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.
HIV infection and Its Epidemiology In the 20 months since the publication of Confronting AIDS, much has been learned about the clinical history of HIV infection and about the dimensions of the epidemic. New information has arisen from continuing clinical observation, surveillance systems, and epidemiological studies. In this chapter, we focus on some important aspects of the disease and the epidemic in areas in which new data have either confirmed or altered initial impressions or in which a deeper understanding of the disease has emerged. These areas include the causative agent of AIDS, the proportion of seropo- sitive persons who will develop AIDS, HIV infection as a continuum of conditions, the modes and efficiencies of HIV transmission, and the preva- lence and incidence of HIV infection and the dimensions of the epi- demic in the United States (global epidemiology is discussed in Chapter 71. HIV: THE ETIOLOGIC AGENT OF AIDS Early in the epidemic, epidemiological analysis of the pattern of the spread of AIDS showed it to be reminiscent of that for hepatitis B virus, an observation that pointed scientists in the right direction in their search for an etiologic agent. In 1983 and 1984 several researchers identified a retrovirus that is now understood to be HIV as the cause of AIDS. The committee believes that the evidence that HIV causes AIDS is scientifically conclusive. That a particular organism causes a disease is demonstrated by a confluence of evidence linking the two: HIV and AIDS have been so 33
34 CONFRONTING AIDS: UPDATE l 988 linked in time, place, and population group. For example, in San Francisco, the examination of frozen blood from a cohort of homosexual men showed the appearance of antibodies to HIV as early as 1978. At that time, the prevalence of HIV infection was probably less than 5 percent in the population of male homosexuals in San Francisco. The first cases of AIDS in homosexual men in San Francisco were detected in 1981 (CDC, 19811. This association between the cumulative incidence of HIV infec- tion and of AIDS cases is the epidemiological pattern that must exist if HIV and AIDS are causally associated: the virus must be newly intro- duced into the population, it must become widely prevalent, and its dissemination must precede the incidence of AIDS (Winkelstein, 19881. The conjunction heralded by the joint appearance of HIV and AIDS has been confirmed by their continued association. HIV seropositivity rates in defined subpopulations of homosexual men in San Francisco and New York City and in IV drug abusers in New York City are associated with later cases of AIDS in the same groups (Curran et al., 19881. In San Francisco, these subpopulations can be further broken down by neigh- borhood of residence, in which the association between HIV seropositiv- ity and AIDS is also high (Winkelstein et al., 1987b). Conversely, AIDS is unknown in populations that are free of HIV antibodies. The virus has been isolated from persons with AIDS; as assay tech- niques have improved, close to 100 percent of affected individuals can be found to harbor the virus (Booth, 19881. The virus is not found in persons who are not at risk for infection. These points are supported by epidemi- ological data from the ongoing San Francisco Men's Health Study, which began in 1984. Among 374 homosexual men who remained uninfected with HIV during the first 30 months of follow-up, no cases of AIDS occurred. Among 36 homosexual men who became infected with HIV during this period, 3 cases of AIDS (8 percent) occurred. Among 399 study subjects who were infected with HIV when they entered the study, 52 (13 percent) developed AIDS. None of the heterosexual men in the study acquired HIV infection, and none developed AIDS. The probability that this distribution might have occurred by chance is less than one in a million (Winkelstein, 1988~. Perhaps the clearest evidence linking HIV to AIDS is to be found in the tragic results of blood transfusions in the United States and around the world. The transmission of HIV in contaminated blood and blood products has been clearly linked to AIDS (Curran et al., 1984~; in the United States, over 1,500 reported cases of AIDS are associated with blood transfusions. Since routine screening of the blood supply for antibodies to HIV began in 1985, HIV transmission by this route has practically disappeared. Nevertheless, 13 recipients from 7 donors who initially tested negative for HIV antibodies are known to have acquired
HIV INFECTION AND ITS EPIDEMIOLOGY 35 HIV infection between March 1985 and October 1987 (Ward et al., 19881. The 7 donors, who were tested at the time of donation, probably had negative test results because testing occurred soon after infection and before the development of detectable antibodies. On later retesting, however, all 7 donors were found to have detectable HIV antibodies in their blood. Of the 13 recipients, 1 developed AIDS, and 3 developed HIV-related illnesses. Of the 3 developing illness, 1 was an infant twin who received transfusions shortly after birth; her fraternal twin, who received no transfusions, remained healthy. Thus, 13 people with no other risk factors became infected, and 4 of them developed the illness after receiving transfusions from donors who were initially thought to be free of infection. After careful investigation, the donors were found, in fact, to have HIV infection. The causal role of HIV in AIDS is also supported by the high risk (30 to 50 percent) of perinatal HIV transmission from an infected mother to her infant (CDC, 1987b) and the subsequent diagnosis of AIDS in the infected infants. The pathogenesis of HIV infection how the organism causes dis- ease is still incompletely understood. Several mechanisms have been proposed for the profound immunodeficiency that results from HIV infection, including the aggregation of uninfected and infected T lympho- cytes into multinucleated syncytia that subsequently die, the infection of stem cells, and the inhibition of lymphocyte functions by viral products. A complete understanding of a disease's pathogenesis, however, is not a prerequisite to knowing its etiology. PROPORTION OF INFECTED INDIVIDUALS WHO WILL DEVELOP AIDS As epidemiological cohorts of HIV-infected individuals are observed over time, a larger and larger proportion of seropositive persons has been seen to develop AIDS. The available data suggest that the great majority of HIV-infected persons will eventually progress to AIDS in the absence of effective therapy to slow or halt the infection's progression. The cohort of individuals that has been studied longest in relation to AIDS is a group of gay men in San Francisco who were enrolled in a study of hepatitis B virus vaccine in the late 1970s. As part of the study, blood samples were collected, from which serum was saved and frozen. Because this group of men was later found to be at high risk for AIDS, samples of the frozen serum were analyzed for HIV infection, and infected individuals have been followed for clinical and laboratory evi- dence of AIDS. Almost no cases of AIDS occurred during the first 2 years after infection was discovered. After 8~/2 years, more than 40 percent of
36 CONFRONTING AIDS: UPDATE 1988 the infected cohort has developed AIDS; a similar proportion has developed symptoms of HIV infection and is expected to progress to AIDS. Statistical modeling of the incidence of AIDS in this cohort predicts the possibility that 100 percent will develop AIDS within 13 years after initial infection (G. W. Rutherford, San Francisco Department of Public Health, personal communication, 1988~. The analysis of another cohort of 288 seropositive homosexual men in San Francisco who were seropositive when the study began shows that 22 percent have developed AIDS after 3 years of observation. Another 19 percent have clinical symptoms of infection, and an additional 24 percent demonstrate laboratory evidence of immunologic compromise. Projec- tions for this cohort are that 50 percent of the men will develop AIDS within 6 years of observation (or probably 9 years of infection) and that many more will develop the disease in subsequent years (Moss et al., 1988). Data from individuals infected with HIV through blood transfusions and data from persons with hemophilia suggest that the rate of progres- sion from HIV infection to AIDS increases with age. The exception to this pattern is newborns, who have the highest progression rate of all age groups (Eyster et al., 1987; Medley et al., 19871. The progression rate in adults with hemophilia appears to be similar to that in male homosexuals (Goedert and Blattner, in press). THE SPECTRUM OF HIV INFECTION In grappling with a new disease, especially one that quickly assumes epidemic proportions, terminology and definitions become vital for clin- ical management of patients, data gathering and research, and decisions about coverage and reimbursement. In 1982 CDC developed a definition of AIDS for surveillance purposes that relied on the presence of oppor- tunistic infections and malignancies; in August 1987 the definition was revised to incorporate two other syndromes indicative of AIDS: dementia and wasting syndrome (see Appendix B). Yet fairly early in the epidemic, it became apparent that many infected individuals who suffered from clinical symptoms and laboratory abnormalities signaling the presence of HIV infection did not meet the CDC criteria for the disease. For example, persistent generalized Iymphadenopathy (PGL) was thought to be asso- ciated with an increased risk of developing AIDS, especially when combined with oral candidiasis and certain laboratory abnormalities. Another group of patients displayed other chronic symptoms of AIDE fever, weight loss, night sweats, chronic diarrhea, and fatigue and a high proportion of this group also exhibited laboratory abnormalities. Even so, these patients did not fit what had become the standard definition of the
