveillance samples to a USDA laboratory. A 4-oz. sample is submitted for analysis, and the USDA laboratory analyzes 100 grams (n = 41. The confirmation sampling program is a "spot check" program for frozen and o

APPLICATION TO FOODS AND FOOD INGREDIENTS 245 liquid products; the frequency of sampling is related to the surveillance sampling rate. For example, if a plant is on a 100% or a one in two surveillance sampling program for frozen or liquid products, then one confirmation sample per week is submitted for analysis. For a surveillance sampling program of one in four, two confirmation samples per month are submitted. For a surveillance level of one in eight, one confirmation sample per month is analyzed. For dried products the frequency of con- firmation samples is one per week for yellow products (whole egg or yolk), two per month for spray dried egg whites, and one per month for pan-dried egg whites. The USDA provides for retesting of lots found positive for Salmonella (USDA, 19771. Plants with known Salmonella problems are not permitted to resample. When permission is granted to resample, product is divided into sublets of not more than 100 containers or not more than 3,000 pounds. The number of containers to be drilled for a sample is the square root of each sublet. A 100-g sample is analyzed from each sublet, and each sublot is analyzed separately. The product may be released when the retest results of all the sublets are negative for Salmonella. If the retest result of any one sublot is positive for Salmonella, the entire original lot or day's production for the product involved is repasteurized and retested. Probabilities Associated with the "Salmonella Standard" The National Research Council (NRC) Committee on Salmonella ca- tegorized foods according to degree of risk to the consumer (NRC, 19691. Egg products, most conservatively, would be placed in Category III as they have historically been considered potential sources of salmonellae in finished products. Furthermore, during their use egg products may be exposed to conditions permitting the growth of salmonellae. Since egg products are subjected to a pasteurizing step capable of destroying sal- monellae, they are subject, in theory, to only two of the three hazard potentials delineated in the NRC report. One might argue that pasteurized egg products are subject to post-pasteurization contamination or, on the other hand, that they are consumed by high-risk populations. Accordingly, a case may be made for placing them in a higher hazard category, namely I or II. Thus, these products could be placed in Category I, II, or III, depending upon one's point of view. According to the NRC report, Cat- egory III products are to be analyzed by a sampling scheme involving the analysis of thirteen 25-g samples, all of which must be negative. Given these results, there is a 95% probability of one organism or less in 125 g of product. The USDA surveillance sampling plan for dried eggs, based

246 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA upon the analyses of three 25-g samples per lot, falls far short of NRC recommendations. Indeed, based upon the assumption of random distri- bution of salmonellae within a particular lot, the probability of accepting a lot with an average Salmonella contamination of one organism per 125 grams is 55% (see Figure 9-31. Thus a lot of the quality that should be rejected according to the NRC Category III sampling plan would be ac- cepted by the USDA sampling plan 55% of the time. And this, in fact, is the USDA surveillance sampling plan that is used to determine how frequently a given processor must conduct Salmonella tests on his prod- ucts. The statistics are only slightly more favorable using the sampling plan prescribed for liquid or frozen eggs wherein a single 100-g sample is analyzed (n = 4 instead of n = 31. With respect to certification samples, the USDA program similarly falls short of the NRC recommendations. Here, the number of samples analyzed (n) is determined by the number of shipping containers within a lot. The virtue of this approach, of course, lies in the fact that the larger the number of units, the more individuals at risk, ignoring the unit weight of the containers themselves. But for the smallest lot (50 containers or less), n = 4, thus employing a consumer's risk far greater than contemplated by the NRC committee. For the largest lot (greater than 1,500 shipping containers of dried eggs), n = 20, the sampling plan is somewhere be- tween a Category II and Category III, according to the NRC report. For liquid and frozen eggs, the sampling plan for the largest lot involves the analyses of three 100-g composite samples (n = 12), which falls short of the stringency recommended by the NRC. In terms of assessing the microbiological safety of individual lots~of~rdd~ it is difficult to justify the stringency of varying sampling plans depending upon lot size (see Chapter 6~. Likewise, the USDA confirmation sampling program falls far short of NRC recommendations. A single 100-g sample (n = 4) is analyzed. Since the USDA surveillance sampling program is based upon somewhat unsound statistical premises, it may be argued that the total surveillance program is in question. The credibility of the USDA certification, sur- veillance and confirmation sampling plans could be significantly improved simply by increasing the weight of the samples analyzed in connection with this program. If one classifies these products as Category III, then the analysis of a single composite 325-g sample (n = 13) would give 95% probability of 1 Salmonella or less in 125 g of product. If one were to consider these products in Category II, then the analysis of a total of 725 g of product would give a 95% probability of 1 organism or less in 250 g. The present USDA sampling schemes fall far short of those rec

APPLICATION TO FOODS AND FOOD INGREDIENTS 1 .00 0.90 0.80 O 0.70 0.60 cat UJ c' ~ 0.50 IL o 0.40 m m O 0.30 0.20 0.10 O n = 3 x = 0  1 - 1 1 1 1 1 0 2 4 6 8 10 12 AVERAGE NUMBER OF SALMONELLAE PER TEN 25-g SAMPLES 247 FIGURE 9-3 Operating characteristic curve homogeneous distribution of sal- monellae contamination. SOURCE: Recommendations by the Subcommittee on Sampling and Methodology for Salmo- nellae in Eggs, Egg Products and Prepared Mixes, Inter-Industry Committee on Salmonellae in Foods (Anonymous, 1966), p. 124. Reprinted from Food Technology. Copyright A) 1966 by Institute of Food Technologists.

248 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA ommended by the NRC and reliance upon them carries far too great a consumer risk. Fortunately there has been only one outbreak of human salmonellosis traced to egg products since the Egg Products Inspection Act (CDC, 1977a,b). This outbreak was caused by dried egg albumen, contaminated with Salmonella infantis. The product was an ingredient of a "Precision Isotonic Diet," a formula for oral or tube feedings used in hospitalized individuals of all ages. One cannot, however, attribute the good recent public health record of egg products to the effectiveness of the USDA sampling programs. The fact is that producers of these products have employed far more stringent sampling plans than those prescribed by the USDA. Furthermore, users of egg products have, through their purchase specifications, required that products be analyzed according to the sam- pling plans recommended by the NRC Committee on Salmonella. The dramatic decrease in egg-associated outbreaks attests to the success of mandatory pasteurization of egg products, as well as the Egg Products Inspection Program of USDA. The subcommittee recommends the con- tinuation of the USDA program, with the intensification of the Salmonella sampling and testing indicated in this chapter. IS Anonymous 1966 Recommendations for sampling and laboratory analysis of eggs, egg products and prepared mixes. A report. Food Technol. 20(4):121-129. Bergquist, D., A. Kraft, O. Cotterill, and H. Maguire 1984 Eggs and Egg Products. In Compendium of Methods for the Microbio- logical Examination of Foods, M. L. Speck, ed. Washington, D.C.: APHA. CDC (Center for Disease Control) 1977a Salmonella infantis California, Colorado. Morb. Mort. Weekly Rpt. 26:41. 1977b Follow-up on Salmonella infantis United States. Morb. Mort. Weekly Rpt. 26:84. Cohen, M. L., and P. A. Blake 1977 Trends in foodborne salmonellosis outbreaks: 1963-1975. J. Food Protect. 40:798- 800. NRC (National Research Council) 1969 An Evaluation of the Salmonella Problem. Committee on Salmonella. Washington, D.C.: National Academy of Sciences. U.S. Congress 1970 Eggs Products Inspection Act. P.L. 95-597 (H. R. 19888). Dec. 29. USDA (U.S. Department of Agriculture) 1975 Section 8 Sampling for bacteriological, chemical and physical testing. AMS-PY Instruction No. 910 (Egg Products) 1. Implementation of Egg Products Inspection Act. Washington, D.C.: Agricultural Marketing Service, USDA. 1977 Implementation of Egg Products Inspection Act. Section 9-Handling Salmonella Positive Samples. Revised. Washington, D.C.: Agricultural Marketing Service.  1980 Regulations governing the inspection of eggs and egg products. Code of Federal Regulations 7 CFR Part 2859.

APPLICATION TO FOODS AND FOOD INGREDIENTS G. FISH, MOLLUSCS, AND CRUSTACEANS Sensitivity of Products Relative to Safety and Quality Safer 249 Fish, molluscs, and crustaceans can acquire pathogenic microorganisms or toxins from the natural aquatic environment, from sewage-contaminated harvesting areas, and from contamination by workers, utensils, and equip- ment during harvesting, processing, distribution, and food preparation. Even freshly caught fish and shellfish from unpolluted waters may contain pathogens such as Clostridium botulinum type E and Vibrio para- haemolyticus. For many of the fish-borne and shellfish-borne diseases, faulty harvesting or postharvesting practices are necessary to cause an outbreak of disease. Included in this category are outbreaks caused by C. botulinum, V. parahaemolyticus, Staphylococcus aureus, hepatitis A, scombroid poisoning, Salmonella, Shigella, and Clostridium perfringens. For botulism, staphylococcal intoxication, or C. perfringens enteritis time- temperature abuse of a heated seafood usually is involved; for scombroid poisoning time-temperature abuse of the raw fish usually is involved. For V. parahaemolyticus infection time-temperature abuse of raw-contami- nated or heated-recontaminated seafood is necessary. Salmonella, Shi- gella, and hepatitis A may enter seafood either from contaminated waters, from contamination post-harvesting, or from contamination post-heating and cause disease in humans. Fish and shellfish that have ingested certain types of toxic dinoflagellates can be directly responsible for disease out- breaks. Paralytic shellfish poisoning (PSP) and ciguatera belong to this category. From 1970-1978 fish, molluscs, and crustaceans were responsible for 7.4, 1.9, and 1.4%, respectively, of the foodborne disease outbreaks in the United States (Bryan, 19801. The more prominent disease outbreaks associated with these products were: for fish: ciguatera, scombroid poi- soning, and botulism; for molluscs: paralytic shellfish poisoning, hepatitis A, and V. parahaemolyticus; and for crustaceans: V. parahaemolyticus and staphylococcal intoxication. The agents responsible for these diseases are discussed in detail in Chapter 4. They are mentioned here relative to the use of and need for microbiological criteria as related to safety of fishery products. Hepatitis A has been associated with eating raw or undercooked shell- fish harvested from estuaries polluted with sewage. In the United States sanitary control of the shellfish industry is primarily based upon classi- fication and control of the harvest areas through comprehensive sanitary survey of the shoreline, microbiological monitoring of growing area waters, and prohibition of harvesting from areas not meeting "approved" growing

250 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA area criteria. Routine control procedures are based upon guidelines in the National Shellfish Sanitation Program (NSSP) Manual of Operations (US- DHEW, 19651. Shellfish safety has been predicated upon the level of coliforms or fecal coliforms present in the water and the direct relationship of these organisms to known sources of pollution. The only microbiol- ogical standard for shellfish meats developed by the NSSP is for APC and fecal coliforms in the product as received at the wholesale market. Fresh or frozen shellfish are considered to be satisfactory if the fecal coliform MPN does not exceed 230 per 100 g and the APC at 35°C (95°F) is not more than 500,000 per g. Historically, it appears that strict adherence to NSSP recommendations for the control of shellfish growing areas gen- erally has resulted in the production of safe shellfish where the hazard has been directly associated with sewage outfalls. Even so, for continued assurance of safety of shellfish, further refinements in the criteria are needed to reflect the impact of quality of growing waters, particularly with reference to the presence of viruses. Since Escherichia cold is the best indicator for fecal contamination, testing water and shellfish for this organism by rapid direct plating methods (see Chapter 5) may be more effective than testing for coliforms or fecal coliforms. Ciguatera is a chemical intoxication resulting from eating fish (usually barracuda, grouper, jack, and red snapper) that have ingested toxic dino- flagellates. In the United States, the illness is usually associated with fish caught in tropical or subtropical waters of areas such as those around Hawaii and Florida. Simple, reliable tests to detect toxic fish are not yet available. An understanding of the ecology and epidemiology of the dis- ease is helpful in predicting to some extent which fish and what areas are likely sources of intoxication. Scombroid poisoning results from eating scombroid fish such as tuna in which histidine has been converted to histamine by microbial activity. This problem can be controlled best by proper harvesting (control of time in net) and postharvesting practices, i.e., proper cooling of fish. FDA (1982) will take regulatory action against canned albacore, skipjack, or yellowfin tuna that contains 20 mg or more of histamine per 100 g. Tuna with between 10-20 mg of histamine per 100 g that shows a second in- dicator of decomposition (spoilage odors or honeycomb appearance) is also subject to regulatory action. In view of the adverse reaction of humans to certain levels of histamine, the above criteria (defect action levels) appear useful. Additional information is needed about the toxic substances involved in scombroid poisoning and on the levels of histamine associated with human toxicity. A recently developed rapid screening test for his- tamine (Lerke et al., 1983) perhaps could be applied to monitor incoming scombroid fish and thus reduce the incidence of scombroid poisoning.

APPLICATION TO FOODS AND FOOD INGREDIENTS 251 Paralytic shellfish poisoning (PSP) is caused by eating shellfish that have ingested certain dinoflagellates, primarily Gonyoulax species. PSP is best controlled by monitoring shellfish for toxins (mouse assay test) during the months of May to October, when this problem occurs most frequently. When the concentration of PSP in representative samples equals or exceeds 80 fig per 100 g of edible portion of raw shellfish, the state shellfish control authority prohibits harvesting in that area (USPHS, 19751. In view of the serious health hazards associated with the ingestion of saxitoxin and other related toxins a microbiological criterion for PSP appears fully justified. With increased interest in and application of shell- fish aquaculture, serious consideration should be given to potential PSP problems. V. parahaemolyticus gastroenteritis is not a serious public health prob- lem in the United States. V. parahaemolyticus can become a problem when a seafood, usually crustaceans, is grossly abused in preparation and storage such as by time-temperature abuse of raw or undercooked product or by cross-contamination of a cooked product and subsequent storage within a temperature range at which this organism multiplies. Thorough cooking, prevention of cross-contamination, and proper cold storage will prevent V. parahaemolyticus gastroenteritis. A microbiological criterion involving V. parahaemolyticus does not appear useful or practical at the present. In addition, there is some uncertainty about the precision and accuracy of present methodologies (see Chapter 41. Vibrio cholerae has in recent years become a concern in the United States, particularly with reference to the consumption of raw or under- cooked shellfish (Blake et al., 1980; Bryan, 1980; CDC, 19811. Accu- mulating evidence (Colwell et al., 1977, 1981; Kaper et al., 1979; Blake et al., 1980; Hood et al., 1981) suggests that V. cholerae (O1 and non- O1) is a component of the autochthonous flora of brackish water and estuarine and salt marshes of coastal areas in temperate regions and poses a potential danger to public health. Measures to minimize this risk include strict adherence to the NSSP regulations for harvesting shellfish, proper refrigerated storage of seafoods, adequate heat treatment, and avoidance of cross-contamination. Other Vibrio species associated with the marine environment, such as v. vuluificus, v. fluvialis, v. mimicus, and V. hollisae are also associated with human disease (Hickman et al., 19821. Their role as causative agents of foodborne illness is not clear. Studies need to be conducted about their ecology, epidemiology, and role in foodborne illness. Botulism from ingestion of fishery products such as smoked and canned products can be controlled best by strict adherence to proper processing and storage practices. For example, smoked fish must be properly heated

252 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA during smoking and then refrigerated so as not to allow growth of C. botulinum. Thermal processing of products in hermetically sealed con- tainers must be adequate and container integrity must be maintained (see sections on heated canned foods and Chapter 41. Conflicting reports have appeared in the literature regarding the botulism hazard of raw fish stored under refrigeration in vacuum packages or in modified gaseous atmospheres (Eyles and Warth, 1981; Eklund, 1982; Lee et al., 1983; Lindsay, 1983; Llobrera, 19831. Some investigators claim that raw fish stored in modified gaseous atmospheres becomes or- ganoleptically unacceptable before toxin can be detected; others have shown that toxicity can occur without overt signs of spoilage. Whether or not the product will have spoiled before toxin is present most likely depends upon a number of conditions such as fish species, film perme- ability, gaseous atmosphere, time-temperature profile of storage, initial number of C. botulinum spores, initial number of bacteria (particularly psychrotrophic spoilage bacteria), and the judgment of the evaluator. In- formation on the effect of most of these variables on toxin production is inadequate. Thorough studies are needed to evaluate the potential hazard of refrigerated storage of raw fish in vacuum packages and in modified gaseous atmospheres. Until such time that the safety of this storage method for raw fish is validated, this practice is not recommended by this sub- committee because of its potential health risks. S. aureus, Salmonella, Shigella, and C. perfringens have occasionally been responsible for foodborne disease outbreaks from consumption of fishery products. Proper postharvesting practices can greatly reduce the likelihood of such outbreaks. Preventive procedures for illness caused by S. aureus and C. perfringens include not holding heated fishery products for prolonged periods within a temperature range at which these organisms can grow. For prevention of salmonellosis, avoid harvesting fish and shellfish from polluted waters and heat the seafood thoroughly. To prevent shigellosis, good personal hygiene is an important measure. The FDA imposed a Salmonella standard as an index of failure to apply good manufacturing practices in order to detain contaminated imported shrimp and frog legs from the U.S. market. Frog legs often are contam- inated by Salmonella because of cross-contamination during handling and processing (Nickelson et al., 19751. Application of the HACCP system including monitoring of processing techniques and cleaning and sanitation of equipment and utensils can effectively reduce Salmonella on commer- cially processed frog legs (Nickelson et al., 19751. APC and E. cold are suggested as parts of microbiological guidelines to monitor critical control points in processing frog legs for example, to monitor exposure of legs to intestinal contents during removal of legs from the frog.

