Background: There is a controversy about self-monitoring of blood glucose (SMBG) in patients with type 2 diabetes who are not using insulin. Randomized trials are limited in duration, size, and validity.
Methods: Systematic search for randomized trials and observational studies published since 1990. For inclusion studies had to report on SMBG in type 2 diabetes managed with oral hypoglycaemic agents and/or diet alone, HbA(1c) or clinical outcome, have at least 50 patients and be of at least 6 months' duration.
Results: Three randomized trials with 1000 patients were included, though all had interventions differing in the amount of education on SMBG, and in the population studied. The two larger studies had statistically significantly lower HbA(1c) levels with SMBG. Thirteen observational studies had information on over 60,000 patients. Smaller studies had lower initial HbA(1c) and showed no association between SMBG and laboratory or clinical improvement. Larger studies tended to have higher initial HbA(1c) and did show an association between SMBG and laboratory or clinical improvement. Overall, improvement in glycaemic control with SMBG tended to be seen in studies with initial HbA(1c) above 8%.
Conclusions: It is likely that SMBG is beneficial in some circumstances, for example as an educational tool, for patients with type 2 diabetes not using insulin who have poor glycaemic control. More information is needed at the level of the individual patient, rather than group means, and about timing and frequency of monitoring, response to those results, what constitutes effective patient education, and long-term clinical outcomes.