|
|
| |||||||||||||||||||||||||||||||||||||||
Background and Methods The smoking of "crack" cocaine is thought to be associated with high-risk sexual practices that accelerate the spread of infection with the human immunodeficiency virus (HIV). We studied 2323 young adults, 18 to 29 years of age, who smoked crack regularly or who had never smoked crack. The study participants, recruited from the streets of inner-city neighborhoods in New York, Miami, and San Francisco, were interviewed and tested for HIV. This report presents the findings for the 1967 participants (85 percent) who had never injected drugs.
Results Of the 1137 crack smokers, 15.7 percent were positive for HIV antibody, as compared with 5.2 percent of the 830 nonsmokers (prevalence ratio adjusted for the city, 2.4; 99 percent confidence interval, 1.7 to 3.6). The prevalence of HIV was highest among the crack-smoking women in New York (29.6 percent) and Miami (23.0 percent). In these two cities, of the 283 women who had sex in exchange for money or drugs, 30.4 percent were infected with HIV as compared with 9.1 percent of the 286 other women (prevalence ratio, 3.1; 99 percent confidence interval, 1.9 to 5.1); of the 91 men who had anal sex with other men, 42.9 percent were infected with HIV as compared with 9.3 percent of the 582 men who did not have anal sex with other men (prevalence ratio, 4.7; 99 percent confidence interval, 3.0 to 7.4). In multivariable analyses, these high-risk sexual practices accounted for the higher prevalence of HIV infection among the crack smokers, as compared with those who did not smoke crack. Women who had recently had unprotected sex in exchange for money or drugs were as likely to be infected as men who had had sex with men (40.9 percent vs. 42.9 percent).
Conclusions In poor, inner-city communities young smokers of crack cocaine, particularly women who have sex in exchange for money or drugs, are at high risk for HIV infection. Crack use promotes the heterosexual transmission of HIV.
When inhaled, vaporized cocaine base gains rapid access to the bloodstream, producing a sudden, intense euphoria that subsides rapidly, leaving the user with a craving for more of the drug3,4,10,11. Addiction to crack cocaine is said to progress more rapidly than addiction to opiates or other forms of cocaine3,11,12. The epidemic of crack use, which has been accompanied by increases in violence and crime, has also been associated with the exploitation of addicted women, who are often induced to provide sex in exchange for crack or money to buy crack8. Crack-smoking women who exchange sex for money or drugs often lack control over their working conditions and suffer severe degradation2,8,9.
Increasing evidence suggests that the widespread use of crack cocaine has increased the spread of sexually transmitted diseases, including infection with the human immunodeficiency virus (HIV), because of high-risk sexual practices among crack users13,14,15,16,17,18. To investigate the relation of crack smoking to the HIV epidemic, we conducted a cross-sectional study of 2323 young crack smokers and nonsmokers in three large cities. This report presents the findings for the 1967 study participants (85 percent) who had never injected drugs.
Methods
Recruitment of Study Participants
We selected inner-city communities in three cities with high prevalences of illicit drug use and HIV infection: eastern and central Harlem in New York, Liberty City and Overtown in Miami, and Bayview-Hunter's Point in San Francisco1,2,8,13. The institutional review board of the participating center in each city approved the study. Eligible participants were 18 to 29 years of age and were either regular smokers or nonsmokers of crack cocaine. Regular smokers were defined as those who had smoked crack at least 3 days each week during the 30 days before recruitment; nonsmokers were defined as those who had never smoked crack. People who had smoked crack at some time in the past but were not doing so regularly at the time of their recruitment were excluded. We chose to study people under the age of 30 years because the crack epidemic has affected this age group disproportionately1,7,8,9,10,11,12,13 and because the risk of HIV infection in this age group would have developed primarily after the beginning of the crack epidemic1,2,3,4,5,6.