APPLICATION TO FOODS AND FOOD INGREDIENTS Quality 253 Faulty harvesting, processing, and distribution methods can result in microbiological activities leading to loss of quality and subsequent spoilage of fish, molluscs, and crustaceans and of products derived from them. The use of meaningful microbiological end-product criteria to assess qual- ity in raw fishery products appears extremely difficult, costly, and hard to enforce. These products are derived from very large numbers of fish and shellfish species, which differ in composition, are harvested from waters differing in temperature, salinity, and degree of pollution, and are subjected to a variety of processing and handling techniques. A more realistic approach to prevent inferior quality (because of microbial activ- ities) products from entering market channels would be to inspect the raw materials at the plant for traits such as odor and appearance (see Chapter 5) and to assure that subsequent processing and handling steps meet Good Manufacturing Practices. A system of self-certification, through a HACCP system designed for a particular industry, should be implemented and enforced. Agencies such as the National Marine Fisheries Service, and the Food and Drug Administration and trade organizations should assist plants in developing a HACCP system. With periodic monitoring, a viable and effective system for insuring quality would be created. In recent years, FDA has applied organoleptic evaluation and indole content (see Chapter 5) to check imported shrimp for decomposition. This evaluation is not applied to domestically processed shrimp because FDA has access to the plants. Other tests that are associated with quality of fishery products such as trimethylamine, total volatile nitrogen, indole, and ethanol are discussed in Chapter 5. Need for Microbiological Criteria The safety and quality of fish and fishery products can best be assured by application of the HACCP system. For fresh products, monitoring of critical control points should consist primarily of inspection of incoming materials for odor and appearance, proper temperature control (refriger- ation and/or freezing), cleaning and sanitizing of equipment and utensils, and handling of product by employees. In addition, for molluscan shellfish it should include analysis of the growing water and the product (at whole- sale) for fecal coliforms and of the product for saxitoxin and related toxins. For processed tuna, histamine is the agent of concern (see previous dis- cussion on scombroid poisoning). For cooked ready-to-eat products such as cooked peeled deveined shrimp and crabmeat, microbiological guide

254 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA lines for the finished products should include APC, E. coli, and S. aureus. These parameters are useful to evaluate faulty processing and/or handling practices such as inadequate heating, cross-contamination with raw prod- uct, contamination from workers, and inadequate refrigeration which may create health hazards. For all processed fish and seafoods, the HACCP system is recommended as the basis for assuring safety and quality and within this system microbiological guidelines should be implemented where needed to monitor critical control points. An examination of microbiological criteria that exist for fishery products in the various states shows that the most common criterion is that for shellfish as developed by the NSSP, or one closely resembling that cri- terion (Martin and Pitts, 19811. However, some states and local regulatory agencies have additional microbiological criteria for fish and seafoods (Wehr, 19821. Unfortunately, most of these criteria are not based on sound data or experience and are impractical from the standpoint of compliance or enforcement. If there is a need for additional microbiological criteria these should be based on properly designed and executed studies. Recently, quality standards have been recommended for frozen crab cakes, frozen fish cakes, and frozen fishsticks (FDA, 1980, 19814. How- ever, these products constitute neither a health hazard nor a serious quality problem. In addition, the methods used to establish "m" and "M" for these microbiological criteria are inconsistent with the principles on which the 3-class plan alas established (Clark, 1982; ICMSF, 1974; see Chapter 61. The limit "M" represents the level at which a food is considered defective and unacceptable because (1) the number of organisms is so high that shelf-life is unacceptably short, (2) the microbiology is indicative of un- acceptable sanitary conditions during manufacturing or handling, or (3) the microbiological data indicate that the product is unsafe or is a health hazard. "M" for frozen crab cakes, fish cakes, and fishsticks was so selected that it exceeded 99% of the survey population with 99% confi- dence. In addition, data used for these criteria were based on bacterio- logical surveys conducted at the retail level. Because of the potential for changes in the microbiological characteristics of the food between man- ufacture (at the plant level) and the time of purchase, it would be difficult to place the responsibility for an unsatisfactory condition at the retail level on unsatisfactory processing or handling at the manufacturing level. (For further discussion of microbiological quality standards, see Chapter 2.) Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated To implement the recommended HACCP system to assure safety and quality of fishery products, an up-to-date HACCP system should be de

APPLlCATlON TO FOODS AND FOOD lNGREDlENTS 255 veloped for the various products. For some products such as frozen breaded shrimp (see Chapter 8), breaded fish products (see Chapter 8), and pro- cessed blue crab (Phillips and Peeler, 1972; National Blue Crab Industry Association, 1982), extensive information is already available. GMP recommendations such as those developed by the Tri-State Sea- food Committee (National Blue Crab Industry Association, 1982) should be useful to processors. Where Criteria Should Be Applied As described above, for certain products, microbiological criteria are applied to the growing water (shellfish); for others, they are applied at the processing plant level. Purchase specifications may be useful to food service operators to monitor incoming products as part of their HACCP program unless the manufacturer can provide data about the quality and safety of the products. References Blake, P. A., D. T. Allegra, J. D. Snyder, T. J. Barrett, L. McFarland, C. T. Caraway, J. C. Feeley, J. P. Craig, J. V. Lee, N. D. Puhr, and R. A. Feldman 1980 Cholera-a possible endemic focus in the United States. New Engl. J. Med. 302:305- 309. Bryan, F. L. 1980 Epidemiology of foodborne diseases transmitted by fish, shellfish and marine crus- taceans in the United States, 1970-1978. J. Food Prot. 43:859-876. CDC (Centers for Disease Control) 1981 Cholera Texas. Morb. Mort. Weekly Rpt. 30:389-390. Clark, D. S. 1982 International perspectives for microbiological sampling and testing of foods. J. Food Prot. 45:667-671. Colwell, R. R., J. Kaper, and S. W. Joseph 1977 Vibrio cholerae, Vibrio parahaemolyticus, and other vibrios: occurrence and distri- bution in Chesapeake Bay. Science 198:394-396. Colwell, R. R., R. J. Seidler, J. Kaper, S. W. Joseph, S. Garges, H. Lockman, D. Maneval, H. Bradford, N. Roberts, E. Remmers, I. Huq, and A. Huq 1981 Occurrence of Vibrio cholerae serotype Of in Maryland and Louisiana estuaries. Appl. Environ. Microbiol. 41:555-558. Eklund, M. W. 1982 Significance of Clostridium botulinum in fishery products preserved short of steril- ization. Food Technol. 36(12): 107-112, 115. Eyles, M. J., and A. D. Warth 1981 Assessment of the risk of botulism from vacuum-packaged raw fish: A review. Food Technol. Austral. 33(11):574-580. FDA (Food and Drug Administration) 1980 Frozen fish sticks, frozen fish cakes, and frozen crab cakes; Recommended micro-  biological quality standards. Federal Register 45 (108):37524-37526.

256 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA 1981 Frozen fish sticks, frozen fish cakes, and frozen crab cakes; Recommended micro- biological quality standards. Federal Register 46 (113):31067-31068. 1982 Defect action levels for histamine in tuna; availability of guide. Federal Register 47 (178):40487-40488. Hickman, F. W., J. J. Farmer, III, D. G. Hollis, G. R. Fanning, A. G. Steigerwalt, R. E. Weaver, and D. J. Brenner 1982 Identification of Vibrio hollisoe sp. nov. from patients with diarrhea. J. Clin. Mi- crobiol. 15:395-401. Hood, M. A., G. E. Ness, and G. E. Roderich 1981 Isolation of Vibrio cholerae serotype O1 from the Eastern oyster, Crassostrea vir- ginica. Appl. Environ. Microbiol. 41:559-560. ICMSF (International Commission on Microbiological Specifications for Foods) 1974 Microorganisms in Foods. 2. Sampling for microbiological analysis: Principles and specific applications. Toronto: University of Toronto Press. Kaper, J., H. Lockman, R. R. Colwell, and S. W. Joseph 1979 Ecology, serology and enterotoxin production of Vibrio cholerae in Chesapeake Bay. Appl. Environ. Microbiol. 37:91-103. Lee, D. A., M. Solberg, D. Furgang, and J. J. Specchio 1983 Time to toxin detection and organoleptic deterioration in Clostridium botulinum in- oculated fresh fish fillets during modified atmosphere storage. Paper (No. 483) pre- sented at the 43rd Annual IFT Meeting, New Orleans. Lerke, P. A., M. N. Porcuna, and H. B. Chin 1983 Screening test for histamine in fish. J. Food Sci. 48:155-157. Lindsay, R. C. 1983 Safety and technology of modified-atmosphere packaging of fresh fish. Paper (No. 152) presented at the 43rd Annual IFT Meeting, New Orleans. Llobrera, A. T. 1983 Bacteriological safety assessment of Clostridium botulinum in fresh fish and shellfish packaged in modified atmosphere containing CO2. Ph.D. dissertation, Texas A&M University, College Station. August. Martin, R. E., and G. T. Pitts 1981 Handbook of State and Federal Microbiological Standards and Guidelines. Washing- ton, D.C.: National Fisheries Institute. National Blue Crab Industry Association 1982 Tri-State Seafood Committee GMP Recommendations. Washington, D. C.: National Blue Crab Industry Association. Nickelson, R., L. E. Wyatt, and C. Vanderzant 1975 Reduction of Salmonella contamination in commercially processed frog legs. J. Food Sci. 40:1239-1241. Phillips, F. A., and J. T. Peeler 1972 Bacteriological survey of the blue crab industry. Appl. Microbiol. 24:958-966. USDHEW (U.S. Department of Health, Education and Welfare) 1965 National Shellfish Sanitation Program. Manual of Operations. Part 1 Sanitation of Shellfish Growing Areas. PHS Pub. 33 (Revised 1965). Washington, D.C.: U.S. Government Printing Office.  USPHS (U.S. Public Health Service) 1975 National Shellfish Safety Program. Federal Register 40 (119):25916-25935. Wehr, H. M. 1982 Attitudes and policies of governmental agencies on microbial criteria for foods an update. Food Technol. 36(9):45-54, 92.

APPLICATION TO FOODS AND FOOD INGREDIENTS H. FRUITS AND VEGETABLES Sensitivity of Products Relative to Safety and Quality 257 Raw fruits and vegetables are not common causes of foodborne illnesses in the United States (CDC, 1971; 1972; 1973; 1974; 1976a,b; 1977; 1979; 1981a,b). The high acidity of many fruits inhibits the growth of bacteria pathogenic for humans and the edible portions often are protected from contamination by a skin or thick rind. Although the vegetables that are eaten raw, such as in salads, may yield APCs of 107/g or greater (Fowler and Foster, 1976), they have not presented a serious public health problem in the United States and Canada because the microorganisms are mainly saprophytic species (Splittstoesser, 19701. However, in countries where polluted water or raw sewage are used for irrigation or fertilization of crops, enteric pathogens such as salmonellae and parasites are common contaminants (Ercolani, 1 976; Tamminga et al., 19781. Certain imported fruits and vegetables, then, might be potential sources of foodborne illness. A recently recognized potential problem has been the presence of Clos- tridium botulinum spores on raw potatoes that will be baked in foil. It has been hypothesized that a botulism outbreak occurred because foil-wrapped baked potatoes that were used in a potato salad had been held at room temperature for a number of days. Laboratory studies showed that botu- linum spores can survive baking and that vegetative growth and toxin production can occur in the foil-wrapped product (Seals et al., 19811. Mung bean and other vegetable sprouts that are often consumed raw develop high populations of bacteria during the sprouting process. Al- though sprouts usually are negative for foodborne pathogens (Patterson and Woodburn, 1980), at least one outbreak has been attributed to Bacillus cereus (Portnoy et al., 19761. Nonsterile processed fruits and vegetables, such as frozen and desiccated products, have rarely been responsible for foodborne illnesses. The few cases that are reported each year have generally been due to contamination or mishandling by persons in food service establishments or home kitchens (see for example CDC-, 1981b). Commercially canned fruits and vegetables have an excellent public health record (see sections on canned foods). Home-canned vegetables, on the other hand, have been a major cause of botulism in the United States with 48 outbreaks occurring during the period 1970 to 1979 4CDC, 1971; 1972; 1973; 1974; 1976a,b; 1977; 1979; 1981a,b). The application of microbiological criteria would not have reduced the incidence of bot- ulism since the outbreaks were due to underprocessing in the home.

258 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Need for Microbiological Criteria Raw Products There is little use for microbiological criteria for fresh fruits and veg- etables at the present time. However, future changes in irrigation and fertilization practices in this country or changes in the source of imported produce could mandate testing for certain pathogens or indicator organ- isms. Nonsterile, Shelf-Stable ProdFucts . ~ Because of an excellent public health record, standards are not needed for foods such as frozen or desiccated fruits and vegetables. Guidelines, on the other hand, may aid in the promotion of Good Manufacturing Practices. These fruits and vegetables commonly receive a treatment at some processing step that destroys most of the epiphytic microflora and field contaminants. High viable counts on the finished product, therefore, usu- ally reflect insanitary equipment, although actual microbial growth on the food also may occur (Splittstoesser, 1973~. Most fruits and vegetables are succulent products whose solubles provide an excellent medium for a wide variety of microorganisms. The nutrients may collect in equipment areas difficult to clean (such as the interior of slicers), and significant microbial buildup may occur. Lactic acid bacteria often predominate on vegetable- processing lines although a variety of other mesophilic organisms including coliforms and Geotrichum make up a portion of the normal microflora (Splittstoesser and Mundt, 1976J. Yeasts, molds, and aciduric bacteria predominate in fruit-processing plants, but other organisms including col- itorms may be present in low numbers.  Aerobic plate counts are useful for the assessment of sanitation and manufacturing practices for low-acid vegetables. The enumeration of yeasts and molds and lactic acid bacteria is useful for fruits. Critical control points for sampling would include conveyor belts and unit operations such as size graders, cutters, slicers, and filling machines. Analyzing for di- acetyl, a microbial metabolic product, has also served as a useful criterion for citrus products (Hill and Wenzel, 19571. Since coliforms, including fecal coliforms such as Klebsiella pneumonias, are a part of the normal microflora of the raw product and processing lines, there is little reason to include them in microbiological criteria (Splittstoesser et al., 1980a). Testing for Escherichia colt, on the other hand, may be informative since this organism usually is not present on these foods. Because of an excellent public health record, the routine testing of

APPLICATION TO FOODS AND FOOD INGREDIENTS 259 frozen and dried fruits and vegetables for common foodborne pathogens is not justified. Studies indicate that staphylococci are absent or are present only in low numbers and that salmonellae are usually absent (Splittstoesser et al., 1965; Splittstoesser and Segen, 19701. Commercially Sterile Fruits and Vegetables As discussed in Chapter 5, Howard mold count and rot fragment count procedures are used mainly to detect the inclusion of moldy materials in canned fruit and tomato products, and the enumeration of Geotrichum candidum hyphae has been usec; to detect insanitary processing machinery in vegetable canneries (AOAC, 19801. These microscopic count criteria are useful because viable counts cannot, of course, be applied to a ther- mally processed food. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated The microbiology of frozen vegetables depends upon the vegetable type and the method of processing (Barnard et al., 1982; Duran et al., 1982; Splittstoesser and Corlett, 19801. For example, green beans usually show higher aerobic plate counts than do peas, and French-style green beans are more heavily contaminated than are the cut variety. Presently lacking are modern data that would relate the microbiology of these foods to Good Manufacturing Practices. This information can be generated only through extensive factory surveys. Similar studies would be needed for frozen fruits as well as for nonsterile fruits and vegetables that are preserved by other means. The Howard mold count and rot fragment standards have been used for many years (FDA, 19781. Although G. candidum has long been promul- gated by FDA as an indicator of insanitation in tomato processing (Ei- senberg and Cichowicz, 1977), more data would be needed (again, developed from extensive factory surveys) before meaningful limits based on filament counts could be established for this and other vegetables (Splittstoesser et al., 1980b). A number of states have guidelines for horticultural products such as frozen potatoes, frozen onion rings, and pasteurized fruit juices (Weird 19821. There is little evidence to justify them. Where Criteria Should Be Applied Microbiological guidelines would usually be applied to frozen and dried fruits and vegetables following packaging. The samples can therefore be

260 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA taken from the processing plant, warehouses, or retail markets. While market samples would show the microbiological condition of a food as purchased by the consumer, they might not permit an interpretation as to the reason for the particular findings. When the objective of the guideline is to monitor a critical control point in the process line, samples would be collected at or following this point. References AOAC (Association of Official Analytical Chemists) 1980 Official Methods of Analysis, 13th Ed. Washington, D.C.: AOAC. Barnard, R. J., A. P. Duran, A. Swartzentruber, A. H. Schwab, B. A. Wentz, and R. B. Read, Jr. 1982 Microbiological quality of frozen cauliflower, corn and peas obtained at retail markets Appl. Environ. Microbiol. 44:54-58. CDC (Centers for Disease Control) 1971 Foodborne Outbreaks. Annual Summary 1970. Atlanta: Centers for Disease Control. 1972 Foodborne Outbreaks. Annual Summary 1971. Atlanta: Centers for Disease Control. 1973 Foodborne Outbreaks. Annual Summary 1972. Atlanta: Centers for Disease Control. 1974 Foodborne and Waterborne Disease Outbreaks. Annual Summary 1973. Atlanta: Cen ters for Disease Control. 1976a Foodborne and Waterborne Disease Outbreaks. Annual Summary 1974. Atlanta: Cen ters for Disease Control. 1976b Foodborne and Waterborne Disease Outbreaks. Annual Summary 1975. Atlanta: Cen ters for Disease Control. 1977 Foodborne and Waterborne Disease Outbreaks. Annual Summary 1976. Atlanta: Cen ters for Disease Control. 1979 Foodborne and Waterborne Disease Outbreaks. Annual Summary 1977. Atlanta: Cen ters for Disease Control. 1981a Foodborne Disease Outbreaks. Annual Summary 1978. Atlanta: Centers for Disease Control. 1981b Foodborne Disease Outbreaks. Annual Summary 1979. Atlanta: Centers for Disease Control. P., A. Swartzentruber, J. M. Lanier, B. A. Wentz, A. H. Schwab, R. J. Barnard, Duran, A. and R. B. 1982 Eisenberg, 1977 Ercolani, G. L. 1976 Bacteriological quality assessment of fresh marketed lettuce and fennel. Appl. En- viron. Microbiol. 31:847-852. FDA (Food and Drug Administration) 1978 The food defect action levels. Food and Drug Administration, HFF-342, Washington, D.C.  Fowler, J. L., and J. F. Foster 1976 A microbiological survey of three fresh green salads Can guidelines be recom- mended for these foods? J. Milk Food Technol. 39:111-113. Read, Jr. Microbiological quality of five potato products obtained at retail markets. Appl. Environ. Microbiol. 44:1076-1080. W. V., and S. M. Cichowicz Machinery mold Indicator organism in food. Food Technol. 31(2):52-56.

