Fig 3.

A: Boxplots of OGTT comparing sickle cell trait status. By OGTT, PD/DM was prevalent in 32/52 (61.5%) of participants with homozygous α⁺ AT, 90/186 (48.4%) of those with heterozygous α⁺ AT, and 89/193 (46.1%) of those without α-thalassemia (Table 3). Compared to those without α-thalassemia, participants with homozygous α⁺ AT had increased odds of being diagnosed with PD/DM by OGTT (adjusted OR 2.00, 95% CI: 1.04, 3.84) (Table 3). However, there was overall no evidence that α-thalassemia was associated with diabetes diagnosis by OGTT (overall p = 0.11 in adjusted analysis). Compared to those without α-thalassemia, the median OGTT glucose was highest among homozygotes (S2 Table) and the distributions of OGTT results were shifted upward among those with homozygous α+ AT (Fig 3B). B: Boxplots of OGTT comparing α-thalassemia status. Those with hemizygous, homozygous or heterozygous G6PD deficiency had PD/DM prevalence of 16/24 (66.7%), 2/4 (50.0%), and 12/40 (30.0%) respectively, while in those with normal G6PD activity, PD/DM prevalence was 181/363 (49.9%) (Table 3). Compared to those without G6PD deficiency, heterozygous females had decreased odds of PD/DM diagnosis (adjusted OR 0.47, 95% CI: 0.22, 0.99, Table 3). Median OGTT glucose was highest among hemizygotes (8.6 (7.4:9.6) mmol/L) and lowest among heterozygotes (7.5 (6.8:8.0) mmol/L) when compared to those without G6PD deficiency (7.7 (7.0:8.7) mmol/L) (S2 Table). The distribution was shifted upward among hemizygotes while in heterozygotes the distribution was shifted downward compared to those not deficient (Fig 3C). C: Boxplots of OGTT comparing G6PD deficiency status. No interaction was observed between HIV status and SCT (p = 0.61), α-thalassemia (p = 0.80), or G6PD deficiency (p = 0.35) with regard to PD/DM diagnosis by OGTT. Sensitivity analyses were consistent with the presented results.