Self-Monitoring of Blood Glucose Can Be Effective in Type 2 Diabetes Only If It Serves a Clearly Identified Purpose: Response to Court

The use of self-monitoring of blood glucose (SMBG) in type 2 diabetic patients is still a matter of debate, as documented by the letter from Dr. Court (1). Despite its recommendation for all diabetic patients by the American Diabetes Association, the evidence supporting its effectiveness in improving glycemic control is questionable. A recent meta-analysis of all randomized trials on this topic failed to show any benefit for patients practicing SMBG (2). Previous observational studies were also unable to document a relation between frequency of SMBG and metabolic control (3).

From this point of view, the results from Karter et al. (4) are the first to show, in a highly homogeneous setting, a positive association between SMBG practice and metabolic control, irrespective of the treatment. These data are not confirmed by our results (5) or those from the recently published third National Health and Nutrition Examination Survey (NHANES III) (6). We believe that the large number of centers involved in our study, as well as in the national sample of the NANHES III, represent a strength rather than a limitation because they provide a true picture of diabetes care, which is without any doubt much more heterogeneous than that described in the article by Karter et al.

To take into account the inter- and intracenter variability, we applied appropriate multilevel models, thus adjusting for the correlation between observations relative to patients enrolled by the same center. The comparability of HbA_1c levels was made possible by widely accepted mathematical transformations; in previous analyses, we have shown that HbA_1c levels in the very same population were strongly associated with physicians’ beliefs, as well as with known clinical correlates (7).

We agree that our study does not exclude the possible benefit of SMBG in type 2 diabetes; on the contrary, it clearly shows that SMBG can be associated with better metabolic control in those patients able to self-adjust insulin doses, thus stressing the crucial role played by patient education. On the other hand, when the information deriving from SMBG cannot be readily used by the patient or in the absence of clear guidelines on the actions to be undertaken in the presence of high blood glucose levels in individuals not treated with insulin, this practice can be related to psychological harm and feelings of powerlessness, as our data clearly show.

We believe that a deeper knowledge of the features of diabetes care in the Kaiser Permanente Medical Care Program of Northern California, which made SMBG so successful in determining lower HbA_1c levels (even in those patients treated with diet alone), would be of great interest in understanding the transferability of these results in other, more heterogeneous, clinical settings.