Incidence, risk factors and interventions addressing mortality associated with HIV and TB infections among children and adults in Kenya

Tuberculosis (TB) and infection with human immuno-deficiency virus (HIV) are leading infectious causes of morbidity particularly in sub-Saharan Africa, which is disproportionately affected by the two epidemics. Kenya is one of the 30 countries with the highest HIV and TB burden in the world. Kenya’s HIV prevalence was estimated to be 4,900 per 100,000 in 2019 while TB prevalence was 558 per 100,000 population in a 2016 survey. Estimates of mortality from TB and HIV were based on modeling of civil registration and vital statistics (CRVS) which may not be accurate in low- and middle-income countries due to under-reporting, poor documentation of causes of death and lack of verification of causes of death. In this thesis, we provide updated estimates of TB and HIV mortality using a variety of data sources including routine surveillance, post mortem surveillance for HIV infection and minimally invasive tissue sampling (MITS). We also evaluated two mortality preventive interventions.

Our mortality studies documented a higher-than-expected HIV prevalence (28.5%) among 846 decedents received by high-volume mortuaries in western Kenya; 17% of deaths were attributable to HIV-infection. Half of HIV-infected decedents had viral load of ≥1,000 copies/milliliter. Among 456 decedents whose causes of death were determined by an expert panel, HIV/AIDS was the leading underlying cause of death (HIV cause-specific mortality rate: 251/100,000 population); TB was ranked 12th among immediate causes of death. Among 176 decedents aged under-five years, malnutrition, malaria, and HIV were the top three underlying causes of death. HIV prevalence among 176 decedents aged under-five years (14%) twenty-fold higher than the estimated prevalence among children in the population (0.7%); 96% of HIV-infected decedents were virally non-suppressed. While Mycobacterium tuberculosis was not detected among under-five decedents, five decedents (four HIV-infected) had been on TB treatment antemortem. Epidemiological analysis of national TB surveillance data showed that 28% of children aged <14 years and 35% of adults on TB treatment were co-infected with HIV. The case fatality ratio during TB treatment was 4% in children and 6% in adults. The risk of death was five-fold higher among HIV-infected children who were not on ART compared to HIV-uninfected children; 42% of deaths during TB treatment were attributed to HIV-infection.

In a retrospective analysis of data from an integrated TB/HIV active case-finding intervention conducted in Kisumu County, the odds of new HIV diagnoses among TB-symptomatic adults attending private not-for-profit facilities were thrice as high as those attending government facilities. An evaluation of the isoniazid preventive therapy (IPT) cascade among 856 HIV-infected children newly enrolled on HIV care over a four-year period revealed high coverage of screening for IPT eligibility (98%), but suboptimal initiation (68%) and completion (78%).

Our mortality studies demonstrate that TB and HIV remain leading causes of death in Kenya despite the availability of effective interventions. Preventing mortality from these diseases would substantially contribute to a reduction in overall mortality in the population. Existing interventions against the two diseases should be evaluated to identify gaps that continue to contribute to excess mortality from TB and HIV.