|
|
| |||||||||||||||||||||||||||||||||||||||
Background Among patients infected with human immunodeficiency virus type 1 (HIV-1), early and widespread use of prophylactic regimens against Pneumocystis carinii is changing the pattern of illnesses related to the acquired immunodeficiency syndrome (AIDS).
Methods We conducted a subcohort analysis of 844 men with AIDS (87 percent of whom have since died) from a prospectively followed cohort of 2592 HIV-1-infected homosexual men.
Results A total of 138 men received prophylaxis before the diagnosis of AIDS, but 39 (28 percent) nevertheless had P. carinii pneumonia at some time. Only four illnesses occurred more frequently in men who received P. carinii prophylaxis before the onset of AIDS: Mycobacterium avium complex disease, which developed in 33.4 percent, as compared with 17.3 percent of the 706 men who did not receive early prophylaxis; wasting syndrome (18.4 percent vs. 6.4 percent); cytomegalovirus disease (44.9 percent vs. 24.8 percent); and esophageal candidiasis (21.3 percent vs. 12.8 percent). Collectively, these four diseases accounted for the initial AIDS-related illness in 42.7 percent of those who received prophylaxis before the onset of AIDS, as compared with 10.7 percent of those who did not. During the three 6-month periods before the diagnosis of AIDS (0 to 6, >6 to 12, and >12 to 18 months), the geometric mean CD4+ cell counts were 48, 87, and 147 per cubic millimeter, respectively, in men who received prophylaxis against P. carinii, as compared with 118, 211, and 279 per cubic millimeter in those who did not.
Conclusions M. avium complex disease, esophageal candidiasis, wasting syndrome, and cytomegalovirus disease are more common in HIV-infected patients who have received prophylaxis against P. carinii than in those who have not. Prophylaxis may delay the first AIDS illness for 6 to 12 months.
Clinical investigators need to know which AIDS-related illnesses should now be targeted for diagnosis and further prophylaxis and treatment. Since P. carinii prophylaxis is effective,3,4,5,6,7,8,9 priorities should be based on the frequency of diseases that develop in patients receiving this therapy. In this report from the Multicenter AIDS Cohort Study, we describe and quantify AIDS-related illnesses in homosexual men receiving P. carinii prophylaxis. We also compare the CD4+ counts within the six-month period before the onset of the AIDS-defining illness in men who received P. carinii prophylaxis and in those who did not.
Methods
The Multicenter AIDS Cohort Study12 recruited 4954 homosexual men without AIDS from the Baltimore-Washington, D.C., area, Chicago, Los Angeles, and Pittsburgh in 1984 and 1985 and an additional 625 between 1987 and 1991; 2592 of the men were HIV-1-seropositive before 1992. Every six months, each man underwent an evaluation consisting of a physical examination and a CD4+ lymphocyte count. The use of zidovudine and P. carinii prophylaxis (aerosolized pentamidine, trimethoprim-sulfamethoxazole, or dapsone) was ascertained at these visits. For the purpose of this analysis, AIDS and AIDS-related illnesses were defined according to the criteria established by the Centers for Disease Control and Prevention (CDC) in 198713. Illnesses and deaths were continually monitored by means of medical reports and reports by study subjects, with confirmation by a physician or a review of medical records. A subject was considered to have received P. carinii prophylaxis (or zidovudine) before the diagnosis of AIDS if he reported such treatment at any visit before the diagnosis of the initial AIDS-related illness and was considered to have begun prophylaxis after the diagnosis of AIDS if such treatment was first reported after that diagnosis. The AIDS-related illnesses that we documented were P. carinii pneumonia, Kaposi's sarcoma, Mycobacterium avium complex disease, esophageal candidiasis, cytomegalovirus disease, wasting syndrome, lymphoma, cryptosporidiosis, leukoencephalopathy, HIV-1-related dementia, cryptococcal disease, toxoplasmosis, and disseminated herpes simplex, as well as all other illnesses that establish a diagnosis of AIDS, according to the 1987 CDC criteria13.
