HIV infection is prevalent among substance abusers.1-3 In the United States, 18% of people living with HIV/AIDS have injection drug use as the mode of HIV transmission. Among injection drug users with HIV, 50% of men and 40% of women will be diagnosed with AIDS in less than 12 months after diagnosis of HIV infection.4 Noninjection drug use, including cocaine, crack-cocaine, and heroin sniffing, is also known to contribute to the spread of HIV through risky sexual behaviors including multiple sex partners, inconsistent condom use, and exchanging sex for drugs or money.3,5
Studies of the effects of illicit drug use on HIV disease progression before widespread use of highly active antiretroviral therapy (HAART) have shown an independent association between drug use with mortality and diagnosis of AIDS,6 whereas others found an association with the development of AIDS-defining illnesses but failed to find a relationship with mortality or specific effects on CD4 cell percentage.7-9 Studies that distinguished among the types of drugs used by people positive for HIV found that cocaine, and particularly crack-cocaine, increased the risk of progression to AIDS,9 with 1 study showing a faster decline of CD4 cell count over a short, 3- to 4-month period in those who used cocaine compared with those who did not.10
Two recent studies of HIV-1-positive women have strengthened the argument for an association between cocaine use and acceleration of HIV disease. Duncan et al11 in a cross-sectional analysis of a cohort of 80 African American women who were crack-cocaine smokers found that CD4+ cell counts were lower for a given HIV viral load with increased reported crack-cocaine use or urinary cocaine levels. Additionally, a multicenter study of women with HIV recently reported a 3.6 times greater risk for AIDS-related death in persistent crack users and a higher proportion reporting AIDS-defining illness during follow-up. Using random effects logistic regression analysis, they showed that persistent and intermittent crack-cocaine use was significantly associated with CD4 cell counts less than 200 cells per microliter and HIV-1 viral load greater than 100,000 copies per milliliter, independent of HAART.12
In the era of HAART, research on the effects of drug use on HIV disease progression has been hampered by the difficulty in separating the effects of medication nonadherence seen in active drug users from the physiologic effects of illicit drugs on disease progression. Numerous studies have shown poor antiretroviral (ARV) medication adherence in illicit drug users, particularly during periods of persistent substance use; therefore the effect that illicit drug use has on disease progression and mortality has been ascribed to poor medication adherence and decreased compliance with treatment-monitoring and clinical visits.13-16
Despite these findings, the evidence of more rapid disease progression and mortality with illicit drug use in the pre-HAART era suggests that drug use may affect disease progression independently of ARV adherence. In addition, there is compelling evidence from both animal and in vitro studies that illicit drugs and alcohol have a direct effect on intracellular HIV multiplication, progression to AIDS, and death.17-19 It is clear that further study into the specific modes of action of illicit drugs on HIV disease progression is needed. The objective of our study was to elucidate further the relationship between specific substances of abuse and disease progression determined by CD4+ cell count decline and viral load in the HAART era in a cohort of HIV-seropositive drug users in Miami.
A prospective, longitudinal study was conducted on a cohort of 222 HIV-infected drug users recruited between March 2002 and December 2005 for a 30-month study in Camillus House, Miami, FL. Participants were eligible for the study if they had documented HIV-seropositive status, were aged 18 years or older, and were active drug users (determined by urine drug toxicology). The study protocol was approved by the Florida International University Institutional Review Board, and all participants gave written informed consent before enrollment in the study.
Participants were followed monthly for 30 months. Demographics were collected at the initial screening visit. Biannually, blood was drawn for assessment of CD4+ cell count and HIV viral load, physical examination was performed, and urine samples for drug toxicology were collected. Questionnaires on history of alcohol and drug use in the preceding 6 months and report of health care utilization and current HIV medication use were administered at this time.
Participant Examination Protocol
Physical examination and medical history were performed by a nurse-practitioner after determination of eligibility and every 6 months thereafter. The medical history, administered verbally to participants, included a history of prescribed medication use in the previous 6 months, including ARV use. Participants were asked if they had been taking ARV medications as prescribed in the previous 6 months. If they answered in the affirmative, they were asked the type of ARV medications they were taking, and a review of records from the previous 6-month interview was used to determine any change in ARV use. Adherence with ARV medications was determined utilizing questionnaires regarding history of ARV medication use and attainment of undetectable HIV-1 RNA viral load (<400 copies/mL). For analysis, the participants were classified at each study visit as reporting adherence with HAART versus all others.
