What the Cohort Studies Say
September 8, 2000
Still, the basic issue put forth by Farber, Pasquarelli and Maggiore needs to be answered: If some unknown factor or factors unrelated to treatment reduced the number of AIDS patients, HAART could be receiving undeserved credit for the drop in deaths. On the other hand, if it can be shown that HAART has substantially improved patients' survival it is at least partly responsible for the good news.
The critical question, then, is: Is there evidence that HAART has improved the survival of HIV/AIDS patients? According to a leading denialist, University of California chemist David Rasnick, "It may come as a surprise that there is not even one study in the vast scientific, medical literature that shows that . . . a group of HIV-positive adults or children who take the anti-HIV drugs live longer or better quality lives than a similar group of adults or children who are HIV-positive and do not take the drugs."(11)
In fact there is an abundance of such evidence. Some, from clinical trials, has been discussed in detail in medical articles and at conferences. But clinical trials, conducted on limited numbers of patients for a relatively short time, with care often provided by physicians with more HIV expertise than average doctors, might not reflect what happens to most patients.
Real-world information on the impact of HAART in daily practice comes from what are known as cohort studies, which follow the experiences of specific groups of patients over extended periods of time. A number of large, prospective cohorts, specifically set up to track both the natural course of HIV infection and the effects of treatment and behavioral factors, have now reported results covering the pre- and post-HAART eras. Additionally, a number of individual hospitals and clinics have reported on the impact of HAART on their patients.
The results from these cohorts, covering tens of thousands of patients from a wide range of locations and backgrounds, have been astonishingly consistent despite differing methodologies: When HAART is introduced, opportunistic infections and deaths drop. Patients on anti-HIV therapy do better than those on no therapy, and those on regimens involving more drugs do better than those on fewer. Most of these analyses, by focusing on deaths among patients already diagnosed with AIDS, are not affected by any overall reduction in the number of AIDS cases, whether due to reduced HIV transmission or some unknown factor.
One of the world's largest AIDS cohorts is the CDC's Adult/Adolescent Spectrum of Disease Project. The ASD project began in 1990 and has enrolled over 49,000 participants at 93 hospitals and clinics in nine cities. As of January 1998, 19,565 had an AIDS diagnosis by the 1993 definition.
During that period 9,280 deaths were recorded, and researcher Amy McNaghten and colleagues included in their analysis all except 188 deaths caused by murder, suicide or drug overdose. Average survival time after diagnosis increased in the later years of the study, coinciding with a shift from monotherapy (a single antiretroviral, such as AZT alone, or ddI alone) to two-drug regimens, and later to three-drug HAART combinations. All anti-HIV regimens improved survival compared to no treatment, with more intensive regimens producing greater improvement. Patients on three-drug combinations had a 1.6 times lower risk of death than those on dual therapy and a 2.5 times lower risk of death than those on monotherapy.(12)
The ASD researchers later reported that incidence of AIDS- defining opportunistic infections in the whole study population of over 49,000 patients plummeted when HAART came into common use in 1996. Strikingly, 46 percent of PCP cases after 1996 occurred in people who had never been in HIV/AIDS care.(13)
One of the most-cited reports came from the HIV Outpatient Study, which has followed over 3,500 patients in eight U.S. cities since 1992. Researchers analyzed data for all who had ever had a CD4 count below 100 (considered most vulnerable for opportunistic infections or death) from 1994 through June, 1997. Use of protease-inhibitor-containing regimens among these 1,255 patients went from two percent in mid-1995 to 82 percent by June, 1997.
Mortality (deaths per 100 person-years) remained roughly constant in 1994 and 1995, then dropped abruptly in the second quarter of 1996 and continued dropping. To determine the effect of treatment, investigators classified patients by type of therapy: no antiretrovirals, nucleoside analogue monotherapy, nucleoside combination therapy, and combination therapy including a protease inhibitor. Patients on no anti-HIV treatment were 1.5 times as likely to die as those on monotherapy, 2.9 times as likely to die as those taking combination nucleosides and 4.5 times as likely to die as those on protease inhibitor combinations. The risk of serious opportunistic infections was reduced in a nearly identical pattern.(14)
Strikingly similar results were reported by the EuroSIDA cohort, a prospective observational cohort that began recruiting patients from across Europe in May 1994. In November 1998 researchers reported on all 4,270 patients enrolled who were over age 16 and had a CD4 count below 500. Through March 1998, 1,215 had died.
As in the HIV Outpatient Study, the death rate was analyzed by treatment category. The results, published in The Lancet, are broken down into six-month periods, and the correlation between more intensive regimens and fewer deaths is consistent and dramatic. The lowest death rate recorded in any period for patients on no treatment was 50.3 per 100 person-years, while for those on one antiretroviral the death rate never rose above 22.3 per 100 person-years. On two drugs deaths never rose above 7.9 per 100 person years and on three or more drugs the highest rate recorded was 3.9 per 100 person-years. In other words, the lowest death rate for patients on no anti-HIV drugs was 13 times the highest death rate recorded for those on three or more. The researchers further noted that "in any given 6-month period, the death rate among patients taking protease inhibitors was much lower than among those not taking protease inhibitors."(15)
The EuroSIDA researchers also examined opportunistic infection incidence for HAART and non-HAART patients. Patients with CD4 counts below 200 were over three times as likely to have an opportunistic infection if they weren't on HAART.(16)
Several other large European cohorts have reported similar results, including the Swiss HIV Cohort,(17) the Italian HIV Seroconverter Study(18) and the Italian Register for HIV Infection in Children.(19)
Copyright 2000 by John S. James. Permission granted for noncommercial reproduction, provided that our address and phone number are included if more than short quotations are used.
This article was provided by AIDS Treatment News. It is a part of the publication AIDS Treatment News.