Outreach workers identified eligible crack smokers and nonsmokers on the streets and in public places through observation and informal conversation and brought those interested in participating to a study center. There, prospective participants were assured that the information they provided would not influence their reimbursement for participation in the study and would not be released to anyone who was not a member of the study team. Confidentiality certificates granted under section 301(d) of the Public Health Service Act (42 U.S.C. 241[d]) allowed us to provide broad protection of confidentiality. Those who gave written informed consent were interviewed, and blood samples were obtained. Participants received a small stipend ($10 to $15) at the time of the initial visit and again when they returned for their test results and post-test counseling.
Tests for HIV and Syphilis
Blood samples from the participants were tested for antibodies to HIV and syphilis by local, licensed laboratories. The Centers for Disease Control and Prevention (CDC) performed additional tests for syphilis (fluorescent treponemal-antibody absorption test with CDC reference reagents) and herpes simplex virus type 2 (type-specific Western blot assay for antibody to glycoprotein G19). A reactive test for syphilis was defined as a reactive result on the treponemal test, indicating infection with syphilis at any time, past or present.
Statistical Analysis
For the data analysis, women who reported having ever received money or drugs in exchange for sex (oral, vaginal, or anal) with men were classified as having engaged in sexual work. Those who reported having received money or drugs in exchange for vaginal or anal sex with men in the previous 30 days without using a condom on each occasion were classified as having engaged in recent, unprotected sexual work. Because of the low prevalence of HIV among the participants recruited in San Francisco, the examination of factors associated with HIV serologic status was restricted to participants in New York and Miami.
Data were analyzed with the use of Statistical Analysis Software (SAS Institute, Cary, N.C.), version 6. Distributions of categorical variables were compared by the chi-square test, and prevalence ratios and test-based confidence intervals were calculated. Mantel-Haenszel techniques for summarizing relative risks across strata were used to adjust for the city in which a participant was recruited20. Distributions of continuous variables were compared by the Wilcoxon rank-sum test. Because of the large sample and the large number of comparisons, a P value of less than 0.01 was considered to indicate statistical significance, and 99 percent confidence intervals are presented.
Multivariable modeling was used to assess the extent to which the sexual practices associated with HIV infection accounted for the higher prevalence of HIV infection among the crack smokers. Because of the large number of variables, exploratory analyses were conducted first, with classification and regression trees21 used to select the sexual practices that most strongly predicted HIV serologic status. A logistic-regression model was then constructed to estimate the risk of HIV infection associated with crack smoking, with adjustment for the selected sexual practices. Demographic factors that changed the estimate of risk associated with crack smoking by more than 5 percent were also included in this model. A second model was constructed with the same adjustment procedures to examine the association between sexually transmitted diseases and HIV infection. In this model we examined the two variables most closely related to HIV infection in the univariate analyses: herpes infection reported by the participant and a reactive test for syphilis.
Results
From January 1991 through September 1992, our outreach workers approached 2902 people whom they considered eligible for the study; 123 (4 percent) refused to participate, and 245 (8 percent) did not appear at the study center. Of the 2534 people who came to the study centers, 165 (7 percent) were found to be ineligible on rescreening, and 46 (2 percent) refused to participate, leaving 2323 eligible participants. The 1967 participants (85 percent) who reported never having injected drugs are the subject of this report. They included 1137 regular smokers of crack and 830 people who had never smoked crack.
Demographic Characteristics
The crack smokers and the nonsmokers were similar in age (median, 26 and 24 years, respectively) and sex (51 percent of both groups were men), but the crack smokers were more likely than the nonsmokers to be African American and less likely to be Hispanic or white (P<0.01). Eighty-seven percent of the crack smokers and 81 percent of the nonsmokers were non-Hispanic African Americans, 8 percent of the smokers and 9 percent of the nonsmokers were Hispanics, 3 percent of the smokers and 5 percent of the nonsmokers were black Hispanics, and 1 percent of the smokers and 5 percent of the nonsmokers were non-Hispanic whites.
