AIDS among older children and adolescents in Southern Africa: projecting the time course and magnitude of the epidemic

Abstract

Objective: An AIDS epidemic among older children and adolescents is clinically apparent in Southern Africa. We estimated the likely scale and time course of the epidemic in older survivors of vertical HIV infection.

Design: We modelled demographic, HIV prevalence, mother-to-child transmission and child survival data to project HIV burden among older children in two Southern African countries at different stages of severe HIV epidemics. Using measured survival data for children, we estimate that 64% of HIV-infected infants are fast progressors with median survival 0.64 years and 36% are slow progressors with median survival 16.0 years. We confirmed model validity by comparing model predictions to available epidemiological data.

Findings: Without treatment, HIV prevalence among 10-year-olds in South Africa is expected to increase from 2.1% in 2008 to 3.3% in 2020, whereas in Zimbabwe, it will decrease from 3.2% in 2008 to 1.6% in 2020. Deaths among untreated slow progressors will increase in South Africa from 7000/year in 2008 to 23 000/year in 2030, and in Zimbabwe from 8000/year in 2008 to peak at 9700/year in 2014. Drugs to prevent mother-to-child transmission could reduce death rate in 2030 to 8700/year in South Africa and to 2800/year in Zimbabwe in 2014.

Conclusions: A substantial epidemic of HIV/AIDS in older survivors of mother-to-child transmission is emerging in Southern Africa. The lack of direct observations of survival in slow progressors has resulted in failure to anticipate the magnitude of the epidemic and to adequately address the clinical needs of HIV-infected older children and adolescents. Better HIV diagnostic and care services for this age group are urgently required.

Figure 1

Proportion of children surviving to different ages. Data from Rakai, red triangle (no error bars); Kampala, blue dots; Abidjan, brown dots; Nairobi, black square; Durban, purple square; and Kigali, orange square [8]; Harare, dark blue diamond [41]; Africa, light green dots [42]; developed countries, black dots [26]. The fitted (red line) line is the sum of an exponential for the fast progressors and a Weibull (pink line) for the slow progressors as described in the text.

Figure 2

The prevalence (red), incidence (green) and mortality (blue) of HIV among adults in South Africa (A) and Zimbabwe (B). Data from ante-natal clinic surveys (red dots) scaled to give the prevalence in all adults.

Figure 3

A: Projected prevalence of HIV among vertically-infected individuals aged 10 years in South Africa (blue) and Zimbabwe (red). Figure 3B and C: Prevalence of vertically-infected HIV-positive children by time and age in South Africa (B) and Zimbabwe (C). Lines at five year intervals: blue 1995; pink 2000; orange 2005; brown 2010; purple 2015; green 2020. In B the age-distribution of children receiving ART from the Catholic Bishops in South Africa is given by the blue dots and line, scaled vertically to match the value at age zero. In C the age-distribution of 16,588 children in care between the ages of 5 and 20 years and infected with HIV is given by the brown dots scaled to match the distribution for 2008. The assumption is that individuals received no PMTCT or ART.

Figure 4

Projected number and age profile of deaths due to HIV infection by time and age in South Africa and Zimbabwe. A and C: South Africa; B and D: Zimbabwe. A and B: deaths in thousands, summed over all ages, on a logarithmic scale among fast-progressors (blue), slow-progressors (red) and adults (green). For fast and slow progressors the upper line assumes no PMTCT, the middle line assumes 80% coverage of single-dose nevirapine for PMTCT by 2010, and the lower line assumes 80% coverage of ART for PMTCT by 2010. C and D: deaths (per year of age and time) in thousands among vertically-infected slow progressors assuming no PMTCT or ART. Lines given at five year intervals: blue 1995; red 2000; green 2005; brown 2010; black 2015; pink 2020.

Figure 5

Observed prevalence (blue dots and 95% confidence limits) of HIV from national prevalence surveys in Botswana 2004 (A), South Africa 2005 (B) and Swaziland 2006 (C) and predicted prevalence (red line) from the model described in the text, assuming a median survival of 16 years. The different shapes of the projected curves reflect differences in the timing and magnitude of the respective adult epidemics in the years preceding the survey data.