One of the most remarkable characteristics of HIV infection is its very long incubation period. Identifying the factors influencing the incubation period is important to our understanding of the natural history of the disease. On an individual basis, predicting HIV disease progression is helpful in deciding on the best moment to start antiviral therapy. This is an important decision as the available evidence indicates that such treatment should be continued for life and has considerable toxic side-effects.
HIV disease progression among homosexual men
The best way to get information about AIDS survival and death is through prospective cohort studies of HIV-infected individuals with well-documented dates of seroconversion. Such studies were started worldwide in the 1980s and thus have limited follow-up time. Many HIV-infected individuals are now being treated with antiretrovirals. To describe the long-term survival after HIV infection, we collected data about 362 homosexual men who participated in hepatitis B vaccine trials in Amsterdam, New York City and San Francisco at the end of the 1970s and were followed from 1978 through 1995, before the availability of highly active antiviral therapy (HAART) ( Koblin et al. 1999 ). These men seroconverted during follow-up or were seropositive at entry into the studies. The expected date of seroconversion was calculated for each subject conditional upon the date of the last HIV-negative and the first HIV-positive test, using the cohort specific sero-incidence data.
The median survival time to death was 12.1 years (95% confidence interval (CI): 11.4–12.9) and did not differ between cities. At 15 years after seroconversion, 37.2% survived (95% CI: 31.4–42.9). The annual risk of dying increased at a constant rate until 8 years after seroconversion, when it levelled off, suggesting that there is a group who is relatively resistant to progression. However, as the numbers are small, the confidence intervals are wide especially at the end of the follow-up. The levelling off could also reflect some treatment improvements, particularly prophylaxis for opportunistic infections. In multivariate analysis only older age at seroconversion was significantly associated with more rapid progression to death. Survival in the most recent calendar period was not significantly different from that in earlier periods, suggesting a limited effect of treatment at least through 1995.
The effect of age on hiv disease progression
Recently a British initiative brought together data from 38 follow-up studies in Europe, North America and Australia. This collaborative study ( CASCADE study 2000) contains data on 13030 HIV-1 infected individuals whose date of seroconversion could be reliably estimated. Almost a third of the participants were homosexual men and about a quarter were persons with haemophilia; 89% were males. The participants seroconverted in the period 1977–96 and the median age at seroconversion was 27 years (range < 1–90). The incidence of AIDS and the mortality increased strongly with age at seroconversion. Survival was 12.5 years (95% CI: 12.1–12.9) for the age group 15–24, and 7.9 years (95% CI: 7.4–8.5) for those aged 45–54. For AIDS survival the corresponding figures were 11.0 and 7.7 years. The annual mortality rate increased by a factor 1.5 for each 10-year increase in age at seroconversion. There was no appreciable effect of exposure category on survival. For AIDS such an effect could be shown but this was explained by the high incidence of Kaposi sarcoma among homosexual men.
Does gender play a role?
To examine the effect of gender on disease progression, we recently compared the CD4 trajectory in male and female HIV-positive injecting drug users (IDU) for whom the dates of a last negative and first positive HIV test was known ( Prins et al. 1999 ). This European Seroconverter Study comprised 221 women and 443 men from eight ongoing cohort studies. The CD4 marker path was modelled on the square root scale by regression analysis for repeated measurements. The median age of both men and women was around 25 years. Women seroconverted, developed AIDS and died at higher CD4 cell counts than men. Declines in CD4 count were not affected by gender, except for the trajectory close to AIDS, when the decline became steeper for men than women. These gender differences are important: guidelines for the initiation of HAART are based on studies among men and this study predicts that women reach this threshold later than men, which may unfavourably influence progression rates.
HIV disease progression and pre-aids mortality in injecting drug users
Among drug users the natural history of HIV infection is characterized by the high mortality rate. In the Amsterdam Cohort Study among Drug Users the probability of dying after 7 years of follow-up was about 5% for HIV-negative non-IDU, about 12% for HIV-negative IDU and 44% for HIV-positive IDU ( van Haastrecht et al. 1996 ). For the general population in the same age group in Amsterdam this figure would be around 1%. Among HIV-negatives the most important causes of death are suicide, overdose and pneumonia while among the positives it is also AIDS, liver failure and endocarditis. Among HIV-infected IDU, mortality before being diagnosed with AIDS (pre-AIDS mortality) was found to increase with a decreasing CD4 count.