Characteristics of Crack Smokers
The crack smokers were socially and economically disenfranchised, although in some respects no more so than the nonsmokers (Table 1). Most of the crack smokers were unmarried, and many had children, but in most cases their children did not live with them. Although nearly half had graduated from high school, more than two thirds supported themselves principally by means of public assistance or illegal activity. Sixty-three percent lived on less than $500 per month, and almost 20 percent were living on the streets. One third had been incarcerated in the previous 12 months. Nearly half the men and one quarter of the women reported having had sex willingly by the age of 13 years. Nineteen percent of the crack-smoking women reported having been raped in the preceding 12 months.
|
Sexual Practices and Sexually Transmitted Diseases
Crack smokers of both sexes were more likely than nonsmokers to report high-risk sexual practices and a history of sexually transmitted diseases (Table 2). The crack smokers had had more sexual partners than the nonsmokers (median for the men, 47 vs. 20 partners, P<0.001; median for the women, 25 vs. 5 partners, P<0.001). Crack-smoking women were more likely than nonsmoking women to have engaged in sexual work at any time and to have engaged in recent, unprotected sexual work (Table 2). Of the 310 women who had engaged in recent sexual work, 97 percent had engaged in vaginal sex, 62 percent in oral sex, and 5 percent in anal sex; in exchange, 96 percent had received money, 52 percent had received crack, and 1 percent had received other drugs. Crack-smoking men were more likely than nonsmoking men to have had anal sex with a male partner and to have had 50 or more such partners (Table 2). Of the 105 men who had ever had anal sex with another man, 33 of 87 crack smokers had done so only in exchange for money or drugs, as compared with none of 18 nonsmokers (P = 0.002).
|
HIV Seroprevalence
The prevalence of HIV infection varied widely according to the city of recruitment, sex, and crack smoking (Table 3). The prevalence was highest among crack-smoking women in New York and Miami (29.6 and 23.0 percent, respectively) (Table 3).
|
|
These four sexual practices (sexual work, recent, unprotected sexual work, anal sex between men, and homosexual anal intercourse with 50 or more male partners) accounted for the higher prevalence of HIV infection among crack smokers. After adjustment in a logistic-regression model for these practices, recruitment in New York, African American race, and homelessness, the prevalence of HIV infection was similar among crack smokers and nonsmokers (adjusted odds ratio, 1.1; 99 percent confidence interval, 0.6 to 2.0).
HIV infection was more prevalent among the participants who had had sexually transmitted diseases, particularly genital ulcer diseases (Table 4). These associations persisted when men who had sex with men were excluded from the analyses. A reactive syphilis test (adjusted odds ratio, 2.3; 99 percent confidence interval, 1.4 to 3.6) and a history of herpes (adjusted odds ratio, 3.6; 99 percent confidence interval, 1.1 to 11) remained significantly associated with HIV infection after adjustment in a logistic-regression model for the four high-risk sexual practices and African American race.
Discussion
Although the epidemiologic relation between crack smoking and HIV infection has been reported anecdotally and in several relatively small, local, clinic-based studies,17,18,22 our study provides estimates of the high prevalence of HIV infection in three inner-city communities and the modes of transmission among active crack smokers. Several caveats, however, deserve mention. First, we recruited the study participants on the streets and in public places specifically to obtain a sample of active drug users from the community, including those who may not have been in contact with substance-abuse treatment programs, medical services, or the criminal-justice system. Recruitment from the streets, however, results in an overrepresentation of members of visible populations, such as people who live or spend their time on the streets, including people engaging in sexual work or drug dealing; less visible drug users (those with daytime jobs and those who use drugs in private) are underrepresented.
Second, to examine the effects of heavy crack use, we deliberately recruited people who were smoking crack regularly at the time of the study and excluded former crack smokers and those who smoked infrequently. Thus, we have no information about people who may be able to smoke crack without becoming heavy users.
Third, our data on drug use and sexual behavior are based on the participants' own reports; we have no way to verify the accuracy of those reports. However, our interviewers believed they obtained honest answers most or all of the time from 94 percent of the participants. Data reported by study participants are generally accurate when they are collected in settings that the participants perceive as safe, with assurances that their responses will not be used against them23,24.