HIV INFECTION AND ITS EPIDEMIOLOGY 37 disease, although some of them seemed to develop AIDS at a rapid pace. They were described as having AIDS-related complex (ARC), and the ARC clinical syndrome was eventually incorporated in a CDC definition (although it was never used as a basis for case reporting). Clinicians noted, however, that even this definition failed to include some patients who appeared to be at high risk for progressing to AIDS. A third, more broadly defined syndrome was termed the AIDS-related condition. Today, with a better understanding of the natural history of HIV infection and with more precise laboratory assessments of disease pro- gression, the committee believes that the term ARC is no longer useful, either from a clinical or a public health perspective, and that HIV infection itself should be considered a disease. It is more accurate to describe HIV infection as a continuum of conditions, ranging from the acute, transient, mononucleosis-like syndrome associated with seroconversion, to asymp- tomatic HIV infection, to symptomatic HIV infection, and, finally, to AIDS, a spectrum that encompasses a great variety of clinical symptom- atology. The terms ARC and PGL do not have the precise prognostic implications they were once thought to have. For instance, it is now known that the presence of persistent, generalized lymphadenopathy in and of itself does not imply a worse prognosis than HIV seropositivity. For clinical (treatment or research) purposes, a patient can be more accurately described by a combination of a description of symptoms and laboratory evidence of immune dysfunction rather than by terms such as ARC or PGL. Experience with cohorts of infected individuals indicates that a major- ity of HIV-infected individuals shows some evidence of progressive immunodeficiency and is likely to develop AIDS in the absence of effective therapy. AIDS, a dramatic and devastating syndrome, caught the attention of physicians and public health officials earlier than the milder manifestations of HIV infection, but it is now clear that AIDS is end-stage HIV infection. Like many other progressive disease processes, both infectious and noninfectious, HIV has an asymptomatic period that varies in length. Viewing HIV infection as a disease is important because it may eventually be amenable to treatment. The drug zidovudine (i.e., AZT) has been shown to prolong the life of AIDS patients; it and other drugs are currently being tested to determine whether they also halt or slow disease progression in infected asymptomatic individuals. If an effective therapy is found, HIV infection will need to be treated early, just as diseases such as gonorrhea are often diagnosed and treated in asymptomatic infected patients. Even though no treatment is available, diagnosing HIV infection is still important now so that opportunistic infections and malignancies
38 CONFRONTING AIDS: UPDATE 1988 ACUTE ASYMPTOMATIC STAGE SEROPOSITIVE TRANSITION LATE 4 ' it, - - VIRUS ANTIBODY - AIDS \ /\ VIRUS ,_ - _ ~ /CD4 CELLS / WEEKS YEARS FIGURE 1 The course of disease from HIV infection to AIDS. Source: Courtesy of David Baltimore, Whitehead Institute for Biomedical Research, Cambridge, Massachusetts. can be recognized as early as possible. Many treatments for these complications are more effective and less toxic when initiated early. Considering HIV infection a disease is important to other aspects of the AIDS crisis. From a public health perspective, the population of most interest is the group infected with the virus, because these persons are capable of infecting others. In addition, as discussed in Chapter 5, medical care coverage should be based on symptoms associated with HIV infection rather than on arbitrary definitions of when "disease" begins. A terminology that reflects the progression of the disease from the initial, acute stage of infection to asymptomatic HIV infection and finally to symptomatic HIV infection and AIDS (Figure 1) would be useful for clinical treatment and for society's management of the disease. CDC has developed a classification system (see Appendix B) that might form the basis for such a terminology. MODES AND EFFICIENCIES OF HIV TRANSMISSION Epidemiological data continue to support the observation that HIV transmission is limited to sexual contact, the sharing of contaminated
HIV INFECTION AND ITS EPIDEMIOLOGY 39 needles and syringes, exposure to infected blood or blood products, transplantation of infected organs or tissue, and transmission from mother to child either across the placenta or during delivery. A recent follow-up investigation of more than 1,100 AIDS cases that were initially reported to CDC as having no identified risk factors has shown that transmission in these individuals was also limited to the recognized routes (Castro et al., 19881. Finally, additional data from studies of health care workers (CDC, 1988d), nonsexual household contacts (Friedland and Klein, 1987), and insect bites (CDC, 1986) all support the conclusion that HIV is not transmitted by casual contact or insect bites. A change in HIV transmis- sion modes would be biologically unprecedented in a virus. There is no evidence that HIV is capable of such a change. Heterosexual Transmission It has been clearly documented that HIV infection can be transmitted from men to women and from women to men through vaginal and anal intercourse (Fischl et al., 1987; Goedert et al., 1987; Padian et al., 1987a; Peterman et al., 19884. So far, however, the heterosexual spread of the virus in the United States has been confined mainly to persons whose sexual partners acquired HIV by other means for example, by sharing contaminated needles and syringes or from blood transfusions. Evidence to date shows that the spread of infection among heterosex- uals has been rather slow in instances in which neither partner can be classified in a known risk category (CDC, 1988b). In nine seroprevalence surveys of heterosexual men and women attending sexually transmitted disease (STD) clinics in six cities, the prevalence of HIV infection ranged from O to 2.6 percent (CDC, 1987b). STD clinics treat individuals in the community who, because of their sexual behavior, are most likely to be infected with HIV. In studies conducted in clinics in which data were collected during personal interviews and not through self-administered questionnaires, and in which seropositive individuals were reinterviewed to obtain better information about their risk status, the prevalence of HIV infection ranged from O to 1.2 percent among persons with no known risk factors. The results obtained from large-scale studies of over 36 million blood donations and 1.5 million military personnel (in which there are indications of the self-exclusion of persons at high risk) show that the overall prevalence of HIV has been less than 1 percent in these popula- tions for the period 1985 to 1987. This low prevalence among heterosex- uals (compared to the 20 to 50 percent prevalence among male homosex- uals) appears to indicate that the virus is not spreading rapidly in populations that are considered to be primarily low-risk groups. To become complacent in the face of this apparent trend would be a mistake, however. Heterosexual transmission of the virus is an estab-
40 CONFRONTING AIDS: UPDATE 1988 fished fact; although the numbers are small, cases acquired through heterosexual transmission are the fastest growing group of AIDS cases in the United States. Indeed, in parts of Africa, heterosexual transmission of HIV is great enough to sustain AIDS in an epidemic status (see Chapter 7~. It is useful to review the African experience with AIDS and attempt to pinpoint conditions that may augur changes in the patterns of disease spread in the United States. It is believed that, in Africa, HIV infection appeared in great numbers first in the heterosexual community and that prostitution has played a major role in its spread. Prostitution is not uncommon in some urban areas in central and east Africa, and the prevalence of HIV infection is quite high (25 to 88 percent) among the prostitutes tested in some of those areas (Kreiss et al., 1986; Plot et al., 1988~. Case-control studies have also shown that sexual activity with female prostitutes is more common among men with AIDS than among controls; African patients with AIDS also report contact with more heterosexual partners than do controls (Quinn et al., 19861. On the other hand, homosexuality and IV drug abuse do not play a major role in HIV transmission in Africa (Plot et al., 1988~. In addition, STDs, in particular, genital ulcers, are fairly prevalent in some sexually active populations in Africa and are associated with an increased risk of infection, perhaps by providing a more direct portal of entry into the bloodstream. The contamination of the African blood supply and frequent exposure to unsterilized needles and syringes in both medical settings and ritual practices may also be important factors in the spread of AIDS among the African heterosexual population. Furthermore, African heterosexual adults show chronically activated immune systems more frequently than American heterosexual men, which may be a factor that increases their susceptibility to HIV infection (Quinn et al., 19871. The pattern of disease spread in the United States has been much different. Here, the epidemic began in a few cities and within a closed community male homosexuals in which high-risk behaviors were prac- ticed (multiple partners and receptive anal intercourse). These behaviors enhanced the rapid spread of HIV infection within that community, which also had high STD rates, another factor that may have increased the risk for HIV infection. The observation that the spread of HIV infection into the heterosexual community appears to be much slower suggests that one or more of the following may be true: (1) there has been relatively little sexual contact between this pool of infected men and heterosexuals; (2) heterosexuals probably change partners less frequently than homosexual men; and (3) vaginal intercourse may not spread the virus as easily as anal intercourse. Consequently, HIV infection in the heterosexual population in the United States has been somewhat contained.
