University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Short-Term Risk of AIDS or Death in People Infected with HIV-1 Before Antiretroviral Therapy in South Africa: A Longitudinal Study
Global Health Sciences Literature Digest
Published November 3, 2006
Journal Article

Badri M, Lawn SD, Wood R. Short-Term Risk of AIDS or Death in People Infected with HIV-1 Before Antiretroviral Therapy in South Africa: A Longitudinal Study. Lancet. 2006 Oct 7;368(9543):1254-9.


To investigate the short-term risk of AIDS or death in patients who had no access to ART or who were given zidovudine alone.

Study Design

This is a longitudinal cohort study based on data from the Cape Town AIDS Cohort (CTAC) that examined risk of disease progression and death in nine groups defined by WHO stage of disease (1 or 2 combined, 3 and 4) and three baseline CD4 groups (<200 cells/mm3, 200-350, >350). Data were collected from 1992 through 2005.


The CTAC study took place at tertiary public-sector HIV clinics in Cape Town, South Africa.


Patients with HIV and living in urban areas around Cape Town were referred to tertiary HIV clinics from primary healthcare facilities. Participants were eligible for inclusion in the study if they were not currently taking ART (except for zidovudine monotherapy), and if a current CD4-cell count was available at the start of any six-month period of assessment.


There was no intervention in this study.

Primary Outcome

Three primary outcomes were assessed: Six-month risk of death, progression to AIDS, and progression to a combined endpoint of a new AIDS-defining illness or death (AIDS/death).


Of all patients recruited into CTAC, 1399 were eligible for this study. The population was 61% male with a median CD4 cell count of 216. The majority (54%) were in WHO clinical stage 1 or 2, and 4% were on zidovudine monotherapy at entry. The median follow-up time was 17 months. Overall incidence of death was 0.1 per patient per year; 14% of deaths occurred in patients in WHO stage 1 or 2, 38% in those in WHO stage 3, and 48% in those in stage 4. Six-month risk of death: The six-month risk of death ranged from 1.2% (95% CI 0.5%-2.3%) in those with the CD4 counts >350 cells/mm3 and in WHO stage 1 or 2 to 22.2% (95% CI 17.9%-27.1%) among patients with less than 200 cells/mm3 and in WHO stage 4. Progression to AIDS: In 1009 patient-years of follow-up (involving only participants who presented without AIDS), 111 new AIDS cases were diagnosed, and six-month risk of progression to AIDS was lowest among those with >350 cells/mm3 and in WHO stage 1 or 2 (0% risk) and highest among those with <200 cells/mm3 and in WHO stage 3 (20%). Progression to AIDS/death: The risk of AIDS/death in 964 patient-years of follow-up ranged from 1.2% for patients with highest CD4 cell count and lowest stage to 23.6% for those with lowest CD4 count and highest stage. Risk of all three endpoints increased as CD4 counts fell from 200-350 and from 200 cells per microliter. The risk of death in people with stage 3 disease was about double that of those with stage 1 or 2 disease, as was the risk in those with stage 4 disease compared to those with stage 3 disease for any CD4 count level.


The authors conclude that the combined use of WHO clinical stage and CD4 cell count can be used to accurately assess short-term risk of AIDS and death and that use of CD4 counts can help refine risk estimates in the various WHO stages of disease. This will lead to more accurate timing of antiretroviral initiation. Overall incidence and six-month risk of AIDS was similar to a combined European and Australian cohort (CASCADE). However, the risk of AIDS was 1.9 times higher than in the CASCADE study in those in the 200-350 cell CD4 stratum, which suggests higher rates of progression among African than European and Australian patients in this CD4 stratum.

Quality Rating

Using the Newcastle-Ottawa (NOS) criteria for evaluating cohort studies, based on selection, comparability, and outcomes, this study receives an adequate quality rating. The comparison cohorts (strata of CD4 count and WHO clinical stage) were drawn from the same population; however, the authors did not control for additional potential confounding factors. The ascertainment of stage and CD4 count through testing at the tertiary clinic was adequate. The primary outcomes were assessed independently, through clinic records. The authors did not report on the baseline clinical or demographic characteristics of those lost to follow up. In addition to the NOS criteria, the authors mention the following limitations:

  • they did not have access to cause-specific mortality and
  • their results may be partly attributable to informative censoring, that is, differential censoring in various strata. However, in additional analyses they found no evidence of informative censoring.

In Context

While other studies have shown that WHO clinical stage can be used as effective criteria for initiation of ART, and can also provide such information independent of CD4 cell count,(1,2) this study shows that combined use of WHO stage and CD4 count can be used at the population level to accurately assess short-term risk of death and AIDS.

Programmatic Implications

Because WHO clinical stage is the criterion most often used in low-resource settings in deciding on initiation of ART, it is important to refine estimates of the short-term risk of death. The results of this study suggest that the addition of CD4 cell counts to WHO clinical stage can result in better estimates of short-term risk of AIDS. For instance, six-month risk of AIDS or death in participants in WHO state 1 or 2 was almost seven times higher (7.0% v. 0.5%) in those with <200 CD4 cells/mm3 compared to those with >350 cells. Furthermore, the higher risk found in patients with 200-350 cells/mm3 compared to the CASCADE cohort suggests that earlier treatment may be warranted. The authors note that these results are likely to be generalizable to other sub-Saharan countries due to the similarity of the spectrum of AIDS-defining illnesses; however, caution must be used as countries may vary with respect to other diseases and standards of healthcare.


  1. Badri M, Wilson D, Wood R. Effect of highly active antiretroviral therapy on incidence of tuberculosis in South Africa: a cohort study. Lancet. 2002 Jun 15;359(9323):2059-64.
  2. Badri M, Bekker LG, Orrell C, Pitt J, Cilliers F, Wood R. Initiating highly active antiretroviral therapy in sub-Saharan Africa: an assessment of the revised World Health Organization scaling-up guidelines. AIDS. 2004 May 21;18(8):1159-68.