Despite these potential limitations, several important findings emerge from our data. The high prevalence of HIV infection that we found among participants who had never injected drugs and had never engaged in male homosexual anal intercourse demonstrates the extensive heterosexual transmission of HIV in these communities of low-income minority groups affected by both the crack and HIV epidemics. This heterosexual spread is supported by the associations we found between HIV infection and both sexual work and genital ulcer diseases, as well as by the relatively high prevalence of HIV (9 percent) and the 1:1 male:female ratio of prevalence among the participants who reported that they did not engage in high-risk sexual practices.
Overall, the prevalence of HIV infection was 2.4 times higher among the crack smokers than among the nonsmokers. When high-risk sexual practices were accounted for, crack smoking was not significantly associated with HIV infection, indicating that the higher prevalence of HIV infection among the crack smokers was due to the greater frequency of such practices in this group. Thus, crack smoking appears to lead to the transmission of HIV through its association with high-risk sexual practices. Although previous studies have reported that oral sex is the most common type of sex exchanged for crack,8 the women in our study who engaged in sexual work received money more often than crack, and provided vaginal sex more often than other types of sex. Although many of these women did not always use condoms with their clients, many did, with a considerable reduction in the prevalence of HIV infection.
Sexually transmitted diseases were widespread among the crack smokers; 80 percent reported having had such a disease. Outbreaks of syphilis among crack smokers have been reported25,26. Sexually transmitted diseases may increase the likelihood of HIV transmission27. Two genital ulcer diseases, herpes and syphilis, were strongly and independently associated with HIV infection in our multivariable analysis that controlled for high-risk sexual practices. Thus, our data support the possibility that the spread of sexually transmitted diseases among crack smokers facilitates the spread of HIV.
The associations between crack smoking and high-risk sexual practices, sexually transmitted diseases, and HIV infection were all stronger among the women than among the men, and the prevalence of HIV infection was considerably higher among the crack-smoking women than among the crack-smoking men. Thus, these intersecting epidemics are particularly dangerous for young women. The women who exchanged sex for money or drugs in the communities we studied had a very high risk of HIV infection, similar to that for men who have sex with men. During its first decade, the HIV epidemic predominantly affected men who have sex with men, injection-drug users, and their sexual partners and children. Our data demonstrate that young, crack-smoking women in poor African American and Hispanic communities who have sex in exchange for money or drugs are a new and sizable group at very high risk for HIV infection.
Our findings provide a strong rationale for preventive efforts aimed at crack smokers in communities affected by the dual epidemics of crack smoking and HIV infection. Interventions are needed to prevent addiction to crack, treat those who are addicted, and reduce the frequency of high-risk sexual practices among crack users, particularly women. Although some suggested that crack use would decline as the devastating consequences of addiction to the drug became widely recognized,28 the use of crack remains widespread and a major cause of morbidity and violence in many urban areas29. Emergency room visits related to the use of cocaine, which increased more than 10-fold from 1980 to 1990, rose an additional 50 percent from 1990 to 199229. Our outreach workers had little difficulty finding more than 1000 active crack smokers for this study.
Preventive efforts directed at crack smokers must take into account the social and economic conditions associated with drug use. We have shown that active, heavy users of crack cocaine can be recruited on the street through outreach methods and the offer of a small stipend; indeed, our interviewers found that many were eager to talk about their lives. But our data also indicate the severe and desperate conditions in which the study participants lived. In that world the lack of food, shelter, and safety and separation from loved ones are more immediate problems than the risk of a chronic infection that may result in death many years in the future8,30,31. To be effective, drug treatment and efforts to reduce the risk of HIV infection will require broad social and economic interventions that address these related needs. Without effective interventions, crack use is likely to result in continued heterosexual transmission of HIV, catalyzing the spread of the epidemic from men who have sex with men and injection-drug users to the heterosexual population.
