Women with HIV who regularly use crack are three times more likely to die

Gus Cairns
Published: 25 June 2008

A US survey of 1686 women with HIV on antiretroviral treatment has found that the 29% who regularly or intermittently used crack cocaine were nearly 60% more likely to develop an AIDS-defining illness, and the 3.2% who used it persistently were three times more likely to die.

It also found that persistent users had baseline HIV viral loads that were on average three times higher than intermittent or non-users, and that this persisted throughout the study despite antiretroviral treatment. This was partly due to lower levels of adherence to HAART, but HIV disease progression in crack users and mortality in persistent users remained higher even when figures were adjusted for reported adherence and baseline viral load and CD4 count.

This was the first longitudinal study of a large group of women to confirm what most, but not all cross-sectional studies have found – that usage of crack and cocaine seems to exacerbate the effects of HIV, independently of treatment status and adherence.

The Women’s Interagency HIV Study (WIHS) is a prospective cohort study of HIV disease progression among 2058 women with HIV at six HIV centres in Chicago, Los Angeles, the San Francisco Bay Area, Washington DC and Brooklyn and Bronx in New York. The current study looked at 1686 women who made at least two study visits between April 1996 and September 2004. Study visits were every six months but women did not have to have two consecutive ones to be included in the study. Women made an average of 12 study visits, with an average follow up time of 18-20 months. Persistent crack users only made an average of six study visits but had a longer follow-up time of 33 months.

At the study baseline 483 women (28.6%) admitted using crack. These were divided into 429 intermittent users (25.4%), who only reported crack use at some study visits, and 54 persistent users (3.2%), who reported crack use at every visit. At each clinic visit the intermittent users were further divided into ‘intermittent abstinent’ users who had used crack in the past but were not currently and ‘intermittent active’ users who reported current use but had not done at every previous visit.

Fifty-six per cent of the women were African American, 24% Latina and 20% white or of other ethnicity. Crack users were more likely to be black (70% of crack users were African American), to be high school drop outs (especially the persistent users) and to have problematic alcohol use alongside crack use – 10% of non-users of crack, 27% of intermittent users and 32% of persistent users had a drink problem too.

At the start of the study 73% of the women were either currently on HAART or had been in the past, but only 32% of the persistent crack users had exposure to antiretroviral drugs. Adherence to HAART in the study participants was not high, though the definition of high adherence was the fairly exacting one of more than 95% of doses taken since the last visit, at every visit. Twenty-nine per cent of non-users of crack, 16% of intermittent users and only 7% of persistent users (four women) reported high adherence as so defined.

Persistent crack users had higher viral loads and lower CD4 counts at baseline. The average viral load at the start of the study was 11,000 copies/ml in non-users, 10,300 copies/ml in intermittent users and 34,000 copies/ml in persistent users, while baseline CD4 counts were 364, 433 and 257 cells/mm3 respectively. Forty per cent of persistent crack users had baseline CD4 counts under 200 compared with 29% of non-users and 22% of intermittent users.

There was a high death rate among participants, with 25% of the women dying, of any cause, during the study period – 419 deaths in all. Of these 47% were AIDS-related, 33% non-AIDS related and the remainder indeterminate.

Mortality was dramatically higher among persistent crack users. Thirty-seven of the 54 persistent users died during the study – 68 per cent. This may be partly due to selection bias: women who died had less time to stop using crack and become classed as intermittent users. The estimated survival rates by 3,000 days were 89% for non-users, 90% for intermittent users, but only 65% for persistent users. The higher death rate among persistent users remained (odds ratio, 3.61) when figures were adjusted for age, race, income, education, drinking, HAART adherence and CD4 count and viral load at baseline.

Just under one-third of the women (32.3% or 543 women) developed a newly-acquired AIDS-defining illness during the study. Progression to AIDS was significantly greater among both intermittent crack users (42%) and persistent users (39%). Both persistent and intermittent crack use remained associated with progression to AIDS after adjusting for HAART adherence, drinking, socioeconomic status, baseline viral load and baseline CD4 count. In the adjusted analysis intermittent users were 57% more likely, and persistent users 65% more likely, to develop AIDS than non-users.

