Abstract
Aims HbA(1c) is reported to underestimate glycaemia in people living with HIV (PLHIV). There is not an internationally agreed screening method for diabetes. The primary aim was to identify which tests are performed to diagnose and monitor diabetes in PLHIV. Secondary aims were to identify whether prevalence or incidence of diabetes differs according to marker of glycaemia and how figures compare in PLHIV compared to people without.
Methods Electronic databases were searched for studies investigating diabetes in PLHIV, not pregnant, aged >= 18 years. Narrative analysis and descriptive statistics were used to describe which markers of glycaemia, and their frequency, were employed in the diagnosis and monitoring of diabetes in PLHIV. Diagnostic studies provided prevalence or incidence of diabetes.
Results In all, 45 of 1028 studies were included. Oral glucose tolerance test (OGTT), fasting glucose (FG), HbA(1c) and Fructosamine were used to investigate diabetes. In total, 27 studies described diagnosing diabetes, 14 using OGTT, 12 FG and 7 HbA1c. All 18 studies monitoring diabetes used HbA1c. Prevalence ranged from 1.3% to 26% and incidence 2.9% to 12.8%. Studies using glucose and HbA(1c) reported HbA(1c) to diagnose fewer people with diabetes, monitoring studies found HbA(1c) to underestimate glycaemia levels. Controlled studies demonstrate diabetes was more common in PLHIV.
Conclusion OGTT was used most frequently to diagnose diabetes, and HbA(1c) to monitor known diabetes. Prevalence and incidence varied depending on marker of glycaemia used. Studies reported a discrepancy in accuracy of HbA(1c) in PLHIV, to address this, well-designed, prospective studies, providing individual-level data on HbA(1c) levels and an additional marker of glycaemia in PLHIV are needed.