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.
B Serologic and V~rolo~c Testing The standard tests used to define individuals who have been exposed to HIV detect antibodies to the virus in the serum. Antibodies to HIV can be detected by several techniques, including enzyme-linked immunosorbent assays (ELISA), immunofluorescent assays (IFA), and Western blot analysis. Each of these techniques, when performed by expert techni- cians, is very accurate at detecting antibody either to the whole virus or to viral subcomponents. In contrast to some viral infections, HIV induces antibodies that do note in most cases, appear to effectively neutralize the establishment or consequences of viral spread in an infected host. Therefore, most patients with positive tests for HIV antibodies are considered to be simultaneously and actively infected by HIV. The resulting concerns about the equation of seropositivity with extant infection, continuing transmissibility, risk of disease in an infected individual, and issues of social stigmatization have caused HIV serologic testing to be very controversial. The tests currently in use attempt to measure specific antibodies to proteins or polyproteins produced as a result of infection with HIV. The virus's gag and end genes encode for the predominant viral antigen to which antibodies detected by today's tests are directed. The gag gene, which encodes the protein constituents of the viral core, initially produces a 55-kilodalton (kc) polyprotein that is present in large amount in virus-infected cells. This nonglycosylated protein is subsequently cleaved to form pl7 (a phosphoprotein), p24, and gag peptides. Although p24 and ply are detectable in both extracellular virus and disrupted virus-infected 304
APPENDIX B 305 cells, the 55-kd precursor protein is not present in significant amounts in virus harvests used to prepare antigen. Antibodies to HIV core protein p24 and its group antigen precursor are thought to appear earliest following infection and are readily detected in a number of ELISA tests. The ens gene encodes for a polyprotein of about 90 kd in its nongly- cosylated form. Since it has numerous glycosylation sites, it migrates as a glycoprotein of about 160 kd in electrophoretic analyses and is found as such in infected cells. This glycoprotein gives rise to two principal proteins, gpl20 and gp41. Both are present in infectious virus particles and infected cells. Antibodies to these proteins are thought to appear somewhat later than core antibodies and are present in most sera from HIV-infected individuals. Additional immunoreactive viral gene products include those nonstruc- tural proteins encoded by sor, a short open reading frame of unspecified function; tat, which is responsible for certain critical aspects of viral expression; and 3'-orf, another open reading frame that encodes for a 27-kd protein of unknown function. The proteins encoded by the pol gene of HIV, which catalyze essential enzymatic processes in viral replication, are strongly immunogenic and recognized by sera from the vast majority of infected persons. The distribution of antibodies against the proteins encoded for by these genes is becoming better understood as newer and potentially improved tests are being developed using recombinant DNA technologies to produce specific viral proteins. While all are potential substrates for improved serologic tests, so far no patterns of serologic reactivity have been found to correlate with disease stage or prognosis. The configuration of the ELISA test most frequently employed in serologic analyses involves coating plastic microtiter wells or plastic beads with HIV antigen and adding test serum in various dilutions. An antigen-antibody reaction is detected by the use of so-called second-stage antibodies, which react with any human antibodies remaining bound to the viral antigens in the ELISA plate. The second-stage antibodies are modified to facilitate their detection by conjugation with enzymes such as horseradish peroxidase or alkaline phosphatase. If antibodies in the serum tested are bound to viral antigens, then the antihuman antibody will bind to the antiviral antibodies, if present, and the attached enzyme will be free to catalyze a chemical reaction after the addition of the appropri- ate substrate. The extent of the reaction is detected calorimetrically. A control serum is used. "Positives" are distinguished from "negatives" on the basis of the relative degree of absorbance of the test serum and control. Where the cutoff point is set affects the sensitivity and specificity of the serologic test. While the ELISA test can identify the presumptive presence of anti- bodies against HIV in a serum sample, another test, the Western blot