APPLICATION TO FOODS AND FOOD INGREDIENTS 261 Hill, E. C., and F. W. Wenzel. 1957 The diacetyl test as an aid for quality control of citrus products. 1. Detection of bacterial growth in orange juice during concentration. Food Technol. 11:240-243. Patterson, J. E., and M. J. Woodburn 1980 Klebsiella and other bacteria on alfalfa and bean sprouts at the retail level. J. Food Sci. 45:492-495. Portnoy, B. L., J. M. Goepfert, and S. M. Harmon 1976 An outbreak of Bacillus cereus food poisoning resulting from contaminated vegetable sprouts. Am. J. Epidemiol. 103:589-593. Seals, J. E., J. D. Snyder, T. A. Edell, C. L. Hatheway, C. J. Johnson, R. C. Swanson, and J. M. Hughes 1981 Restaurant-associated type A botulism: Transmission by potato salad. Am. J. Epi- demiol. 113:436-444. Splittstoesser, D. F. 1970 Predominant microorganisms on raw plant foods. J. Milk Food Technol. 33:500-505. 1973 The microbiology of frozen vegetables. Food Technol. 27(1):54-56,60. Splittstoesser, D. F., J. Bowers, L. Kerschner, and M. Wilkison 1980b Detection and incidence of Geotrichum candidum in frozen blanched vegetables. J. Food Sci. 45:511-513. Splittstoesser, D. F., and D. A. Corlett, Jr. 1980 Aerobic plate counts of frozen blanched vegetables processed in the United States. J . Food Protect . 43 :717-719. Splittstoesser, D. F., and J. O. Mundt 1976 Fruits and vegetables. In Compendium of Methods for the Microbiological Exami- nation of Foods. M. L. Speck, ed. Washington, D.C.: American Public Health Association. Splittstoesser, D. P., and B. Segen 1970 Examination of frozen vegetables for salmonellae. J. Milk Food Technol. 33:111- 113. Splittstoesser, D. F., G. E. R. Hervey II, and W. P. Wettergreen 1965 Contamination of frozen vegetables by coagulase positive staphylococci. J. Milk Food Technol. 28: 149- 151. Splittstoesser, D. F., D. T. Queale, J. L. Bowers, and M. Wilkison 1980a Coliforrn content of frozen blanched vegetables packed in the United States. J. Food Safety 2:1-11. Tamminga, S. K., R. R. Beumer, and E. H. Kampelmacher 1978 The hygienic quality of vegetables grown in or tentative survey. J. Hyg., Camb. 80: 143- 154. imported into the Netherlands: A Wehr, H. M. 1982 Attitudes and policies of governmental agencies on microbiological criteria for foods- An update. Food Technol. 36(9):45-54, 92.  I. FRUIT BEVERAGES Fruit beverages include single-strength juices, juice concentrates, and drinks that contain a small percentage of fruit juice plus added sugar, acids, flavors, and colors. They may be preserved as frozen concentrates,

262 EVALUATION OF THE ROLE OF MICROBlOLOGlCAL CRITERIA by refrigeration, as pasteurized shelf-stable products, or with the addition of benzoate and/or sorbate. Many of the criteria and principles useful for fruit beverages also apply to sweetened and acidified beverages that contain no fruit juice. Sensitivity of Products Relative to Safety and Quality Fruit beverages have had an excellent public health record. Most possess a low pH, which prevents growth of pathogenic bacteria, and many juice products are given a thermal process that kills microorganisms of spoilage and public health significance. While it is believed that enteric pathogens usually die off rapidly when in a highly acidic environment, Salmonella typhimurium has been shown to survive when inoculated into apple juice (Goverd et al., 1979) and nonpasteurized cider was responsible for a large outbreak of salmonellosis (CDC, 19751. Heat-resistant molds, mainly species of Byssochlamys, Aspergillusfish- eri, and Penicillium vermiculatum may cause spoilage of pasteurized fruit beverages because their ascospores are sufficiently resistant to survive the usual heat treatments that are given to these foods (Splittstoesser and Splittstoesser, 1977; Van der Spuy et al., 19751. The spores are present in orchards and vineyards (Splittstoesser et al., 1971), and thus are intro- duced into the beverage via the fruit. It is not uncommon to find low numbers of viable ascospores in juice concentrates. Usually only a limited amount of mold growth occurs in a canned fruit beverage due to the low levels of oxygen present. However, off-flavors may be apparent and some heat-resistant molds may produce mycotoxins such as patulin (Rice et al., 19771. Need for Microbiological Criteria Standards are not needed because fruit juice beverages are rarely re- sponsible for foodborne illnesses. Guidelines are useful for the assessment of good manufacturing prac- tices. Counts of yeasts and lactic acid bacteria in most fruit juices (Luthi, 1959; Murdock, 1976) and levels of diacetyl in citrus juices (Hill and Wenzel, 1957) often can be correlated with conditions of sanitation. Trace levels of patulin in apple juice are indicative of the pressing of fruit rotted by Penicillium expansum and other molds (Lindroth and Niskanen, 19781. A criterion for patulin in apple juice might be useful. Testing for coliforms is of limited value because they constitute a part of the normal processing line microflora and their presence does not indicate fecal contamination (Dack, 1 955; Patrick, 1 9531.

APPLICATION TO FOODS AND FOOD INGREDIENTS 263 Purchase specifications that limit heat-resistant mold spores in juice concentrates are useful. Also, specifications for mesophilic bacteria, yeasts, and molds in sugar to be used in nonpasteurized beverages are applicable (see Chapter 9, Part O). Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated While companies may possess a wealth of data that relate processing conditions to the microbiology of their products, little information is avail- able in the literature. Therefore, extensive surveys of fruit beverage pro- cessing lines might be required if a guideline were to be applied. Where Criteria Should Be Applied Guidelines can be applied to frozen, nonsterile products such as fruit juice concentrates at all stages of processing including the end product. With thermally processed beverages, on the other hand, analyses are most useful at critical control points prior to pasteurization. Beverages to be treated with preservatives are usually analyzed prior to packaging. Ingre- dients subjected to specifications should be tested after receipt by the purchaser. References CDC (Center for Disease Control) 1975 Salmonella typhimurium outbreak traced to a commercial apple cider. Morb. Mort. Weekly Rpt. 24:87. Dack, G. M. 1955 Significance of enteric bacilli in foods. Am. J. Pub. Health 45:1151-1156. Goverd, K. A., F. W. Beech, R. P. Hobbs, and R. Shannon 1979 The occurrence and survival of coliforms and salmonellae in apple juice and cider. J. Appl. Bacteriol. 46:521-530. Hill, E. C., and F. W. Wenzel 1957 The diacetyl test as an aid for quality control of citrus products. 1. Detection of bacterial growth in orange juice during concentration. Food Technol. 11 :240-243. Lindroth, S., and A. Niskanen 1978 Comparison of potential patulin hazard in home made and commercial apple products. J. Food Sci. 43:446-448. Luthi, H. 1959 Microorganisms in noncitrus juices. Pp. 221-284 in Advances in Food Research. Vol. 9. E. M. Mrak and G. F. Stewart, eds. New York: Academic Press. Murdock, D. I. 1976 Fruit drinks, juices, and concentrates. In Compendium of Methods for the Micro- biological Examination of Foods. M. L. Speck, ed. Washington, D. C.: American  Public Health Association.

264 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Patrick, R. 1953 Coliform bacteria from orange concentrate and damaged oranges. Food Technol. 7:157-159. Rice, S. L., L. R. Beuchat, and R. E. Worthington 1977 Patulin production by Byssochlamys spp. in fruit juices. Appl. Environ. Microbiol. 34:791-796. Splittstoesser, D. F., and C. M. Splittstoesser 1977 Ascospores of Byssochlamys fulva compared with those of a heat resistant Aspergillus. J. Food Sci. 42:685-688. Splittstoesser, D. F., F. R. Kuss, W. Harrison, and D. B. Prest 1971 Incidence of heat-resistant molds in eastern orchards and vineyards. Appl. Microbiol. 21 :335-337. Van der Spuy, J. E., F. N. Matthee, and D. J. A. Crafford 1975 The heat resistance of moulds Penicillium vermiculatum Dangeard and Penicillium brefeldianum Dodge in apple juice. Phytophylactica 7(3):105-107. J. LOW-ACID CANNED FOODS The U.S. Food and Drug Administration defines a low-acid canned food as one with an equilibrium pH >4.6 and a water activity >0.85. There is an exclusion to pH 4.7 for tomatoes and tomato products (FDA, 19831. Low-acid canned foods are packed in hermetically sealed contain- ers. These products must be processed to achieve commercial sterility, i.e., a condition achieved by application of heat that renders such food free of viable forms of microorganisms having public health significance, as well as any microorganisms of non-health significance capable of re- producing in the food under normal nonrefrigerated conditions of storage and distribution (FDA, 1983~. The degree of heat treatment applied varies with the physiochemical nature of the product. For example, canned low-acid vegetables and un- cured meats receive a "botulinum cook" (see below). Lesser heat treat- ments are applied to shelf-stable canned cured meats as well as to foods in which reduced water activity provides a barrier to the growth of spore- forming bacteria. These treatments may also be applied to aseptically processed low-acid canned foods wherein commercially sterilized cooled product is filled into presterilized containers followed by aseptic hermetical sealing with a presterilized closure in a sterile environment. Sensitivity of Products Relative to Safety and Quality Safer Botulism is the prime hazard in low-acid canned foods. If spores of Clostridium botulinum survive the thermal process and thereafter grow and produce toxin in the canned product, and if the food is not adequately

APPLICATION TO FOODS AND FOOD INGREDIENTS 265 heated to destroy the toxin before it is eaten, consumption may lead to botulism, an often fatal disease. About 30 billion cans of low-acid food are consumed in the United States each year. From 1940 to 1982 there were 6 botulism outbreaks with 8 deaths caused by low-acid commercially canned foods in the United States. During this period there were an ad- ditional 7 instances in which the presence of C. botulinum or its toxin was detected in low-acid canned foods but from which no illness resulted. Of the 13 incidents, 5 were attributed to postprocessing contamination (NFPA-CMI, 1984). The excellent public health history of low-acid canned foods is largely attributable to the application of the 1 2-D concept. From the classic studies of Esty and Meyer (1922) and Townsend et al. (1938), it was determined that the time required at 121 . 1 °C (250°F) to reduce the survivor population of C. botulinum spores in phosphate buffer by a factor of 12 decimal units was 2.45 min (For of 2.45~. This work was based upon studies using the most heat-resistant strain of C. botulinum. Determination of processing schedules for the destruction of C. botulinum in other heating menstra, e.g., a canned vegetable, requires inoculated-pack studies in the food under consideration. On the basis of such studies, necessary time/tem- perature parameters for a 12-D process on the specific food are established. Shelf-stable canned cured meats receive heat processes far below that dictated by the 12-D process, e.g., expressed as F-values, the heat treat- ment frequently corresponds to Fo = 0.4 to 0.6, but FO-values as low as 0.05 are used. These processes alone do not account for the excellent public health record of these low-acid products. Rather the products' safety with respect to botulism is due to the combined effects of heat injury, curing salts, and the low incidence of C. botulinum spores in the raw materials (see ICMSF, 19801. Quality Though application of the 12-D process has resulted in the production of products with excellent public records, this heat treatment does not assure commercial sterility. There are many nontoxigenic sporeforming bacteria with D values much greater than C. botulinum. Among them are thermophilic anaerobes, including the flat-sour group (Bacillus stearo- thermophilus), the gaseous-spoilage group (Clostridium thermosacchar ~Fo is a term used to denote the number of minutes a product is exposed to 250°F. In practice, the total heat process applied to a product is determined and through mathematical calculations this is equated to equivalent minutes at 250°F.

266 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA olyticum), and the sulfide stinkers (Desulfotomaculum nigrificans). Also included are mesophilic putrefactive anaerobes, e.g., Clostridium spo- rogenes. To avoid serious economic losses due to spoilage, the severity of heat processes generally applied to low-acid canned foods is greater than that necessary to assure destruction of C. botulinum, i.e., the gen- erally acknowledged 12-D process. Even with the application of more severe heat processes, it is necessary to rigidly control the level of these sporeforming bacteria in raw materials, if severe economic losses due to product spoilage are to be avoided (APHA, 1976, 1984; see also parts M and O of this chapter). Need for Microbiological Criteria The presence of C. botulinum in processed low-acid canned foods is expected to occur so infrequently that no conceivable sampling plan for microbiological testing of finished product or by incubation test is a prac- tical means of assuring safety. The Food and Drug Administration (FDA, 1973a) suggests finished product incubation and evaluation for aseptically processed canned foods, the primary purpose of which is to detect sterility breaks in the aseptic canning system. The U.S. Department of Agriculture requires incubation of approximately 1 container of meat and poultry containing low-acid canned foods per 1,000 processed as a check against gross underprocessing or container failure. The safety of low-acid canned foods is based upon the establishment of heat processes capable of destroying C. botulinum, or in the case of shelf-stable cured meat, the control of the interrelated factors preventing C. botulinum outgrowth in the cured meat environment. The control of spoilage depends upon the establishment of heat processes capable of destroying spoilage organisms, some of which are more resistant than C. botulinum, and/or controlling the numbers of these organisms through judicious selection of raw materials. The control of safety and stability is accomplished through the appli- cation of the HACCP system. In no phase of food processing is monitoring of critical control points so essential to the safety and stability of finished products as it is in the manufacture of low-acid canned foods. Indeed, monitoring, the extent of which has been reviewed by Ito ( 1974), is subject to federal regulations (FDA, 1973a,b; FPI, 19751. Physical and chemical tests are performed during production to ensure that all factors necessary to the application of the established safe processes have been adequately controlled. Numerous checks and tests are made at various critical control points, including the adequacy of ingredient blending, determination of

APPLICATION TO FOODS AND FOOD INGREDIENTS 267 consistency, ratio between solids and liquids, weight of product placed in the container, amount of head space, adequacy of double seam, time and temperature during sterilization, quality of cooling water, and post- processing handling of cooled cans. There is no intent above to indicate all the points that are monitored. Rather, the object is to emphasize the importance of checking each critical control point to ensure that the es- tablished procedures have been properly carried out. This necessitates specification of the method for measuring each parameter, determination of satisfactory limits for each test, determination of the frequency with which tests and checks should be employed, and recording of the results of these tests. If a defect is observed, remedial action should be taken and documented. The U.S. Food and Drug Administration requires that operators of retorts, aseptic processing and packaging systems, product-formulating systems, and container closure inspectors be under the supervision of a person who has satisfactorily completed prescribed courses approved by the FDA Commissioner. At the same time, FDA inspectors are trained in the elements of the HACCP system as it relates to low-acid canned foods. The net result is a cost-effective approach to the control of safety and stability based upon both industry and regulatory knowledge of effective control of critical control points. It is the opinion of this subcommittee that the HACCP system as applied to low-acid canned foods should serve as a model for the rest of the food industry. The identification of critical control points and the establishment of effective monitoring systems for these foods evolved as a result of joint government/industry efforts. Were the same approach to be taken in other phases of the food industry, similar cost-effective solutions to the problems of food safety and stability might be achieved. Training of inspectors in the HACCP approach should likewise be extended to other types of food- processing operations. Where Criteria Should Be Applied Microbiological criteria to determine the adequacy of raw materials can and are in most cases incorporated as a part of purchase specifications. The application of National Food Processors Association (formerly the National Canners Association) criteria to such commodities as sugar and starch has a well-established and effective history in the food industry. Microbiological criteria are also useful for monitoring critical control points such as cooling water and equipment surfaces.