The proportion of women with CD4 counts under 200 cells/mm3 stayed at 25% in intermittent users, went down throughout the course of the study from 29% at baseline to 17% at the end of the study in non-users, and varied erratically between 23% and 45% in persistent users. Similarly the proportion of women with HIV viral loads over 100,000 copies/ml went down from 17% at baseline to 2% in non-users at the end of the study and from 17% in intermittent users to 8%, but in persistent users, after an initial decline from 47% to 3%, it then varied between 8% and 27% for the rest of the study.

Crack use was the single most important influence on both viral load and CD4 count in the study. Women who had used crack in the past but not currently were 67% more likely to have a CD4 count under 200 and 45% more likely to have a viral load over 100,000 copies/ml than non users; current but non-regular users were 98% and 58% more likely to have CD4s under 200 cells/mm3 and a viral load over 100,000 copies/ml respectively; while persistent users were 82% more likely to have a CD4 count under 200 cells/mm3 and 224%, or over three times, more likely to have a viral load over 100,000 copies/ml (all figures adjusted for other variables).

The only other significant influences on CD4 count and/or viral load were persistent problem drinking (drinkers were nearly twice as likely to have a viral load over 100,000 copies/ml, though there was no influence on CD4 count), length of HIV diagnosis, being Latina (69% more likely to have less than 200 CD4s), low income (20% more likely to have a low CD4 count and 26% more likely to have a high viral load), age (42% more likely to have a low CD4 count for every 10 year increase in age; no effect on viral load) and finally adherence: women with over 95% adherence were less than half as likely to have a CD4 count under 200 cells/mm3 and less than a third as likely to have a viral load over 100,000 copies/ml.

However, persistent crack use was even more likely to predict a high viral load than high adherence was to predict a low one.

One interesting finding was that low levels of HAART adherence and high morbidity and mortality rates were not associated with low levels of health care. At their last interview, 100% of participants reported seeing a healthcare provider in the last six months and 93% said they saw the same provider consistently, including 94% of the persistent crack users.

Are the study findings due to direct effects of crack on immune status and HIV replication in users? Previous studies, the authors point out, have shown that cocaine causes immune alterations in T-cells, inhibits the functions of other immune cells like macrophages and neutrophils, suppresses cell-signalling chemicals (cytokines), and increases HIV replication in cells in the test tube.

Recent studies have also found that cocaine increases the permeability of the blood-brain barrier, thus increasing the vulnerability of the central nervous system to HIV infection, and that crack users develop chronic lung disease because of inhaling crack contaminants. There was a predominance of respiratory diseases in the women who developed AIDS-defining conditions; 18% developed bacterial pneumonia, 10% PCP and 4% TB.

On the other hand, this study cannot rule out other explanations for higher morbidity and mortality among persistent crack users such as greater sexual risk taking, poorer diet, homelessness, and real as opposed to reported adherence. Its greatest weakness is that researchers did not take urine or blood samples to measure true crack and alcohol exposure and similarly took no direct measures of HIV drug levels or other direct ways of monitoring HAART adherence and exposure. Research in prevention studies has shown that self-reported adherence is not always reliable. Although the results are highly suggestive that crack and cocaine directly worsen HIV infection, it is still an open question as to whether it is the drugs or the lifestyle of drug users that cause more damage.


Cook JA et al. Crack cocaine, disease progression, and mortality in a multicenter cohort of HIV-1 positive women. AIDS 22:1355-1363. 2008.

Community Consensus Statement on Access to HIV Treatment and its Use for Prevention

Together, we can make it happen

We can end HIV soon if people have equal access to HIV drugs as treatment and as PrEP, and have free choice over whether to take them.

Launched today, the Community Consensus Statement is a basic set of principles aimed at making sure that happens.

The Community Consensus Statement is a joint initiative of AVAC, EATG, MSMGF, GNP+, HIV i-Base, the International HIV/AIDS Alliance, ITPC and NAM/aidsmap