In the European Seroconverter Study among IDU the pre-AIDS mortality was studied in a competing risk model, right censoring the seroconverters either at the date of pre-AIDS death, AIDS or loss to follow-up, whichever date was first ( Prins et al. 1997 ). The study population consisted of 664 HIV-positive IDU with documented intervals of seroconversion, 107 of whom died, 57 of causes other than AIDS. Pre-AIDS causes of death were overdose/suicide (49%), natural causes such as bacterial infections (40%) and unintentional injuries/unknown (11%). At 7 years from seroconversion 15% of the 664 IDU were estimated to have died without AIDS and 17% to have developed AIDS. Overall pre-AIDS mortality increased with a decreasing CD4 count. Evaluating cause-specific mortality, only pre-AIDS mortality from natural causes (i.e. all deaths excluding deaths from suicide, overdose, unintentional injuries and unknown causes) appeared to be associated with time since seroconversion as well as immunosuppression. So, from these studies it appears that many HIV-positive IDU never ‘reach’ AIDS because they die from HIV-related (and also unrelated) causes before they develop AIDS.
Recently we found – in a collaborative study with the Royal Free and University College Medical School in London – that pre-AIDS deaths also occur in HIV-infected haemophilic men and in homosexual men, although at a much lower rate than among IDU, and are also partially related to HIV disease progression ( Prins et al. 2000 ). By 10 years after seroconversion, 7.3% of the haemophilic men had died without AIDS and 38.2% had developed AIDS. These figures were 20.2% and 30.5% for IDU and 8.0% and 55.0% for homosexual men. Because of the longer life expectancy due to HAART, pre-AIDS deaths among HIV-infected individuals are likely to further rise in the future.
Does hiv-subtype play a role?
Nearly all studies on the natural history of HIV-1 have been conducted in industrialized countries where HIV-1 subtype B predominates. Elsewhere non-B subtypes are circulating and the question is whether the natural history of HIV-1 differs by subtype. This is also of importance for industrialized countries as we and colleagues have shown that the non-B subtypes are regularly imported into industrialized countries mainly among heterosexuals ( Op de Coul et al. 1998 ). Recently, Kanki et al. (1999) presented evidence that progression may differ by HIV-1 subtype: in a prospective study among female sex workers in Senegal, women with a non-A subtype were 8 times more likely to develop AIDS than those infected with subtype A. However, the numbers in this study were small and it is debatable whether analysing the groups by A and non-A is appropriate. Swedish investigators compared the progression rate by HIV-1 subtype and found no difference in the rate of CD4 cell decline between individuals infected with subtypes A, B, C or D over a mean observation period of 44 months ( Alaeus et al. 1999 ). This is confirmed in an Israelian study where the progression rate in Ethiopian Jews infected with HIV-1 subtype C was compared with that of Israelis infected with subtype B ( Galai et al. 1997 ). In a study in Uganda, the progression rate to AIDS was 22% after 5 years which is similar to the one in developed countries ( Morgan et al. 1997 ). As the great majority of HIV-1 infections worldwide are caused by non-B subtypes, this remains an important issue and is currently being investigated by several research groups including ours (in Ethiopia).
The role of genetic factors
Besides CD4, certain chemokine receptors such as CCR5 and CXCR4 are essential coreceptors for HIV infection, and genetic mutations in these chemokine receptors can have a major impact on progression rates ( Stewart 1998). For example, about 15% of whites are heterozygous for a Δ32-bp deletion mutation in the gene encoding for the CCR5 receptor, resulting in a reduced CCR5 expression at the cell surface level. These individuals have a delayed HIV-1 disease progression, which was also found in persons heterozygous for a mutation in the CCR2b receptor gene that occurs in both whites and blacks ( de Roda Husman et al. 1997 ; van Rij et al. 1998 ). Interestingly, we found no effect on disease progression to AIDS, death or a CD4 count < 200/μL in injecting drug users with the above-described mutations in the CCR5 and the CCR2b genes ( Schinkel et al. 1999 ). Immunologic differences between IDU and other risk groups may account for the observed lack of effect and further studies will try to elucidate the responsible mechanism. Finally, several investigators have found that polymorphic products of genes in the HLA region exert substantial influence on the progression rate, which is not surprising as HLA plays an important role in the processing of viral peptides and their presentation to cytotoxic T cells and CD4 cells ( Kaslow et al. 1996 ; Keet et al. 1999 ).