Secondary Logo

Journal Logo

Epidemiology and Social Science

Effect of Hard-Drug Use on CD4 Cell Percentage, HIV RNA Level, and Progression to AIDS-Defining Class C Events Among HIV-Infected Women

Thorpe, Lorna E*†; Frederick, Margaret; Pitt, Jane§; Cheng, Irene; Watts, D Heather; Buschur, Shelley; Green, Karen#; Zorrilla, Carmen**; Landesman, Sheldon H††; Hershow, Ronald C‡‡

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: November 1st, 2004 - Volume 37 - Issue 3 - p 1423-1430
doi: 10.1097/01.qai.0000127354.78706.5d


The length and variability of the HIV incubation period, from infection to active disease, suggest that multiple factors influence the course of disease progression.1,2 Most in vitro and animal studies have shown decreases in immune function after exposure to hard drugs, namely, heroin, cocaine, and methadone.3–10 Researchers also have documented in vitro evidence that heroin and cocaine upregulate HIV expression.7,11–13

In contrast, most epidemiologic studies show no effect of drug use on HIV disease progression. However, studies designed to measure the impact of drug use face a number of methodological difficulties. Patterns of drug use tend to be highly variable, and it is unclear whether current or long-term drug use patterns play a more important role in influencing health status. Most follow-up studies capture drug use practices at 1 point in time and examine participants for later health outcomes; only a few studies have accurately characterized drug use changes over time.14–16 In addition, most studies rely on self-reported measures,17–30 raising questions about exposure accuracy. Other studies resort to indirect assessments by comparing HIV disease progression between known injection drug users and homosexual male cohorts, without directly measuring drug use exposure.17–22 These 2 risk groups are often socioeconomically and behaviorally disparate, potentially confounding any observed association between drug use and progression.

This study examined the relationship between hard-drug use and 4 outcome measures—CD4 cell percentage, HIV RNA level, progression to a first AIDS-defining class C event, and mortality—in a large national cohort of HIV-infected women. This study used periodic urine screening to validate self-reported drug use, and the longitudinal design allowed for repeated drug use assessments. To our knowledge, no study has examined the association between drug use and HIV disease progression exclusively in women. Women are an important subgroup for 2 reasons. First, drug use is more prevalent in HIV-infected women than in HIV-infected men in the United States,31 and the question of whether drug use has an effect on HIV disease progression is potentially more pertinent to women. Second, in examining the association between drug use and disease progression, less potential for confounding exists among HIV-infected women than among HIV-infected men because demographic disparities between women infected heterosexually and via injection drug use are not as great as between male homosexual and male injection drug use populations.32



The Women and Infant Transmission Study is a prospective, multicenter study of the natural history of HIV infection among pregnant women and their infants. Subject recruitment began in December 1989. HIV-infected pregnant women were recruited from 7 pediatric and obstetric clinics located in Chicago, IL, New York, NY (2 centers), San Juan, PR, Boston, MA, Worcester, MA, and Houston, TX. Eligible women enrolled in the study during pregnancy or at delivery. Visits were scheduled at <18, 25 ± 2, and 34 ± 2 weeks’ gestation, at delivery, and at 2 and 6 months after birth. Subsequent visits were every 6 months through 3 years after birth and yearly after that. The institutional review board at each institution accepted the study protocol, and written informed consent was obtained from all participants.

Analyses for this study were restricted to all women enrolled in the Women and Infant Transmission Study between December 1989 and December 1995 for whom self-reported drug use data were available, and each woman was followed for up to 5 years. Follow-up extended though June 30, 2001. Delivery, the first visit with comprehensive enrollment, was designated as baseline for this study. Women with multiple deliveries during the study entered the cohort at their first Women and Infant Transmission Study pregnancy only.

Flow cytometry was performed in concordance with AIDS Clinical Trials Group flow cytometry guidelines.33 Viral RNA was quantified according to standard procedures using quantitative RNA polymerase chain reaction analysis (Amplicor HIV-1 Monitor Test, Roche Diagnostic Systems, Branchburg, NJ).34 Urine specimens for toxicology assays were collected 3 times: at intake, during labor, and at 6 months after birth. Urine was analyzed by immunoassay for cocaine, heroin/opiates, methadone, marijuana, alcohol, and other drugs (EMIT [for methadone], Syva Co., Palo Alto, CA; Abuscreen RIA [for other drugs], Roche Diagnostic Systems). Positive screenings were confirmed by gas chromatography or mass spectrometry if a woman reported not using drugs. In June 1997, investigators initiated an intensive retrospective documentation of class C diagnoses in study participants before enrollment up through their last visit. Beginning in July 1997, class C diagnoses were prospectively documented at each visit interval.