306 APPENDIX B analysis, permits the documentation of antibodies to specific viral pro- teins and thus a more specific level of resolution of serologic reactivity. The Western blot analysis is basically an immunoelectrophoretic test in which viral proteins from purified disrupted virus are fractionated by size using polyacrylamide gel electrophoresis. The fractionated viral proteins are then transferred to nitrocellulose paper to permit subsequent immu- nologic detection. Control molecular-weight standards are included to identify migration of proteins of various sizes. Samples being tested for antibody are added to the strip, followed by appropriate stages of inoculation and washing. Enzyme-linked antihuman IgG globulin is added and incubated. Then an appropriate substrate is added and the enzyme catalyzes the calorimetric reaction, which detects bound antibody. Again, the extent of the reaction is measured by the intensity of the color produced. There has been considerable variation in what different laboratories have interpreted as positive in Western blots. In the past the presence of antibody to the gag p24 alone was considered positive by the Centers for Disease Control (CDC) criteria. With accumulating experience of HIV Western blot determination, it became apparent that many patients who have only this antibody appear to represent false positives. Recently, a consensus has developed that serum specimens are considered specifi- cally reactive with HIV if antibody to the following proteins is demon- strated in the presence or absence of other bands: (1) p24 and p41, (2) p24 and pS5, (3) p41. If only p24 antibody is demonstrated, the reaction is considered equivocal and must be repeated on the same serum sample. Other tests for antibody detection include the radioimmunoprecipita- tion and cytoplasmic or membrane immunofluorescence tests. Both tests work well but are less well suited for screening purposes and appear to be more appropriate for use in research laboratories. In good hands, im- munofluorescent testing is as accurate as ELISA testing, but it is much harder to standardize and therefore is not used as often. In the first generation of the ELISA tests made available, the intact virus was used to detect antibodies. In more recent versions, recombinant antigens are employed. Even with the first generation of test kits, accuracy has been very high. Sensitivity and specificity (see below) are in the 95 to 99 percent range. PERFORMANCE CHARACTERISTICS OF THE TESTS In evaluating the utility of a test, the terms "sensitivity," "specificity," "prevalence," "predictive value," "gold standard," and "cutoff point" are often used. The sensitivity of a test is the percentage of infected
APPENDIX B 307 persons who will have a positive test. This often varies depending upon the stage or severity of the disease. With HIV infection, seropositivity appears to rise slowly after initial infection and then to remain relatively stable. Estimates of the true sensitivity of the HIV test are hampered by the lack of a "gold standard" an independent verification of the presence of infection. For example, if one could isolate virus in all cases, one would be able to determine the exact sensitivity of a serologic test. In the case of HIV, however, the virus isolation techniques are far from 100 percent sensitive. Many investigators calculate sensitivity using small numbers of samples that are not generally representative. Also, the use of sera from AIDS patients to determine the sensitivity of a test may be very misleading if the test is applied to low-risk populations to detect early infections. Given the better appreciation of reactivity with certain partic- ularly immunogenic viral proteins, it has become increasingly important to determine the protein content of the different antigens used for the test and to relate this to what is being learned about changes in the antibody profile to various proteins over time. Specificity is the percentage of uninfected persons who have a negative test. One would like to have 100 percent specificity, especially for an infection with the implications of HIV infection. The way the antigen is prepared may affect specificity, in that one may be measuring antibodies present in the test sera reacting with cellular products that contaminate viral preparations. This may be particularly troublesome in patients who use drugs or have chronic illnesses. Again, failure to use a large number of broadly representative sera may give falsely high estimates of speci- ficity. For example, measurements of specificity in well populations may underestimate the problems encountered when the test is later used in alternate populations. The cutoff point is the point above which one calls a test positive and below which one calls it negative. No test is 100 percent sensitive and 100 percent specific, so one is usually trading off either sensitivity or specificity. One sets the cutoff point to maximize one or the other, depending on whether it is important to detect all those who are infected and then sort out true positives from false positives or whether one is willing to miss a certain number of true positives in order to minimize the risk of false positives. Chapter 4 describes general aspects of the uses of these tests, their applications in improving the safety of the blood supply, their emerging use as an indicator of infection in individuals, and the problems associated with these uses. Chapter 6 contains recommendations regarding desirable future efforts in this area.
308 APPENDIX B BIBLIOGRAPHY American Medical Association Council on Scientific Affairs. 1985. Status report on the acquired immunodeficiency syndrome: Human T-cell lymphotropic virus type III testing. JAMA 254:1342-1345. National Institutes of Health. 1985. Workshop on experience with HTLV-III antibody testing: Update on screening, laboratory and epidemiologic conditions. Bethesda, Md., July 31. National Institutes of Health. 1986. Program and abstracts: Impact of routine HTLV-III antibody testing on public health. Bethesda, Md., July 7-9. Silberner, J. 1986. AIDS blood screens: Chapters 2 and 3. Science News 130:56-57.