Temporal evolution of HIV sero-discordancy patterns among stable couples in sub-Saharan Africa

Abstract

Introduction: Objective was to examine the temporal variation of HIV sero-discordancy in select representative countries (Kenya, Lesotho, Mali, Niger, Tanzania, and Zimbabwe) in sub-Saharan Africa at different HIV epidemic scales. A sero-discordant couple is defined as a stable couple (SC) in which one partner is HIV-positive while the other is HIV-negative.

Methods: A deterministic compartmental mathematical model was constructed to describe HIV transmission dynamics. The model was pair-based, that is explicitly modeling formation of SCs and infection dynamics in both SCs and in single individuals. The model accommodated for different forms of infection statuses in SCs. Using population-based nationally-representative epidemiologic and demographic input data, historical (1980-2014) and future (2015-2030) trends of sero-discordancy and other demographic and epidemiologic indicators were projected throughout HIV epidemic phases.

Results: As the epidemics emerged, about 90% of SCs affected by HIV were sero-discordant. This proportion declined to 45%-88% at epidemic peak and stabilized as the epidemics started their natural decline. The largest reductions in sero-discordancy were in high HIV-prevalence countries. As the epidemics further declined with antiretroviral therapy (ART) scale-up, the proportion of sero-discordant couples among HIV-affected couples was projected to increase to 70%-92% by 2030. The proportion of sero-discordant couples among all SCs increased as the epidemics emerged and evolved, then peaked at 2%-20% as the epidemics peaked, and then declined as the epidemics declined to reach 0.3%-16% by 2030.

Conclusions: Sero-discordancy patterns varied with the evolution of the epidemics, and were affected by both epidemic phase and scale. The largest variations were found in high HIV-prevalence countries. The fraction of stable couples that are sero-discordant, as opposed to being sero-concordant positive, was projected to increase with ART scale-up and further HIV incidence decline over the coming two decades. These findings inform strategic planning and resource allocation for interventions among sero-discordant couples.

Fig 1. Flow-chart depicting the basic structure of the used pair-based HIV model.

Fig 2

(A) Conceptual diagram illutrating the population in reproductive age by partnership status: not engaged in a stable couple (singles), sero-concordant HIV-negative stable couple, HIV sero-discordant stable couple, and sero-concordant HIV-positive stable couple. (B) Characterization of the four generic phases of the HIV epidemics in sub-Saharan Africa. The sero-discordancy patterns in each country were described and analyzed within the context of these phases of each country’s epidemic.

Fig 3. Model predicted proportion of HIV-infected individuals by partnership status: Not engaged in stable couple, sero-discordant stable couple, and sero-concordant positive stable couple.

Countries are shown in order of increasing HIV prevalence. (A) Niger, (B) Mali, (C) Tanzania, (D) Kenya, (E) Zimbabwe, and (F) Lesotho.

Fig 4. Model predicted proportion of stable HIV sero-discordant couples among all stable couples with at least one HIV-infected individual in the couple (P_discord) for six representative countries in sub-Saharan Africa.

Countries are shown in order of increasing HIV prevalence: (A) Niger, (B) Mali, (C) Tanzania, (D) Kenya, (E) Zimbabwe, and (F) Lesotho. The black lines show model predictions while the red asterisks show Demographic and Health Surveys data points and their 95% confidence interval [17].