The impact of acute illness on HbA1c determination of undiagnosed diabetes

Abstract

To improve diabetes screening efforts, the American Diabetes Association now recommends haemoglobin A(1c) (HbA(1c)) as a diagnostic test, increasing access to patients found in acute care environments. However, the influence of acute illness and care on HbA(1c) levels has not been well studied. To address this, we evaluated for intra-patient differences in HbA(1c) assessed in the emergency department (ED) and after recovery from the acute illness. Adult patients with no known history of diabetes were tested for HbA(1c) during an ED and scheduled follow-up visit. HbA(1c) differences between the two visits were compared using limits of agreement with 95% confidence intervals. The frequency of individuals who changed diagnostic categories (using ≥6.5% to classify newly diagnosed diabetes) from ED to follow-up was determined. A total of 589 patients were included with a mean age of 50 years, and 57/589 (9.7%) had an ED HbA(1c) ≥ 6.5% with the average follow-up visit 45 days after the ED visit. The mean ED HbA(1c) was 5.67% (±0.86), and the follow-up HbA(1c) was 5.65% (±0.89), (difference -0.0129%, 95% limits of agreement -0.740, 0.714). The ED and follow-up HbA(1c) was highly correlated (r² = 0.829). Although on follow-up almost all patients were classified in the same diagnostic category as in the ED, 17 patients had an HbA(1c) ≥ 6.5% in the ED and an HbA(1c) < 6.5%. On follow-up most patients (14/17) still fell in an abnormal range (6.0-6.5%). The HbA(1c) value is not substantially affected by acute illness and is feasible as a screening assay for diabetes in the acute care setting such as an ED.