At each visit, staff administered a survey that included questions regarding type and frequency of drug use. For this analysis, hard-drug use was assessed using 3 different window periods: at baseline (during pregnancy), in the past year (to allow for intermittent use in time-varying analyses), and at any time during the study. For each, a woman was defined as a hard-drug user if (1) she reported the use of cocaine, crack, heroin, or other opiates (including methadone), (2) she reported engaging in injection drug use, or (3) her urine was positive for any of these drugs.35

Statistical Analyses

This study examined the association between hard-drug use and 4 outcomes: 2 markers of HIV disease progression (CD4 cell percentage and HIV type 1 RNA load) and 2 clinical outcomes (time to first clinical class C diagnosis and time to all-cause mortality). The analysis used CD4 cell percentages rather than absolute cell counts because of their lesser variability and because they are not subject to dilution from increasing plasma volume during pregnancy.36 HIV RNA levels (copies/mL) were log transformed to achieve a normal distribution, and samples with RNA levels below the limit of detection (400 copies/mL) were assigned a value of 400 copies/mL. Associations were examined using generalized estimating equation37 models to calculate the average change in CD4 cell percentage and HIV RNA level (significance level, α = 0.05). Fixed-time covariates examined were age at enrollment, race, study site, baseline absolute CD4 cell count, and number of subsequent Women and Infant Transmission Study pregnancies. Time-varying covariates included cigarette smoking, alcohol consumption, and antiretroviral therapy (ART) status (no ART, monotherapy, combination therapy [≥2 drugs without highly active compound], and highly active ART [≥3 drugs, including at least 1 highly active compound]). Variables were included in the final models if they were statistically significant in the univariate analysis (α = 0.05). All analyses were conducted using PROC GENMOD in SAS Analysis Software (SAS version 6.12; SAS Institute, Inc., Cary, NC).

We then examined the association between hard-drug use and 2 clinical outcomes (time to first class C diagnosis38 and all-cause mortality) using Cox proportional hazards models.39 We modified the class C definition to exclude CD4 cell counts of <200/mm3 to avoid redundancy with the earlier CD4 cell marker analysis and to focus on opportunistic illness. Data were stratified by enrollment date (before and after March 1, 1994) to account for changes in medical practice following the AIDS Clinical Trials Group O76 protocol. In predicting time to first class C diagnosis, death before a class C diagnosis was censored because cause of death could not be assigned in most cases. Prior study findings suggested that this would provide the most conservative estimate of the effect of hard-drug use,40 because drug users tend to die of other causes (eg, drug overdose, violence, and liver disease) more than nonusers do. Proportionality of the hazards function was examined using partial residuals.41 Analyses were conducted using PROC PHREG in SAS Analysis Software.


Of 1342 HIV-infected women enrolled through January 1, 1996, 1148 (86%) had self-reported data on drug use during pregnancy and were included in these analyses. Nearly one half (45%) of the women enrolled were African American, 36% were Hispanic, and 14% were white. The mean age at delivery was 27 years (range, 14–45 years). ART use during pregnancy was common (51%) but almost universally limited to the singular use of zidovudine (98%). The use of hard drugs during pregnancy, measured by self-report and urine toxicology assay, was high (40%), and approximately one half of the women reported smoking cigarettes (50%) or drinking alcohol (47%) while pregnant.

A comparison of self-reported hard-drug use and urine toxicology assay results among women found high overall concordance. In this analysis, 82% (823) of 1005 women with available toxicology assay results were concordant between their urine screening and self-reported cocaine use, 88% of women had concordant results for heroin use, and 98% had concordant results for methadone use.

Descriptive Trends in ART and Hard-Drug Use Exposure

The median length of follow-up among participants was 21.3 months (range, 0–60.0 months), and the median number of visits was 6 (range, 1–13). ART use increased during the study, and the proportion of women receiving no therapy or monotherapy declined as treatment regimens began to include protease inhibitors (Fig. 1).

