HAART continues to improve survival, but not all HIV risk groups benefit equally

The prognosis of HIV-positive people improved significantly with the introduction of HAART in 1997, according to an analysis of 22 cohorts published in the October 17^th edition of The Lancet, and it continues to be sustained. However, compared to the pre-HAART era injecting drug users now have a risk of death four times greater than that of gay men, but individuals aged over 45 no longer seem to have an increased risk of progressing to AIDS.

Investigators involved in the CASCADE collaboration (Concerted Action on SeroConversion and AIDS Death in Europe) analysed data from 22 prospective and retrospective cohorts in Europe, Australia and Canada to see if the initial falls in AIDS deaths following the introduction of HAART in 1996/97 were being maintained. Although changes in CD4 cell count and viral load after the use of HAART have been intensively examined, the effect of demographic factors on prognosis in the HAART-era have been little explored. The CASCADE investigators designed their study to address this issue.

The demographic variables which the investigators wished to examine were age at seroconversion, HIV risk category, sex, and presentation with acute HIV infection before 1997 (pre-HAART), 1997-98 (limited HAART use) and 1999 – 2001 (widespread HAART use).

Data from 7740 individuals whose approximate date of HIV-seroconversion were analysed by the investigators. A total of 2000 individuals (26%) had died. The proportion of person-time spent on HAART increased significantly from 0.5% in 1995 to 22% in 1997, 40% in 1999 and 51% in 2001. At the same time, the proportion of person time spent naïve to therapy fell from 63% in 1995 to 25% in 2001. The proportion of time spent taking non-HAART regimens also fell from 26% in 1995 to 7% in 2001.

This increased use of HAART led to a substantial fall in the risk of progression to AIDS. Compared with the period before 1997, the hazard ratio (HR) of disease progression was .46 (95% CI, 0.38 – 0.55) in 1997, falling to 0.13 (95% CI, 0.09 – 0.21) by 2001 (trend, p

These reductions in progression to AIDS did not, however, fall constantly year on year. Rather, there was a sharp initial drop in the risk of disease progression in 1997 (HR 0.43 compared with previous years), followed by a more moderate but significant linear decline to 2001 (P=0.003).

Reductions in the risk of death were very similar to those for the reduction in the risk for disease progression, with once again, a sharp initial drop in observed in 1998, followed by smaller year on year risk reduction to 2001 (p

Investigators found strong evidence that the effect of exposure category on AIDS prognosis changed over the period of the study (p

The impact of age on prognosis also changed significantly over the period analysed by the CASCADE team. Before 1997 individuals aged over 45 were at a higher risk of AIDS (HR 2.03, 95% CI, 1.67 – 2.47, p

There was no evidence that the role of gender in relation to the risk of disease progression changed over time (p=0.93).

Although the risk of death fell for all categories, the greatest fall in mortality was seen in gay men (HR 0.09, 95% CI 0.06 – 0.14, p

Overall gender did not effect the risk of death, but the investigators found that male injecting drug users were at significantly greater risk of death than female drug users (male drug users HR4.59 versus female HR 2.99, p

The effect of age at seroconversion on the risk of death did not change significantly over the study period (p=0.63).

Commenting on their findings the investigators say, “of note is that as survival time has increased substantially in the HAART era, the relative importance of demographic factors has changed.” In particular injecting drug users appear to have much shorter survival time in the HAART era compared to gay men. Although this could in part to explained by a higher burden of co-infections such as hepatitis C virus, the fact that the investigators also found that drug users were at higher risk of progressing to AIDS suggests that HIV-related causes also have a role.

Several earlier studies have found that injecting drug users are less likely to access HAART. By contrast the CASCADE data found that the proportion of person-time spent on HAART with a viral load above 1000 copies/mL was 42% for injecting drug users and 27% for other risk categories.

Older age at seroconversion seems, the investigators note, to have become less important as a prognostic factor since the introduction of HAART. However, the age differences in the risk of mortality were preserved.

The investigators conclude by calling for longer follow-up to establish the long-term impact of HAART, and particularly to see if mortality in HIV-positive individuals continues to fall, level off, or indeed increase. In addition such follow-up will be able to determine if older people do have poorer survival in the HAART era.

Further information on this website

Prognosis - factsheet

Factors affecting disease progression - overview

Treatment response at six months best guide to long-term prognosis - news story

Response to HAART in first 18 months determines five-year prognosis - news story