LAUSR

Objective: Debate is ongoing on the relationship between HIV infection and type 2 diabetes mellitus with increased risk and no difference described in European populations. A systematic review and meta-analysis were conducted to 1) determine the pooled prevalence of diabetes and prediabetes among people living with HIV (PLWH) in Africa, and 2) examine the magnitude of diabetes and prediabetes by sociodemographic characteristics, body mass index, definitions used for diabetes and prediabetes diagnoses, and HIV-related characteristics. Design and method: A comprehensive search of major databases was conducted for original research articles published between 2000 and 2021. Relevant studies in English and French were included, irrespective of study design, data collection techniques and diagnostic definitions used. Study-specific estimates were pooled using random effects models to generate the overall prevalence of diabetes and prediabetes for each diagnostic definition and relevant sub-groups. Data analyses were conducted using the R statistical software and the “meta’’ package. Results: Of the 2211 records initially screened, 54 were selected from 346 full-text articles assessed for eligibility. Most included studies were from South Africa (n = 15), Tanzania (n = 7), Ethiopia (n = 6), with four each from Cameroon, Malawi and Uganda. All studies were cross-sectional by design and clinic-based, except for four population-based studies. The proportion of men ranged from 16–52%. Mean/median age ranged from 29–62 years. Diabetes prevalence was 4.8% (95%CI: 4.02–5.59) among 101,704 participants; biochemical testing combined with self-reported diabetes yielded a higher prevalence (oral glucose tolerance test: 6.2%, 95%CI: 2.5–11.3; fasting blood glucose: 6.4%, 95%CI: 4.9–8.1) than self-reported diabetes prevalence alone (3.7%, 95%CI: 2.5–5.2). Prediabetes prevalence was 13.7% (95%CI: 10.21–17.59) among 13,502 participants. There were no statistically significant differences in diabetes or prediabetes prevalence by sub-groups examined, including CD4 count, and antiretroviral treatment status and duration. Conclusions: Diabetes prevalence in PLWH in Africa probably reflects that of general populations with similar influences. The lower prevalence of self-reported diabetes vs. biochemical testing with self-report possibly demonstrates suboptimal diabetes detection in this vulnerable group who are in regular contact with healthcare services. The high prevalence of prediabetes likely foreshadows a substantial increase in future diabetes in African PLWH.