. ART use among HIV-infected women that was measured at postpartum visits (1989–1998).

Hard-drug use was often a highly intermittent behavior. Only 26% (295) of the women reported recent hard-drug use at each of their visits; almost half (47%) of the women reported using hard drugs at some time during the study. When demographic characteristics, health behaviors, and follow-up were compared, women using hard drugs at any time during the study were more likely to be older, to be non-Hispanic, to smoke cigarettes, and to drink alcohol and less likely to use zidovudine during pregnancy. They were also more likely to be lost to follow-up (Table 1).

Table 1
Table 1:
Baseline Factors and Follow-Up Experience of Hard-Drug–Using and Nonusing Women During the Course of the Study

Hard-Drug Use Effects on CD4 Cell Percentage and HIV RNA Level

Mean CD4 cell and HIV RNA trajectories for hard-drug users and nonusers were mostly similar across study visits (Fig. 2). Mean values were significantly higher among drug users than nonusers at 2 of 13 visits for CD4 cell percentage and 1 of 13 visits for HIV RNA level. In univariate models, hard-drug use during the study was not associated with change in either CD4 cell percentage (P = 0.83) or HIV RNA level (P = 0.22) over time.

. Mean CD4 cell percentages (A) and log HIV RNA levels (B) per clinic visit among HIV-infected women by any use of hard drugs during the study (1989–1998).

Hard-drug use did not exert an observable influence on CD4 cell percentage or HIV RNA level in multivariate analyses either (Table 2). Significant predictors of higher CD4 cell percentage, other than baseline CD4 cell percentage, were ART use and smoking. Those who reported smoking during follow-up had, on average, CD4 cell percentages that were 0.7% higher than nonsmokers. ART use was predictive of a lower mean log HIV RNA level.

Table 2
Table 2:
%Multivariate Analyses Examining the Effect of Hard-Drug Use on Repeated Measures of CD4 Cell Percentage, HIV RNA level, Time to Class C Diagnosis, and Death Among HIV-Infected Women, Adjusting for Time and Other Covariates (n = 1148)

Progression to Class C Events

Analyses were then conducted to determine the relative hazard of developing a class C diagnosis for hard-drug users compared with nonusers (Table 2), controlling for the same set of covariates. Fifty-one women attained a class C end point before enrollment and were thus excluded. Hard-drug users were more likely to have had a class C event before enrollment than were nonusers (6.8% vs. 2.5%, respectively; odds ratio, 2.86; 95% confidence interval, 1.55–5.29), due to a greater number of recurrent pneumonia and tuberculosis events compared with nonusers (10 recurrent pneumonia events vs. 1 and 7 tuberculosis events vs. 3, respectively).

As of June 2001, 90 women in the study developed a first class C end point; herpes simplex infection was the most common end point (24), followed by recurrent pneumonia (17), tuberculosis (16), Pneumocystis carinii pneumonia (8), and esophageal candidiasis (6). All other events were infrequent (≤5 events). Hard-drug users did progress faster to a class C event than did nonusers (Fig. 3). Baseline CD4 cell percentage and use of ART were the only other significant predictors of progression to a class C diagnosis, with monotherapy having a paradoxically positive association. Cigarette smoking was not associated with risk for a class C diagnosis.

. Kaplan–Meier survival curves for time to a first class C diagnosis or death among HIV-infected women stratified by any use of hard drugs during the study (1989–1998). A, Univariate; B, adjusted for covariates from final Cox proportional hazards model.

Results of a proportional hazards regression analysis for all-cause mortality indicated no elevation in mortality among hard-drug users (Table 2). In this model, the protective effect of ART was evident for combination therapy and highly active ART, but no association was observed for monotherapy.


In this large cohort of HIV-infected women, we found no effect of hard drugs on virologic or immunologic markers of HIV disease progression, nor did hard-drug–using women progress more rapidly to death than nonusers. However, a greater number of hard-drug–using women did develop class C events than nonusers, both before enrollment and during the study period. Most common diagnoses included recurrent pneumonia or tuberculosis, 2 bacterial infections that tend to emerge at higher CD4 cell counts than other AIDS-defining events15; both have been previously associated with drug use, independent of HIV infection.42–44 Hard-drug users were also at higher risk for herpes simplex infection, the single most common AIDS-defining illness in this cohort. Many studies have documented that drug-using women, in particular those using crack cocaine, have elevated sexual risks compared with either drug-using men45,46 or non–drug-using women.46,47 Although drugs like cocaine and heroin may not have a direct biologic effect on immunologic and virologic parameters in HIV-infected women, their use probably exacerbates the acquisition and presentation of certain class C opportunistic illnesses.

