Diagnostic accuracy of Hba1c Test in diagnosis of Diabetes and Prediabetes in a Multiethnic Population at Aga Khan University Hospital, Nairobi.

Introduction: Diabetes mellitus is a metabolic disease characterized by hyperglycaemia secondary to impaired insulin secretion, impaired insulin action or both. In Kenya, an estimated 750,000 people are diabetic with more than 50% estimated to be unaware of their status. Conventionally, diagnosis has been based on plasma glucose levels. In 2010, Haemoglobin A1c (HbA1c) was approved by the American Diabetes Association (ADA) and World Health Organisation (WHO) for diagnosis of diabetes. However, different studies have shown variations in diagnostic performance of HbA1c. Research question: What is the diagnostic sensitivity and specificity of HbA1c in diagnosis of diabetes and prediabetes at Aga Khan University Hospital, Nairobi? Primary objective: To determine the diagnostic sensitivity and specificity of HbA1c based on the ADA cut-off values, in diagnosis of diabetes and prediabetes at Aga Khan University Hospital, Nairobi Secondary objective: To determine the correlation of HbA1c and fasting plasma glucose levels in diagnosis of diabetes and prediabetes. To determine the HbA1c cut-off level optimal for diagnosis of diabetes and prediabetes in the study population. To determine the association between anthropometric characteristics, and HbA1c Methodology: Cross-sectional study that recruited participants from those directed to undertake fasting plasma glucose at Aga Khan University Hospital, Nairobi in 2016-2017. Sample size: One hundred and fifty seven non replicate samples were evaluated. Data collection and statistical analysis: HbA1c and fasting plasma glucose levels were determined for each participant. Receiver operating characteristic curves were constructed and the area under curve calculated for HbA1c to determine its sensitivity and specificity. Results: Fifteen study participants (9.6%) were diagnosed with diabetes based on criteria combining fasting plasma glucose and HbA1c while fifty five (35.5%) had prediabetes. ROC curves found the HbA1c cut-off of ≥6.5% to have high sensitivity (93.3%) and specificity (98.6%) in diagnosis of diabetes. HbA1c was also shown to have high sensitivity and specificity (87.3% and 86.9% respectively) in diagnosis of prediabetes using the range of 5.7-6.4% as recommended by ADA Conclusion: HbA1c showed high diagnostic accuracy in diabetes and prediabetes using the current cut-offs recommended by ADA. Based on these findings, this study recommends adoption of HbA1c in the national guidelines for diagnosis of prediabetes and diabetes