268 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA References APHA (American Public Health Associations) 1976 Compendium of Methods for the Microbiological Examination of Foods. M. L. Speck, ed. Washington, D.C.: APHA. 1984 Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed., M. L. Speck, ed. Washington, D.C.: APHA. Esty, J. R., and K. F. Meyer 1922 The heat resistance of spores of B. botulinus and allied anaerobes. XI. J. Infect. Dis. 31 :650-663. FDA (Food and Drug Administration) 1973a Thermally processed low-acid foods packaged in hermetically sealed containers. Part 128B (remodified as Part 113). Federal Register 38(16):2398-2410, Jan. 14. 1973b Emergency permit control. Part 90 (recodified as Part 109), Federal Register 38(92):12726-12721. May 14. 1983 Thermally processed low-acid foods packaged in hermetically sealed containers. Gen eral provisions. Definitions. Code of Federal Regulations 21 CFR 113.3. FPI (Food Processors Institute) 1975 Canned Foods Principles of thermal process control in container closure evaluation, 2nd Ed. National Canners Association, eds. Berkeley, Calif.: FPI. ICSMF (International Commission on Microbiological Specifications for Foods) 1980 Microbial Ecology of Foods. Vol. 1. Factors affecting life and death of organisms. New York: Academic Press. Ito, K. 1974 Microbiological critical control points in canned foods. Food Technol. 28(9):46-48. NFPA-CMI (National Food Processors Association Can Manufacturers Institute, Container Integrity Task-force) 1984 Botulism risk from post-processing contamination of commercially canned foods in metal containers. J. Food Prot. 47(10):801-816. Townsend, C. T., J. R. Esty, and F. C. Baselt 1938 Heat-resistant studies on spores of putrefactive anaerobes in relation to determination of safe processes for canned foods. Food Res. 3:323-346. K. ACID CANNED FOODS The U.S. Food and Drug Administration defines acid canned foods as those with a natural pH of ~ 4.6 (below 4.7 for tomatoes and tomato products). An acidified food is a low-acid food to which acidify or acid fondest have been added. It has a water activity > 0.85 and a finished equilibrium pH of ~ 4.6 (FDA, 1983a). Acid and acidified canned foods packed in hermetically sealed con- tainers are rendered commercially sterile by the application of heat suf- ficient to produce a food free from microorganisms capable of growth in  the food at normal, nonrefrigerated conditions. The heat treatments are not designed to inactivate spores of Clostridium botulinum because this organism cannot grow at pH 4.6 or below. Example treatments are the

APPLICATION TO FOODS AND FOOD INGREDIENTS 269 "Bacillus coagulans cook" for tomato juice of 0.7 min at 121. 1°C (250°F) or equivalent and the hot-fill and hold processes for products with greater acidity (pH 3.7 to 4.5) such as tomato paste, puree, and ketchup (90.6°C/ 195°F for 3-5 min). The heat treatment is designed to inactivate vegetative forms and sporeformers capable of multiplication in the product. The acid content of the food is relied upon to act as a barrier to growth of spores such as C. botulinum that are not inactivated by the relatively mild thermal treatment. Sensitivity of Products Relative to Safety and Quality Safety Acid and acidified canned foods prepared commercially in the United States have a remarkable record with respect to safety. This record un- doubtedly is due to careful control of pH below the limits of growth of most pathogens, including C. botulinum. Quality Certain acid and acidified canned foods are susceptible to spoilage by flat-sour microorganisms such as Bacillus coagulans. Care must be taken to use ingredients with low initial counts of this organism and to destroy them with thermal treatment or limit their ability to reproduce by choice of a suitably low pH and/or high titratable acidity. Need for Microbiological Criteria Acid canned foods are considered less critical than low-acid canned foods with respect to control of manufacture. However, the U.S. Food and Drug Administration in 1979 promulgated Good Manufacturing Prac- tices for Acidified Foods (FDA, 1983b). Many of the controls, records, and procedures listed for low-acid canned foods must also be followed for acidified foods. The critical issues unique to successful processing of acid and acidified canned foods are: (1) destruction of microorganisms capable of repro- duction in the food and (2) provision of a pH/acidity barrier to the growth of those microorganisms that survive. This barrier must control not only the acid-generating flat-sour types that would cause spoilage of economic (but not public health) concern but also acid utilizing microorganisms whose reproduction might result in the elevation of pH to the point where a pathogenic sporeformer such as C. botulinum could grow. Criteria,

270 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA usually purchase specifications, should be applied to ingredients such as starch and sugar to prevent the use of raw materials that could lead to these types of spoilage. Where Criteria Should Be Applied The critical control points in the manufacture of acid and acidified canned foods are much the same as those already outlined for low-acid canned foods. Selection of ingredients of low spore load is important. The thermal process is of great importance to prevent loss of product due to spoilage. Control of pH and the provision of sound containers are critical. The presence and growth of microorganisms that can elevate the pH of the food and permit reproduction of pathogens must be precluded. References FDA (Food and Drug Administration) 1983a Acidified foods. General provisions. Definitions. Code of Federal Regulations. Title 21: part 114.3. 1983b Acidified foods. General provisions. Good manufacturing practices. Code of Federal Regulations 21 CFR 114.5 L. WATER ACTIVITY-CONTROLLED CANNED FOODS The Code of Federal Regulations requires that a low-acid canned food have a pH greater than 4.6 and a water activity (aw) greater than 0.85 (FDA, 19831. If a canned food has a pH greater than 4.6 and aw greater than 0.85, it should receive a C. botulinum cook. There are, however, a number of foods in hermetically sealed containers with pH >4.6 and aw >0.85 that do not receive a C. botulinum cook because the severe heat treatment would jeopardize product acceptability. Examples include cheese spreads and canned cakes and breads, as well as a multitude of syrups and toppings. The reason for this seeming con- tradiction of the regulation lies in the fact that C. botulinum and most other sporeformers are unable to grow at aw levels below 0.93. Therefore, if enough heat is applied to inactivate all vegetative forms and the aw is controlled below 0.93, shelf-stable products result. However, manufac- turers of these products must register their plants, file their processes, and demonstrate that spores of microorganisms not destroyed by the thermal treatment are incapable of growth in the food under normal nonrefrigerated conditions of storage and distribution. In fact, parameters such as pH, aw, heat, and possibly others may interact to result in a stable product. The long safety record of processed cheese

APPLICATION TO FOODS AND FOOD INGREDIENTS 271 spreads attests to the efficacy of multiple inhibitory factors. Research is now under way to identify and quantitate the interaction of pH, aw, salt and phosphate concentrations, and possibly other factors that contribute to the stability of cheese spreads. Sensitivity of Products Relative to Safety and Quality Safety Canned foods of reduced aw have had a remarkable safety record in the United States. This record is the result of careful application of the avail- able knowledge on water activity control of microorganisms. Quality Canned foods of reduced aw are susceptible to spoilage by organisms that gain entry as a result of lack of container integrity. Some of the post- process contaminants, such as certain yeasts and molds, can grow in environments of low water activity. Need for Microbiological Criteria and Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Because reduced water activity is the barrier to growth of those micro- organisms surviving the mild thermal treatments given these foods, control of water activity is critical. However, the level of reduced water activity needed for commercial sterility may vary depending on food composition. For example, C. botulinum has not been demonstrated to grow under ideal conditions in laboratory media below a water activity of 0.93, the exact level for inhibition depending upon the humectant used to control available water. Therefore, it is reasonable to expect that complex foods of varying ingredient mixtures combined with sublethal heat treatments will prevent the growth of C. botulinum at water activities somewhat higher than 0.93. There is a need to destroy vegetative forms, particularly Staphylococcus aureus, yeasts, and molds in all thermally processed, water activity-con- trolled canned foods. Where Criteria Should Be Applied Careful control of finished product water activity is essential for the low-acid foods that do not receive the botulinum process. This control must include a thorough procedure to monitor and standardize the accuracy o

272 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA and precision of the instrument used to measure water activity. Checks on each batch of product prepared are indicated. Microbiological criteria are not applicable. References FDA (Food and Drug Administration) 1983 Thermally processed low-acid foods packaged in hermetically sealed containers. Gen- eral provisions. Definitions. Code of Federal Regulations CFR 21 113.3. M. CEREALS AND CEREAL PRODUCTS Raw cereal grains include wheat, corn, oats, rye, rice, barley, and millet. When processed into flours, meals, flakes, grits, etc., they serve as the basic materials in bakery items, breakfast foods, dry mixes, pastas, and refrigerated doughs. They also are used as ingredients in foods such as canned goods, confectioneries, meat products, and baby foods (APHA, 19761. Soybeans, which are pulses, are included in this section because many soy products are similar to those produced from cereals. Sensitivity of Products Relative to Safety and Quality Grains and Milled Grain Products Under normal growing and harvesting conditions, cereal grains are exposed to a wide variety of microorganisms indigenous to the environ- ment (APHA, 1976; Hesseltine et al., 1978; ICMSF, 1980; Swain et al., 19811. The types and numbers are related to: 1. soil type and species of plant environmental conditions during growth, harvest, transport, and stor age 3. grain quality 4. extent of contamination from man, animals, equipment surfaces and storage facilities 5. abusive handling such as permitting water uptake with subsequent microbial growth ~ , Grains normally contain bacterial populations of about 106/g. A wide variety of species are present, including sporeformers, lactic acid bacteria, pseudomonads, and coliforms (Rogers et al., 1915; APHA, 1976; ICMSF, 19801. These microorganisms are usually of little concern because their growth is prevented by the low water activity of the unprocessed grain.

APPLICATION TO FOODS AND FOOD INGREDIENTS 273 However, many may carry through and survive the milling process and some may be potential spoilers of cereal products. For example, flour may be contaminated with heat-resistant bacterial spores, particularly ther- mophiles, which may survive the process given canned foods. Flour also may be contaminated with mucoid variants of Bacillus subtilis and Bacillus licheniformis which can cause "ropy" bread (Pylor, 1973; ICMSF, 19801: When the spores of these species survive baking, their subsequent ger- mination and growth in the bread results in ropiness, which consists of a soft and sticky interior, brown discoloration, and a peculiar odor. Yeast and mold populations on cereal grains often are as high as 104/g. Various molds, mainly species of Alternaria, Fusarium, Helmin- thosporium, and Cladosporium have the ability to invade kernels while the plants are growing. When grains have been properly dried to 13% water, these field fungi will slowly die off during storage. When grains contain higher levels of moisture due to inadequate drying or storage under wet conditions, molds are the principal spoilage organisms. Members of the Aspergillus glaucus group will grow at water activities as low as 0.73 and thus commonly spoil some of the lower moisture grains. Cereal grains and their milled products have seldom been implicated as sources of foodborne disease. However, low numbers of pathogenic bacteria including salmonellae, Bacillus cereus, Clostridium perfringens, and Clostridium botulinum may be introduced from soil, feces, and other environmental sources. B. cereus, a soil organism commonly found on grains, has been the cause of foodborne disease when cooked rice dishes have been held for extended periods under conditions suitable for spore germination and veg- etative growth (ICMSF, 1980; Bryan et al., 1981~. Soybeans and soy products have, at times, been responsible for sal- monellosis in animals and humans (Elliott, 1967; ICMSF, 19801. Com- mercially prepared tofu (soybean curd) was the source of an outbreak of yersiniosis involving 87 cases; two strains of Yersinia enterocolitica were isolated (Aulisio et al., 19831. Mycotoxins are another potential health hazard since they are produced by some of the fungi that are present on cereals and their milled products. Aflatoxins are of particular concern because they are potent carcinogens; Aspergillus flaws and Aspergillus parasiticus are common contaminants of certain grains (Busby and Wogan, 1979; Labuza, 19831. The prevention of mycotoxins in milled cereal products depends upon the application of HACCP principles (see Chapters 1 and 10), which includes the prevention of mold growth on grains after harvest and the careful inspection of incoming shipments. The observation of fluorescence when kernels are examined under black light indicates mold presence and possible aflatoxin

274 EVALUATION OF THE ROLE OF MICROBIOLOGIC CRITERIA contamination (see Chapter 5), while the finding of pink kernels suggests that vomitoxin (2-deoxydivalenol) may be present (Trenholm et al., 1 98 1 ). Pasta Products There are two basic types of pasta products. The first, noodles, is composed of mainly wheat flour, enrichment nutrients, and eggs or egg yolks, the latter at a level of not less than 5.5% of the total solids content (FDA, 1982~. The second, macaroni, spaghetti, vermicelli, and related pastas' is similar but does not contain the eggs. The ingredients are mixed with water to form a stiff dough of approx- imately 30% moisture, which then is extruded, shaped, cut, and dried. The dough is not heated during its preparation and, traditionally, the subsequent steps are conducted at temperatures that will support microbial growth. Drying to ~ 13% moisture for example, may be at 35 to 40°C (95 to 104°F) for a period as long as 18 hours. Since the dough provides an excellent growth medium, high microbial populations can result, par- ticularly if good sanitation is not practiced during the blending and ex- trusion steps (APHA, 1976; ICMSF, 1980; Swartzentruber et al., 1982~. Newer drying procedures, which involve several stages and higher tem- peratures, have minimized the opportunity for microbial growth. For ex- ample, in one process macaroni is dried 40 minutes at 71°C (159.8°F) followed by 4 hours at 74°C (165.2°F). Staphylococcus aureus and salmonellae can survive the manufacturing process given pasta products (Lee et al., 1975; Walsh et al., 1974; Walsh and Funke, 1975) . The heat resistance of S. aureus and Salmonella anatum has been shown to increase during the drying process, particularly within the intermediate moisture range, thereby increasing the probability that some cells will survive (Hsieh et al., 1976~. Market surveys have revealed that pasta products may contain viable staphylococci and Salmonella (Walsh and Funke, 1975; Rayman et al., 1981; Swartzentruber et al., 1982.) These pathogens die off relatively rapidly during storage of the dried products (Lee et al., 1975) and will be killed by normal cooking although staphylococcal enterotoxins would remain (Denny et al., 1966 Lee et al., 19754. . . Pastries Pastry products are baked doughs that are filled or topped with meat- containing gravies, custards, creams, imitation cream fillings, sauces, and/ or meringues. These materials are often excellent growth media for spoil- age and pathogenic microorganisms (Bryan, 1976; ICMSF, 19801. An

APPLICATION TO FOODS AND FOOD INGREDIENTS 275 exception is the butter cream filling composed of sucrose, water, and shortening that does not support microbial growth due to its low water activity. The different pastry products are distributed and retailed at am- bient, refrigerated, or freezer temperatures. Cream- and custard-filled pastries have been implicated in numerous foodborne disease outbreaks; S. aureus usually has been the causative organism although Salmonella have been responsible at times (see for example CDC, 1981a,b). In recent years the incidence of foodborne ill- nesses from pastries has decreased as a result of educational programs; improvements in manufacturing, sanitation, and refrigeration; and changes in formulations. In the 1940s cream-filled pastries were implicated in 15% of the outbreaks, but by the 1970s they accounted for only 2% (Bryan, 1976), and no foodborne disease outbreaks were attributed to them in 1978 and 1979 (CDC, 1981a,b). Need for Microbiological Criteria Cereal grains and their milled products have seldom been implicated as vehicles of foodborne disease and thus there does not appear to be a need for bacterial standards. On the other hand, because these foods are susceptible to mold growth, regulatory action levels for aflatoxins exist and are appropriate. Similar criteria for other mycotoxins may also be necessary if future studies show that there is a health hazard. Soy products that do not receive a heat treatment should be free of salmonellae and Yersinia. Because grains, flour, grits, and related items are essentially raw ag- ricultural products, and there is little opportunity for microbial growth during their processing, the microbiology of these foods usually does not correlate well with manufacturing practices. Guidelines for these foods, therefore, have little application. Specifications may be useful when flour or some other grain product is used as a food ingredient, e.g., limits on thermophilic spores. Ready- to-eat cereal products, such as bakery items, should be free of enteric pathogens and toxins. The application of the HACCP concept is the best means for assuring that these foods will not be vehicles for foodborne illnesses. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Information that can be utilized for the development of microbiological criteria is available in the literature (APHA, 1976; ICMSF, 1978, 1980;

276 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Rayman et al., 19811. It is advisable for manufacturers to develop addi- tional data through appropriate microbiological surveillance studies. Where Criteria Should Be Applied Criteria for mycotoxins should be applied to grain shipments as received at the mill. Microbiological specifications and guidelines for milled prod- ucts, and especially for foods that contain them, might best be applied at the factory processing level as a component of an ongoing HACCP pro- gram. These criteria should be used for monitoring in-process critical control points as well as finished products for safety, wholesomeness, and stability. References APHA (American Public Health Association) 1976 Compendium of Methods for the Microbiological Examination of Foods. M. L. Speck, ed. Washington, D.C.: APHA. Aulisio, C.C.G., J. T. Stanfield, S. D. Weagant, and W. E. Hill 1983 Yersiniosis associated with tofu consumption; serological, biochemical and patho- genicity studies of Yersinia enterocolitica isolates. J. Food Prot. 46:226-230,234. Bryan, F. L. 1976 Public health aspects of cream-filled pastries. A review. J. Milk Food Technol. 39:289- 296. Bryan, F. L., C. A. Bartleson, and N. Christopherson 1981 Hazard analysis in reference to Bacillus cereus of boiled and fried rice in Cantonese- style restaurants. J. Food Prot. 44:500-512. Busby, W. F., and G. N. Wogan 1979 Foodborne Mycotoxins and Alimentary Mycotoxicoses. In Food-borne Infections and Intoxications. 2nd Ed., H. Riemann and F. L. Bryan, eds. New York: Academic Press. CDC (Centers for Disease Control) 1981a Foodborne Disease Outbreaks. Annual Summary 1978. Atlanta: Centers for Disease Control. 1981b Foodborne Disease Outbreaks. Annual Summary 1979. Atlanta: Centers for Disease Control. Denny, C. B., P. L. Tan, and C. W. Bohrer 1966 Heat inactivation of staphylococcal enterotoxin A. J. Food Sci. 31:762-767. Elliott, R. P. 1967 Bacteriological Problems in the Manufacture of Oilseed Proteins. A presentation at the Conference on Engineering of Unconventional Protein Production. Santa Barbara, Calif.: American Institute of Chemical Engineers. FDA (Food and Drug Administration) 1982 Macaroni and Noodle Products. Code of Federal Regulations 21 CFR 139. Hesseltine, C. W., R. F. Rogers, and R. J. Bothast 1978 Microbiological study of exported soybeans. Cereal Chem. 55(3):332-340.