HIV INFECTION AND ITS EPIDEMIOLOGY 41 Similarly to the appearance of disease among homosexual men, HIV infection also ~ame~onounced in communities of IV drug abusers who practiced high-risk behaviorsin this instance, frequent drug injections and the sharing of contaminated drug injection equipment. These behav- iors are the functional equivalents of frequent receptive anal intercourse and multiple sexual partners among homosexual men. Here also, the spread of infection was rapid but contained (Robertson et al., 1986; Des Jarlais et al., 1988~. The potential for the spread of infection beyond the IV drug-abusing population is discussed below. - Will HIV infection reach epidemic proportions in the `'general" het- erosexual population in the United States, and are the conditions neces- sary for such an epidemic already in place? Sustaining the spread of the disease requires a "chain of transmission" from individuals practicing high-risk behaviors to their partners and from them to individuals with no known risks. This chain of transmission would have to include sufficient numbers of infected women interacting with men who would not other- wise be at high risk. Such a reservoir of infected women might be created in several ways: one mechanism is bisexuality; another probably more significant avenue is IV drug abuse (Guinan and Hardy, 1987; Moss et al., 1987~. Of all IV drug abusers, 90 percent are heterosexuals, and 30 percent are women. Moreover, between 30 and 50 percent of female IV drug abusers have engaged in prostitution. Thus, there exists the possi- bility that a pool of infected prostitutes might be created (whose source of infection is the sharing of contaminated needles and syringes). HIV infection could then enter the heterosexual community from male cus- tomers of female prostitutes. To date, most of the cases of AIDS among heterosexuals have resulted from IV drug abuse, and the number of infected addicts is growing. Moreover, seroprevalence among heterosexuals with no known risk factors is higher in areas of the country in which seroprevalence among IV drug abusers is high. This correspondence means that IV drug abusers play a pivotal role in the spread of HIV to adults through heterosexual transmission (and to infants through perinatal transmission). For 1987, a 30 percent increase in syphilis was reported in the United States (CDC, 1988c), primarily among heterosexuals. Higher rates were reported for blacks and Hispanics than for whites. In addition, the areas reporting the largest absolute increases in syphilis cases (i.e., Florida, New York City, and California) were also areas that have high rates of HIV infection. The increases in syphilis cases suggest that behavior that increases the probability of HIV infection among heterosexuals is not being effectively curtailed. In sum, the evidence to date is that heterosexual HIV transmission occurs from men to women and from women to men through vaginal and
42 CONFRONTING AIDS: UPDATE 1988 anal intercourse. The virus is capable of spreading among heterosexuals, but so far the prevalence of infection in the heterosexual population with no known risks for infection is low. Yet the extent of future heterosexual spread is uncertain. A "window of opportunity" apparently exists for preventing the further spread of infection to the heterosexual population. Efficiencies of Transmission The modes of HIV transmission are well documented. What is not as clear is how easily or how "efficiently" HIV is transmitted by a particular route if an individual is exposed. Specially designed epidemiological studies provide information that helps to estimate the probability of HIV transmission by the various known routes. Blood Transfusions. The efficiency of this transmission route can be estimated using studies of the recipients of blood from donors who were subsequently found to have AIDS or HIV antibodies. Between 66 and 100 percent of blood transfusion recipients became infected if donors either tested positive for antibodies to HIV or later became antibody positive or developed AIDS (Ward et al., 1987~. Furthermore, recipients were more likely to become infected if the transfusion occurred close to the time the donor developed symptoms. All recipients of blood transfusions became infected if the donors developed AIDS within 23 months of the donation (Ward et al., 1987~. Thus, the large dose of the virus a transfusion represents, coupled with this particular route, appears to be quite efficient as a transmission path. Perinatal Transmission. Risk of transmission can also be estimated from studies that evaluate the risk that an HIV-infected pregnant woman will deliver an infected infant. The results from such studies suggest that the probability of HIV transmission from mother to infant ranges from 30 to 50 percent (CDC, 1987b). Some studies suggest that the risk of transmission is higher for infants born to mothers who have symptoms of HIV infection during pregnancy or who show evidence of immunosuppression (Mok et al., 1987; Nzilambi et al., 1987; Plot et al. 1988). IV Drug Abuse. Information on HIV transmission through the sharing of contaminated needles and syringes is hard to gather because of the illicit nature of IV drug abuse. However, several studies have shown that once HIV is introduced into a community, its spread is rapid among IV drug abusers and a majority of them soon becomes infected (Novick et al., 1986; Robertson et al., 1986; Des Jarlais et al., 1988~. In New York City, where there are large numbers of infected IV drug abusers, the patterns of needle-sharing behavior include the practice of renting used needles and other drug paraphernalia in "shooting galleries" in which IV
HIV INFECTION AND ITS EPIDEMIOLOGY 43 drug abusers gather (Friedland and Klein, 19871. Studies of IV drug abusers have shown an association between HIV seropositivity and both the frequency of drug injections and the sharing of drug injection equipment (Chaisson et al., 1987b; Marmor et al., 19871. Homosexual Transmission. The risk of HIV transmission from recep- tive anal intercourse between homosexual men has been estimated, although partner tracing among homosexuals can be difficult in situations in which there have been multiple sexual partners (Grant et al., 1987~. Cohort and case-control studies of homosexual men (Darrow et al., 1987; Kingsley et al., 1987; Moss et al., 1987; Winkelstein et al., 1987) show that the risk of HIV infection is greatest for persons who engage in receptive anal intercourse. The risk of infection is less for partners who engage in insertive anal intercourse, and the risk appears even lower for oral receptive intercourse. Heterosexual Transmission. Estimates of the risk of heterosexual transmission have been derived from studies of the sex partners of infected persons. In this study design, an index case (the infected person) is identified, and the antibody status of his or her sexual partner is determined at entry and observed over time. In several studies of female partners of IV drug abusers, the risk of infection was reported to be about 50 percent (Curran et al., 1988~. Studies of the male sex partners of female IV drug abusers found similarly large risks