Return to 2001 Abstacts List

129. Selecting a Parsimonious Blood Glucose Monitoring Strategy for Insulin-treated Veterans with Type II Diabetes Mellitus

RM Hoffman, Albuquerque VAMC; CS Wendel, Tucson VAMC; C Dalton, Tucson VAMC; KD Adam, Albuquerque VAMC; WC Duckworth, Phoenix VAMC; JH Shah, Tucson VAMC; SU Bokhari, Phoenix VAMC; MT Montagnini, Phoenix VAMC; GH Murata, Albuquerque VAMC

Objectives: Many insulin-treated veterans with type II diabetes test their blood glucose (BG) only once or twice daily. The purpose of this study was to determine the optimal testing times for patients who cannot comply with recommendations for more intensive monitoring.

Methods: Subjects were randomly selected using pharmacy records at the Albuquerque, Tucson and Phoenix VAMCs. Subjects were eligible if, during the preceding two months, they received no new prescriptions for oral hypoglycemic agents and insulin doses were not increased by more than 10 units or 15%. Subjects were instructed to monitor blood glucose 4 times daily for 8 consecutive weeks. Electronically-stored glucometer readings were downloaded at 4 and 8 weeks. Hemoglobin A1c (A1c) was measured after 8 weeks. We evaluated testing strategies by: a) the correlation between the 8-week mean BG and A1c; b) the proportion of hypoglycemic events (BG<=60 mg/dL) that were detected; c) the proportion of hyperglycemic events (BG>=400mg/dL) that were detected; and d) the proportion of all hypo- and hyperglycemic events combined that were detected.

Results: Glucometer data were obtained from 74 subjects, who completed 84% of the protocol measurements and averaged 3.6 tests daily. Testing compliance was highest before breakfast (94.6%) and before dinner (84.4%). The correlation between A1c and overall mean BG was 0.738. Hypoglycemia was detected in 2.03% of samples and hyperglycemia in 0.93% of samples. We evaluated once daily testing at 4 testing times: before breakfast, before lunch, before dinner, and at bedtime. The correlations between mean glucose and A1c for these testing times were 0.634, 0.606, 0.658 and 0.665, respectively. Testing before lunch captured the highest proportion of hypoglycemic events (34.2%) and combined events (29.5%). Testing at bedtime captured the highest proportion of hyperglycemic events (51.5%). We evaluated twice daily testing using two testing strategies: before breakfast and before dinner (Option A) and before lunch and at bedtime (Option B). Option A was based on the testing times with the highest compliance while Option B was based on the testing times detecting the highest proportions of hypo- and hyperglycemic events. Correlations between A1c and mean glucose were similar for Options A and B (0.731 and 0.722, respectively). Options A and B also captured a similar proportion of hypoglycemic events (51.1% and 48.9%, respectively). However, Option B captured a higher proportion of hyperglycemic events (70.8% vs. 29.2%) and combined events (55.8% vs. 44.2%) than Option A.

Conclusions: The optimal strategy for once-daily testing depends upon the primary objective for monitoring. Patients prone to hypoglycemia should test before lunch, while those who are poorly controlled should test at bedtime. For twice-daily testing, testing before lunch and at bedtime captured more hypo- and hyperglycemic events than testing before breakfast and before dinner.

Impact: If the results of this study are confirmed, a significant proportion of patients testing less frequently than recommended may need to change their BG monitoring regimen.