Comparison of different cardiovascular risk tools used in HIV patient cohorts in sub-Saharan Africa; do we need to include laboratory tests?

IntroductionCardiovascular disease (CVD) is the leading cause of death globally, representing 31% of all global deaths. HIV and long term anti-retroviral therapy (ART) are risk factors for development of CVD in populations of people living with HIV (PLHIV). CVD risk assessment tools are currently being applied to SSA populations, but there are questions about accuracy as well as implementation challenges of these tools in lower resource setting populations. We aimed to assess the level of agreement between the various cardiovascular screening tools (Data collection on Adverse effects of anti-HIV Drugs (D:A:D), Framingham risk score, WHO risk score and The Atherosclerotic Cardiovascular Disease Score) when applied to an HIV ART experienced population in Sub-Saharan Africa.MethodsThis study was undertaken in an Anti-Retroviral Long Term (ALT) Cohort of 1000 PLHIV in care who have been on ART for at least 10 years in urban Uganda. A systematic review was undertaken to find the most frequently used screening tools from SSA PLHIV populations; these were applied to the ALT cohort. Levels of agreement between the resulting scores (those including lipids and non-lipids based, as well as HIV-specific and non-HIV specific) as applied to our cohort were compared. Prevalence Bias Adjusted Kappa was used to evaluate agreement between tools.ResultsOverall, PLHIV in ALT cohort had a median score of 1.1-1.4% risk of a CVD event over 5 years and 1.7-2.5% risk of a CVD event over 10 years. There was no statistical difference in the risk scores obtained for this population when comparing the different tools, including comparisons of those with lipids and non-lipids, and HIV specific vs non-HIV specific.ConclusionThe various tools yielded similar results, but those not including lipids are more feasible to apply in our setting. Long-term cohorts of PLHIV in SSA should in future provide longitudinal data to evaluate existing CVD risk prediction tools for these populations. Inclusion of HIV and ART history factors to existing scoring systems may improve accuracy without adding the expense and technical difficulty of lipid testing