Validity of HbA1c in Diagnosing Diabetes Among People with Sickle Cell Trait in Tanzania

BACKGROUND: Hemoglobin A1c (HbA1c) measures the average of the past three month's glucose concentration and is recommended for diagnosing and monitoring diabetes. However in people with shorter red blood cells life spans like those with sickle cell trait (SCT), the test may underestimate the prevalence of diabetes. There are no data on the extent of this problem in East African region where the prevalence of SCT approaches 22%.

OBJECTIVE: To compare diabetes diagnosis outcomes among adults with and without SCT by HbA1c and 2hrs oral glucose tolerance test (OGTT).

METHODOLOGY: This was a cross sectional study conducted among patients who were previously recruited in the Chronic Infection, Cormobidities and Diabetes in Africa (CICADA) study, a cohort study investigating risk factors for diabetes in North western Tanzania during 2016 to 2021. Participants were included in this study if aged ≥18 years and were residents of Mwanza. After identification of eligible participants, stored blood samples were analyzed for SCT and other hemoglobinopathies i.e. thalassemia and G6PD deficiency through gene extraction, PCR and gel electrophoresis. Demographic, and anthropometric data as well as HbA1c and OGTT, hemoglobin, and lipids test results were available as part of CICADA study.

Data were managed and analysed in stata. Student ttest was used for comparison of continuous variables while for categorical variables chi squire test was used. To investigate the validity of HbA1c, regression models were used to evaluate the association between SCT and diabetes diagnosis by HbA1c and OGTT separately. P< 0.05 indicated significant differences.

RESULTS: 480 participants were included. Their mean age was 40.8 (±11.8) years, 292 (60.8%) were females, 128 (26.7%) had SCT and no sickle cell disease(SCD) which is homozygous trait observed. Those with SCT had lower body mass index (BMI) ((21.6(±4.3) vs. 22.5 (±5.0), P= 0.01) and lower HbA1c (5.2% vs. 5.9%, P<0.0001) compared to those without SCT.

In multivariable logistic regression analysis adjusted for sex, age, BMI, another hemoglobinopathy which is thalassemia , G6PD deficiency , hemoglobin and lipids , participants with SCT were 88% less likely to be diagnosed with diabetes by HbA1c compared to those without SCT (OR=0.12, 95% CI (0.1,0.2), P <0.001). In contrast to logistic regression model adjusted for the same variables as above, SCT was not associated with diabetes diagnosis by OGTT (1.44, 95% CI (0.9, 2.3), P= 0.12).

CONCLUSION: When compared with OGTT the findings shows that HbA1c systematically underestimate the prevalence of diabetes among people with SCT. The use of HbA1c for diabetes diagnosis should be done with caution especially in areas with high prevalence of SCT like in North Western Tanzania. Further research is needed to optimize the use of HbA1c in diagnosing and monitoring diabetes among people with SCT.