Previous studies comparing clinical outcomes between HIV-infected injection drug users and homosexual men have either shown no difference17–20 or slower progression rates among injection drug users than homosexual men.21,22 Studies conducted exclusively with injection drug users have also observed no association between frequency or persistence of drug use and AIDS incidence or mortality.15,16,23–26 Most of these studies used the 1987 AIDS definition, which did not include recurrent pneumonia, tuberculosis, or CD4 cell counts of <200/mm3 as AIDS-defining events.48 Studies using the 1993 definition include the immunologic event of a CD4 cell count of <200/mm3, the most common AIDS-defining event. We chose to exclude the criterion of a CD4 cell count of <200/mm3 and only examine progression to actual clinical events because we already demonstrated that drug users and nonusers had similar CD4 cell trajectories; inclusion would have masked any potential differences in progression to clinical illnesses.

Several studies have determined plasma viral load to be an independent predictor of disease progression,49–52 yet only 1 prior study examined the effects of drug use on HIV RNA trajectories, using data for 149 study participants; the study showed no significant differences between injection drug users and nonusers.53 Our study is the first prospective analysis using repeated measures of drug exposure to examine their effect on trends in virologic parameters, and our results suggest that hard-drug use does not contribute to increased HIV RNA levels.

We found a consistent, positive association between cigarette smoking and CD4 cell percentage yet no protective effect between smoking and a class C diagnosis. Several studies have identified a positive association between smoking and CD4 lymphocyte profiles in both HIV-infected and uninfected populations54–56; however, there appears to be no protective effect on the rate of progression to AIDS.57–59

We found ART use, in particular highly active ART, to have beneficial effects on CD4 cell percentage and HIV RNA level, as well as on mortality, but it was not associated with lower risk for developing a class C diagnosis. Indeed, monotherapy was positively associated with faster progression, a paradoxical association observed in earlier HIV studies often ascribed to preferential prescription of ART for persons with advanced disease.60

This study has several potential limitations. First, HIV seroconversion dates were unknown for women, and systematic differences in duration of infection between drug users and nonusers may have confounded any observed associations. To minimize this possibility, however, we adjusted for baseline CD4 cell counts in the multivariate proportional hazards models, a standard practice in seroprevalent cohort studies.

Second, drug use was only periodically confirmed using urine toxicology assays, and a woman’s self-reported drug use exposure may have been misclassified on visits without toxicology screening. The high concordance between self-reported use and screening results, however, suggests that self-reports were valid. Third, although we examined hard drugs using 3 different exposure window periods (at baseline, in the past year, and any time during the study), it is possible that critical drug use effects occur in a shorter time frame (ie, very recent hard-drug use may impact CD4 cell or HIV RNA markers). We chose to select a 1-year window period because of the intermittent nature of drug use and because shorter window periods might introduce biases if women stopped using drugs when sick. Last, loss to follow up was substantial, and hard-drug users were more frequently lost to follow-up than nonusers. Median retention times differed by an average of 1 visit (6 months), suggesting that potential biases associated with selective dropouts were minimimal.61

The lack of an apparent effect of hard-drug use on CD4 cell percentage or HIV RNA level in our study contradicts results from available animal and in vitro studies yet is consistent with most prior epidemiologic evidence. Using comprehensive drug use ascertainment and a sample restricted to women only, we identified that hard-drug–using HIV-infected women are at higher risk for developing nonfatal opportunistic infections, particularly recurrent pneumonia, pulmonary tuberculosis, and herpes simplex virus infection. Although perhaps not life threatening, these illnesses can be debilitating and are communicable to others. Clinicians should have a heightened index of suspicion for these illnesses when treating HIV-infected patients who use hard drugs.