Venous blood was drawn for CD4+ cell count, viral load, and blood chemistry. Drug use was assessed every 6 months using a drug use questionnaire administered by research staff that detailed type, frequency, mode of administration, and quantity of cigarette, alcohol, and illicit drugs in the previous 6 months. Urine was collected for drug toxicology.
Lymphocyte phenotype was determined with a 4-color immunophenotyping panel of monoclonal antibodies. Differential counts were determined using a Coulter MaxM hematology instrument and corroborated with cytocentrifuge smears. Viral load was obtained by the reverse transcriptase-polymerase reaction using the Roche Amplicor reagents and protocol.
Descriptive statistics were used to characterize the population (N = 222). Data were assessed for normality of distribution, and transformations performed on variables when appropriate. HIV viral load was log transformed to normalize the distribution. Differences in baseline categorical characteristics by drug use were assessed using χ2 test, continuous variables with nonnormal distribution were analyzed using Mann-Whitney U test, and continuous variables with normal distribution were evaluated using the Student t test.
The association between substance use and time to decline of CD4 cell count to ≤200 cells per microliter over a 30-month period was evaluated longitudinally in participants who had a CD4+ cell count greater than 200 cells per microliter at baseline (n = 130), using Cox proportional hazards model. In the Cox analysis, CD4 cell count ≤200cells per microliter was defined as a nonrecurrent endpoint. Only data from participants who did not have missing CD4 cell count data were entered in the analyses. These analyses were conducted separately for crack-cocaine, powder cocaine, heroin, marijuana, and alcohol. Baseline CD4+ cell count, years since diagnosis of HIV, age, and gender as time-fixed variables, and ARV medication use as a time-dependent variable, were controlled for in this analysis. Similar analyses were run using the major drug use combinations.
Consistency over time of the use of drugs found to be predictive of CD4+ cell count decline was evaluated using descriptive statistics and correlations. The proportion of participants who reported use of the drug at baseline and continued the use in 80% or more of their follow-up visits was calculated for participants who had data from at least 2 assessment visits. Individual Pearson correlations were run between drug use reported at the baseline visit with drug use reported at each subsequent visit.
Drugs found to have a significant effect on CD4+ cell count decline were further assessed for an association with HIV viral load over time using generalized estimating equation models as described by Zeger and Liang,20 which are able to handle data sets with missing values. These models included age, gender, and years since HIV diagnosis as fixed-time covariates and HAART use and CD4 cell count as time-varying covariates. In addition, a comparison of the proportion of participants reporting ARV use whose HIV RNA viral load was <400 copies per milliliter, according to type of drug use, was assessed with data from the baseline, 6-month, and 12-month visits, using χ2 test. All analyses were conducted using the SPSS-15 for Windows statistical software package (Pearson Prentice Hall, Inc., Upper Saddle River, NJ).
Population and Baseline Characteristics
We recruited 222 HIV-infected drug users. The annual attrition rate was 8%. The population was 77% black, 13.1% white Hispanic, and 6.8% white non-Hispanic. Seventy-three percent were male with a mean age of 42 ± 7.4 years. At the baseline visit, median (range) CD4 cell count was 314 (2-1254) cells per microliter and median (range) HIV viral load was 11,884.5 (399-750,001) copies per milliliter, with 63% reporting ARV use. Mean (SD) income was $324 ($638.56), and rate of homelessness was 49%. (Table 1) The proportion of participants on ARVs was consistent throughout the study ranging from 63% to 67%. Median CD4 cell count ranged from 266 to 315 cells per microliter and HIV viral load from 2561 to 27,000 copies per milliliter through the 30 months of the study, with no directional trend of increase or decline.
The most common drugs of abuse at baseline in the population were crack-cocaine (50%), marijuana (35%), and powder cocaine (14%); and 83% of the population smoked cigarettes. A small portion of the population (6%) used heroin, with less than 1% using methadone or amphetamines. Injection drug use was uncommon (4%). Alcohol was consumed by 54.5% of the population with beer the most commonly consumed alcoholic beverage (48%); only 6% of participants consumed wine or liquor. The highest rate of heavy drug/alcohol use, defined as more than daily use, was found for crack-cocaine (16.1%) and alcohol (13.1%). There was significant multiple substance use, with 45% of those using alcohol, 32% of those using crack-cocaine, and 55% of those using cocaine and also using marijuana. Seventy-three percent of cocaine users and 65% of crack-cocaine users also drank alcohol.
