The natural history of HIV-1 infection and preparations for HIV vaccine trials in Tanzania

Abstract

In order to study the natural history of HIV infection in Tanzania, a total of 1887 hotel workers were recruited between 1989 and 1998 in Dar es Salaam. From 682 subjects selected for lymphocyte subset determinations, the median yearly decline of the CD4+ T-lymphocyte counts and percentages among HIV-1 seroprevalents was found to be -21.5 cells/μ1 and -1.3% and among the HIV-1 seroincidents was -22.0 cells/μ1 and -1.5% respectively. The mean duration to a CD4' T lymphocyte level corresponding to definition of AIDS among the HIV-1 seroincidents was 13.3 years or 11.8 years for CD4' counts or percentage respectively.

Compared to HIV sero-negatives, HIV-1 sero-positive hotel workers had higher rates of clinical events that occurred at CD4 count levels similar to those reported in developed countries before the introduction of ART. Of the 85 total deaths, 65 (76.5%) were HIV-1 infected. Thus infection with HIV in Tanzania results into similar immune deterioration and consequent clinical manifestations at CD4 count levels as earlier reported from the developed countries, and is associated with significant mortality. Similar management recommendations can therefore be used.

Fifty HIV-1 infected workers had >3 CD4 cell determinations and were followed for >5 years. Their selection was not biased for the healthiest individuals. With a criterion of maintaining a CD4 cell count ≥500 cells/μl, 8 (16,0%) of them were long- term non-progressors (LTNP). With a definition of maintaining a CD4 cell slope ≤-10 (loss of not more than 10 cells/μl/year), 13 (26.0%) were long-term slow progressors, (LTSP). Their mean (median) baseline CD4 Cell Count of 547.6 (530) cells/μ1 was higher than that of 403.4 (400) cells/μ1 among the progressors, p=0.009. Eleven of the 13 LTSP (84.6%) had a baseline CD4 cell count <500 cells/μ1. Thus slow immune progression is not uncommon in Tanzania, and repeated CD4 counts are required to evaluate patients for accurate prognosis. Furthermore studies on such individuals may yield useful information for HIV intervention strategies, including vaccine development.

Prevalence of sexually transmitted infections (STI's) at baseline among the workers was 16.6% (236/1422) and 15.5% (72/465) in males and females respectively, p=0.573. Serological evidence of active syphilis was present in 12.8% of males and 6.7% of females, p<0.001 HIV-1 prevalence for the whole study population at the end of 1998 was 16.0% for men and 21.5% for women. p=0.009. HIV-1 incidence was 1.9/100 person years at risk (PYAR), and by gender was 1.8 and 2.57100 PYAR for males and females respectively, p=0.011. HIV-1 incidence over the first 4 years was 2.22/100 PYAR compared to 1.50/100 PYAR during the last 5 years, p=0.038. A nonsignificant trend of a declining prevalence of subjects with STI's was observed. In multivariate analysis, the significant factor associated with acquiring HIV-1 infection was the presence of STI's, with a RH of 1.75 (95% CI 1.21-2.54). Hence there was a close association between incidence of STI's and HIV-1, and it was possible to reduce the incidence of both STI's through the syndromic management of STI's and VCT for HIV-1 respectively.

In preparation for future HIV vaccine trials, an open cohort of 2850 Police officers was recruited beginning 1994 of whom 2733 (96%) consented to be tested for HIV. The overall HIV-1 seroprevalence at recruitment was 13.8% (378/2733). Females had a significantly higher HIV-1 seroprevalence, 18.0% (551306), as compared to males, 13.3% (323/2427), p<0.05. Extramarital sex was common and this was coupled with low condom use. The overall crude HIV-1 incidence was 1.99/100 PYAR; 1.96 and 2.24/100 PYAR for males and females respectively. The overall prevalence and incidence of active syphilis were 3.1% (88/2850) and 0.86/100 PYAR (26 of 3149), respectively. Males had a higher prevalence of active syphilis, 84 of 2525 (3.3%) than females, five of 325 (1.5%), p=0.09. It was concluded that there was high-risk sexual practice including low condom use in this cohort of police officers. The incidence and prevalence of HIV infection were high, and thus this was a potential cohort for HIV vaccine related studies.