Principal investigators, study coordinators, program officers, and funding were as follows: Clemente Diaz and Edna Pacheco-Acosta (University of Puerto Rico, San Juan, PR; U01 AI 34858); Ruth Tuomala, Ellen Cooper, and Donna Mesthene (Boston/Worcester Site, Boston, MA; 9U01 DA 15054); Jane Pitt and Alice Higgins (Columbia Presbyterian Hospital, New York, NY; U01 DA 15053); Sheldon Landesman, Edward Handelsman, and Gail Moroso (State University of New York, Brooklyn, NY; HD-3-6117); Kenneth Rich and Delmyra Turpin (University of Illinois at Chicago, Chicago, IL; U01 AI 34841); William Shearer, Susan Pacheco, and Norma Cooper (Baylor College of Medicine, Houston, TX; U01 HD 41983); Joana Rosario (National Institute of Allergy and Infectious Diseases, Bethesda, MD); Robert Nugent (National Institute of Child Health and Human Development, Bethesda, MD); Vincent Smeriglio and Katherine Davenny (National Institute on Drug Abuse, Bethesda, MD); and Bruce Thompson (Clinical Trials & Surveys Corp., Baltimore, MD, N01 AI 85339). Scientific Leadership Core: Kenneth Rich, PI (1 U01 AI 50274-01). Additional support was provided by local clinical research centers as follows: Baylor College of Medicine, Houston, TX; NIH GCRC RR00188; Columbia University, New York, NY; NIH GCRC RR00645.