This analysis is based on 844 men in whom AIDS was diagnosed as of January 7, 1992, and who were followed until April 18, 1993, by which time 87 percent had died.
Descriptive and statistical methods for the analysis of categorical and continuous data (contingency tables, t-tests, linear regressions, and logistic regressions) were programmed with the use of SAS14. Because of skewed data, the logarithmic transformation of CD4+ lymphocyte counts was used, and geometric mean counts (which are close to the medians) are reported here.
Results
Among the 844 subjects with AIDS, 138 (16.4 percent) reported receiving P. carinii prophylaxis and 255 (30.2 percent) reported using zidovudine before the diagnosis of AIDS; only 1 man reported receiving P. carinii prophylaxis but not zidovudine. Of those not receiving P. carinii prophylaxis before the diagnosis of AIDS, 87 (12.3 percent) began primary prophylaxis against P. carinii pneumonia after the diagnosis of an AIDS-defining illness. Table 1 shows the annual changes in the incidence of AIDS, the proportion of cases in which P. carinii pneumonia was the initial AIDS-related illness, the proportion of cases involving prior use of zidovudine and P. carinii prophylaxis, and the mean CD4+ counts within the six months before the diagnosis of AIDS.
|
From 1985 to 1988, the mean CD4+ count during the six months before the onset of AIDS ranged from 100 to 129 per cubic millimeter. In 1989, when the use of P. carinii prophylaxis before the diagnosis of AIDS became more common, the mean CD4+ count within the six months before the onset of AIDS began to decline (P<0.001), dropping to a low of 64 per cubic millimeter by 1991. The proportion of AIDS cases diagnosed on the basis of P. carinii pneumonia also started to decline in 1989 (P<0.001), from 47 percent in 1988 to 25 percent in 1991.
Table 2 shows that in all periods before the diagnosis of AIDS, the mean CD4+ count was significantly lower in the 138 men who reported receiving P. carinii prophylaxis before the diagnosis of AIDS than in the 706 who did not (P<0.001 for all comparisons). During the six months before the diagnosis of AIDS, the mean CD4+ count was 48 per cubic millimeter in the men who received P. carinii prophylaxis (interquartile range, 24 to 97). Among the men who did not receive prophylaxis, the mean count during the 6 months before the diagnosis of AIDS was 118 per cubic millimeter, a value that lay between the means observed for two earlier 6-month periods (>6 to 12 months and >12 to 18 months) before the diagnosis of AIDS in the men receiving P. carinii prophylaxis (87 and 147 per cubic millimeter, respectively). These results suggest that, as compared with the group of men who did not receive prophylaxis, the group of men who did contracted the first AIDS-related illness after 6 to 12 months' additional loss of CD4+ cells.
|
|
Four diseases were significantly more common as initial AIDS-related illnesses among the men who received prior P. carinii prophylaxis: M. avium complex disease (in 13.0 percent of those who received prophylaxis vs. 1.8 percent of those who did not, P<0.001), wasting syndrome (10.9 vs. 0.8 percent, P<0.001), cytomegalovirus disease (9.4 vs. 3.1 percent, P = 0.004), and esophageal candidiasis (9.4 vs. 5.0 percent, P = 0.04). These four illnesses collectively constituted 42.7 percent of initial AIDS-related illnesses in those who received prior P. carinii prophylaxis, as compared with 10.7 percent in those who did not receive prior prophylaxis. This difference of 32 percentage points is close to the previously noted figure of 31.8 percentage points (from 46.3 percent to 14.5 percent) for the decline in the incidence of P. carinii pneumonia as an initial AIDS-related illness among men who received P. carinii prophylaxis.