of infection, although the numbers of male partners tested were small. In these studies, HIV transmission by the sharing of contaminated needles and syringes cannot be ruled out. The risk of transmission is lower for female partners of hemophiliacs and bisexual men and for partners of transfusion-infected persons than it is for male or female partners of IV drug abusers (Padian, 1987; Padian et al., 1987a; Curran et al., 1988; De Gruttola and Mayer, 1988; Johnson, 19881. In studies of the wives or female partners of hemophiliacs, the risk of infection was about 10 percent. Studies of female partners of bisexual men reported a risk of transmission of around 25 percent. Risks of similar magnitude have been found in studies of the spouses of transfusion-infected persons. In a recent study, of the 55 wives who had sexual contact with their infected partners, 10 (18 percent) became seropositive (Peterman et al., 1988~. In this study, the risk of infection was not related to the number of sexual contacts a woman had with her infected spouse; in fact, seropositive wives reported fewer sexual contacts and were somewhat older than seronegative wives. This result suggests that, in addition to behavioral factors, biological factors probably play a role in determining how easily HIV is transmitted. There may be differences in transmissibility as a result of changes in the infectiousness of the infected individual over time. Thus, heterosexual
44 CONFRONTING AIDS: UPDATE 1988 contact during periods of high infectiousness may be more likely to transmit the virus than contact during periods of low infectiousness. A similar finding has been reported from a cohort study of partners of infected hemophiliacs. In this study, the best predictor of HIV trans- mission was the absolute number of T-helper lymphocytes in the hemophiliacs, suggesting that, as their immune systems became more suppressed, they were more likely to infect their sex partners, regard- less of the frequency of sexual contact or the duration of their infection (Goedert et al., 1987~. Another finding from the investigation of spouses of transfusion- infected persons was a higher rate of transmission from men to women than from women to men. Whereas 10 of 55 wives (18 percent) became seropositive through sexual contact with their infected husbands, only 2 of 25 husbands (8 percent) became seropositive through sexual contact with their infected wives. Although this difference was not statistically significant, the finding that HIV transmission may be more efficient from men to women than from women to men has been reported in other studies (Padian, 1987; De Gruttola and Mayer, 19881. Needle-Stick Transmission. Studies among health care workers of accidental needle-stick injuries or cuts with sharp objects provide infor- mation on the risk of HIV infection by this route. As of December 31, 1987, a collaborative surveillance study conducted by CDC had followed 489 health care workers who sustained needle-stick exposures to infected blood and for whom both acute- and convalescent-phase serum samples were obtained. Three of the 489 health care workers (0.6 percent) had seroconverted within 6 months of exposure. Two other prospective studies of health care workers in the United States are also assessing the risk of HIV transmission from accidental needle-stick exposure. As of April 15, 1988, a similarly designed study at the University of California at San Francisco had reported that 1 of 180 health care workers (0.5 percent) seroconverted after at least 6 months of follow-up (J. L. Gerberding and H. F. Chambers, personal communication [updating Gerberding et al., 19871, 1988~. As of April 30, 1988, the National Institutes of Health had obtained both acute and convalescent serum samples at least 6 months after exposure for 108 health-care workers with needle-stick injuries; none had seroconverted (D. K. Henderson, per- sonal communication [updating Henderson et al., 1986], 1988~. The results from these studies, as well as from studies conducted in England and Canada, suggest that the risk of transmission from needle-stick exposure is less than 1 percent and probably closer to 0.5 percent. In most instances, these risks are associated with one episode of exposure, in contrast to the studies of sexual partners, which involve multiple expo- sures over time.
HIV INFECTION AND ITS EPIDEMIOLOGY 45 Relative Efficiencies of HIV Transmission Several investigators have estimated the probability of HIV transmis- sion for different modes of transmission (e.g., per episode of heterosexual intercourse or per screened blood transfusion), but these estimates are uncertain because the calculations are based on very limited information (Grant et al., 1987; Padian et al., 1987b; Hearst and Hully, 1988; Ward et al., 1988~. Neither is it possible to rank routes of transmission from the greatest to the least efficient with complete certainty. Factors other than the route appear to determine whether transmission of the virus occurs; these factors include the dose (inoculum size) of virus transferred, the frequency of exposure, differences in host susceptibility, variation in infectiousness of an infected person over time, possible differences in virulence among HIV isolates, and the presence of other sexually transmitted diseases or other cofactors. In addition, study designs differ, which makes it difficult to compare results. However, some preliminary conclusions can be drawn. The recipi- ents of infected blood transfusions are at very high risk of infection, as are children born to infected mothers. Studies of infected IV drug abusers also report high rates of infection for this group, suggesting that the sharing of contaminated needles and syringes combined with frequent injections carries a high risk of infection. The sexual partners of IV drug abusers have a greater risk of becoming infected than the sexual partners of individuals who were infected by other routes, suggesting that the mode of transmission may be either heterosexual transmission or the unacknowledged sharing of contaminated needles and syringes. Studies of the sex partners of individuals who were infected by routes other than IV drug abuse show much lower risks of infection, as do studies of the risk of infection from accidental needle- stick exposure. It is important to note that, although the risk of infection from one episode of heterosexual sex and one needle-stick injury may be roughly equivalent, studies of sexual transmission involve repeated exposures over the course of years, whereas needle- stick injury studies usually involve one or a few episodes. Although it seems intuitively correct that risk increases with increased exposure, other factors (for instance, the apparent inability of a particular infected person to transmit the virus) may intervene. In sexual transmission, the risk of infection is greater for the receptive partner in anal intercourse than for the insertive partner. Vaginal inter- course is probably less efficient than receptive anal intercourse as a transmission route; preliminary evidence also suggests that, in vaginal intercourse, infectivity from men to women is somewhat greater than from women to men.
