Although much has been made of the increased survival seen in the age of HAART, Swiss researchers have determined that the Swiss HIV-positive population taken as a whole still have a much lower life expectancy – with a risk of death of 3% per year – compared to the general Swiss population, although there are huge variations based, mainly, on whether or not the individual is taking HAART and/or injecting drugs. The study appears in the September 3rd issue of the journal AIDS.
The study covers the eleven years from January 1990 to December 2001, and includes a cohort of almost 11,000 individuals. Almost half (46%) of the cohort acquired HIV through injection drug use (IDU), 32% were men who acquired HIV through sex with another man (MSM), 18% acquired HIV through heterosexual sex, and 4% either acquired HIV in some other way (e.g. blood transfusion), or their route of transmission is unknown.
During the study period, 3630 (33%) individuals died; the median age of death varied from 33 years in IDU women to 41 years in non-IDU men. Importantly, the yearly risk-of-death declined significantly from 13% in the pre-HAART years (1990-1995), to 3% in the post-HAART years (1998-2001).
To ascertain just how this compares with the general Swiss population, the investigators calculated standardised mortality ratios (SMRs). These compare the number of actual deaths with the number of deaths one would expect to see in the same age group and gender in the general Swiss population, and are often used by life insurance companies to calculate the mortality risk of certain populations.
In the pre-HAART era, a person in the Swiss HIV cohort was almost 80 times more likely to die than someone of the same age and gender in the general Swiss population. In the post-HAART era, someone on HAART in the Swiss HIV cohort was only ten times more likely to die, and 15 times more likely to die if not on HAART.
However, there is much variation of SMRs within the cohort, particularly when it comes to IDUs. Prior to HAART, there wasn’t such a great difference in the SMRs of IDUs (98.2) and non-IDUs (69.2). However, in the post-HAART era, whilst non-IDUs on HAART had an SMR of 6.8, IDUs on HAART had an SMR of 26.7. The authors suggest that this could be due to poorer adherence and more frequent non-AIDS-related deaths, including overdose and hepatitis C-related liver disease.
Although there is a difference between the SMRs of men and women in this study, the authors point out that this is because in Switzerland, men between the ages of 20 and 45 have a higher mortality rate than women of the same age, and that previous studies have shown that HIV-positive men and women are of equal risk of dying. However, this study found that non-IDU men on HAART had the lowest SMR of 5.2 – i.e. they were just over five times more likely to die than someone of their age in the general Swiss population.
The investigators also looked the survival function, which is based on life tables in which the survival of each individual is projected on the basis of current age-specific mortality rates. This hypothesised the life expectancy of someone who was infected with HIV at the age of 20. It found that, compared with the general Swiss population – who have a 90% chance of living to 60 years of age – a non-IDU HIV-positive person on HAART has a 50% chance of living to 60. The chance is much reduced for IDUs, whose probability of living to 60 is less 20%. Without HAART, no-one with HIV was expected to reach 60.
It should be pointed out that the study only analysed the mortality of HIV-positive people in general, and did not differentiate between HIV-related and other causes, and that these estimates are specific only to Switzerland. The authors say that although their SMRs for people on HAART are similar to those reported in the French APROCO cohort study, “calculation of SMRs does not completely remove the influence of differences in population composition and therefore these SMRs are not directly comparable with those presented here...”
Further complicating the analysis was the fact that 2290 (21%) individuals dropped out of the study. Women were more likely to be lost to follow up than men, and IDUs were more likely to dropout than non-IDUs. The researchers used the last-known CD4 count to estimate whether these individuals were likely to have died or survived during the study period, and used the median between that analysis and an assumption that all dropouts died 14 months after leaving the study.
The authors conclude that “during the period in which HAART has been available, mortality remains higher in HIV-positive patients than in the general Swiss population,” however, “the differences are substantial among subgroups.”
They add that “this observed mortality pattern probably does not apply to a newly HIV-infected individual,” since things may improve with the advent of new, more powerful and less toxic drugs, or may deteriorate “with increased resistance to therapeutic drugs, and an increased impact of side effects. As new treatments for HIV appear, it will continue to be important to identify differential response to those treatments across demographic groups.”
Keiser O et al. All cause mortality in the Swiss HIV cohort study from 1990 to 2001 in comparison with the Swiss population. AIDS 18: 1835-1843, 2004.