The sum of the shaded and unshaded portions of each bar in Figure 1 represents the cumulative occurrence of the illness in question as either the initial or a later AIDS-related condition. Only 12 percent of the men who did not receive prior prophylaxis and 14 percent of those who did were living at the time of analysis and hence were at risk for contracting additional illnesses. Furthermore, because at least 15 months had passed since the diagnosis of AIDS, many of these survivors were likely to die soon. Thus, the cumulative occurrence of each illness shown in Figure 1 should only slightly underestimate the total probability that the illness would have occurred before death from AIDS.
Of the men who did not receive P. carinii prophylaxis before the diagnosis of AIDS, 87 (12.3 percent) began primary P. carinii prophylaxis after the AIDS-defining illness had developed. If we had excluded these men from the analysis, the remaining group would, by artifact alone, probably have had fewer subsequent AIDS-related illnesses. Those who died soon after the development of the initial AIDS-related disease had less time to begin P. carinii prophylaxis or to contract subsequent AIDS-related illnesses. Nevertheless, a comparison of the cumulative occurrence of AIDS-related illnesses in subjects who received prophylaxis before the diagnosis of AIDS and those who did not produces a conservative bias. It is more difficult to detect differences that can be attributed to P. carinii prophylaxis with this comparison than it would be with a comparison of pure groups.
Although 39 (28.3 percent) of the men who received prophylaxis before the diagnosis of AIDS nevertheless contracted P. carinii pneumonia at some point, this proportion was much smaller (P<0.001) than the 61.3 percent of those with no prior prophylaxis (including those who received prophylaxis after but not before the diagnosis of AIDS) who contracted P. carinii pneumonia at some point. The nine illnesses (Figure 1) that did not differ statistically between the two groups with respect to their occurrence as the AIDS-defining illness also did not differ with respect to their occurrence at any time.
In contrast, for each disease that occurred significantly more often as the initial AIDS-related illness in the men who received P. carinii prophylaxis before the diagnosis of AIDS, the difference was even larger when any occurrence of the disease was considered. Individually, the frequencies of any occurrence of the disease in the men with prior prophylaxis as compared with those without were as follows: 44.9 percent and 24.8 percent, respectively, for cytomegalovirus disease (P<0.001); 33.4 percent and 17.3 percent for M. avium complex disease (P<0.001); 21.3 percent and 12.8 percent for esophageal candidiasis (P<0.01); and 18.4 percent and 6.4 percent for wasting syndrome (P<0.001). Men who received P. carinii prophylaxis before the diagnosis of AIDS had a 74.6 percent chance of contracting at least one of these four illnesses.
Discussion
Although our results confirm other reports that prophylaxis against P. carinii pneumonia is effective,3,4,5,6,7,8,9 P. carinii pneumonia was the initial AIDS-related illness in 14.5 percent of the men who received prophylaxis and was eventually seen in at least 28.3 percent of those who received prophylaxis before the development of AIDS. This suggests that more effective long-term protection against P. carinii pneumonia is needed and that the failure of current prophylactic regimens warrants investigation.
Pentamidine was used by 78 percent of the men in our study who received prophylaxis, making it impossible to compare the various prophylactic regimens here. Other researchers have reported that P. carinii pneumonia occurred less frequently in patients using trimethoprim-sulfamethoxazole than in those using pentamidine,7 but a similar advantage was not observed with dapsone8.
Comparisons of CD4+ counts before the diagnosis of AIDS (Table 2) provide an estimate of the net time by which prophylaxis delays the onset of the initial AIDS-related illness. These comparisons suggest that by preventing P. carinii pneumonia, prophylaxis delays the first AIDS-related illness by 6 to 12 months.
We examined whether the results in Table 2 might in part reflect a selection bias, because P. carinii prophylaxis is not recommended until the CD4+ count is below 200 per cubic millimeter,9 and thus the men who contracted AIDS-related illnesses when their CD4+ counts were above this level did not have the chance to use prophylaxis. The mean CD4+ count in the six months before the diagnosis of AIDS for those who did not receive prophylaxis was 118 per cubic millimeter, which was similar to the overall mean CD4+ count (114 per cubic millimeter) in the corresponding six-month period before prophylaxis was recommended (1985 to 1988), suggesting that the selection bias was not strong.