46 CONFRONTING AIDS: UPDATEl988 PREVALENCE AND INCIDENCE OF HIV INFECTION IN THE UNITED STATES The importance of accurate descriptions of the prevalence and inci- dence of HIV infection, both at present and for the future, cannot be overstated. Defined cases of AIDS are only the clinical end stage of the devastating effects produced by HIV infection. The description of HIV infection by demographic characteristics and other distinguishing features helps determine which groups to target for intervention strategies to prevent the further spread of infection. HIV Prevalence in Groups at Recognized Risk In November 1987, CDC summarized current knowledge of the prev- alence and incidence of HIV infection for various segments of the United States population according to age, sex, race or ethnic group, and geographic area (CDC, 1987b). The report reviewed data obtained from several sources (including federal agencies, health departments, and medical centers) on the prevalence of HIV infection as measured by the presence of HIV antibodies in the blood (i.e., seroprevalence). The observed prevalence of HIV infection is highest in those risk groups that account for the majority of AIDS cases reported to CDC. Still, caution is needed in interpreting prevalence data. For example, the prevalence of HIV infection may be seen to vary in STD clinics in different geographic areas because the background prevalence in any two communities may be different. Other problems in comparing data arise from differences in questionnaire design, the inclusion or exclusion of symptomatic individuals in reports of seroprevalence, and differences in the demographic characteristics of the individuals being tested. Further, reported prevalence may be higher or lower than the true prevalence for a given group depending on who "walked in the door" (i.e., most of these surveys are based on self-selected samples). Estimates of the prevalence of HIV infection in homosexual and bisexual men based on data from 23 cities range from 10 to 70 percent, with most estimates falling between 20 and 50 percent (CDC, 1987b). Prevalence is highest in cohorts of homosexual men in San Francisco. Yet the data probably overestimate the true prevalence of HIV infection in this group because most of the respondents to these surveys were persons who were either seeking medical attention for STDs or who were concerned that their past or present sexual behavior had placed them at risk (Curran et al., 19881. The populations of IV drug abusers appear to be less mobile than the population of homosexual men, as larger differences in HIV prevalence are reported by geographic area. Surveys consistently show very high preva-
HIV INFECTION AND ITS EPIDEMIOLOGY 47 fence (50 to 60 percent) in major East Coast cities with geographic or close cultural connections to New York City and northern New Jersey; prevalence is much lower (less than 5 percent) in other areas of the country (CDC, 1987b). Most of the surveys measuring prevalence in IV drug abusers are conducted at facilities for chronic heroin abuse treatment. It is thought that only 10 to 20 percent of the estimated 1.2 million drug abusers in the United States are currently in treatment and that those not in treatment may be habitual users whose risk for HIV infection is even greater (CDC, 1987b). The prevalence of HIV infection for persons with hemophilia ranges from 15 percent to more than 90 percent, depending on the type and severity of hemophilia and, in turn, the amount of clotting factor received (CDC, 1987b). Persons with severe hemophilia A have the highest prevalence (approxi- mately 70 percent), whereas persons with hemophilia B or mild hemophilia A have a somewhat lower prevalence (approximately 35 percent). Within these clinical categories, however, prevalence is uniform throughout the country, reflecting the distribution of clotting factor concentrate received before 1985. Only hemophiliacs who seek treatment are tested; conse- quently, the prevalence reported here may be an overestimate of the true prevalence for hemophiliacs as a group (CDC, 1987b). The prevalence of HIV in female prostitutes in the United States varies from 0 percent to more than 50 percent. Seropositivity is higher in black and Hispanic prostitutes than in white prostitutes. The differences in prevalence appear to be related to the extent of IV drug abuse in the groups tested and the background HIV prevalence in IV drug abusers in the area (CDC, 1987b). In studies of seroprevalence conducted among persons who are hexers sexual sex partners of HIV-infected persons but who have no other identi- fiable risk factors for HIV infection, prevalence ranged from less than 10 percent to 60 percent (CDC, 1987b). As noted earlier in this chapter, surveys in STD clinics of heterosexual men and women who do not belong to any risk group and who do not have partners in any risk group report prevalences ranging from 0 percent to 2.6 percent, depending on the population studied and the method of data collection. Seroprevalence is higher among hetero- sexuals in areas in which seroprevalence in IV drug abusers is high. However, such studies may overrepresent the true prevalence of HIV among heterosexuals because people surveyed in STD clinics may be more sexually active than the "general" heterosexual population. HIV Prevalence Among Selected Segments of the General Population The prevalence of HIV infection in the population at large has been estimated primarily from studies of various special populations: blood
48 CONFRONTING AIDS: UPDATE 1988 donors, civilian applicants to the military, Job Corps entrants, sentinel hospital patients, and newborn infants (whose antibody status at birth reflects the presence or absence of antibodies in their mothers) (CDC, 1987b). More than 36 million blood or plasma donations in the United States have been tested for HIV antibodies since 1985. HIV prevalence among first-time donors for the period 1985 to 1987 was 0.04 percent. Prevalence was much higher for men than for women and higher for blacks and Hispanics than for whites. Since October 1985, blood samples from over 1.5 million applicants for military service have also been tested for HIV antibodies. The prevalence of HIV infection increases with age for applicants between the late teens and late twenties. Prevalence by birth year cohorts also increased from the first screening period (1985-1986) to the second (1986-1987~. As with blood donors, seroposi- tivity was higher for men than for women and higher for blacks and Hispanics than for whites. The overall prevalence (for October 1985 to September 1987), adjusted for the age, sex, and racial and ethnic composition of the U.S. adult population aged 17 to 59 years, was 0.14 percent. Since March 1987, HIV antibody screening has been conducted for new members of the Job Corps who participate in residential training programs. This program recruits rural and inner-city disadvantaged youths aged 16 to 21. Provisional data from the first 25,000 entrants showed a seroprevalence of 0.33 percent. The prevalence of infection in the nation as a whole is probably higher than what has been observed in blood donors, applicants to the military, and Job Corps entrants, as persons at highest risk for infection are probably underrepresented. To avoid the self-selection bias associated with volunteer programs, anonymous HIV antibody testing has also recently begun on selected hospital patients (excluding AIDS cases and other conditions related to HIV infection) at sentinel hospitals. Based on the first 8,668 test results, the age- and sex-adjusted prevalence of infection was 0.32 percent. This sample represents hospitalized patients who are at low risk for infection. In addition, the hospitals selected to participate in the program may service specialized segments of the community; therefore, the data collected are not representative of all hospitalized patients. Several states have begun programs to assess the prevalence of HIV infection in women of childbearing age by testing for HIV antibodies in their newborns. Maternal antibodies against HIV cross the placenta and are therefore present in the baby's blood. A baby with antibodies to HIV may or may not itself be infected; however, the presence of antibodies in the baby's blood always indicates that the mother is infected. Neonatal blood specimens are routinely collected in hospitals to test for metabolic disorders; the test for HIV antibodies has been added to this program. A recent study has reported that 1 of every 476 women (0.2 percent) giving