1. Sheppard H, Lang W, Ascher M, et al. The characterization of non-progressors: long-term HIV-1 infection with stable CD4+ T cell levels. AIDS. 1993;7:1159–1166.
2. Munoz A, Wang MC, Bass S, et al. Acquired immunodeficiency syndrome (AIDS)-free time after human immunodeficiency virus type 1 (HIV-1) seroconversion in homosexual men. Am J Epidemiol. 1989;130:530–539.
3. Donahoe R, Nicholson J, Madden J, et al. Coordinate and independent effects of heroin, cocaine, and alcohol abuse on T cell E-rosette formation and antigenic marker expression. Clin Immunol Immunopathol. 1986;41:254–264.
4. McDonough R, Madden J, Falek A, et al. Alteration of T and null lymphocyte frequencies in the peripheral blood of human opiate addicts: in vivo evidence for opiate receptors on T lymphocytes. J Immunol. 1980;125:2539–2543.
5. Shavit Y, Lewis J, Terman G. Opioid peptides mediate the suppressive effect of stress on natural killer cell cytotoxicity. Science. 1984;223:188–190.
6. Peterson P, Gekker G, Brummit C, et al. Suppression of human peripheral blood mononuclear cell function by methadone and morphine. J Infect Dis. 1989;159:480–487.
7. Peterson P, Sharp B, Gekker K, et al. Morphine promotes the growth of HIV-1 in human peripheral blood mononuclear cell cocultures. AIDS. 1990;4:869–873.
8. Szabo I, Rojavin M, Bussiere J, et al. Suppression of peritoneal macrophage phagocytosis of Candida albicans by opioids. J Pharmacol Exp Ther. 1993;267:703–706.
9. Adler M, Geller E, Rogers T, et al. Opioids, receptors and immunity. Adv Exp Med Biol. 1993;335:13–20.
10. Hung C, Lefkowitz S, Geber W. Interferon inhibition by narcotic analgesics. Proc Soc Exp Biol Med. 1973;142:106–111.
11. Peterson P, Gekker G, Chao C, et al. Cocaine potentiates HIV-1 replication in human peripheral blood mononuclear cell cocultures. J Immunol. 1991;146:81–84.
12. Bagasra O, Pomerantz R. Human immunodeficiency virus type 1 replication in peripheral blood mononuclear cells in the presence of cocaine. J Infect Dis. 1993;168:1157–1164.
13. Rosenberg Z, Fauci A. Immunopathogenesis of HIV infection. FASEB J. 1991;5:2382–2390.
14. Lyles C, Margolick J, Astemborski J, et al. The influence of drug use patterns on the rate of CD4+ lymphocyte decline among HIV-1-infected injecting drug users. AIDS. 1997;11:1255–1262.
15. Selwyn PA, Alcabes P, Hartel D, et al. Clinical manifestations and predictors of disease progression in drug users with human immunodeficiency virus infection. N Engl J Med. 1992;327:1697–1703.
16. The Italian Seroconversion Study. Disease progression and early predictors of AIDS in HIV-seroconverted injecting drug users. AIDS. 1992;6:421–426.
17. Prins M, Veugelers P. Comparison of progression and non-progression in injecting drug users and homosexual men with documented dates of HIV-1 seroconversion. AIDS. 1997;11:621–631.
18. Pezzotti P, Rezza G, Lazzarin A, et al. Influence of gender, age and transmission category on the progression from HIV seroconversion to AIDS. JAIDS. 1992;5:745–747.
19. Biggar R. AIDS incubation in 1891 HIV seroconverters from different exposure groups. AIDS. 1990;4:1059–1066.
20. Spijkerman I, Langendam M, Veugelers P, et al. Differences in progression to AIDS between injection drug users and homosexual men with documented dates of seroconversion. Epidemiology. 1996;7:571–577.
21. Carre N, Deveau C, Belanger F, et al. Effects of age and exposure group on the onset of AIDS in heterosexual and homosexual HIV-infected patients. AIDS. 1994;8:797–802.
22. Pehrson P, Lindback S, Lidman C, et al. Longer survival after HIV infection for injecting drug users than for homosexual men: implications for immunology. AIDS. 1997;11:1007–1012.
23. Mientjes GH, Miedema F, van Ameijden EJ, et al. Frequent injecting impairs lymphocyte reactivity in HIV-positive and HIV-negative drug users. AIDS. 1991;5:35–41.
24. Margolick J, Munoz A, Vlahov D, et al. Longitudinal changes in T-lymphocyte subsets in a cohort of intravenous drug users with prevalent or incident HIV-1 infection. JAMA. 1992;267:1631–1636.
25. Alcabes P, Schoenbaum EE, Klein RS. Correlates of the rate of decline of CD4+ lymphocytes among injection drug users infected with the human immunodeficiency virus. Am J Epidemiol. 1993;137:989–1000.
26. Pezzotti P, Galai N, Vlahov D, et al. Direct comparison of time to AIDS and infectious disease death between HIV seroconverter injection drug users in Italy and the United States: results from the ALIVE and ISS studies. JAIDS. 1999;20:275–282.
27. Des Jarlais DC, Friedman SR, Marmor M, et al. Development of AIDS, HIV seroconversion, and potential cofactors for T4 cell loss in a cohort of intravenous drug users. AIDS. 1987;1:105–111.
28. Galai N, Vlahov D, Margolick JB, et al. Changes in markers of disease progression in HIV-1 seroconverters: a comparison between cohorts of injecting drug users and homosexual men. JAIDS. 1995;8:66–74.
29. Kaslow R, Blackwelder W, Ostrow D, et al. No evidence for a role of alcohol or other psychoactive drugs in accelerating immunodeficiency in HIV-1 positive individuals. JAMA. 1989;261:3424–3429.
