Assessing impaired hypoglycemia awareness in type 1 diabetes: Agreement of self-report but not of field study data with the autonomic symptom threshold during experimental hypoglycemia

OBJECTIVE - The aim of our study was to determine the agreement of two noninvasive methods, a self-report and a field study method, for the assessment of impaired hypoglycemia awareness with a gold standard criterion of hypoglycemia awareness, the autonomic symptom threshold during experimental hypoglycemia. RESEARCH DESIGN AND METHODS - A total of 19 type 1 diabetic patients completed a standardized questionnaire to assess impaired hypoglycemia awareness and performed a hand-held computer (HHC) study to assess their recognition of hypoglycemic episodes occurring during 2-4 weeks. Patients subsequently underwent a stepped hypoglycemic clamp to study responses to standardized hypoglycemia. Diagnoses of impaired hypoglycemia awareness were based on the separate self-report questions, a composite self- report score, and three different cutoff levels for the percentage of accurately recognized hypoglycemic episodes during the field study agreement of these noninvasive measures with the hypoglycemic clamp measure were tested by calculating kappa values, sensitivity, and specificity. RESULTS - The composite self-report score agreed reasonably well with the hypoglycemic clamp measure (kappa 0.49, sensitivity 66.7%, and specificity 85.7%) and showed a better agreement than the separate self-report questions. The HHC criterion of impaired hypoglycemia awareness did not agree with the hypoglycemic clamp criterion at any of the cutoff levels tested. CONCLUSIONS - The composite self-report tested in this study is a reasonably reliable assessment method for the diagnosis of impaired hypoglycemia awareness, using the physiological definition of an absence of autonomic symptoms at a blood glucose level of 3 mmol/l. In contrast, the recognition of hypoglycemic events in everyday life as measured using the HHC method is not related to the hypoglycemic clamp criterion.