HIV INFECTION AND ITS EPIDEMIOLOGY 49 birth in Massachusetts was antibody positive during the period December 1986 to June 1987. The prevalence of HIV infection differed according to the type and location of the maternity hospitals. Prevalence was highest in inner-city hospitals (0.8 percent), lower in mixed urban and suburban hospitals (0.25 percent), and lowest in suburban and rural hospitals (0.09 percent) (Hoff et al., 1988~. In New York, the prevalence of HIV infection among women delivering babies in hospitals in the five New York City boroughs between November 1987 and February 1988 was 1.45 percent; the prevalence of HIV infection among women delivering babies in hospitals outside the metropolitan area was 0.18 percent (Novick et al., 1988). Incidence of New Infections Data on the incidence (the number of new infections over time) of HIV infection are more difficult to obtain than prevalence data, but they are crucial for longer term projections of the course of the epidemic. Evidence from eight cohort studies of gay men suggest a lower HIV incidence rate in that population for 1985-1987 than in the earlier part of the decade (CDC, 1987b). This observed decline in the incidence of infection may be attributed to several factors, but it is consistent with reports of a decline in other sexually transmitted diseases in this group (CDC, 1988c) as well as changes in sexual behavior (Winkelstein et al., 1987a). Serologic screening of blood and plasma donors and heat treat- ment of factor concentrate, as well as efforts to exclude donors at high risk, have also reduced the rate of new infection among transfusion recipients and hemophiliacs since 1985. In contrast, HIV incidence appears to be increasing in IV drug abusers in New York City and San Francisco (Chaisson et al., 1987a; Des Jarlais et al., 1987; Schoenbaum et al., 1987~. These data suggest that the epidemic of HIV infection in the United States may be viewed as a series of overlapping smaller epidem- ics, each with its own dynamics and time course (Curran et al., 1988~. National Estimates of HIV Infection In 1986, CDC estimated the size of various segments of the population that were known to be infected (i.e., male homosexuals, IV drug abusers, hemophiliacs, heterosexuals with no known risks), as well as the preva- lence of HIV infection for each of these groups. It then calculated from these estimates that 1 to 1.5 million people in the United States were currently infected with HIV. In November 1987, CDC reviewed these estimates and modified them slightly based on new information about the size of the various populations and new seroprevalence data for these
50 CONFRONTING AIDS: UPDATE 1988 groups. In retrospect, the 1986 estimates made by CDC appear to have been too high. CDC now estimates that between 945,000 and 1.4 million Americans currently are infected with HIV. The major limitation of both the original and the revised estimates is the unknown size of the homosexual population that engages in at-risk behaviors (CDC, 1987b). CDC will continue to update national estimates of the prevalence of HIV infection as more information is gathered. Other groups and investigators have also estimated the prevalence of HIV infection in the United States; these estimates have been both higher and lower than those made by CDC, ranging from 400,000 to 2.2 million for the end of 1987 (De Gruttola and Lagakos, 1987; Harris, 19874. Such estimates provide an overall picture of the magnitude of the epidemic; however, seroprevalence and incidence data on specific groups at risk are more important because they offer the necessary information to target prevention strategies and eval- uate their effectiveness in curbing the epidemic. The Program of HIV Surveys and Studies CDC has responded to the urgent need to monitor the spread of HIV infection by instituting a series of seroprevalence studies and surveillance systems (Dondero et al., 19881. In approximately 30 metropolitan areas in the United States, blood samples will be routinely collected from persons treated at STD clinics, drug abuse treatment centers, family planning and women's health clinics, and tuberculosis clinics, as well as from selected hospital admissions and newborns. These studies will provide local officials with information on HIV prevalence so that interventions can be designed to control HIV infection in specific settings. The surveys of newborns will provide some of the most valuable information because sample selection is unbiased; the entire population of childbearing moth- ers is included. As noted earlier, testing for HIV antibodies, which will occur in approximately 30 states, will be added to already existing programs that routinely test newborns for metabolic disorders. These surveys will, therefore, be population based and, by providing informa- tion on the antibody status of newborns, will reflect the prevalence of infection in mothers delivering in these hospitals. In addition to these activities, studies of HIV infection will continue in civilian applicants to the military services, active duty military personnel, blood donors, and Job Corps entrants. Surveys of HIV prevalence will also be conducted in other populations of special interest such as patients from emergency rooms, patients using other hospital services, students on college campuses, and prisoners. The National Center for Health Statistics (NCHS) will also conduct a study to determine the feasibility of a nationwide household seroprevalence survey. In addition, NCHS also
HIV INFECTION AND ITS EPIDEMIOLOGY 51 plans to include anonymous HIV antibody testing of an estimated 17,000 blood specimens collected from adults over a 6-year period as part of the National Health and Nutrition Examination Survey. It is important to note that some of these surveys will be "blinded" (i.e., HIV antibody testing will be done on blood specimens collected for other purposes with personal information on the individual permanently removed) to avoid the uninterpretable impact of self-selection bias. Other surveys, however, will be nonblinded. In these settings, volunteer par- ticipants will be interviewed to evaluate risk factors for HIV transmis- sion. AIDS CASES IN THE UNITED STATES In October 1986, when Confronting AIDS was published, approxi- mately 24,500 cases of AIDS had been reported to CDC. As of May 1988, 62,200 cases of AIDS had been reported since June 1981, and 35,051 of these had ended in death (CDC, 1988a). An additional 10 to 20 percent of cases are believed to have been missed by the surveillance system. The number of cases reported each year continues to increase, although the rate of increase is less steep than it was earlier in the decade. Cases have been reported from all 50 states and the District of Columbia. Since the publication of Confronting AIDS, the distribution of cases by risk group as well as by sex, race, age, and geographic area has not changed substantially: 63 percent of cases are homosexual or bisexual men not known to have abused IV drugs, 19 percent are heterosexual IV drug abusers, 7 percent are both male homosexuals and IV drug abusers, 1 percent are patients with hemophilia and related disorders, 4 percent are persons who acquired the disease through heterosexual contact, 3 percent are recipients of blood transfusions, and 3 percent are cases in which risk information is undetermined because it is incomplete (patients have died, refused to be interviewed, or have been lost to follow-up) or the patients are still under investigation. This 3 percent also includes men reporting contact with a prostitute and patients with no identifiable risk factor. Of the 981 cases of AIDS among children that had been reported to CDC by May 1988, 77 percent are offspring of a parent with AIDS or at high risk for AIDS. Of the remaining pediatric cases, 6 percent are children with hemophilia, 14 percent are transfusion recipients, and 4 percent are children for whom risk information cannot be determined. Over the past 2 years, the largest increases in new cases have been observed in two groups: heterosexual partners of HIV-infected individu- als and children whose mothers abuse IV drugs or are sexual partners of men at high risk. There is an overrepresentation of blacks and Hispanics in both of these groups. The only group showing a steady decline in AIDS