Relative Risk for Decline of CD4 Cell Count to ≤200 Cells per Microliter by Substance Use
The effect of substance use (marijuana, powder cocaine, crack-cocaine, heroin, and alcohol) reported at the baseline visit on the rate of decline of CD4+ cell count to less than 200 cells per microliter, longitudinally over 30 months, was assessed in 130 participants who had CD4+ cell count greater than 200 cells per microliter at baseline and did not have missing CD4 cell count data. A participant was classified as a “user” of a substance if they reported any use in the previous 6 months and a “nonuser” if they reported no use of the substance in the previous 6 months. Only crack-cocaine use was significantly associated with time to event after controlling for baseline CD4+ cell count, ARV medication use as a time-dependent variable, years since self-reported HIV diagnosis, age, and gender. (Table 2, Fig. 1). Crack-cocaine use doubled the relative risk, of CD4 decline to ≤200 cells per microliter [hazard ratio (HR) = 2.14; 95% confidence interval (CI): 1.08 to 4.25, P = 0.029]. Results for heroin users were not conclusive, however, due to the small sample size of heroin users for this analysis (n = 8). The interaction between gender and drug use was tested and found to be nonsignificant. Due to the large number of participants who used both crack-cocaine and alcohol, we further controlled for alcohol in the model and found that crack-cocaine use remained a significant risk factor for faster decline of CD4+ cell count (HR = 2.093; 95% CI: 1.029 to 4.257, P = 0.041). Additionally, we evaluated crack-cocaine use as a risk factor for CD4+ cell count decline to less than 200 cells per microliter in those who were not on ARVs at baseline (n = 53) and found crack-cocaine quadrupled the relative risk for CD4+ cell decline (HR = 3.946; 95% CI: 1.049 to 14.85, P = 0.042) in this group.
To evaluate the effects of multisubstance use on relative risk of decline of CD4+ cell count to ≤200 cells per microliter, the major substance use combinations were entered into the Cox proportional hazards model. A participant who indicated at the baseline visit any use of 2 substances in the previous 6 months was classified as a “user” of that multisubstance combination. The multisubstance combinations occurring at baseline in 30% or more of participants were tested. These included marijuana and cocaine use, marijuana and crack use, marijuana and alcohol use, cocaine and alcohol use, and crack-cocaine and alcohol use. Only those who used both crack-cocaine and marijuana showed a significant increase in risk (HR = 2.42; 95% CI:1.042 to 5.617, P = 0.04) independent of baseline CD4+ cell count, HAART use, age, gender, and years since HIV diagnosis.
Consistency of Crack-Cocaine Use Over Time
One hundred ten participants reported crack-cocaine use at baseline, and 96 participants had at least 2 assessment visits. Of these, 80.2% reported continued crack-cocaine use or tested positive for cocaine in urinalysis in at least 80% of their follow-up visits. Significant correlations were found between baseline crack-cocaine use and crack-cocaine use at the 6-month (R = 0.447, P < 0.001), 12-month (R = 0.453, P < 0.001), 18-month (R = 0.298, P < 0.001), and 24-month visit (R = 0.426, P < 0.001). The correlation weakened at 30 months (R = 0.202, P = 0.074).
General Estimating Equation Model of Crack-Cocaine Use and Viral Load Over 30 Months
Participants were classified as a crack-cocaine “user” or “nonuser” at each of their follow-up visits. Participants who reported any use of crack-cocaine in the previous 6 months were classified as a “user,” and those who reported no use of crack-cocaine in the previous 6 months were classified as a “nonuser.” Age, gender, and years since diagnosis of HIV were included as fixed-time covariates and HAART use and CD4 cell count as time-varying covariates. As Table 3 indicates, crack-cocaine use was a significant predictor of higher viral load over the 30-month duration of the study in multivariate analysis (β = 0.345, P = 0.019).
ARV Use and HIV Viral Load Control by Substance Use
To evaluate the effect of substance use on adherence to ARV treatment, we compared the percentage of participants on ARVs who had controlled HIV viral load (HIV viral load < 400 copies/mL) by type of substance used. Analysis was performed separately for each assessment visit and is reported for the baseline, 6-month, and 12-month visits in Table 4, although similar results were obtained at all visits. The mean percentage of participants on ARVs with viral load control in the first year was 28.6%. At every assessment visit, a significantly lower proportion of those who used crack-cocaine had controlled viral load compared with those who did not use crack-cocaine. There were no significant differences in viral control in those who used powder cocaine, marijuana, alcohol, or combined crack-cocaine and marijuana when compared with those who did not use those drugs.
