HIV infection and anaemia do not affect HbA1c for the detection of diabetes in black South Africans - evidence from the Durban Diabetes Study

Thomas R Hird, Uttara Partap, Pravi Moodley, Fraser J Pirie, Tonya M Esterhuizen, Brian O'Leary, Mark I McCarthy, Elizabeth H Young, Manjinder S Sandhu, Ayesha A Motala

Research output: Contribution to journalArticlepeer-review


OBJECTIVE: South Africa has a high burden of HIV infection and anaemia. These conditions may cause glycated haemoglobin (HbA1c ) to over- or underestimate glycemia; however, this has not been comprehensively investigated in African populations. We assessed the association of anaemia, HIV infection and antiretroviral therapy (ART) with HbA1c , and implications for the detection and diagnosis of diabetes, in a black South African population.

RESEARCH DESIGN AND METHODS: In this population-based cross-sectional study in eThekwini municipality (Durban), South Africa, we assessed HbA1c and conducted oral glucose tolerance tests (OGTTs), HIV diagnostic tests and full blood count measurements among 1067 participants without a history of diabetes diagnosis. Linear regression was used to examine differences in HbA1c by anaemia (comparator: no anaemia), or HIV and ART (comparator: no HIV) status. HbA1c-based diabetes prevalence was compared with OGTT-based prevalence among individuals with anaemia and with untreated and ART-treated HIV.

RESULTS: In adjusted analyses, normocytic and microcytic anaemia were associated with higher HbA1c compared with no anaemia, whilst macrocytic anaemia and ART-treated HIV were associated with lower HbA1c compared with no anaemia and no HIV respectively. However, magnitudes of association were small (range: β=-3.4 mmol/mol or -0.31%, P<0.001 [macrocytic anaemia] to β=2.1 mmol/mol or 0.19%, P<0.001 [microcytic anaemia]). There was no significant difference in diabetes prevalence based on HbA1c or OGTT among individuals with anaemia (2.9% versus 3.3%, P=0.69), untreated HIV (1.6% versus 1.6% P=1.00) or ART-treated HIV (2.9% versus 1.2%, P=0.08).

CONCLUSIONS: Our results suggest that anaemia and HIV status appear unlikely to materially affect the utility of HbA1c for diabetes detection and diagnosis in this population. Further studies are needed to examine these associations in sub-Saharan African populations.

Original languageEnglish
Pages (from-to)e14605
JournalDiabetic Medicine
Publication statusE-pub ahead of print - 24 May 2021

Cite this