52 CONFRONTING AIDS: UPDATE 1988 incidence over the past 2 years has been children with transfusion- associated AIDS. This decline is attributed to the screening of blood and blood products that began in early 1985 and to the rather short incubation period of 12 months or less observed for children with transfusion- associated AIDS. It is now thought that more than 80 percent of HIV infection in children can be directly linked to IV drug abuse in the mother or father. THE DEMOGRAPHIC IMPACT OF AIDS AIDS has already begun to alter the demographic characteristics of New York City and San Francisco. A disease that was virtually unknown to Americans 8 years ago, AIDS is now the leading cause of death in New York City among men aged 25 to 44 and women aged 25 to 34. In 1986, mortality from AIDS was the eighth leading cause of years of potential life lost before the age of 65 in the United States (CDC, 1988b). Recent data from New York City indicate that 1 of every 66 infants born between November 1987 and February 1988 tested positive for HIV antibodies, reflecting the prevalence of HIV infection in women of childbearing age in that city (Novick et al., 1988~. In San Francisco, approximately 50 percent of the male homosexual population is infected with the virus, suggesting the possible future devastation of a large component of the city's population. In 1986, New York City and San Francisco accounted for approximately 40 percent of all AIDS cases; by 1991 these two cities will account for less than 20 percent of cases nationwide (Morgan and Curran, 1986), suggesting that other metropolitan areas will soon face major economic and demographic losses. AIDS cases occur in higher proportions in black and Hispanic popula- tions than in white populations (on the West Coast, the proportion is 3 times higher in black and Hispanic than in white populations and 12 times higher on the East Coast), mainly as a result of higher HIV prevalence in black and Hispanic IV drug abusers and their sex partners and offspring. Recent data also suggest that the virus is spreading more rapidly among blacks and Hispanics at risk than among other population groups, especially in Northeastern cities, suggesting that the future composition of AIDS cases will consist primarily of poor, urban minorities. FUTURE RESEARCH NEEDS Epidemiological studies are the main source of information on the prevalence and incidence of HIV infection and AIDS, the modes and efficiencies of HIV transmission, the proportion of infected individuals who progress to AIDS, serologic markers of disease progression, and the
HIV INFECTION AND ITS EPIDEMIOLOGY 53 distribution of behaviors associated with increased exposure to HIV. Epidemiological studies have also provided some of the strongest evi- dence for the association between HIV infection and AIDS. Whether or not these studies are prevalence or incidence surveys, cohort or case- control in design, they provide essential data to understand and control the epidemic. Although much has been learned about the epidemiology of HIV infection, more research is needed to address its many unanswered questions. The committee therefore strongly urges continued epidemiolog- ical research in support of appropriate prevention and control measures. CDC must be provided with the necessary funding to ensure that personnel, space, and technical resources are adequate to the task of continuing epidemiological research. REFERENCES Booth, W. 1988. A rebel without a cause of AIDS. Science 239: 1485-1488. Castro, K. G., A. R. Lifson, C. R. White, T. J. Bush, M. E. Chamberland, A. M. Lekatsas, and H. W. Jaffe. 1988. Investigations of AIDS patients with no previously identified risk factors. J. Am. Med. Assoc. 259:1338-1342. CDC (Centers for Disease Control). 1981. Kaposi's sarcoma and Pneumocystis pneumonia among homosexual menNew York City and California. Morbid. Mortal. Wkly. Rep. 30:305-308. CDC. 1986. Acquired immunodeficiency syndrome (AIDS) in western Palm Beach County, Florida. Morbid. Mortal. Wkly. Rep. 35:609-612. CDC. 1987a. Antibody to human immunodeficiency virus in female prostitutes. Morbid. Mortal. Wkly. Rep. 36:157-161. CDC. 1987b. Human immunodeficiency virus infection in the United States: A review of current knowledge. Morbid. Mortal. Wkly. Rep. 36(suppl. 6):1-48. CDC. 1987c. Revision of the CDC surveillance case definition for acquired immunodefi- ciency syndrome. Morbid. Mortal. Wkly. Rep. 36(suppl. 1):3S-1SS. CDC. 1988a. AIDS weekly surveillance reportUnited States, May 16. Atlanta, Gal: CDC. CDC. 1988b. Changes in premature mortality United States, 1979-1986. Morbid. Mortal. Wkly. Rep. 37:47-48. CDC. 1988c. Continuing increase in infectious syphilis United States. Morbid. Mortal. Wkly. Rep. 37:35-38. CDC. 1988d. Update: Acquired immunodeficiency syndrome and human immunodeficiency virus infection among health-care workers. Morbid. Mortal. Wkly. Rep. 37:229-239. Chaisson, R. E., D. Osmond, A. R. Moss, H. W. Feldman, and P. Bernacki. 1987a. HIV, bleach, and needle sharing. Lancet 1:1430. Chaisson, R. E., A. R. Moss, R. Onishi, D. Osmond, and J. R. Carlson. 1987b. Human immunodeficiency virus infection in heterosexual intravenous drug users in San Fran- cisco. Am. J. Public Health 77:169-172. Curran, J. W., D. N. Lawrence, H. Jaffe, J. E. Kaplan, L. D. Zyla, M. Chamberland, R. Weinstein, K.-J. Lui, L. B. Schonberger, T. J. Spira, W. J. Alexander, G. Swinger, A. Ammann, S. Solomon, D. Auerbach, D. Mildvan, R. Stoneburner, J. M. Jason, H. W. Haverkos, and B. L. Evatt. 1984. Acquired immunodeficiency syndrome (AIDS) associ- ated with transfusions. N. Engl. J. Med. 310:69-75.
54 CONFRONT7NGAiDS:UPDATE~g88 Curran, J. W., H. W. Jaffe, A. M. Hardy, W. M. Morgan, R. M. Selik, and T. J. Dondero. 1988. Epidemiology of HIV infection and AIDS in the United States. Science 239:610-616. Darrow, W. W., D. F. Echenberg, H. W. Jaffe, P. M. O'Malley, R. H. Byers, J. P. Getchell, and J. W. Curran. 1987. Risk factors for human immunodeficiency virus (HIV) infections in homosexual men. Am. J. Public Health 77:479-483. De Gruttola, V., and S. W. Lagakos. 1987. The value of doubling time in assessing the course of the AIDS epidemic. Paper prepared for the Institute of Medicine Workshop on Modeling the Spread of Infection with Human Immunodeficiency Virus and the Demographic Impact of Acquired Immune Deficiency Syndrome, Washington, D.C., October 15-17. De Gruttola, V., and K. H. Mayer. 1988. Assessing and modeling heterosexual spread of the human immunodeficiency virus in the United States. Rev. Infect. Dis. 10:138-150. Des Jarlais, D. C., S. R. Friedman, M. Marmor, H. Cohen, D. Mildvan, S. Yancovitz, U. Mathur, W. El-Sadr, T. J. Spira, J. Garber, S. T. Beatrice, A. S. Abdul-Quader, and J. L. Sotheran. 1987. Development of AIDS, HIV seroconversion, and potential cofactors for T4 cell loss in a cohort of intravenous drug users. AIDS 1: 105-111. Des Jarlais, D. C., S. R. Friedman, and R. L. Stoneburner. 1988. HIV infection and intravenous drug use: Critical issues in transmission dynamics, infection outcomes, and prevention. Rev. Infect. Dis. 10:151-158. Dondero, T. J., M. Pappaioanou, and J. W. Curran. 1988. Monitoring the levels and trends of HIV infection: The Public Health Service's HIV Surveillance Program. Public Health Rep. 103:213-220. Eyster, M. E., M. H. Gail, J. O. Ballard, H. Al-Mondhiry, and J. J. Goedert. 1987. Natural history of human immunodeficiency virus infections in hemophiliacs: Effects of T-cell subsets, platelet counts, and age. Ann. Intern. Med. 107:1-6. Fischl, M. A., G. M. Dickinson, G. B. Scott, N. Klimas, M. A. Fletcher, and W. Parks. 1987. Evaluation of heterosexual partners, children, and household contacts of adults with AIDS. J. Am. Med. Assoc. 257:640-644. Friedland, G. H., and R. S. Klein. 1987. Transmission of the human immunodeficiency virus. N. Engl. J. Med. 317:1125-1135. Gerberding, J. L., C. E. Bryant-LeBlanc, K. Nelson, A. R. Moss, D. Osmond, H. F. Chambers, J. R. Carlson, W. L. Drew, J. A. Levy, and M. A. Sande. 