Demographic Parameters and Drug Use
There was no significant difference in age, baseline CD4+ cell count, self-reported ARV use, prevalence of homelessness, and mean income between those who used crack-cocaine and those who did not (not shown). A comparison of those who used both marijuana and crack-cocaine to those who did not also showed no significant differences.
Active substance use is associated with nonadherence to ARV medications, resulting in failure of viral suppression and reduced CD4 cell recovery.13,14,21 Recent research that evaluated the effect of drug use on disease progression adjusting for the confounding effects of HAART adherence, however, suggests that drug use may also accelerate disease progression independent of ARV noncompliance.22 Our results from an active drug using population show a significant acceleration of decline of CD4+ cell count to a level considered diagnostic for AIDS and elevated viral load in crack-cocaine users independent of reported HAART use over 30-month duration. This was not the case with powder cocaine, heroin, alcohol, or marijuana users. Definitive results cannot be made about heroin due to the small sample size of heroin users. Moreover, controlling for alcohol use in the model, crack-cocaine remained a significant factor for faster decline in CD4 cell count. Viral load over a 30-month period was significantly higher in crack users than those who did not use crack independent of HAART use over time. Although a significantly lower proportion of crack-cocaine users who were on HAART had controlled viral load, suggesting lower adherence to ARV medication, when considering only those not on HAART, crack-cocaine users continued to have a greater risk for HIV disease progression than those who did not use crack-cocaine. The only multidrug combination that significantly increased the risk of disease progression was the crack-cocaine and marijuana combination.
Previous studies have shown that illicit drug use accelerates HIV disease progression measured by an increased risk for developing new Class C opportunistic infections and that it accelerates time to death.7,22 In a cohort from the pre-HAART era, cocaine and hallucinogens were independently associated with increased mortality risk and hallucinogens were also independently associated with time to AIDS.6 Similarly, in a study of 98 HIV-1-seropositive drug users in a period before the introduction of zidovudine, cocaine users had a significantly greater short-term decline of CD4 compared with nonusers.10 To our knowledge, our study is the first long-term longitudinal study to show a specific effect of crack-cocaine on disease progression measured by CD4+ cell decline to ≤200 cells per microliter, in a mixed gender population of active substance users.
Other studies have failed to find an association between crack-cocaine use and CD4 decline. In a study that compared HIV-positive drug users who used cocaine, heroin, speedball, and crack-cocaine before the HAART era, Webber et al found that crack-cocaine was a significant predictor of AIDS hazard, defined using the 1993 Centers for Disease Control and Prevention criteria, but excluding CD4+ cell count ≤200 cells per microliter. In a study of Webber et al,9 however, neither slope of CD4+ cell decline nor survival differed between crack-cocaine users and nonusers. In contrast with our study, where only 4% were injection drug users, the population from the above mentioned study was 77% injection drug users and there was a gender bias in drug of choice, with women more likely to smoke crack-cocaine and with a higher median CD4+ cell count at baseline. In comparison, in our population, there were no significant differences in drug of choice or route of administration between men and women.
We have also found that viral load over a 30-month period was significantly higher in crack users than those who did not use crack independent of HAART use longitudinally. Significantly higher HIV viral load levels have also been found by Lucas et al13,22 in heroin or cocaine users compared with nonusers and decreased HIV viral load suppression has been observed in those who switch from drug nonuse to use. This poorer HIV-1 viral load outcome was attributed to the effect of drug use on ARV adherence. In contrast, Thorpe et al7 found no association between hard drug use, defined as any report of cocaine, crack, heroin, or other opiates, and HIV viral load in a repeated measures analysis of a cohort of women followed longitudinally. In all these studies, comparisons were made between drug users and those not using drugs and no comparisons between types of drug were performed. Our results are consistent with those of Lucas et al,22 in that crack-cocaine users on ARVs had consistently poorer adherence and HIV viral load control; our findings, however, go farther by using repeated measures analysis to show that this relationship remains, even after controlling for ARV use over time. Our mixed gender cohort and analysis comparing the effect of different types of drugs of abuse on disease progression may explain the difference in our findings and those of Thorpe et al.7
According to our data, crack-cocaine use has a multifactorial mode of action in which both direct effects on disease progression and reduced adherence to ARV medication seem to be compounded to accelerate disease progression. Chronic use of illicit drugs is associated with a variety of health problems, which can profoundly affect the HIV-1-infected individual, including effects on the central nervous system, circulatory system, the respiratory system, and other abnormalities.23-25 Furthermore, drugs of abuse may alter immunological indices and tend to stimulate HIV activity in vitro.17-19,26-29 The independent effect of crack-cocaine use from HAART use measured over time on increased risk of decline of the CD4+ cell count, and the significant risk of CD4+ cell count decline with crack-cocaine use in those not on ARVs at baseline, suggests that crack-cocaine may have a physiological action unrelated to HAART use that accelerates disease progression in this population. There is evidence from our data that crack-cocaine use is associated with increased viral load over time, independent of HAART, and that the crack users who were not on HAART had a faster HIV-1 disease progression as determined by CD4 cell decline to ≤200 cells per microliter, than participants who did not use crack-cocaine.