APPLICATION TO FOODS AND FOOD INGREDIENTS 277 Hsieh, F., K. Acott, and T. P. Labuza 1976 Death kinetics of pathogens in a pasta product. J. Food Sci. 41:516-519. ICMSF (International Commission on Microbiological Specifications for Foods) 1978 Microorganisms in Foods. 1. Their significance and methods of enumeration. Toronto: University of Toronto Press. 1980 Microbial Ecology of Foods. Vol. 2: Food Commodities. New York: Academic Press. Lee, W. H., C. L. Staples, and J. C. Olson, Jr. 1975 Staphylococcus aureus growth and survival in macaroni dough and the persistence of enterotoxins in the dried products. J. Food Sci. 40:119-120. Pylor, E. J. 1973 Baking Science and Technology. 2nd Ed., Vol. 1. Chicago: Seibel Publishing. Rayman, M. K., K. F. Weiss, G. W. Riedel, S. Charbonneau, and G. A. Jarvis 1981 Microbiological quality of pasta products sold in Canada. J. Food Prot. 44:746-749. Rogers, L. A., W. M. Clark, and A. C. Evans 1915 The characteristics of bacteria of the colon type occulting on grains. J. Infect. Dis. 17: 137-159. Swain, E. W., H. L. Wang, and C. W. Hesseltine 1981 Heat-Resistant Aerobic Bacterial Spores in Soybeans. Peoria, Ill.: Northern Regional Research Center, ARS, USDA. Swartzentruber, A., W. L. Payne, B. A. Wentz, R. J. Barnard, and R. B. Read, Jr. 1982 Microbiological quality of macaroni and noodle products obtained at retail markets. Appl. Environ. Microbiol. 44:540-543. Trenholm, H. L., W. P. Cochrane, H. Cohen, J. I. Elliot, E. R. Farnworth, D. W. Friend, R.M.G. Hamilton, G. A. Neish, and J. F. Standish 1981 Survey of vomitoxin contamination of the 1980 white winter wheat crop in Ontario, Canada. J. Amer. Oil Chem. Soc. 58(12):992A-994A. Walsh, D. E., and B. R. Funke 1975 The influence of spaghetti extruding, drying and storage on survival of Staphylococcus aureus. J. Food Sci. 40:714-716. Walsh, D. E., B. R. Funke, and K. R. Graalum 1974 Influence of spaghetti extruding conditions, drying and storage on the survival of Salmonella typhimurium. J. Food Sci. 39: 1105- 1106. N. FATS AND OILS Sensitivity of Products Relative to Safety and Quality Fats and oils per se do not support growth of microorganisms. However, when in the presence of moisture and other essential nutrients, these products are subject to degradation by a variety of microorganisms. Such a condition is provided by several processed foods in which fats and oils are the major ingredients. Among such products, mayonnaise, salad dress- ings, peanut butter, margarine, and butter are of concern. Peanut butter differs from mayonnaise, salad dressing, margarine, and butter in that the  water content (aw) is too low to permit microbial growth and spoilage. If moisture is permitted to deposit on the surface of peanut butter, mold growth might, however, be possible.

278 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Each of these products is an emulsion. Butter, margarine, and peanut butter represent foods that are emulsions comprised of oil (or fats as the continuous phase and water as the discontinuous phase. Mayonnaise, salad dressing, and related products are emulsions in which water is the con- tinuous phase and fat the discontinuous phase. The form of emulsion has a profound relationship to microbial stability. With oil-in-water emulsions such as mayonnaise the growth rate of mi- croorganisms is not affected by the water dispersion; only the chemical composition of the water phase plays a role. On the other hand, with water-in-oil emulsions such as butter the water exists as microscopic drop- lets that are dispersed throughout the oil matrix. If microorganisms are present in these droplets, their growth is restricted by limited water and food supply as well as by inhibitory factors in the aqueous phase, such as salt, preservatives, and organic acids (ICMSF, 19801. Mayonnaise and Salad Dressings The microbiology of these products has been reviewed by Smittle ( 1977) and ICMSF (19801. Mayonnaise and salad dressings are defined by stan- dards of identity (FDA, 1978a,b). The safety of these products relates directly to the pH (4.1 or below) and the acetic acid content (approximately 0.3-1.2%) of the moisture phase. Egg is the sensitive ingredient in mayonnaise and salad dressings because of the potential for Salmonella contamination from this source. However, salmonellae, if present, in properly prepared products die in a matter of days due to the low pH and the acetic acid content. Also, growth of Clostridium botulinum, Clostridium perfringens, Bacillus cereus, and Staphylococcus aureus is prevented. Thus the hazard presented by properly prepared products is remote. Unfortunately, the federal standards of identity for mayonnaise and for salad dressings do not specify a pH of 4.1 or below for either of these products. In the case of mayonnaise, the acidity of vinegar or of vinegar diluted with water that is used in its preparation must be not less than 2.5% calculated as acetic acid. Presumably a pH of 4.1 in the aqueous phase of a properly formulated product would be assured. No specific concentration of acetic acid is specified in the standard of identity for salad dressings nor is a pH level specified. Thus, assurance of safety depends primarily upon the manufacturer's understanding of the necessity to produce a product having a pH of 4.1 or below. Neither standard of identity as presently written would guarantee that such a pH level would be obtained. In some mayonnaise manufactured in Europe the acidifying agent is

APPLICATION TO FOODS AND FOOD INGREDIENTS 279 lemon or lime juice or hydrochloric acid instead of vinegar (acetic acid). Citric and hydrochloric acids are far less active than acetic acid against . . microorganisms . In 1976 the World Health Organization cooperated with the Spanish authorities in investigating outbreaks of salmonellosis among passengers on four flights between Las Palmas, in the Canary Islands, and Helsinki. The flights were provisioned with egg mayonnaise. Salmonella typhi- murium was isolated from several patients and one food handler and a batch of mayonnaise used for one of the meals. Many hundreds of victims were severely ill; there were six deaths (Davies and Wahba, 19761. The hazards of improperly prepared mayonnaise are further illustrated by the following: In Denmark in 1975 mayonnaise from a salad factory caused approximated 10,000 cases of salmonellosis. A circular was pre- pared by the Danish Administration recommending procedures for pre- paring mayonnaise so that the pH of the product would be lower than 4.5. These recommendations involved a certain proportion of egg yolk in vinegar and a holding time before use of either room temperature for four days or 40°C (104°F) for at least two hours. Since then no cases of salmonellosis caused by commercially prepared mayonnaise have occurred in Denmark (B. Simonsen, 1980. Personal communication). For a de- scription of additional Danish outbreaks, see ICMSF (19801. Mayonnaise and salad dressings are subject to microbial spoilage (Smit- tle, 1977; ICMSF, 19801. Generally, yeasts and lactobacilli are the spoil- age organisms of concern. They may be introduced originally from contaminated ingredients. Small numbers of them usually are held in check by the acidity of the finished product. However, faulty sanitary practices during manufacture, i.e., failure to properly clean and sanitize pumps, mixing and filling equipment, and other product contact surfaces, may easily introduce numbers sufficiently great to bring about spoilage. Fre- quently, however, malfunctions of equipment rather than improper clean- ing may be the direct cause of spoilage. For example, a faulty seal or gasket on a pump or filler may lead to accumulation of product behind the pump or filler housing. The normal sanitation procedures would require that the face of the pump or filler be cleaned and sanitized, the assumption being made that the portion of the equipment behind the face is sealed by a properly functioning gasket. With malfunction, product accumulates behind the face and may result in inoculation of the product with inor- dinately high numbers of spoilage organisms despite adequate normal cleaning. In this type of operation, periodic breakdown of equipment to determine the proper function, i.e., preventive maintenance, is a critical control point. Thus, the critical control points are equipment sanitation and, particularly, the proper functioning of the equipment. The latter can

280 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA best be controlled by periodic breakdown of various key pieces of equip- ment to verify their proper functioning. Peanut Butter The federal standard of identity for peanut butter (FDA, 1978c) requires that it be made from ground shelled roasted peanuts to which may be added suitable seasoning and stabilizing ingredients not to exceed 10% by weight of the finished product. The fat content must not exceed 55%. The water content is less than 1%. For a review of the microbiology of peanut butter, see ICMSF ( 19801. Lack of available moisture (low aw) in peanut butter prevents growth of microorganisms in the product. Thus, foodborne illness caused by toxigenic microorganisms such as S. aureus, C. perfringens, C. botulinum, and B. cereus is precluded. Occasionally Salmonella as well as other microorganisms occur in pea- nut butter, although no foodborne illness outbreaks have been reported. The source of contaminating organisms is almost always raw peanuts, which may contain Salmonella and a variety of other microorganisms, some of which are of public health concern. The roasting process destroys Salmonella, as well as other non-sporeforming bacteria. If, however, cross-contamination between the raw and roasted peanuts were to occur, it would not be catastrophic, since the occasional Salmonella (much as in inert material) would be "diluted out" in the finished product since the moisture content of the finished product is sufficiently low as to prevent Salmonella growth. Problems arise when processing equipment used after the roaster process is wet-cleaned or if moisture is not completely con- trolled in the processing equipment, i.e., condensate from pipes or leaking water valves permitting water to accumulate on floors or on the surfaces of equipment, which would permit Salmonella growth. Control of the processing environment is the primary critical control point in the manufacture of peanut butter. This control requires physical separation of raw and roasted peanuts to prevent contamination of the processing environment with microorganisms present in the raw peanuts (in particular Salmonella). It requires further the complete control of mois- ture in the processing environment. If wet cleanup becomes necessary, this must be undertaken with extreme care, with provision being made to rapidly dry areas that have been subject to cleaning and sanitizing. Yet another concern is the occurrence of aflatoxin in the raw peanuts. Control is generally accomplished by pretesting incoming shipments of raw peanuts and rejecting those showing aflatoxin levels greater than the accepted level (20 ppb). Further control is generally exercised by electronic

APPLICATION TO FOODS AND FOOD INGREDIENTS 281 sorting of the roasted peanuts, a process that tends to eliminate peanuts with high levels of aflatoxin. Although the likelihood of peanut butter being involved in Salmonella outbreaks appears remote, manufacturers would be well advised to rou- tinely test production batches for Salmonella. Plant guidelines for aerobic plate count, coliform, and Escherichia cold can be useful as indicators of cross-contamination. Margarine and Butter For a review of the microbiology of these products see ICMSF (19801. The microbiological stability of margarine and butter is dependent upon heat treatment (pasteurization) of the ingredients, prevention of postpas- teurization contamination and minute dispersion of water droplets in the fat (continuous) phase of the finished products. In the case of margarine, stability is further enhanced by certain preservatives that are permitted by the standard of identity for this product (FDA, 1978d). Failure to achieve the above conditions may lead to growth of spoilage organisms as well as pathogens. Reported outbreaks involving these products are rare. Whipped butter and high moisture margarines are more susceptible to microbial growth due primarily to poorer dispersion of water in the fat phase. Two recent staphylococcal foodborne illness outbreaks due to whipped butter have occurred (CDC, 1970, 19771. Studies that have demonstrated the ability of S. aureus to grow in mildly salted whipped butter (Minor and Marth, 1972) point up the hazard that this product may present. Current technology of butter and margarine manufacture can virtually assure a relatively stable product with excellent keeping quality. On occasion movement of butter or margarine from refrigerated to warmer environ- ments may lead to the deposition of moisture on exposed surfaces. This can result in development of mold. Need for Microbiological Criteria Mayonnaise and Salad Dressings There is little if any justification for imposing a microbiological criterion for Salmonella or for any other pathogen. However, control of pH is critical to the safety of these products. In view of the many different manufacturers of mayonnaise and salad dressings and especially the ever increasing types of salad dressings available to the public, it would seem prudent to amend the standards of identity for these products to include a requirement for a pH of 4.1 or below in the aqueous phase.

282 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Microbiological guidelines applicable at critical control points of man- ufacture should be useful. In addition to the APC, yeasts and molds and lactobacilli counts are applicable to evaluate plant sanitation and to detect potential spoilage problems (Kurtsman and Smittle, 19841. Direct micro- scopic examination of product is useful when product spoilage is being investigated. Peanut Butter A standard with respect to Salmonella would be applicable for peanut butter, this being a product that is generally consumed without cooking and is, indeed, a product that is apt to be consumed by the very young and the aged, groups highly susceptible to Salmonella infection. Margarine and Butter The rarity of involvement in outbreaks indicates that little would be served by imposing any regulatory criteria for butter and margarine, not withstanding the current USDA recommended criteria (see Chapter 8, Table 8-4~. However, in-plant guidelines for APC and for S. aureus, particularly in the case of whipped butter, should be applied at critical control points of manufacture. Where Criteria Should Be Applied Microbiological guidelines are most usefully applied at the processing plant level at the critical control points for each of the products as identified earlier in this section. References CDC (Center for Disease Control) 1970 Staphylococcal food poisoning traced to butter Alabama. Morb. Mort. Weekly Rpt. 19:271. 1977 Presumed staphylococcal food poisoning associated with whipped butter. Morb. Mort. Weekly Rpt. 26:268. Davies, R. F., and A. H. Wahba 1976 Salmonella infections of charter flight passengers. Report on a visit to Spain (Canary Islands) February 26th-March 2, 1976. Copenhagen: World Health Organization, Regional Office, Europe. FDA (Food and Drug Administration) 1978a Mayonnaise. Code of Federal Regulations 21 CFR 169.140. 1978b Salad Dressings. Code of Federal Regulations 21 CFR 169.150. Washington, D.C.: U.S. Government Printing Office.

APPLICATION TO FOODS AND FOOD INGREDIENTS 283 1978c Peanut Butter. Code of Federal Regulations 21 CFR 164.150. Washington, D.C.: U.S. Government Printing Office. 1978d Margarine. Code of Federal Regulations 21 CFR 166.110. Washington, D.C.: U.S. Government Printing Office. ICMSF (International Commission on Microbiological Specifications for Foods). 1980 Microbial Ecology of Foods. Vol. 2. Food Commodities. New York: Academic Press. Kurtzman, C. P., and R. B. Smittle 1984 Salad dressings. In Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed. M. L. Speck, ed. Washington, D.C.: APHA. Minor, T. E., and E. H. Marth 1972 Staphylococcus aureus and enterotoxin A in cream and butter. J. Dairy Sci. 55: 1410- 1414. Smittle, R. B. 1977 Microbiology of mayonnaise and salad dressing: A review. J. Food Prot. 40:415- 422. O. SUGAR, COCOA, CHOCOLATE, AND CONFECTIONERIES Sensitivity of Products Relative to Safety and Quality These products have been involved only infrequently in foodborne dis- ease outbreaks (ICMSF, 19801. Although the risks are low, the potential of such products to lead to major outbreaks cannot be ignored. Witness to this was the 1973-1974 outbreak of salmonellosis involving chocolate candy (D'Aoust, 19771. Microbial spoilage of these commodities is rare due to low water ac- tivities, although specific problems involving yeasts, molds, and spore- formers have been documented (ICMSF, 19801. The HACCP system should be applied in the production of cocoa, chocolate products, and confections. Critical control points should be monitored microbiologically when indicated. Sugar Because of the source and nature of sugar, microorganisms are intrins- ically present. These organisms include thermophilic sporeformers such as Bacillus stearothermophilus, Bacillus coagulans, and Clostridium ther- mosaccharolyticum. The spores of these organisms, while generally in- nocuous under usual storage conditions, have proved to be a nuisance to the canning industry. Thus, the National Food Processors Association (NCA, 1972) and other groups have developed specifications for bacterial spores in sugar. Liquid sugar is likely to give rise to problems with growth of osmophilic yeasts.