30. Schechter MT, Craib KJ, Le TN, et al. Progression to AIDS and predictors of AIDS in seroprevalent and seroincident cohorts of homosexual men. AIDS. 1989;3:347–353.
31. Hader SL, Smith DK, Moore JS, et al. HIV infection in women in the United States. JAMA. 2001;285:1186–1192.
32. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report. Vol 12(1). Atlanta, GA: US Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention; 2001.
33. Calvelli T, Denny T, Paxton H, et al. Guideline for flow cytometric immunophenotyping: a report from the National Institute of Allergy and Infectious Diseases, Division of AIDS. Cytometry. 1993;14:702–715.
34. Garcia PM, Kalish LA, Pitt J, et al. Maternal levels of plasma human immunodeficiency virus type 1 RNA and the risk of perinatal transmission. N Engl J Med. 1999;341:394–402.
35. Rodriguez EM, Mofenson LM, Chang BH, et al. Association of maternal drug use during pregnancy with maternal HIV culture positivity and perinatal transmission. AIDS. 1996;10:273–282.
36. Koepke J, Landay A. Precision and accuracy of absolute lymphocyte counts. Clin Immunol Immunopathol. 1989;52:19–27.
37. Liang K, Zeger S. Regression analyses for correlated data. Annu Rev Public Health. 1993;14:43–68.
38. Centers for Disease Control and Prevention. 199 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly Rep. 1992;41(RR-17):1–15.
39. Cox DR, Oakes D. Analysis of Survival Data. London: Chapman & Hall; 1984.
40. Prins M, Aguado IH, Brettle RP, et al. Pre-AIDS mortality from natural causes associated with HIV disease progression: evidence from the European Seroconverter Study among injection drug users. AIDS. 1997;11:1747–1756.
41. Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239–241.
42. Reichman LB, Felton CP, Edsall JR. Drug dependence, a possible new risk factor for tuberculosis disease. Arch Intern Med. 1979;139:337–339.
43. Perucci C, Forastiere F, Rapiti E, et al. The impact of intravenous drug use on mortality of young adults in Rome, Italy. Br J Addict. 1992;87:1637–1641.
44. Tumbarello M, Taconelli E, de Gaetano K, et al. Bacterial pneumonia in HIV-infected patients; analysis of risk factors and prognostic indicators. JAIDS. 1998;18:39–45.
45. Stevens S, Estrada A, Estrada B. HIV, sex, and drug risk behavior and behavior change in a national sample of injection and crack cocaine using women. In: Coyle SL, ed. Women, Drug Use and HIV Infection. New York: Haworth Medical Press; 1998.
46. Edlin BR, Irwin KL, Faruque S, et al. Intersecting epidemics—crack cocaine use and HIV infection among inner-city young adults. Multicenter Crack Cocaine and HIV Infection Study Team. N Engl J Med. 1994;331:1422–1427.
47. DeHovitz JA, Kelly P, Feldman J, et al. Sexually transmitted diseases, sexual behavior, and cocaine use in inner-city women. Am J Epidemiol. 1994;140:1125–1134.
48. Centers for Disease Control. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR Morb Mortal Wkly Rep. 1987;36(Suppl 1):1S–15S.
49. Fahey J, Taylor J, Detels R, et al. The prognostic value of cellular and serological markers in infection with HIV-1. N Engl J Med. 1990;322:166–172.
50. Phillips A. Studies of prognostic markers in HIV infection: implications for pathogenesis. AIDS. 1991;6:1391–1394.
51. Mellors J, Rinaldo C, Gupta P, et al. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science. 1996;272:1167–1170.
52. Mellors J, Munoz A, Giorgi J, et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med. 1997;126:946–954.
53. Lyles CM, Dorrucci M, Vlahov D, et al. Longitudinal human immunodeficiency virus type 1 load in the Italian Seroconversion Study: correlates and temporal trends of virus load. J Infect Dis. 1999;180:1018–1024.
54. Craib KJ, Schecchter MT, Montaner JS, et al. The effect of cigarette smoking on lymphocyte subsets and progression to AIDS in a cohort of homosexual men. Clin Invest Med. 1992;15:301–308.
55. Freedman DS, Flanders WD, Barboriak JJ, et al. Cigarette smoking and leukocyte subpopulations in men. Ann Epidemiol. 1996;6:299–306.
56. Park LP, Margolick JB, Giorgi JV, et al. Influence of HIV-1 infection and cigarette smoking on leukocyte profiles in homosexual men. The Multicenter AIDS Cohort Study. JAIDS. 1992;5:1124–1130.
57. Burns DN, Hillman D, Neaton JD, et al. Cigarette smoking, bacterial pneumonia, and other clinical outcomes in HIV-1 infection. JAIDS. 1996;13:374–383.
58. Conley LJ, Bush TJ, Buchbinder SP, et al. The association between cigarette smoking and selected HIV-related medical conditions. AIDS. 1996;10:1121–1126.
59. Galai N, Park LP, Wesch J, et al. Effect of smoking on the clinical progression of HIV-1 infection. JAIDS. 1997;14:451–458.
60. Ahdieh I, Gange SJ, Greenblatt R, et al. Selection by indication of potent antiretroviral therapy use in a large cohort of women infected with human immunodeficiency virus. Am J Epidemiol. 2000;152:923–933.
61. Touloumi G, Pocock SJ, Babiker AG, et al. Impact of missing data due to selective dropouts in cohort studies and clinical trials. Epidemiology. 2002;13:347–355.

HIV; AIDS; disease progression; drug use; CD4 lymphocyte count; HIV RNA level

© 2004 Lippincott Williams & Wilkins, Inc.