We did not find that cocaine in powder form was associated with faster CD4+ cell decline most likely due to a low frequency and prevalence of use. Of those who used powder cocaine, only 10% (3 individuals) reported daily or more than daily use and less than a quarter used the drug more than weekly. In contrast, 35% of crack users reported daily or more than daily usage and almost two-thirds used the drug more than weekly. Additionally, all crack users smoked the drug, whereas half of those using cocaine administered the drug nasally. Although the psychoactive and physiological effects of cocaine and crack-cocaine are similar, there is evidence of a greater risk of causing harm, propensity for dependence, and development of compulsive cocaine use when smoked.30-32
Drug use is not static, and patterns of use and subsequent abstinence may mirror the effects on medication adherence15 and HIV disease progression. Accordingly, Lucas et al22 categorized drug use in an observational cohort of HIV-infected patients as nonuse, intermittent use during abstinent and active use periods, and persistent use. The study found a relationship between periods of use and risk of developing new opportunistic conditions. We recruited only active drug users into this study and found striking consistency of drug use over time in our cohort, with 80% of those reporting crack-cocaine use at baseline also using it and/or testing positive for cocaine in urinalysis at >80% of their following visits. Significant correlations were also found between baseline crack use and reported crack use at the subsequent assessment visits. Because drug use was relatively consistent throughout the study period and active drug use was an inclusion criterion into the study, we do not believe that change in drug use over time was a major influencing factor in this study.
Multiple substance use is common in active drug users, and the synergistic or antagonistic effects of combining several types of drugs or alcohol can be difficult to separate. Our analysis of disease progression using major drugs alone and in combination revealed that drugs not found to affect HIV disease progression alone did not result in increased risk of disease progression in combination. Only participants who used both crack-cocaine and marijuana showed accelerated disease progression, and it is likely that this is related to the significant effect of crack-cocaine.
This study showed that an active drug using population had a significant acceleration of decline of CD4+ cell count to a level considered diagnostic for AIDS and elevated viral load in crack-cocaine users independent of use of HAART over 30-month duration. This seemed to occur due to lower adherence with HAART and an effect independent of HAART as crack-cocaine users not on HAART continued to have a greater risk for HIV disease progression than those who did not use crack-cocaine. These findings make a strong case for the need for targeted interventions to curb crack-cocaine use and to reduce overall drug use in those who are infected with HIV to slow disease progression in people who use crack-cocaine.
We would like to thank all the participants in the study without whom advancement in the management of HIV would not be possible. We also thank the Camillus House of Miami, FL, for providing space and resources without which this study would have not been feasible.
1. Kral AH, Blunenthal RN, Booth RE, et al. HIV seroprevalence among street-recruited injection drug and crack cocaine users in 16 US municipalities. Am J Public Health
2. Vignoles M, Avila MM, Osimani ML, et al. HIV seroincidence estimates among at-risk populations in Buenos Aires and Montevideo: use of the serologic testing algorithm for recent HIV seroconversion. J Acquir Immune Defic Syndr
3. Pechansky F, Woody G, Inciardi J, et al. HIV seroprevalence among drug users: an analysis of selected variables based on 10 years of data collection in Porto Alegre, Brazil. Drug Alcohol Depend
4. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report: HIV infection and AIDS in the United States and Dependent Areas, 2005. Vol 17, Tables 2, 9 & 11
. Rev. ed. Atlanta: US Department of Health and Human Services, CDC; 2007. Available at: http://www.cdc.gov/hiv/topics/surveillance/resources/reports/2005report/
. Accessed December 27, 2007.