284 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Cocoa and Chocolate Products The predominant microorganisms in cocoa are Bacillus, yeasts, and molds (Gabis et al., 1970; ICMSF, 19801. The number of these organisms depends on the inherent contamination and control of the process. The level of heat-resistant spores found in cocoa (Gabis et al., 1970) should be controlled because they may cause spoilage in high moisture products where cocoa is an ingredient, for example, chocolate milk. Salmonella in cocoa (Collins-Thompson et al., 1978; D'Aoust, 1977) make it a sensitive ingredient. When contaminated, the numbers of sal- monellae are usually low, less than 1/g (D'Aoust, 1977; ICMSF, 19801. Milk chocolate made from contaminated cocoa was responsible for over 200 cases in 1973-1974. In this outbreak, the contaminated chocolate contained 20-90 cells of S. eastbourne per 100 g (D'Aoust et al., 1975; Craven et al., 19751. Thus, low levels of Salmonella can cause human infection (see Chapter 4, Table 4-31. Organisms capable of surviving the processes used in chocolate man- ufacture belong primarily to the genus Bacillus (Barrile et al., 1971; Collins-Thompson et al., 19811. The key to low spore counts in chocolate appears to be the quality of the cocoa bean. The low water activity of chocolate does not permit outgrowth of the contaminating organisms but may give rise to spoilage problems when chocolate is used as an ingredient in foods with higher water activities. Sources of Salmonella in chocolate have been traced back to cocoa (D' Aoust, 1977) and milk powder. The final heating (conching) cannot be relied upon to kill all Salmonella in the product. There are no steps in the manufacturing operation that assure destruction of microorganisms. The water activities of both the ingredients and the finished chocolate preclude microbial growth. Wet cleaning should not be employed in chocolate and confectionery manufacture. Confectionery The multitude of ingredients in confectionery including sugar, milk products, egg albumin, gums, nuts, fruits, and spices can give rise to microbial spoilage problems. Many of these are kept in check, however, as a result of low water activity. Fondants and other products vulnerable to spoilage have water activities in the 0.75-0.85 range. The spoilage of such products is usually caused by osmophilic yeasts (ICMSF, 19801. Residual lipolytic enzymes of microbial origin may give rise to soapy off- flavors in chocolate-containing confectionery after extended storage. While sugar has not been shown to be a source of Salmonella, other

APPLICATION TO FOODS AND FOOD INGREDIENTS 285 confectionery ingredients such as dried eggs, milk, coconut, and gelatin are classified as critical raw materials because they may contain Salmonella (D'Aoust, 19771. Also, there are no steps in most confectionery processing to destroy these microorganisms. Another safety concern with confectionery is the presence of mycotoxins in nuts. Confectionery products containing aflatoxin-contaminated nuts in excess of aflatoxin standards are subject to regulatory action. Since it is impossible to remove these toxins by heating, the raw ingredients should be inspected for mycotoxins as part of an ongoing HACCP program (see part S of this chapter). Needs for Microbiological Criteria and Where Criteria Should Be Applied Sugar Microbiological criteria are important for sugar to be used as an ingre- dient in the canning, beverage, and confectionery industries. The canning industry has specifications for thermophilic, flat-sour, and sulfide spoilage spores (NCA, 19721. The bottling industry has developed criteria for mesophilic bacteria, yeasts, and molds (National Soft Drink Association, 19751. The specification for sugar to be used in soft drinks are ~ 10 yeasts and ~ 10 molds per 10 grams of sugar (Anonymous, 19711. These spec- ifications appear adequate for the purposes intended. In the manufacture of confectioneries, liquid sugar is a critical control point because of the potential for growth of osmophilic yeasts. Therefore, microbiological cri- teria are applicable as part of a HACCP program. Cocoa and Chocolate Various microbiological criteria are useful (ICMSF, 19801. For ex- ample, specifications for ingredients of retorted products such as cocoa beverages should include limits for thermophilic spores (NCA, 19721. It is appropriate to examine cocoa and chocolate for Salmonella. Sam- pling plans should be based upon the intended use of these products (NRC, 19691. The HACCP program is recommended for chocolate production. Raw material inspection and testing to ensure that the incoming ingredients are Salmonella-free are necessary components of the HACCP system as the usual processing steps will not destroy Salmonella. The plant environment must be maintained free of Salmonella. Appropriate points should be monitored by regular collection and analyses of environmental samples.

286 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Confectionery Microbiological criteria are appropriate when applied to ingredients used in confectionery products. Guidelines for osmophilic yeasts are applicable for soft-centered chocolates, which are particularly prone to fermentation. Critical ingredients such as cocoa, coconut, dried milk, and egg albumin should be examined for Salmonella. Tolerances for aflatoxins are appli- cable for products containing nuts (see part S). As with the production of chocolate products, the environment is a critical control point (see above). References Anonymous 1971 Beverage Production and Plant Operation. Washington, D.C.: National Soft Drink Association. Barrile, J. C., K. Ostovar, and P. G. Keeney 1971 Microflora of cocoa beans before and after roasting at 150°C. J. Milk Food Technol. 34: 369-371. Collins-Thompson, D. L., K. F. Weiss, G. W. Riedel, and S. Charbonneau 1978 Sampling plans and guidelines for domestic and imported cocoa from a Canadian national microbiological survey. Can. Inst. Food Sci. Technol. J. 11: 177-179. Collins-Thompson, D. L., K. F. Weiss, G. W. Riedel, and C. B. Cushing 1981 Survey of and microbiological guidelines for chocolate and chocolate products in Canada. Can. Inst. Food Sci. Technol. J. 14: 203-208. Craven, P. C., D. C. Mackel, W. B. Baine, W. H. Barker, E. J. Gangarosa, M. Goldfield, H. Rosenfeld, R. Altman, G. Lachapelle, J. W. Davies, and R. C. Swanson 1975 International outbreak of Salmonella eastbourne infection traced to contaminated chocolate. Lancet 1: 788-793. D'Aoust, J. Y. 1977 Salmonella and the chocolate industry. A review. J. Food Prot. 40: 718-727. D'Aoust, J. Y., B. J. Aris, P. Thisdele, A. Durante, N. Brisson, D. Dragon, G. Lachapelle, M. Johnston, and R. Laidley 1975 Salmonella eastbourne outbreak associated with chocolate. Can. Inst. Food Sci. Tech- nol. J. 8:181-184. Gabis, D. A., B. E. Langlois, and A. W. Rudnick 1970 Microbiological examination of cocoa powder. Appl. Microbiol. 20: 644-645. ICMSF (International Commission on Microbiological Specifications for Foods) 1980 Sugar, cocoa, chocolate, and confectioneries. Pp. 778-821 in Microbial Ecology of Foods. Vol. 2. Food Commodities. New York: Academic Press. National Soft Drink Association 1975 Quality Specifications and Test Procedures for Bottlers: Granulated and Liquid Sugar. Washington, D.C.: National Soft Drink Association. NCA (National Canners Association) 1972 Bacterial Standards for Sugar. June 1972. Revised. Washington, D.C.: NCA. NRC (National Research Council)  1969 An Evaluation of the Salmonella Problem. Committee on Salmonella. Publication 1683. Washington, D.C.: National Academy of Sciences.

APPLICATION TO FOODS kD FOOD INGREDIENTS P. SPICES 287 Spices are the dried parts of plants which are used mainly as flavoring agents for foods. Some, such as black pepper in a household shaker, may be added to a food immediately before consumption (table spice), while others may be used as ingredients of foods that will receive further proc- ess~ng. Sensitivity of Products Relative to Safety and Quality AS raw agricultural commodities, uncleaned and untreated spices com- monly contain high microbial populations (APHA, 1976; ICMSF, 1980; Julseth and Deibel, 1974; Schwab et al., 19821. Spice-bearing plants are exposed to a wide variety of microorganisms from the environment in which they are grown and harvested. Furthermore, during drying, often a simple procedure utilizing the sun, significant microbial growth may occur. Numerous species of bacteria, yeasts, and molds constitute the nodal microflora of dried spices (Christensen et al., 1967; Julseth and Deibel, 1974; Schwab et al., 1982), although aerobic sporeforming bac- teria usually predominate (APHA, 1976; ICMSF, 19801. While bacteria generally are not responsible for the spoilage of spices, growth of fungi during storage and shipping may result in the development of off-flavors and in the production of undesirable enzymes (ICMSF, 19801. Spices can be a source of spoilage microorganisms when they are used as seasonings for processed foods. For example, they may introduce heat- resistant spores into canned foods and a variety of spoilage bacteria into refrigerated products such as processed meats (ICMSF, 19801. Bacteria of public health significance, e.g., Clostridium pe~fringens (Krishnaswamy et al., 1971; Leitao et al., 1975; Powers et al., 1975), Bacillus cereus (Kim and Goepfert, 1971; Powers et al., 1976), Esche- richia cold (Kadis et al., 197 1), Staphylococcus aureus (Julseth and Deibel, 1974; Schwab et al., 1982) and Salmonella (ICMSF, 1980; Laidley et al., 1974; Health and Welfare Canada, 1983) may be present but usually in low numbers. Mycotoxin-generating molds also have been isolated from spices (Christensen et al., 1967) and low levels of aflatoxin, usually under 10 Agog, have been detected (Scott and Kennedy, 1975~. Although contaminated spices have not been major causes of foodborne disease, a potential hazard exists, especially if the spice is to be used as a table condiment. Contaminated pepper, for example, was responsible for at least 126 cases of salmonellosis (CDC, 19821.

288 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Need for Microbiological Criteria The end use for a spice will determine the type of criterion that might be needed. Table spices and those used as ingredients of processed foods that do not receive a subsequent lethal treatment should be free of infectious pathogens such as Salmonella; therefore they should be tested for this pathogen. Guidelines may be useful for monitoring processes applied to spices, in particular the treatment used to destroy much of the intrinsic microflora. Exposure to ethylene oxide gas is the common process, although there is concern at present about some of the products that are formed. The En- vironmental Protection Agency has set residue tolerances of 50 ppm eth- ylene oxide in Heated spices (EPA, 1982~. Treatment of spices with ionizing irradiations, a maximum of 1 megarad, has recently been approved (FDA, 1983). References APHA (American Public Health Association) 1976 Compendium of Methods for the Microbiological Examination of Foods. M. L. Speck, ed. Washington, D.C.: APHA. CDC (Centers for Disease Control) 1982 Outbreak of Salmonella oranienburg-Norway. Morb. Mort. Weekly Rpt. 31(48):655- 656. Christensen, C. M., H. A. Fanse, G. H. Nelson, F. Bates, and C. J. Mirocha 1967 Microflora of black and red pepper. Appl. Microbiol. 15:622-626. EPA (U.S. Environmental Protection Agency) 1982 Ethylene oxide; tolerances for residues. Code of Federal Regulations 40 CFR 180.151. FDA (Food and Drug Administration) 1983 Irradiation in the production, processing and handling of food. Final rule. Federal Register (48) 129:30613-30614. July 5. Health and Welfare Canada 1983 Salmonella in spices and chocolate. Canada Diseases Weekly Report 9(9):35-36. ICMSF (International Commission on Microbiological Specifications for Foods) 1980 Microbial Ecology of Foods. Vol. 2. Food Commodities. New York: Academic Press. Julseth, R. M., and R. H. Deibel 1974 Microbial profile of selected spices and herbs at import. J. Milk Food Technol. 37:414- 419. Kadis, V. M., D. A. Hill, and K. S. Pennifold 1971 Bacterial content of gravy bases and gravies obtained in restaurants. Can. Inst. Food Technol. J. 4: 130-132. Kim, H. U., and J. M. Goepfert 1971 Occurrence of Bacillus cereus in selected dry food products. J. Milk Food Technol. 34: 12-15. Krishnaswamy, M. A., J. D. Patel, and N. Parthasarathy  1971 Enumeration of microorganism in spices and spice mixtures. J. Food Sci. Technol. 8: 191-194.

APPLICATION TO FOODS AND FOOD INGREDIENTS 289 Laidley, R., S. Handzel, D. Severs, and R. Butler 1974 Salmonella weltevreden outbreaks associated with contaminated pepper. Epidemiol. Bull. 18:62. Ottawa: Dept. Natl. Health Welfare. Leitao, M. F., I. Delazar~, and H. Mazzoni 1975 Microbiology of dehydrated foods (Coletanea Inst. Technol. Aliment. 5:223-241). Tech. Abstr. from Food Science 7, 9 B72. Powers, E. M., R. Lawyer, and Y. Masuoka 1975 Microbiology of processed spices. J. Milk Food Technol. 38:683-687. Powers, E. M., T. G. Latt, and T. Brown 1976 Incidence and levels of Bacillus cereus in processed spices. J. Milk Food Technol. 39:668-670. Schwab, A. H., A. D. Harpestad, A. Swartzentruber, J. M. Lanier, B. A. Wentz, A. P. Duran, R. J. Barnard, and R. B. Read, Jr. 1982 Microbiological quality of some spices and herbs in retail markets. Appl. Environ. Microbiol. 44:627-630. Scott, P. M., and B.P.C. Kennedy 1975 The analysis of spices and herbs for aflatoxins. Can. Inst. Food Sci. Technol. J. 8: 124-125. Q. YEASTS Sensitivity of Products Relative to Safety and Quality Baker's yeast is grown in open fermenters and thus during propagation may be contaminated with low numbers of lactic acid bacteria and some- times with coliforms including Escherichia cold (Reed and Peppier, 19731. Following the fermentation, the cells are centrifuged and washed several times and then concentrated to about 30% solids using presses or filters. The yeast may then be packaged and marketed as compressed yeast cake or it may be dried at a low temperature to produce active dry yeast. A third product is nutritional yeast, which consists of cells killed by drying at high temperatures on drum dryers. Both live yeast and nutritional yeast have on occasion been contaminated with salmonellae (NRC, 19751. They may not be present in the fermenters but can be introduced at subsequent processing steps. The dead yeast cells that collect on equipment surfaces constitute an excellent growth medium for bacteria and thus soiled equipment may be a significant source of contamination (Reed and Peppier, 19731. Since nutritional yeast is heated to a high temperature during drum drying, the presence of viable sal- monellae represents contamination following this treatment. The spores of certain bacilli are another concern because their intro- duction into dough may result in ropy bread.

290 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA Need for Microbiological Criteria Many yeast companies have established internal guidelines for levels of bacteria, wild yeasts, and molds. These guidelines aid in assessing conditions of sanitation in the fermenters and during subsequent process- ~ng. The testing of yeast products for Salmonella is appropriate, and specific methods have been developed (FDA, 1978~. The presence of the pathogen in nutritional yeast is of special concern since this food may not be heated prior to consumption. Baking, on the other hand, would destroy the sal- monellae introduced with the active dry or compressed yeast. References FDA (Food and Drug Administration) 1978 FDA Bacteriological Analytical Manual. Washington, D.C.: Association of Official Analytical Chemists. NRC (National Research Council) 1975 Prevention of Microbial and Parasitic Hazards Associated with Processed Foods: A Guide for the Food Processor. Committee on Food Protection. Washington, D.C.: National Academy of.Sciences. Reed, G., and H. J. Peppler 1973 Yeast Technology. Westport, Conn.: AVI Publishing. R. FORMULATED FOODS Sensitivity of Products Relative to Safety and Quality Formulated foods have been described as commercially prepared, ready- to-cook or ready-to-eat foods containing major ingredients from two or more commodity categories (ICMSF, 19851. The potential hazards of the major ingredients of these foods are discussed in other sections of this chapter. The combining of these ingredients into a single product presents not only the original hazards of each ingredient, but also the possibility of magnified or additional hazards due to further handling, process ~n- novations, or modification of the environment. For example, the chicken, gravy, vegetables, and spices in a chicken potpie carry their own hazards and also carry additional potential hazards due to increased handling, possible temperature abuse during the time spent in such handling, and the possibility that a microbiological situation of no concern in the original ingredients is of concern in the formulated food. Beef and chicken potpies have been the cause of botulism as a result of gross abuse by the consumer

APPLICATION TO FOODS AND FOOD INGREDIENTS 291 when heated pies were held at room or slightly higher temperatures for extensive periods of time (CDC, 1960; State of California, 1975, 19761. Formulated foods have increased in number, diversity, and volume to meet the public's desire for more convenience. Today less food is being prepared from the basic ingredients in the home kitchen where it is likely to be consumed soon thereafter and where any outbreaks are likely to be limited to a small number of consumers. In contrast, increased amounts of food are now produced in factories and held for some time in distribution channels before they are eaten. Under these conditions there is increased potential for abuse and any foodborne illnesses will most likely affect a much larger number of consumers. The responsibility for preserving safety and keeping quality of foods lies with the processor or formulator of the food. As a result of the increased number and volume of formulated foods, this responsibility has shifted somewhat from the individual in the home to the commercial processor. Because of large differences in composition, preparation, and storage of the many formulated foods, each formulated food or group of foods presents particular circumstances relative to assessment of needs for microbiological criteria (see Chapter 21. Criteria for some formulated foods have been discussed by the ICMSF (1984~. However, no general all- product criteria can be developed. Instead, the potential hazards of each product or group of products must be identified and appropriately moni- tored if microbial growth and resulting health hazard or spoilage are to be prevented during the time the product is under producer control or in appropriate storage. The HACCP systems may include control points in addition to those developed for the separate ingredients. (See for example the discussion of critical control points for frozen foods by Peterson and Gunnerson, 1974.) Means of preventing growth during storage depend on the conditions of processing, the temperature of storage, and the composition of the product. For example, some formulated foods, e.g., beef stew, liquid infant formula, chili, and tamales, are sterilized in a sealed container. These products are shelf-stable and need no more microbiological criteria than do other low-acid canned foods (see part J). Other formulated foods are simple blends of dry ingredients, e.g., dried infant formula, soup mix, or cake mix. These foods also are shelf-stable, but they can be hazardous if an ingredient is contaminated with a path- ogenic organism such as Salmonella. Both the FDA and the international Codex Alimentarius Commission recognize the need for the imposition of stringent standards on formulated foods designed for use by high-risk populations. The FDA is the more stringent of the two in the application of microbiological criteria to this broad group of products.