Supported by grants (U64/CCU204582, U64/CCU404539, and U64/CCU904453) from the Centers for Disease Control and Prevention.
We dedicate this paper to Yolanda Serrano, the founder and former executive director of the Association for Drug Abuse Prevention and Treatment, who died on October 21, 1993. Her heroic work fighting AIDS in impoverished communities of minorities will long be remembered.
Source Information
From the Division of HIV/AIDS, Centers for Disease Control and Prevention, Atlanta (B.R.E., K.L.I., S.D.H.); the Association for Drug Abuse Prevention and Treatment, New York (S.F., Y.S.); the University of Miami, Miami (C.B.M.); Bayview-Hunter's Point Foundation, San Francisco (C.W.); the University of Delaware, Newark (J.A.I.); California State University, Hayward (B.P.B.); and Columbia University, New York (R.F.S.). Presented in part at the First National Conference on Human Retroviruses, Washington, D.C., December 14, 1993.The members of the Multicenter Crack Cocaine and HIV Infection Study Team are listed in the Appendix.
Address reprint requests to Dr. Edlin at the Division of HIV/AIDS (E-45), Centers for Disease Control and Prevention, 1600 Clifton Rd., Atlanta, GA 30333.
References
The Multicenter Crack Cocaine and HIV Infection Study Team includes the following organizations and investigators: the Association for Drug Abuse Prevention and Treatment, New York: executive director -- Yolanda Serrano (deceased); principal investigators -- S. Faruque and R.F. Schilling; coprincipal investigator -- N. El-Bassel; research assistants -- A. Cabrera, Y. Flores, N. Guzman, R. Hogan, N. Melendez, E. Nieves, A. Ocasio, G. Rivera, A. Rizzolo, F. Steele, and S. Turso; data managers -- B. Bidassie and S. Kuo-Hsien; and laboratory director -- D. Strauss; the Comprehensive Drug Research Center, University of Miami, Miami: principal investigators -- C.B. McCoy and J.A. Inciardi; project director -- H.V. McCoy; data and statistical consultants -- D.C. McBride, N. Weatherby, and J.E. Rivers; research assistants -- E. Alonso, G. Aristide, M. Ashely, S. Bowens, D. Buck-Walden, S. Comerford, V. DeVeaux-Shepard, E. Dyer, M. Galvez, J. Griffin, M. Jones, V. LoCascio, L. Magilner, N. Mendez, C. McKay, L. McQueen, C. Miles, R. Miranda, L. Pagan, R.-M. Pierre, R. Salas, L. Seoane, B. Shabazz, and E. Walden; and laboratory directors -- M.A. Fletcher and R. Garcia-Morales; Bayview-Hunter's Point Foundation, San Francisco: principal investigators -- B.P. Bowser and P.E. Evans; project director -- C.O. Word; research assistants -- C. Ballesteros, S. Byrd, W. Curtis, D. Dogan, A. Garner, M. Griffin, C. Hawkins, D. Hunter-Gamble, C. Iregui, M. Justice, M. Lee, M. Lodico, J. McGilroy, V. Patterson, S. Penn, C. Perkins, N. Persaud, C. Richardson, and N. Robertson; and laboratory director -- A. Back (deceased); and the Centers for Disease Control and Prevention, Atlanta: principal investigator -- B.R. Edlin; coprincipal investigators -- K.L. Irwin and S.D. Holmberg; laboratory directors -- S. Larsen and D.S. Schmidt; and data and statistical consultants -- R.H. Byers, D. Ludwig, R. Johnson, L. Wong, and J. Dushku.
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Crack Cocaine and HIV in the Inner City
Sugarman K., Herman M., Fernando D., Edlin B. R., Faruque S., McCoy C. B., Word C. O., Des Jarlais D. C., Padian N.
Extract |
Full Text
N Engl J Med 1995;
332:1233-1235, May 4, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2010 Massachusetts Medical Society. All rights reserved. |
Previous
