• diabetes;
  • Africa;
  • glucose;
  • glycosylated haemoglobin


Aims and methods To examine the reliability of random venous or capillary blood glucose testing, random urine glucose testing, and a current symptom history in predicting a high HbA1c in Type 2 diabetic patients taking oral hypoglycaemic agents in a poorly controlled rural African population.

Results For a cut-off point for HbA1c of ≥ 8%, for random venous plasma glucose of ≥ 14 mmol/L (present in 47.2% of subjects), specificity was 97.1% (95% CI 85.1–99.9), sensitivity 56.8% (48.8–64.5) and positive predictive value (PPV) 98.9% (94.2–99.9). HbA1c≥ 8% is predicted by a random capillary blood glucose of 17 mmol/L (present in 28.4% of subjects) with specificity 100% (90.0–100.0), PPV 100% (93.7–100.0) and sensitivity of 34.3% (27.2–42.1). HbA1c≥ 8% is predicted by the presence of heavy glycosuria (≥ 55 mmol/L) (present in 35.6%) with specificity 94.1% (80.3–99.3), sensitivity of 41.9% (34.1–49.9) and PPV 97.1% (89.9–99.6). Polyuria/nocturia (present in 31.3%) was the only symptom found to be associated with poor control, with a specificity for predicting HbA1c of ≥ 8% of 81.5% (61.9–93.7), PPV 89.1% (76.4–96.4) and sensitivity 30.6% (22.9–39.1).

Conclusions Where resources are short, random glucose testing can be used to detect a significant proportion of those with the worst control with a high degree of specificity enabling primary care staff to modify treatment safely. Where facilities are limited capillary blood or urine testing with reagent strips, may be substituted for venous plasma testing in the laboratory. A symptom history was insufficient to replace biochemical testing, but where this is unavailable, urinary symptoms may be helpful.

Diabet. Med. 19, 195–200 (2002)