292 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA The FDA prescribed (FDA, 1978) analytical procedures for ingredients such as dried yeast, cheese, and pasta frequently used in blended for- mulated foods. While proper use of such formulated foods or their ingre- dients would destroy salmonellae, regulatory concern with respect to cross- contamination or inadequate heat treatment exists. Numerous recalls of dried blended soups and other foods have occurred and standards or im- plied standards for salmonellae have been imposed. The Codex Alimentarius Committee on Food Hygiene concluded that microbiological criteria are not justified in the case of dried soup mixes. However, the committee has established microbiological end product spec- ifications for some products that do require special attention either because they contain pathogenic organisms that may survive cooking procedures or because they are typically eaten by highly susceptible persons such as the very young, the aged, or those who are ill (Codex Alimentarius Com- mission, 1979a). These products include coated or filled dried shelf-stable biscuits for which criteria for coliforms and Salmonella were imposed, and also dried and instant products requiring reconstitution and dried products requiring heating to boiling before consumption for which criteria for APC, coliforms, and Salmonella were established. A few formulated foods can be stored at room temperature because of their high acidity, e.g., mustard dressings, or their low water activity, e.g., certain pastries with butter cream fillings (see part M). Most formulated foods are preserved by storage at low temperatures. Some are simply refrigerated, e.g., meat and seafood salads and sand- wiches, but the majority are frozen, e.g., dishes such as meat potpies, egg rolls, pizza, and enchiladas. Microbiological hazards are diminished if the products are thoroughly heated before consumption. An Example: Ready-to-eat Salads Ready-to-eat salads are one class of formulated foods for which mi- crobiological criteria are often considered. Most, but not all, of these salads have four characteristics in common: (1) they contain at least one raw ingredient, (2) they are ready-to-eat when purchased, (3) they are exposed to contamination during preparation, and (4) they are normally preserved by refrigeration but often suffer temperature abuse. Clearly, products with these characteristics are potential public health problems. Particularly vulnerable are those ready-to-eat salads containing high-pro- tein foods such as diced or chopped meat, seafood, poultry, cheese, boiled eggs, and/or cooked potatoes. Salads made of diced ham, chicken, egg, or potato have been responsible for many outbreaks of foodborne disease during the past several decades (see, for example, CDC, 19811. Salads

APPLICATION TO FOODS AND FOOD INGREDIENTS 293 composed of raw celery, onion, peppers, carrots, or other vegetables are less likely to present microbial hazards. Other types of salads include diced fruit, gelatin with marshmallows, cooked macaroni with cheese or canned peas, and shredded cabbage (core slaw). The variations in com- position are almost limitless. The microbial content of salads (or delicatessen foods, as they are sometimes called) has received a great deal of attention from regulatory agencies. Several microbiological criteria in the form of standards or guidelines have been established (Wehr, 19824. Unfortunately, these cri- teria usually do not reflect the basic differences between products. It is common for an agency to specify a single set of limits and requirements for all salads regardless of the composition of each. For example, the APC limit for all types of salads may be to not exceed 100,000 per gram while the coliform count may be to not exceed 100 per gram. Criteria such as these which fail to recognize the wide diversity of ingredients that go into salads are ill-advised. The natural microflora of raw vegetables, for example, can easily exceed 1,000,000 per gram. These organisms are perfectly harmless, yet salad manufacturers may have to soak vegetables in chlorine solution in order to meet an arbitrary limit of 100,000 per gram. For the same reason, a manufacturer may have to use processed cheese with typically lower counts rather than natural cheese in order to meet a general microbiological criterion. On the other hand, a limit of 100,000 per gram may be high for gelatin salads or other products made entirely of cooked ingredients. In summary, then, microbiological criteria may be useful in revealing health hazards and breaches of good food-handling practice with certain types of salad items. Both the criteria and the types of salads to which they are applied must be carefully selected if they are to be meaningful. The current practice of applying the same criteria to all salads regardless of composition and potential hazard of each salad type reveals a lack of understanding of both the potential health hazards of salads and of ap- propriate solutions. Need for Microbiological Criteria Owing to the wide diversity of composition, preparation, and storage conditions, it is not possible to develop a single set of microbiological criteria that are suitable for all formulated foods. However, criteria have been established and are appropriate for certain classes of formulated foods as discussed above. Establishment of criteria should be preceded by con- sideration of the product as outlined in the foregoing section so as to

294 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA decide what, if any, microbiological criteria would offer protection against health hazards and poor quality. Assessment of Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Factors that determine the potential usefulness of microbiological cri- teria with formulated foods include: 1. The nature and extent of contamination: Ingredients that are tradi- tionally associated with salmonellae, for example, should be subjected to routine testing for these organisms, especially if the product is likely to be consumed without thorough cooking. 2. Likelihood of contamination during preparation: Any product that is subjected to handling can be expected to contain a few coagulase- positive staphylococci; excessive numbers suggest breaches of good man- ufacturing practice. 3. Bactericidal treatmentts) during processing: Commercial pasteuri- zation or sterilization will eliminate all pathogenic organisms from a food. Hence, the application of microbiological criteria to products that have undergone a terminal heat treatment usually is not cost-effective. Efficacy of heating can better be monitored by measuring temperature or enzyme destruction or by other appropriate means. Microbiological criteria may be appropriate, however, if there is opportunity for recontamination with pathogens after heating. 4. Likelihood of growth during storage: Under proper conditions of storage, microorganisms do not grow in a canned (heat sterilized), frozen, or dried product. In fact, bacteria gradually die off during frozen or dry storage. Abuses such as thawing and holding prior to refreezing or in- advertent moistening of a dry product may allow microbial growth. The occurrence of such abuse is sporadic and may occur in any of a number of places in the distribution system. Therefore, routine application of microbiological criteria is not an effective way to detect specific abuse situations. In the production of refrigerated formulated foods, microbiol- ogical guidelines can be usefully applied by the processor to assess po- tential shelf-life. However, the application of such criteria after the product has left the processing facility is not recommended. 5. Susceptibility of consumer: The very young, the aged, and the infirm are known to be more susceptible to infection by salmonellae and other pathogens than are healthy adults. Therefore, special care must be taken with foods intended for infants, hospital patients, residents of nursing homes, and other highly susceptible groups of people at elevated risk.

APPLICATION TO FOODS AND FOOD INGREDIENTS 295 Microbiological criteria often are appropriate for these foods. (For further discussion, see Chapter 3 and Codex Alimentarius Commission, 1979b.) References CDC (Centers for Disease Control) 1960 Botulism. Morb. Mort. Weekly Rpt. 9:2. 1981 Foodbo~ne Disease Outbreaks. Annual Summary 1979. Atlanta: Centers for Disease Control. Codex Alimentar~us Commission 1979a Microbiological specifications for foods for infants and children. Alinorm 79/13. Appendix V. . . 1979b Code of Hygienic Practices for Infants and Children, Appendix V. Report of the 16th Session of the Codex Committee on Food Hygiene, Codex Alimentar~us Commission. Rome: Food and Agriculture Organization. FDA (Food and Drug Administration) 1978 FDA Bacteriological Analytical Manual. Washington, D.C.: Association of Official Analytical Chemists. ICMSF (International Commission on Microbiological Specifications for Foods) 1985 Formulated foods. In Microorganisms in Foods. 2. Sampling for microbiological analyses: Principles and specific applications. 2nd Ed. In preparation. Peterson, A. C., and R. E. Gunnerson 1974 Microbiological critical control points in frozen foods. Food Technol. 28 (9):37-44. State of California (Department of Health Services) 1975 Botulism home canned figs and chicken pot pie. Calif. Morbidity. No. 46. 1976 Type A botulism associated with commercial pot pie. Calif. Morbidity. No. 51. Wehr, H. M. 1982 Attitudes and policies of governmental agencies on microbial criteria for foods an update. Food Technol. 36(9):45-54,92. S. NUTS Sensitivity of Products Relative to Safety and Quality The meats of tree nuts are usually sterile when in the intact shell. Microbial contamination may occur during their harvest, transport, and processing (Hall, 1971; Hyndman, 1963; King et al., 1970; Kokal and Thorpe, 1969; Meyer and Vaughn, 19691. Microorganisms may be intro- duced as a result of insect infestation, e.g., immature almonds are attacked by the navel orange worm. The shells of tree nuts may be contaminated with enteric organisms, especially if they come in contact with the ground. Permitting cattle and other animals to graze in nut groves increases the opportunity for contamination by fecal microorganisms; collecting the nuts on clean canvas minimizes this problem. Nuts exposed to water, either  for washing or tempering, are prone to contamination, especially if there

296 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA has been a separation of the sutures. Microorganisms present on the shell may also be introduced onto meats during the cracking operation. Bacterial growth on nut meats is rarely a problem due to their low aw. An exception is the coconut, which may develop leaks, become contam- inated, and then support growth of various species, including salmonellae, in the coconut milk (Schaffner et al., 19671. When present, salmonellae may survive the drying process given to coconut meat. Elimination of the pathogen depends upon pasteurization in hot water prior to drying. Peanuts (Arachis hypogaea) are subject to contamination from soil fungi, including species that produce aflatoxins. Mold growth may occur prior to harvest or when the peanuts have been dried insufficiently or stored under humid conditions (Marth and Calanog, 19761. Tree nuts such as pecans, almonds, pistachios, and Brazil nuts also may contain aflatoxins. The molds often are introduced via insect penetration. Need for Microbiological Criteria Nut meats are used primarily as ingredients of foods such as bakery products. Because they can be a source of spoilage organisms, purchase specifications for nut meats may contain limits on the microorganisms of concern such as molds. Epidemiological evidence indicates that nut meats are not a significant vehicle of foodborne disease. However, a microbiological standard does exist in that the Food and Drug Administration has taken the position that the presence of Escherichia cold on nut meats is evidence of filth and thus adulteration; apparently growers and processors usually can produce a product that is free of this organism. Salmonellae have been a problem in dried coconut; therefore, these products should be tested for this pathogen. The presence of aflatoxins in nut meats is a chronic problem, especially in peanuts and peanut products. Standards that place limits on the permitted concentration of unavoidable -toxins should be continued (see Chapter 8 and this chapter, part N). References Hall, H. E. 1971 The significance of Escherichia cold associated with nut meats. Food Technol. 25 (3):230-232. Hyndman, J. B. 1963 Comparison of enterococci and coliform microorganisms in commercially produced pecan nut meats. Appl. Microbiol. 11:268-272. King, A. D., Jr., M. J. Miller, and L. C. Eldridge 1970 Almond harvesting, processing, and microbial flora. Appl. Microbiol. 20:208-214.

APPLICATION TO FOODS AND FOOD INGREDIENTS 297 Kokal, D., and D. W. Thorpe 1969 Occurrence of Escherichia cold in almonds of nonpareil variety. Food Technol. 23(2):93- 98. Marth, E. H., and B. G. Calanog 1976 Toxigenic fungi. In Food Microbiology: Public Health and Spoilage Aspects, M. P. DeFigueiredo and D. F. Splittstoesser, eds. Westport, Conn.: AVI Publishing. Meyer, M. T., and R. H. Vaughn 1969 Incidence of Escherichia cold in black walnut meats. Appl. Microbiol. 18:925-931. Schaffner, C. P., K. Mosbach, V. C. Bibit, and C. H. Watson 1967 Coconut and Salmonella infection. Appl. Microbiol. 15:471-475. T. MISCELLANEOUS ADDITIVES A number of food additives used by the food industry are derived from animals, plants, or microorganisms. Three of the larger classes of additives are gums, enzymes, and food colors. Only limited published information is available regarding the microbiology of these substances. Sensitivity of Products Relative to Safety and Quality Gums Gums hydrate in water to form viscous solutions or dispersions and thus exhibit useful suspending, dispersion, and stabilizing properties. They function in foods as emulsifiers, gelling agents, binders, flocculating agents, film farmers, foam stabilizers, release agents, and lubricants. With this range of functional properties, they are widely used in dairy products, sauces, pie fillings, whipped toppings, salad dressings, puddings, and jellies. The sources of these useful polysaccharides are the resinous exudates of trees (acacia, karaya, tragacanth), seeds (locust bean, guar), seaweed (agar, alginates, carrageenan), and microorganisms (xanthan). Starches and pectins from plants and gelatin from animals (Cottrell and Baird, 1980) are also used. Synthetic hydrocolloids include starch derivatives and modified celluloses (Whistler and Zysk, 1978~. The highest microbial populations in the major botanical gums (about 108/g) have been found in raw, unprocessed acacia, carrageenan, and tragacanth (Souw and Rehm, 1975, 19761. Pretreated products such as alginates, carrageenan powder, locust bean gum, and guar flour yielded lower counts. The predominant organisms in the unprocessed gums were the Bacillus species and Streptococcus faecalis; Escherichia coli, how- ever, was not recovered. Coagulase-positive staphylococci with counts of up to 105/g have been found in tragacanth and locust bean; Clostridium

298 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA perfringens has been recovered from these two gums as well as from acacia. Most gums will support bacterial growth when sufficient moisture is present. Enzymes secreted by Bacillus species can degrade gels and gum solutions reducing viscosity. Gums most susceptible to degradation are tragacanth, acacia, karaya, guar, locust bean, carrageenan, and sodium alginates (Souw and Rehm, 1975, 1976~. Enzymes The enzymes added to foods consist mainly of carbohydrates (amylase, cellulase, invertase, pectinase, etc.) and proteases (e.g., papain and rennet). Plants, animals, and microorganisms are sources of the enzymes. Little is known about the microbiology of enzyme preparations. Animal and microbial rennets, the enzymes whose microbiology has been studied most extensively, may yield high counts but appear to be free of pathogenic bacteria (de Becze, 19701. Never has a health problem been traced to the use of an enzyme per se in food processing (Pariza and Foster, 19831. Colors Natural colors, including annatto, anthocyanins, beet red, carotene, carmine, and saffron, have long been added to foods. Except for carmine, which comes from an insect, these colors are derived from plants. They are extracted into oil or aqueous systems from seeds, skins, or flowers. There are virtually no published data about the microbiological quality of natural food colors. Some anthocyanins have limited antimicrobial activity and thus may restrict growth of certain microorganisms. The one recorded disease outbreak traced to a food color was due to salmonellae in carmine. Salmonella cabana was found to be responsible although additional samples of insects contained Salmonella newport and Salmo- nella enteritidis (Lang et al., 19671. Need for Microbiological Criteria Since many of these additives are raw agricultural products, the usual indices such as APCs usually would be of little value. The type of criteria that might be applied would depend upon end use of the additive. For example, additives to be used in canned foods should not be a significant source of heat-resistant bacterial spores. At present there is little published information to suggest that microbiological criteria would be useful for most gums, enzymes, and colors. This also was the conclusion of a

APPLICATION TO FOODS AND FOOD INGREDIENTS 299 committee of FAD/WHO who examined the subject a number of years ago (de Becze, 19701. References Cottrell, I. W., and J. K. Baird 1980 Gums. Pp. 45-66 in Vol. 12 of Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Ed. New York: John Wiley and Sons. de Becze, G. I. 1970 Food enzymes. Critical Reviews in Food Technology 1(4): 479-518. Lang, D. J., L. J. Kunz, A. R. Martin, S. A. Schroeder, and L. A. Thomson 1967 Carmine as a source of nosocomial salmonellosis. N. Eng. J. Med. 276:829-832. Pariza, M. W., and E. M. Foster 1983 Determining the safety of enzymes used in food processing. J. Food Prot. 46:453 468. Souw, P., and H. J. Rehm 1975 IV. Microbiological degradation of three plant exudates and two seaweed extracts. Z. Lebensm. Unters.-Forsch. 159(5): 297-304. 1976 V. Degradation of the galactomannans guar gum and locust bean gum by different bacilli. European J. Appl. Microbiol. 2: 47-58. Whistler, R. L., and J. R. Zysk 1978 Carbohydrates. Pp. 535-555 in Vol. 4 of Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Ed. New York: John Wiley and Sons. U. BOTTLED WATER, PROCESSING WATER, AND ICE Sensitivity of Products Relative to Safety and Quality Drinking water has been and still is an important vehicle for transmitting disease-causing agents (bacteria, viruses, parasites, chemicals) to man. From 1977-1981, 189 outbreaks of water-related diseases involving an estimated 49,453 persons occurred in the United States (CDC, 1979, 1980, 1981, 1982 a,b). As for foodborne disease outbreaks, these figures should not be the basis for firm conclusions about the true incidence of waterborne disease outbreaks as it is most likely many times greater than that reported. Of the 189 outbreaks, 100 (53%) were of unknown etiology and were designated "acute gastrointestinal illness" (AGI). The remaining 89 (47%) outbreaks were of a confirmed etiology: Giardia (31), chemical (27), Shigella (9), Norwalk agent (7), Salmonella (5), Campylobacter (3), Parvo- virus-like agent (3), hepatitis A (2), Vibrio cholerae Of (1), and Rotavirus (11. In none of these outbreaks, however, were bottled waters identified as vehicles. Public interest in pure, better-tasting water has created a large demand for bottled drinking water. Bottled water is defined by FDA (1982a) as