In our study, only four diseases occurred more often in the men who received P. carinii prophylaxis before the diagnosis of AIDS than in the men who did not: M. avium complex disease, wasting syndrome, cytomegalovirus disease, and esophageal candidiasis. Why did the nine other AIDS-related illnesses not occur more frequently in the men whose lives were extended at a lower level of immune function by P. carinii prophylaxis? The prevention of P. carinii pneumonia should not greatly increase the frequency of illnesses that occur at levels of immune function higher than the level at which P. carinii pneumonia occurs. At least two of the nine illnesses, Kaposi's sarcoma and herpes simplex, generally occur at earlier stages of immunosuppression than does P. carinii pneumonia15,16. Other diseases, such as toxoplasmosis, that did not increase in frequency among the men who received prediagnosis prophylaxis are caused by infections that were not common in the study population,17 and such diseases thus remained unlikely to develop, even in the men whose lives were extended by P. carinii prophylaxis.
Preventing or delaying death from P. carinii pneumonia should selectively increase the frequency of illnesses that result from common conditions and occur at lower levels of immune function than does P. carinii pneumonia. Three of the four illnesses that increased in frequency after P. carinii prophylaxis result from organisms that are common in humans (cytomegalovirus and candida) or in the environment (M. avium complex). The fourth illness, wasting syndrome, may be due to physiologic abnormalities directly caused by immune deregulation and cytokine production18. All four illnesses occur at a lower level of immune function than does P. carinii pneumonia16.
Interpretation of the results of our study is limited by its design as a natural-history study. Care must be taken to separate treatment effects from other, possibly unrelated, chronologic changes. The associations we reported persisted when pre-1987 data were excluded from the analysis. This was true even for HIV-1-related dementia and wasting syndrome, which were added to the CDC list of AIDS-defining conditions in 198713. Because of the extensive medical and neurologic follow-up of this cohort, occurrences of wasting syndrome and dementia before 1987 could be identified from medical records, although there was probably some underreporting.
In our study, all medications were prescribed by the patients' physicians or provided as part of an experimental protocol. The use of prediagnosis P. carinii prophylaxis was not strongly associated with demographic characteristics (unpublished data). As indicated earlier, there was only a small selection effect resulting from the practice of not recommending prophylaxis until the CD4+ count dropped below 200 per cubic millimeter. However, prophylaxis was rarely prescribed without the concurrent use of zidovudine. Bone marrow suppression from zidovudine19,20 could conceivably alter the immune system in ways that change the susceptibility to AIDS-related illnesses. However, Table 1 shows that the shifts in AIDS-related illnesses and CD4+ counts before diagnosis occurred not after the increase in the use of zidovudine but after the subsequent increase in the use of prediagnosis P. carinii prophylaxis. Some associations were observed between zidovudine use before the diagnosis of AIDS and both AIDS-related illnesses and CD4+ counts in the men who did not receive prior P. carinii prophylaxis (data not shown). The magnitude of these associations was small, however, and they may in fact be explained by a confounding effect of unreported P. carinii prophylaxis.
Although we believe that the findings of our study can be applied to other groups, such generalizations must be made cautiously. Our study may need to be replicated in populations with other risk factors or with other highly endemic background infections, such as M. tuberculosis.
Clinicians monitoring CD4+ counts in patients should consider that although these counts vary from person to person, among homosexual men receiving P. carinii prophylaxis the initial AIDS-related illness may tend to develop when the CD4+ count is about 50 per cubic millimeter. Furthermore, although Kaposi's sarcoma and P. carinii pneumonia cannot be ignored, patients receiving P. carinii prophylaxis should also be monitored for M. avium complex disease, cytomegalovirus disease, wasting syndrome, and esophageal candidiasis. As forms of prophylaxis and treatment for other AIDS-related diseases continue to be developed and tested,7,8,11,21,22 priority should be placed on these last four illnesses. Prophylactic and therapeutic regimens should be evaluated not only for their efficacy against the target illness but also for the resulting delay in the first clinical manifestation of AIDS and the extension of life.