5. Booth RE, Watters JK, Chitwood DD. HIV risk-related sex behaviors among injection drug users, crack smokers, and injection drug users who smoke crack. Am J Public Health
6. Vittinghoff E, Hessol NA, Bacchetti P, et al. Cofactors for HIV disease progression in a cohort of homosexual and bisexual men. J Acquir Immune Defic Syndr
7. Thorpe LE, Frederick M, Pitt J, et al. Effect of hard-drug use on CD4 cell percentage, HIV RNA Level, and progression to AIDS-defining class C events among HIV-infected women. J Acquir Immune Defic Syndr
8. Margolick JB, Munoz A, Vlahov D, et al. Direct comparison of the relationship between clinical outcome and change in CD4+
lymphocytes in human immunodeficiency virus-positive homosexual men and injecting drug users. Arch Intern Med
9. Webber MP, Schoenbaum EE, Gourevitch MN, et al. A prospective study of HIV disease progression in female and male drug users. AIDS
10. Siddiqui NS, Brown LS Jr, Makuch RW. Short-term declines in CD4 levels associated with cocaine use in HIV-1 seropositive, minority injecting drug users. J Natl Med Assoc
11. Duncan R, Shapshak P, Page JB, et al. Crack cocaine: effect modifier of RNA viral load and CD4 count in HIV infected African American women. Front Biosci
12. Cook JA, Burke-Miller, Cohen MH, et al. Crack cocaine, disease progression, and mortality in a multicenter cohort of HIV-1 positive women. AIDS
13. Lucas GM, Gebo KA, Chaisson RE, et al. Longitudinal assessment of the effects of drug and alcohol abuse on HIV-1 treatment outcomes in an urban clinic. AIDS
14. Arnsten JH, Demas PA, Grant RW, et al. Impact of active drug use on antiretroviral therapy adherence and viral suppression in HIV-infected drug users. J Gen Intern Med
15. Hinkin CH, Barclay TR, Castellon SA, et al. Drug use and medication adherence among HIV-1 infected individuals. AIDS Behav
16. Campa A, Jayaweera DT, Rafie C, et al. When access to antiretroviral for all is not enough. J Public Adm and Manage
17. Meyerhoff DJ. Effects of alcohol and HIV infection on the central nervous system. Alcohol Res Health
18. Bagasra O, Kajdacsy-Balla A, Lishcner HW, et al. Alcohol intake increases human immunodeficiency virus type 1 replication in human peripheral blood mononuclear cells. J Infect Dis
19. Lopez MC, Colombo LL, Huang DS, et al. Modification of thymic cell subsets induced by long-term cocaine administration during a murine retroviral infection producing AIDS. Clin Immunol Immunopathol
20. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics
21. Stein MD, Rich JD, Maksad J, et al. Adherence to antiretroviral therapy among HIV-infected methadone patients: effect of on-going illicit drug use. Am J Drug Alcohol Abuse
22. Lucas GM, Griswold M, Gebo KA, et al. Illicit drug use and HIV-1 disease progression: a longitudinal study in the era of highly active antiretroviral therapy. Am J Epidemiol
23. Isaacs SO, Martin P, Willoughby JH.“Crack” (an extra potent form of cocaine) abuse: a problem of the eighties. Oral Surg Oral Med Oral Pathol
24. Cherubin CE. The medical sequelae of narcotic addiction. Ann Intern Med
25. Louria DB, Hensle T, Rose J. The major complications of heroin addiction. Ann Intern Med
26. Peterson PK, Sharp BM, Gekker G. Morphine promotes the growth of HIV-1 in human peripheral blood mononuclear cell co-cultures. AIDS
27. Donahoe RM, Nicholson JK, Madden JJ, 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
28. Mientjes GH, van Ameijden EJ, van den Hoek AJ, et al. Increasing morbidity without rise in non-AIDS mortality among HIV-infected intravenous drug users in Amsterdam. AIDS
29. Brown SM, Stammel B, Taub RN, et al. Immunologic dysfunction in heroin addicts. Arch Intern Med
30. Levine SR, Brust JC, Futrell N, et al. A comparative study of the cerebrovascular complications of cocaine: alkaloidal versus hydrochloride-a review. Neurology
31. Schwartz RH, Luxenberg MG, Hoffmann NG.“Crack” use by American middle-class adolescent polydrug abusers. J Pediatr
32. Hatsukami DK, Fischman MW. Crack cocaine and cocaine hydrochloride. are the differences myth or reality? JAMA