Mortality rates amongst adult African HIV-infected patients on antiretroviral therapy (ART) are as high as 26% in the first 12 months of treatment. Although some risk factors for death in these patients have been elucidated, the specific causes of death have not. Therefore, there is an urgent need for evidence-based knowledge of underlying causes of early mortality in HIV-infected patients on ART in sub-Saharan Africa. By informing decision-making processes on resource allocation, treatment and care emphasis-areas and planned strategies, such knowledge would provide treatment programs a great opportunity to lower HIV-related mortality while increasing the benefits of ART. The USAID-AMPATH Partnership (AMPATH) has enrolled >80,000 HIV-infected patients across 18 urban and rural care sites in western Kenya, >40,000 of whom have initiated ART. To date, 5,500 patients have been confirmed dead, half of whom were on ART. AMPATH is uniquely placed to perform cause-specific mortality studies because sufficient study subjects are available and AMPATH has the resources necessary for such studies including: (1) the Moi Teaching Referral Hospital (MTRH) medical wards where very sick AMPATH- enrolled patients are admitted;(2) the MTRH Morgue which has cold storage facilities, an autopsy laboratory and 4 full-time pathologists;(3) the AMPATH Medical Record System (AMRS), a fully computerized system containing >34 million observations collected prospectively for >1.2 million visits;(4) high-quality reference laboratories;(5) experienced investigators from the Moi and Indiana University Schools of Medicine;and (6) effective community outreach programs. Finally, autopsies are well accepted in the communities served by AMPATH and are commonly performed on HIV+ patients by Moi University and MTRH pathologists. We propose to perform 400 detailed pathologic autopsies on patients on ART who die within 12 months of initiating treatment (160 in hospital deaths and 240 in home deaths). Biological specimens from subjects will undergo various cultures and laboratory tests. Autopsy findings will be used to establish prevalent causes of death and estimate differences in cause-specific mortality between patients who die in hospital and those who die at home. Because full autopsies are rarely done on distant urban and rural patients, nor always affordable by majority of Kenyans, we plan to develop a HIV-specific verbal autopsy (VA) module, incorporate it into the generic WHO standard VA for adults, and determine its utility by comparing it to pathologic autopsies ("gold standard") in determining causes of death in HIV+ patients on ART. This validated VA will be used long-term to determine causes of death. Additionally, we will set up a Central Review Board (CRB) for the ascertainment of cause-specific mortality. We will provide this CRB with de-identified VA and supplemental patient ante-mortem data from the MTRH wards and AMRS for use in determining the cause of death. We will then evaluate the utility of these adjunct data in increasing the accuracy of the VA as compared to pathologic autopsy.

Public Health Relevance

As availability and access to antiretroviral medicines continues to increase in sub-Saharan Africa, the onus is on HIV treatment and care programs to prevent excess mortality in HIV-infected patients, especially so in those able to access care. Evidence-based knowledge of the types and frequency of specific diseases causing or contributing to death (using pathologic autopsies) in HIV-infected patients already on ART will be important to these programs as they plan and implement strategies to prevent mortality. Because trends in pattern and distribution of diseases change over time, long-term success of such mortality prevention strategies will depend on continuous collection of deaths data through appropriate and affordable means (such as verbal autopsy).

National Institute of Health (NIH)
National Institute of Allergy and Infectious Diseases (NIAID)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1-IDM-P (50))
Program Officer
Livnat, Daniella
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Moi University
Zip Code