Prophylaxis against P. carinii infection, which was previously the predominant cause of AIDS-related illness and death,2 may delay the onset of the first AIDS-related illness by 6 to 12 months. But further prophylaxis against any other AIDS-related illness will probably not be as effective, since no single illness predominates among patients receiving P. carinii prophylaxis. No illness is responsible for more than 20 percent of AIDS diagnoses, and each of six illnesses has an 18 to 45 percent chance of occurring at some point. Although it might be possible to provide prophylaxis or treatment for several illnesses at the same time, the number of regimens a patient can tolerate simultaneously may be limited.
Clearly, specific additional treatments will be advantageous when patients receiving P. carinii prophylaxis contract other illnesses. In our study, however, the 50 men receiving pre-AIDS prophylaxis for whom data were available in the six months before death had a geometric mean CD4+ count of only 15 per cubic millimeter. With CD4+ counts that are already so low, it may not be realistic to expect prophylaxis and treatment against additional AIDS-related illnesses to extend life substantially.
Supported by grants (N01-AI-72634, N01-AI-72676, N01-AI-32535, N01-AI-72631, and N01-AI-72632) from the National Institutes of Health.
We are indebted to Edith Muth and Judy Konig for assistance in the preparation of the manuscript.
Source Information
From the Departments of Epidemiology and Biostatistics, School of Hygiene and Public Health, Johns Hopkins University, Baltimore (D.R.H., A.J.S., H.B.); the Division of Infectious Diseases, Northwestern University Medical School, Evanston, Ill. (J.P.); the School of Public Health, UCLA, Los Angeles (R.D.); the Graduate School of Public Health, University of Pittsburgh, Pittsburgh (R.A.); and the National Institute of Allergy and Infectious Diseases, Bethesda, Md. (R.A.K.). Participating centers and investigators are listed in the Appendix.
Address reprint requests to Dr. Hoover at Johns Hopkins University, 624 N. Broadway, Suite 784, Baltimore, MD 21205.
References
The Multicenter AIDS Cohort Study includes the following centers and investigators, in addition to the authors: Johns Hopkins University, Baltimore (J. Palenicek, H. Armenian, H. Farzadegan, N. Graham, J. Margolick, and J. McArthur); Howard Brown Memorial Clinic-Northwestern University, Chicago (J. Chmiel, B. Cohen, M. O'Gorman, D. Variakojis, J. Wesch, and S. Wolinsky); UCLA, Los Angeles (B. Visscher, D. Besley, I. Chen, J. Dudley, J. Fahey, J. Giorgi, M. Lee, O. Martinez-Maza, E. Miller, P. Nishanian, J. Taylor, and J. Zack); University of Pittsburgh, Pittsburgh (C. Rinaldo, L. Kingsley, J. Becker, P. Gupta, and M. Ho); Data Coordinating Center: Johns Hopkins School of Hygiene and Public Health, Baltimore (A. Munoz, L. Epstein, N. Galai, L. Jacobson, A. Kirby, C. Meinert, K. Nelson, S. Piantadosi, and S. Su); the National Institute of Allergy and Infectious Diseases, Bethesda, Md. (L. Schrager, S. Vermund, and M. VanRaden); and the National Cancer Institute, Bethesda, Md. (D. Seminara).
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Marijuana, the AIDS Wasting Syndrome, and the U.S. Government
Grinspoon L., Bakalar J. B., Doblin R., Abrams D. I., Child C. C., Mitchell T. F.
Extract |
Full Text
N Engl J Med 1995;
333:670-671, Sep 7, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
