Estimates of global, regional, and national incidence, prevalence, and mortality of HIV, 1980-2015: the Global Burden of Disease Study 2015

Abstract

Background: Timely assessment of the burden of HIV/AIDS is essential for policy setting and programme evaluation. In this report from the Global Burden of Disease Study 2015 (GBD 2015), we provide national estimates of levels and trends of HIV/AIDS incidence, prevalence, coverage of antiretroviral therapy (ART), and mortality for 195 countries and territories from 1980 to 2015.

Methods: For countries without high-quality vital registration data, we estimated prevalence and incidence with data from antenatal care clinics and population-based seroprevalence surveys, and with assumptions by age and sex on initial CD4 distribution at infection, CD4 progression rates (probability of progression from higher to lower CD4 cell-count category), on and off antiretroviral therapy (ART) mortality, and mortality from all other causes. Our estimation strategy links the GBD 2015 assessment of all-cause mortality and estimation of incidence and prevalence so that for each draw from the uncertainty distribution all assumptions used in each step are internally consistent. We estimated incidence, prevalence, and death with GBD versions of the Estimation and Projection Package (EPP) and Spectrum software originally developed by the Joint United Nations Programme on HIV/AIDS (UNAIDS). We used an open-source version of EPP and recoded Spectrum for speed, and used updated assumptions from systematic reviews of the literature and GBD demographic data. For countries with high-quality vital registration data, we developed the cohort incidence bias adjustment model to estimate HIV incidence and prevalence largely from the number of deaths caused by HIV recorded in cause-of-death statistics. We corrected these statistics for garbage coding and HIV misclassification.

Findings: Global HIV incidence reached its peak in 1997, at 3·3 million new infections (95% uncertainty interval [UI] 3·1-3·4 million). Annual incidence has stayed relatively constant at about 2·6 million per year (range 2·5-2·8 million) since 2005, after a period of fast decline between 1997 and 2005. The number of people living with HIV/AIDS has been steadily increasing and reached 38·8 million (95% UI 37·6-40·4 million) in 2015. At the same time, HIV/AIDS mortality has been declining at a steady pace, from a peak of 1·8 million deaths (95% UI 1·7-1·9 million) in 2005, to 1·2 million deaths (1·1-1·3 million) in 2015. We recorded substantial heterogeneity in the levels and trends of HIV/AIDS across countries. Although many countries have experienced decreases in HIV/AIDS mortality and in annual new infections, other countries have had slowdowns or increases in rates of change in annual new infections.

Interpretation: Scale-up of ART and prevention of mother-to-child transmission has been one of the great successes of global health in the past two decades. However, in the past decade, progress in reducing new infections has been slow, development assistance for health devoted to HIV has stagnated, and resources for health in low-income countries have grown slowly. Achievement of the new ambitious goals for HIV enshrined in Sustainable Development Goal 3 and the 90-90-90 UNAIDS targets will be challenging, and will need continued efforts from governments and international agencies in the next 15 years to end AIDS by 2030.

Funding: Bill & Melinda Gates Foundation, and National Institute of Mental Health and National Institute on Aging, National Institutes of Health.

Figure 1

Evolution of the HIV epidemic from 1980 to 2015 Global estimates of new HIV infections (A), people living with HIV/AIDS (B), HIV/AIDS deaths (C), and proportion of people living with HIV receiving ART (D). Shaded areas show 95% uncertainty intervals. ART=antiretroviral therapy.

Figure 2

Incidence of new HIV infections from 1980 to 2015, and HIV incidence in 2015 Global number of new HIV infections by region (A). Bars show the mean number of estimated new infections within a given year. Error bars represent 95% uncertainty intervals. Each Global Burden of Disease region is represented by a separate colour. HIV incidence by country (B). We calculated incidence as cumulative new cases of HIV throughout the year divided by the total population at the mid-year. Rates are per 100 000 people. Colour bins correspond to the 0–50th, 50–70th, 70–80th, 80–90th, 90th–92nd, 92nd–94th, 96–98th, 98–99th, and 99–100th percentiles to highlight variation within sub-Saharan Africa. ATG=Antigua and Barbuda. VCT=Saint Vincent and the Grenadines. LCA=Saint Lucia. TTO=Trinidad and Tobago. TLS=Timor-Leste. FSM=Federated States of Micronesia.

Figure 3

Number of people living with HIV receiving ART from 1995 to 2015, and the proportion living with HIV receiving ART in 2015 Number of people living with HIV receiving ART by region (A). Bars represent the mean number of people living with HIV who received ART within a given year. Error bars represent 95% uncertainty intervals. Each Global Burden of Disease (GBD) region is represented by a separate colour. Proportion of people living with HIV receiving ART by country (B). The number of people living with HIV receiving ART and the total number of people living with HIV are year-end point prevalences. ART=antiretroviral therapy. ATG=Antigua and Barbuda. VCT=Saint Vincent and the Grenadines. LCA=Saint Lucia. TTO=Trinidad and Tobago. TLS=Timor-Leste. FSM=Federated States of Micronesia.

Figure 4

Global HIV/AIDS deaths, 2005–15 Global deaths caused by HIV/AIDS resulting in either mycobacterial infection (tuberculosis) or other diseases, by age and sex in 2015 (A); dark shading indicates deaths caused by tuberculosis associated with HIV; light shading indicates deaths caused by other diseases resulting from HIV; error bars show 95% uncertainty intervals. Mean estimates of global and super-regional HIV/AIDS deaths per prevalent case fom 2005 to 2015 (B).

Figure 5

Comparison of GBD 2015 and UNAIDS 2014 estimates Adult HIV prevalence rate (A) and estimates of death caused by HIV/AIDS (B). UNAIDS' published prevalence values are limited to three decimal places. The x and y values of each point are the log transformation of the mean estimates from UNAIDS and GBD, respectively, enabling variation to be seen despite disparate values. Tick-mark labels on the x and y axes are the value of the mean estimate before log transformation (ie, the real value and not the log-transformed value is shown). Locations mentioned in the manuscript are highlighted by plotting the ISO 3 code of the location. Each location is plotted with a different colour by super-region. GBD=Global Burden of Disease. UNAIDS=the Joint United Nations Programme on HIV and AIDS. ZAF=South Africa. KEN=Kenya. NGA=Nigeria. COG=Democratic Republic of the Congo. SLE=Sierra Leone. BDI=Burkina Faso. COD=Congo. SEN=Senegal. SWZ=Swaziland.