300 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA water that is sealed in bottles or other containers and intended for human consumption. The following types of bottled drinking water are available (APHA, 1984~: Spring or Well Water. This water is taken directly from a spring or well and bottled with minimum treatment. 2. Specially Prepared Drinking Water. This is water in which the mineral content has been adjusted and controlled to improve the taste. The source may be a public water supply or a well. 3. Purified Water. This water conforms to the United States Pharma copeia standard (USP, 1980) for purified water with minerals removed to less than 10 mg/l. Water can be "purified" by distillation, ion-exchange treatment, or reverse osmosis. Method of preparation must be indicated. Only water prepared by distillation can be called "distilled water." 4. Fluoridated Water. Fluoride has been added to drinking water at the optimum concentration as set forth in the FDA Quality Standards (FDA, 1982a). No definition or quality standard for "mineral" water has yet been established in the United States. In view of the potential for water-related illnesses in humans, it is essential that bottled water be from a safe source (spring, artesian well, drilled well, municipal water supply, or other source) and that it be pro- cessed, bottled, held, and transported under sanitary conditions. Water containing small numbers of enteric pathogens can cause disease in humans, whereas the same organisms ingested with food may require larger quantities of bacteria. Small amounts of water taken between meals pass the pyloric area with very little delay. Under such conditions enteric pathogens are hardly exposed to the bactericidal effect of gastric juice and reach the duodenum virtually unchanged (Levine and Nalin, 1976; Mossel and Oei, 19751. When the same bacteria are ingested with solid food, intragastric retention times are considerable. This results in a reduction in viable bacterial cells in individuals with normal gastric secretions. In the United States, few published data are available on the micro- biological condition of bottled drinking water or on the incidence of human disease outbreaks resulting from their use. According to two reports (EPA, 1972; Geldreich et al., 1975), the bacteriological quality of freshly bottled water varied greatly from brand to brand and from sample to sample within brand. Only 10% of samples had an initial APC greater than 500/ml. Coliforms were detected in 6 of 129 samples but only two of these samples exceeded the USPHS Drinking Water Standards. One of these samples also contained fecal coliforms and the other contained Pseudomonas aeru

APPLICATION TO FOODS AND FOOD INGREDIENTS 301 ginosa. During storage of bottled water pulsating changes in aerobic plate counts frequently occurred. Over 90% of counts of 10,000 bacteriological analyses of bottled water in California in 1977 were less than 100 per ml at time of bottling (Sheneman, 19831. Over 99.8% of these samples were free of coliforms. Although there is no epidemiological evidence that bottled water pro- cessed in the United States has been a public health problem, bottled water has been cited as a cause of human disease in other parts of the world. For example, bottled noncarbonated mineral water was implicated as one of the primary vehicles involved in a cholera epidemic in Portugal in 1974 (Blake et al., 19771. FDA microbiological standards for bottled water (FDA, 1982a) are based on the presence of coliforms. With the multiple tube fermentation method not more than one unit in a sampling of 10 (subsamples) shall have a MPN of 2.2 or more coliforms per 100 ml and no analytical unit shall have a MEN of 9.2 or more coliforms per 100 ml. With the membrane filter method, not more than one of the analytical units in the sample shall have 4.0 or more coliforms per 100 ml and the arithmetic mean of the coliform density of the sample shall not exceed one coliform per 100 ml. The FDA GMPs for bottled water (FDA, 1982b) require coliform anal- ysis at least once a week of a representative sample from a batch or segment of a continuous production run for each type of bottled drinking water produced during a day's production. Additionally, source water obtained from other than public water systems is to be sampled and analyzed for coliforms at least once each week. In addition, at least once each three months, a bacteriological swab and/or rinse count should be made from at least four containers and closures selected just before filling and sealing. No more than one of the four samples may exceed more than one bacterium per ml of capacity or one colony per cm2 of surface area. All samples shall be free of coliforms. In addition to the federal standards, state and local microbiological criteria and good manufacturing practice codes exist to monitor the pro- duction, processing, and distribution of bottled drinking water (Wehr, 19821. To promote high standards of quality in the bottled water industry, the International Bottled Water Association has published a technical man- ual containing a quality control program to assure compliance with FDA standards (IBWA, 19831. The American Sanitation Institute (ASI) inspects all IBWA-member plants for conformance with the regulations. The indigenous microflora of bottled drinking water usually consists of gram-negative bacteria belonging to genera such as Pseudomonas, Cy- tophaga, Flavobacterium, and Alcaligenes. Although no total count is specified, good-quality drinking water at time of bottling usually contains

302 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA less than 100 bacteria per ml (APHA, 1984~. Higher initial counts represent a lack of good manufacturing practices. The presence of coliform bacteria in bottled water indicates either a lack of good manufacturing practices and/or a potential health problem. Ozone may be applied as a disinfectant just prior to bottling. Some surviving bacteria may multiply in the water after the ozone has dissipated. No one microorganism or group of microorganisms can serve as an ideal indicator of pollution of various types of water. Although many organisms such as Aeromonas, Streptococcus, Escherichia colt, fecal col- iforms, coliforms, sulfite-reducing Clostridium, P. aeruginosa, Vibrio, and E. cold phages have been suggested as potential indicator organisms of drinking water safety, total coliforms appear at the present to be the best indicator organisms (Ptak and Ginsberg, 19771. Processing Water and Ice Water is used extensively in the food-processing industry. It comes in direct contact with major food commodities such as meat, poultry, fish, fruits, vegetables, and cheese curd during washing or chilling operations. In addition, water is used in the cleanup operations of equipment and utensils that come in contact with food. Water also is used as an ingredient in the preparation of foods. Ice is used widely to chill foods such as fish and poultry and is added to some foods as in the preparation of some processed meats. In many food-processing industries, water is chlorinated to control levels of microorganisms, for example in cooling water of canning plants. Chlorination of processing water requires an understanding of needed chlorine concentration, effect of water characteristics such as pH on available chlorine, and proper testing procedures to periodically examine for available chlorine (FPI, 19821. Water used in food production and processing can be a source of spoilage microorganisms and, if obtained from a nonpotable source or subsequently contaminated, it can be a source of pathogens. Several outbreaks of food- borne disease have been associated with or traced to the use of contam- inated water, e.g., salmonellosis from fish (Gangarosa et al., 1968), typhoid fever from canned corned beef (Howie et al., 1968), and yersiniosis from tofu (Nolan et al., 19824. Need for Microbiological Criteria There is no epidemiological evidence to indicate that bottled water as available currently in retail channels offers a significant health hazard to

APPLICATION TO FOODS AND FOOD INGREDIENTS 303 the American public. Therefore, there appears to be little evidence of need for additional or modifications of criteria currently in FDA regulations. However, the commercial vending of bottled water including import sup- plies and the variety of sources from which water for bottling is obtained has proliferated. These recent increases suggest that a periodic reassess- ment should be made of practices in this industry relative to the micro- biological quality and safety of bottled water offered to the public. Water or ice that comes in contact with or becomes part of a food should be from a potable supply and the microbiological criteria for them should meet the standards set for drinking water (Greenberg et al., 19811. Where Criteria Should Be Applied Microbiological examination of bottled drinking water for compliance with standards or guidelines should be camed out on samples collected at the processing plant. Examination of samples for APC at the retail level has little merit. References APHA (American Public Health Association) 1984 Compendium of Methods for the Microbiological Examination of Foods. 2nd Ed., M. L. Speck, ed. Washington, D.C.: APHA. Blake; P. A., M. L. Rosenberg, J. Florencia, J. B. Costa, L. D. P. Quintino, and E. J. Gangarosa 1977 Cholera in Portugal, 1974. II. Transmission by bottled mineral water. Am. J. Epi- demiol. 105 :344-348. CDC (Centers for Disease Control) 1979 Foodborne and Waterborne Disease Outbreaks. Annual summary 1977. Atlanta: Cen ter for Disease Control. 1980 Water-related Disease Outbreaks. Annual Summary 1978. Atlanta: Centers for Disease Control. 1981 Water-related Disease Outbreaks. Annual Summary 1979. Atlanta: Centers for Disease Control. 1982a Water-related Disease Outbreaks. Annual Summary 1980. Atlanta: Centers for Disease Control. 1982b Water-related Disease Outbreaks. Annual Summary 1981. Atlanta: Centers for Disease Control. EPA (U.S. Environmental Protection Agency) 1972 Bottled water study. A pilot survey of water bottlers and bottled water. Washington, D.C.: Water Supply Division, EPA. FDA (Food and Drug Administration) 1982a Quality standards for foods with no identity standards, bottled water. Code of Federal Regulations 21 CFR 103 (as corrected in Federal Register 47(205):47003-47004.) 1982b Processing and bottling of bottled drinking water. Code of Federal Regulations 21 CFR 129.

304 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA FPI (The Food Processors Institute) 1982 Canned Foods, Principles of Thermoprocess Control, Acidification and Container Closure Evaluation, 4th Ed. Washington, D.C.: FPI. Gangarosa, E. J., A. L. Bisno, E. R. Eichner, M. D. Treger, M. Goldfield, W. E. DeWitt, T. Fodor, S. M. Fish, W. J. Dougherty, J. B. Murphy, J. Feldman, and H. Vogel 1968 Epidemic of febrile gastroenter~tis due to Salmonella Java traced to smoked whitefish. Am. J. Pub. Health 58:114-121. Geldreich, E. E., H. D. Nash, D. J. Reasoner, and R. H. Taylor 1975 The necessity of controlling bacterial populations in potable waters Bottled water and emergency water supplies. J. Amer. Water Works Assoc. 67:117-124. Greenberg, A. E., J. J. Conners, D. Jenkins, and M. A. H. Franson 1981 Standard Methods for the Examination of Water and Wastewater, 15th Ed. Wash- ington, D.C.: American Public Health Association. Howie, J. W. 1968 Typhoid in Aberdeen, 1964. J. Appl. Bacter~ol. 31:171-178. IBWA (International Bottled Water Association) 1983 International Bottled Water Association Technical Bulletin, Winter. Alexandria, Va. IBWA. Levine, R. J., and D. R. Nalin 1976 Cholera is primarily waterborne in Bangladesh. Lancet 2:1305. Mossel, D. A. A., and H. Y. Oei 1975 Person-to-person transmission of enters bacterial infection. Lancet 1:751. Nolan, C., N. Hands, J. Ballard, J. Allard, and J. Kobayashi 1982 Outbreak of Yersinia enterocolitica Washington State. Morb. Mort. Weekly Rpt. 31 :562-564. Ptak, D. J., and W. Ginsberg 1977 Bacterial indicators of drinking water quality. Pp. 218-221 in Bacterial Indicators/ Health Hazards Associated with Water. Spc. Techn. Publ. 635. Philadelphia: Amer- ican Society for Testing and Materials. Sheneman, J. 1983 Memorandum from the California Food and Drug Section. Water bottling plants bacteriological analysis summary. In International Bottled Water Association, Tech- nical Bulletin, Winter. Alexandria, Va.: IBWA. USP (United States Pharrnacopeia) 1980 Purified water. P. 851 in The United States Pharmacopeia. Rockville, Maryland: U.S. Pharmacopeial Convention. Wehr, H. M. 1982 Attitudes and policies of governmental agencies on microbial criteria for foods an update. Food Technol. 36(9):45-54, 92. V. PET FOODS Companion animals of man have been a source of diseases, including salmonellosis, in humans (Morse and Duncan, 1974, 1975; Morse et al., 1976; Pace et al., 1977) and contaminated pet foods have at times been incriminated as the original cause (Pace et al., 19771. Interest in the microbiology of these foods is influenced by the above and the fact that  certain pet foods, mainly canned products, are at times consumed by humans.

APPLICATION TO FOODS AND FOOD INGREDIENTS Sensitivity of Products Relative to Safety and Quality 305 Although the true incidence of salmonellosis in animals is unknown (NRC, 1969), Salmonella are widely distributed in warm- and cold-blooded species and have been isolated from dogs, cats, horses, caged birds, turtles, frogs, skunks, raccoons, opossums, and others. The role of pets in the distribution of Salmonella has been recognized in a report by the Com- mittee on Salmonella (NRC, 19691: Of the many routes by which man can acquire salmonellosis, special mention should be made of household pets, including dogs, cats, turtles, chicks and ducklings. . Pet animals can become infected with Salmonella by a wide variety of routes, e.g., through coprophagy, by direct contact with infected animals, through eating diseased carrion and wildlife, and by the consumption of contaminated pet food. The latter is probably the least important source since the present day incidence of Salmonella in commercial pet food is very low (Pace et al., 1977~. Pet foods, which are sold predominantly for dogs and cats, may be marketed as canned, intermediate moisture (aw 0.80-0.90), or dried prod- ucts. Canned pet foods are terminally heat processed in hermetically sealed containers and are commercially sterile. They are subject to the regulations for low-acid canned foods and when in compliance are not of public health concern. Intermediate moisture pet foods and the dry products are given a heat process, generally during extrusion and pelleting, that will destroy the vegetative cells of pathogenic bacteria. The prevention of recontam- ination following heating, then, is the critical control step in their pro- cessing. Monitoring of environmental samples as well as finished product for Salmonella is thus important. Recontamination of dry pet food with Salmonella is of special concern because water is often added to the food prior to feeding. Significant growth of the pathogen can occur if the food is held at ambient temperature for an extended time period following rehydration. A potential would then exist not only for infection of the pet but for cross-contamination of household items such as equipment, utensils, and human foods. Need for Microbiological Criteria Canned pet foods are subject to the regulations for low-acid foods and thus the main applications of criteria are to assure that ingredients are free of bacterial spores that might survive the thermal process. Dry and intermediate moisture pet foods should be free of Salmonella, and a standard exists for this pathogen (U.S. Congress, 19801. Specifi- cations and guidelines for these products are especially useful if applied

306 EVALUATION OF THE ROLE OF MICROBIOLOGICAL CRITERIA at critical control points identified within a HACCP system. Guidelines and specifications serve to: 1. assess suitability, including safety, of incoming ingredients (The elimination of salmonellae from feed ingredients, particularly those of animal origin, would greatly reduce the occurrence of these organisms in finished pet foods, but this goal does not appear to be readily attainable at this time [USDA, 1978] . ); 2. identify acceptable ingredient suppliers; 3. assess control effectiveness at critical control points in manufactur- ~ng; 4. determine the acceptability of a finished product. Information Necessary for Establishment of a Criterion if One Seems To Be Indicated Extensive information is available regarding those feed ingredients that may be contaminated with Salmonella (ICMSF, 1980) and therefore may require specifications that limit this pathogen. Information is also available on the occurrence of Salmonella in pets (Morse, 1978), and in pet foods (D'Aoust, 1978; ICMSF, 1980; Pace et al., 19771. It is advisable for pet food manufacturers to conduct appropriate microbiological surveillance studies that will generate the information required for the development of guidelines. Where Criteria Should Be Applied Analyses for salmonellae might be conducted on the packaged product. Guidelines and specifications would best be applied at the plant processing level as components of an ongoing HACCP program. Their application at critical control points and on the finished product should assist in minimizing the contamination of pet foods with undesirable microorgan- isms. References D'Aoust, J. Y. 1978 Salmonella in commercial pet foods. Can. Vet. J. 19:99-100. ICMSF (International Commission on Microbiological Specifications for Foods) 1980 Microbial Ecology of Foods. Vol 2. Food Commodities. New York: Academic Press. Morse, E. V. 1978 Salmonellosis and pet animals. In Proceedings of the Salmonellosis Seminar. Wash- ington, D.C.: U.S. Department of Agriculture.

APPLICATION TO FOODS AND FOOD INGREDIENTS 307 Morse, E. V., and M. A. Duncan 1974 Salmonellosis An environmental health problem. J. Am. Vet. Med. Assoc. 165: 1015- 1019. 1975 Canine salmonellosis: Prevalence, epizootiology, signs, and public health signifi- cance. J. Am. Vet. Med. Assoc. 167:817-820. Morse, E. V., M. A. Duncan, D. A. Estep, W. A. Riggs, and B. O. Blackburn 1976 Canine salmonellosis: A review and report of dog to child transmission of Salmonella enteritidis. Am. J. Publ. Health 66:82-84. NRC (National Research Council) 1969 An Evaluation of the Salmonella Problem. Committee on Salmonella. Washington, D.C.: National Academy of Sciences. Pace, P. J., K. J. Silver, and H. J. Wisniewski 1977 Salmonella in commercially produced dried dog food: Possible relationship to a human infection caused by Salmonella enteritidis serotype Havana. J. Food Prot. 40(5):317- 321. U.S. Congress 1980 Federal Food, Drug and Cosmetic Act, as amended. Washington, D.C.: U.S. Govt. Printing Office. USDA (U.S. Department of Agriculture) 1978 Recommendations for Reduction and Control of Salmonellosis. A Report of the U.S.  Advisory Committee on Salmonella. Washington, D.C.: USDA.