Is Self-Monitoring of Blood Glucose Appropriate for All Type 2 Diabetic Patients? The Fremantle Diabetes Study: Response to Kolb et al.

Despite the negative results of our community-based study of the link between self-monitoring of blood glucose (SMBG) and glycemia (1), Kolb et al. (2) commented on the overall concordance of our study and several other observational studies. They also suggest that we might have missed a positive impact of SMBG since new users in our cohort were not differentiated from either prevalent users or never users.

In our prospectively followed 531 type 2 diabetic patients (1), 92 who …

D espite the negative results of our community-based study of the link between self-monitoring of blood glucose (SMBG) and glycemia (1), Kolb et al.
(2) commented on the overall concordance of our study and several other observational studies. They also suggest that we might have missed a positive impact of SMBG since new users in our cohort were not differentiated from either prevalent users or never users.
Previously, we have shown in the same longitudinal cohort of 531 patients (5) that the median A1C was reduced by 0.3% when diet-treated patients started OHA and by 1.5% when OHA-treated patients started insulin. The apparent benefit of initiating SMBG might, therefore, reflect intensification of treatment. However, when our patients started SMBG, 85% continued with the same treatment, 12% reduced treatment, and 3% intensified treatment.
The largest randomized controlled trial (6) has found that the glycemic improvement observed in non-insulintreated patients allocated to SMBG occurred in the first 3 months, with a steady state thereafter. These and our longer-term data suggest that SMBG may have only a relatively transient beneficial effect on glycemia. There is a need for strategies that ensure its sustainability.   (1) used data from the Diabetes Control and Complication Trial to investigate the relationship between glycemic variability and the subsequent development of diabetes complications. They report that glucose variability, as measured by a quarterly eight-point glucose profile (excluding the 3 A.M. value because of limited data), was not associated with development or progression of retinopathy or nephropathy. An accompanying editorial by Bolli (2) highlights the potential clinical impact of this finding, stating that "the instant blood glucose at a given time of day is not important, and it does not matter if it is high or low either before or after meals (or vice versa) as long as A1C is at the target value Ͻ7.0%." We believe that these results, and the clinical recommendations that have sprung from them, should be interpreted with caution. While the Diabetes Control and Complication Trial database is large and its data regarding complications extraordinary, quarterly seven-point glucose profiles are unlikely to fully reflect true glycemic variation in these subjects with type 1 diabetes.
Continuous glucose monitors provide the opportunity to capture the magnitude of glycemic variation far better than seven-point glucose profiles. The Diabetes Research in Children Network (DirecNet) Study Group (3) compared simultaneous eight-point glucose profiles over three days with near continuous glucose profiles (values every 5 min) using Medtronic-Minimed CGMS in 161 children and adolescents with type 1 diabetes. The eight-point glucose profiles were measured using One Touch UltraSmart (LifeScan) meter, a device shown to be quite accurate (4). The meal-related glucose excursion measured using eightp o i n t t e s t i n g w a s c a l c u l a t e d b y subtracting premeal from postmeal glucose. The analogous glucose excursion measured with continuous glucose selfmonitoring (CGMS) was calculated as the difference between the premeal CGMS value (corresponding to the time of the eight-point test) and the peak value (within 3 h of the premeal eight-point test). Postprandial excursions were two to three times larger when measured by the CGMS than by eight-point testing. These findings are not surprising as it is unlikely a single glucose measurement would coincide with the postmeal peak. Moreover, a single measurement cannot measure the duration of the postmeal glucose rise.
Given that glucose profiles based on single point-in-time postprandial measurements are a suboptimal measure of glycemic variability, we believe it is premature to discount the potential clinical importance of reducing glycemic variability. Further studies using continuous glucose data will be needed to finally answer this important question.