Self-monitoring of blood glucose in non-insulin-treated diabetic patients: a longitudinal evaluation of its impact on metabolic control

Aims In the framework of a nationwide outcomes research programme, we assessed the impact of self-monitoring of blood glucose (SMBG) on metabolic control over 3 years in patients with Type 2 diabetes mellitus (DM2) not treated with insulin. Methods The study involved 1896 patients who completed, at 6-month intervals for 3 years, a questionnaire investigating SMBG practice. Clinical information was collected by participating clinicians at the same time intervals. The predictive value of SMBG frequency on long-term metabolic control was estimated using multilevel analysis. The impact of SMBG on metabolic control was also evaluated in distinct and homogeneous subgroups of patients showing different likelihood of performing SMBG, identified using a tree-growing technique (RECPAM). Results Overall, 22% of the patients were on diet alone and 78% were treated with oral agents; 41% practiced SMBG >= 1/week (10.3%>= 1/day). The analysis of metabolic control according to the frequency of SMBG failed to show any significant impact of this practice on HbA(1c) levels over 3 years. Similarly, changes in SMBG frequency during the study were not related to significant changes in HbA(1c) levels. RECPAM analysis led to the identification of eight classes, characterized by substantial differences in the likelihood of performing SMBG with a frequency of at least 1/week. Nevertheless, in none of the RECPAM classes identified, did SMBG predict a better metabolic control over 3 years of follow-up. In those RECPAM classes indicating that SMBG was mainly performed to avoid hypoglycaemic episodes, SMBG was associated with a decrease in the frequency of hypoglycaemic episodes during the study. Conclusions In a large sample of non-insulin-treated Type 2 diabetic patients, the performance and frequency of SMBG did not predict better metabolic control over 3 years. We could not identify any specific subgroups of patients for whom SMBG practice was associated with lower HbA(1c) levels during the study.