1987. Risk of transmitting the human immunodeficiency virus, cytomegalovirus, and hepatitis B virus to health care workers exposed to patients with AIDS and AIDS-related conditions. J. Infect. Dis. 156:1-8. Goedert, J. J., and W. A. Blattner. In press. The epidemiology and natural history of human immunodeficiency virus. In AIDS: Etiology, Diagnosis, Treatment and Prevention, 2nd ea., V. T. DeVita, S. Hellman, and S. A. Rosenberg, eds. Philadelphia: Lippincott. Goedert, J. J., M. E. Eyster, R. J. Biggar, and W. A. Blattner. 1987. Heterosexual transmission of human immunodeficiency virus: Association with severe depletion of T-helper lymphocytes in men with hemophilia. AIDS Res. Hum. Retrovir. 3:355-361. Grant, R. M., J. A. Wiley, and W. Winkelstein. 1987. Infectivity of the human immunode- ficiency virus: Estimates from a prospective study of homosexual men. J. Infect. Dis. 156: 189-193. Guinan, M. E., and A. Hardy. 1987. Epidemiology of AIDS in women in the United States: 1981 through 1986. J. Am. Med. Assoc. 257:2039-2042. Harris, J. E. 1987. The AIDS epidemic: Looking into the 1990s. Technol. Rev. 90:59-64. Hearst, N., and S. B. Hulley. 1988. Preventing the heterosexual spread of AIDS. Are we giving our patients the best advice? J. Am. Med. Assoc. 259:2428-2432. Henderson, D. K., A. J. Saah, B. J. Zak, R. A. Kaslow, H. C. Lane, T. Folks, W. C. Blackwelder, J. Schmitt, D. J. LaCamera, H. Masur, and A. S. Fauci. 1986. Risk of
HIV INFECTION AND ITS EPIDEMIOLOGY 55 nosocomial infection with human T-cell lymphotropic virus type III/lymphadenopathy- associated virus in a large cohort of intensively exposed health care workers. Ann. Intern. Med. 104:644-647. Hoff, R., V. P. Berardi, B. J. Weiblen, L. Mahoney-Trout, M. L. Mitchell, and G. F. Grady. 1988. Seroprevalence of human immunodeficiency virus among childbearing women. Estimation by testing samples of blood from newborns. N. Engl. J. Med. 318:525-530. Johnson, A. M. 1988. Heterosexual transmission of human immunodeficiency virus. Br. Med. J. 296:1017-1020. Kingsley, L. A., R. Kaslow, C. R. Rinaldo, Jr., K. Detre, N. Odaka, M. VanRaden, R. Detels, B. F. Polk, J. Chmiel, S. F. Kelsey, D. Ostrow, and B. Visscher. 1987. Risk factors for seroconversion to human immunodeficiency virus among male homosexuals. Lancet 1:345-349. Kreiss, J. K., D. Koech, F. A. Plummer, K. K. Homles, M. Lightfoote, P. Plot, A. R. Ronald, J. O. Ndinya-Achola, L. J. D'Costa, P. Roberts, E. N. Ngugi, and T. C. Quinn. 1986. AIDS virus infection in Nairobi prostitutes. Spread of the epidemic to east Africa. N. Engl. J. Med. 314:414-418. Marmor, M., D. C. Des Jarlais, H. Cohen, S. R. Friedman, S. T. Beatrice, N. Dubin, W. El-Sadr, D. Mildvan, S. Yancovitz, U. Mathur, and R. Holzman. 1987. Risk factors for infection with human immunodeficiency virus among intravenous drug abusers in New York City. AIDS 1:39-44. Medley, G. F., R. M. Anderson, D. R. Cox, and L. Billard. 1987. Incubation period of AIDS in patients infected via blood transfusion. Nature 328:719-721. Mok, J. Q., C. Giaquinto, A. De Rossi, I. Grosch-Worner, A. E. Ades, and C. S. Peckham. 1987. Infants born to mothers seropositive for human immunodeficiency virus. Prelimi- nary findings from a multicentre European study. Lancet 1:1164-1168. Morgan, W. M., and J. W. Curran. 1986. Acquired immunodeficiency syndrome: Current and future trends. Public Health Rep. 101:459-465. Moss, A. R. 1987. AIDS and intravenous drug use: The real heterosexual epidemic. Br. Med. J. 294:389-390. Moss, A. R., D. Osmond, P. Bacchetti, J.-C. Chermann, F. Barre-Sinoussi, and J. Carlson. 1987. Risk factors for AIDS and HIV seropositivity in homosexual men. Am. J. Epidemiol. 125:1035-1047. Moss, A. R., P. Bacchetti, D. Osmond, W. Krampf, R. E. Chaisson, D. Stites, J. Wilber, J.-P. Allain, and J. Carlson. 1988. Seropositivity for HIV and the development of AIDS or AIDS related condition: Three-year follow-up of the San Francisco General Hospital cohort. Br. Med. J. 296:745-750. Novick, D. M., M. J. Kreek, D. C. Des Jarlais, T. J. Spira, E. T. Khuri, J. Ragunath, V. S. Kalyanaraman, A. M. Gelb, and A. Miescher. 1986. Abstract of clinical research findings: Therapeutic and historical aspects. Pp. 318-320 in Proceedings of the 47th Annual Scientific Meeting, Committee on Problems of Drug Dependence (1985), L. Harris, ed. Rockville, Md.: National Institute on Drug Abuse. Novick, L. F., D. Berns, R. Stricof, and R. Stevens. 1988. New York State l~epartment of Health newborn seroprevalence study. Interim report draft. Albany. March 15. Nzilambi, N., R. W. Ryder, F. Behets, H. Francis, E. Bayende, A. Nelson, J. M. Mann, et al. 1987. Perinatal HIV transmission in two African hospitals. P. 158 in Abstracts of the Third International Conference on AIDS, Washington, D.C., June 1-5. Padian, N. S. 1987. Heterosexual transmission of acquired immunodeficiency syndrome: International perspectives and national projections. Rev. Infect. Dis. 9:947-960. Padian, N., L. Marquis, D. P. Francis, R. E. Anderson, G. W. Rutherford, P. M. O'Malley, and W. Winkelstein, Jr. 1987a. Male-to-female transmission of human immunodeficiency virus. J. Am. Med. Assoc. 258:788-790.
56 CONFRONTING AIDS: UPDATE Eggs Padian, N., J. Wiley, and W. Winkelstein. 1987b. Male-to-female transmission of human immunodeficiency virus (HIV): Current results, infectivity rates, and San Francisco population seroprevalence estimates. P. 171 in Abstracts of the Third International Conference on AIDS, Washington, D.C., June 1-5. Peterman, T. A., R. L. Stoneburner, J. R. Allen, H. W. Jaffe, and J. W. Curran. 1988. Risk of human immunodeficiency virus transmission from heterosexual adults with transfusion- associated infections. J. Am. Med. Assoc. 259:55-58. Piot, P., F. A. Plummer, F. S. Mhalu, J.-L. Lamboray, J. Chin, and J. M. Mann. 1988. AIDS: An international perspective. Science 239:573-579. Quinn, T. C., J. M. Mann, J. W. Curran, and P. Piot. 1986. AIDS in Africa: An epidemiologic paradigm. Science 234:955-963. Quinn, T. C., P. Piot, J. B. McCormick, F. M. Feinsod, H. Taelman, B. Kapita, W. Stevens, and A. S. Fauci. 1987. Serologic and immunologic studies in patients with AIDS in North America and Africa. The potential role of infectious agents as cofactors in human immunodeficiency virus infection. J. Am. Med. Assoc. 257:2617-2621. Robertson, J. R., A. B. V. Bucknall, P. D. Welsby, J. J. K. Roberts, J. M. Inglis, J. F. Peutherer, and R. P. Brettle. 1986. Epidemic of AIDS related virus (HTLV-III/LAV) infection among intravenous drug abusers. Br. Med. J. 292:527-529. Schoenbaum, E. E., P. A. Selwyn, D. Hartel, R. S. Klein, K. Davenny, and G. H. Friedland. 1987. HIV seroconversion in intravenous drug abusers: Rate and risk factors. P. 117 in Abstracts of the Third International Conference on AIDS, Washington, D.C., June 1-5. Ward, J. W., D. A. Deppe, S. Samson, H. Perkins, P. Holland, L. Fernando, P. M. Feorino, P. Thompson, S. Kleinman, and J. R. Allen. 1987. Risk of human immunodeficiency virus infection from blood donors who later developed the acquired immunodeficiency syn- drome. Ann. Intern. Med. 106:61-62. Ward, J. W., S. D. Holmberg, J. R. Allen, D. L. Cohn, S. E. Critchley, S. H. Kleinman, B. A. Lenes, O. Ravenholt, J. R. Davis, M. G. Quinn, and H. W. Jaffe. 1988. Transmission of human immunodeficiency virus (HIV) by blood transfusions screened as negative for HIV antibody. N. Engl. J. Med. 318:473-478. Winkelstein, W., Jr. 1988. Epidemiological observations on the causal nature of the association between infection by the human immunodeficiency virus and the acquired immunodeficiency syndrome. Paper presented at the Scientific Forum on the Etiology of AIDS, American Foundation for AIDS Research, Washington, D.C., April 9. Winkelstein, W., Jr., M. Samuel, N. S. Padian, J. A. Wiley, W. Lang, R. E. Anderson, and J. A. Levy. 1987a. The San Francisco Men's Health Study. III. Reduction in human immunodeficiency virus transmission among homosexual/bisexual men, 1982-86. Am. J. Public Health 76:685-689. Winkelstein, W., Jr., D. M. Lyman, N. Padian, R. Grant, M. Samuel, J. A. Wiley, R. E. Anderson, W. Lang, J. Riggs, and J. A. Levy. 1987b. Sexual practices and risk of infection by the human immunodeficiency virus. The San Francisco Men's Health Study. J. Am. Med. Assoc. 257:321-325.