Glycemic control was not affected by self-monitoring of blood glucose in type 2 diabetes.

Abstract

Questions In patients with newly diagnosed type 2 diabetes, does self-monitoring of blood glucose (SMBG) improve glycemic control? How do SMBG and no monitoring compare for scores on psychological indices? Methods Design Randomized controlled trial (Efficacy of Self Monitoring On glycemic control in Newly diagnosed type 2 diabetes [ESMON]). Allocation Unclear allocation concealment.* Blinding Blinded (laboratory staff).* Follow-up period 1 year. Setting Outpatient hospital diabetes clinics in Altnagelvin, Belfast City, Causeway, and Ulster, Northern Ireland. Patients 184 patients <70 years of age (mean age 59 y, 60% men) with newly diagnosed type 2 diabetes. Exclusion criteria were secondary diabetes, previous SMBG, use of insulin, major illness in the past 6 months, chronic kidney or liver disease, and alcohol misuse. Intervention SMBG (n =96) or no monitoring (n =88). The SMBG group made 4 fasting and 4 postprandial capillary blood glucose measurements per week and were provided dietary and exercise suggestions for high readings. All patients attended a structured education program. Outcomes Changes in hemoglobin (Hb) A1c levels, scores on psychological indices (depression, anxiety, treatment satisfaction, and attitude; measured on a 100-point scale), body mass index (BMI), rates of hypoglycemia, and use of oral hypoglycemic drugs. Patient follow-up 98% (intention-to-treat analysis). Main results Patients in the SMBG group scored 6% higher on the depression subscale than did the no-monitoring group (P =0.01). Groups did not differ for changes in HbA1c level (Table), anxiety, treatment satisfaction, attitude, BMI, rates of hypoglycemia, and use of oral hypoglycemic drugs. Conclusion Compared with no monitoring, self-monitoring of blood glucose did not improve glycemic control but increased depression in patients with newly diagnosed type 2 diabetes. Self-monitoring of blood glucose vs no monitoring in newly diagnosed type 2 diabetes Outcome at 1 y Monitoring No monitoring Mean difference (95% CI) HbA1c level (%) 6.9 6.9 0.07 (0.25 to 0.38) Hb = hemoglobin; CI defined in Glossary. Commentary Perhaps 20% of people with diabetes use insulin. The other 80% are managed with diet, lifestyle, and oral hypoglycemic drugs; this group incurs most of the diabetes-related morbidity, mortality, and costs. It is no wonder so much time, effort, and expense is spent on improving glycemic control. SMBG has been recommended to achieve greater quality and quantity of life for patients with type 2 diabetes. Indeed, observational studies suggest that 5 to 10 years of regular SMBG reduces mortality by 51% (1). Similarly, economic projections based mostly on epidemiologic data report that SMBG is cost-effective because it reduces morbidity and adds 6 months to life expectancy (2). Fortunately, we no longer need to rely on opinions and observational studies because there is new, more valid evidence of the effect of SMBG on glycemic control, quality of life, and costs. First, does SMBG improve glycemic control in patients who do not use insulin? Not much. A clinically worthwhile reduction in HbA1c level is 0.5%; previous meta-analysis suggests that the pooled effect of SMBG on HbA1c level was <0.4% (3). However, the quality of pooled studies was less than adequate, and the more rigorous the study, the smaller the effect (3). This was confirmed by the DiGEM trial that compared 2 SMBG strategies with usual care and reported nonsignificant changes in HbA1c levels of approximately 0.15% (4). DiGEM had 3 limitations: patients did not have new diabetes; diabetes was already well-controlled (mean HbA1c level 7.5%); and 85% of potentially eligible patients were excluded. The ESMON trial overcame these limitations: patients had new diabetes; diabetes was not well-controlled (mean HbA1c level 8.7%); and only 13% of potentially eligible patients were excluded. In this well-executed and adequately powered trial, all patients received identical education, and oral hypoglycemic drugs and insulin were given according to standardized protocols. SMBG patients received additional instructions and were asked to test 8 times per week. After 1 year, HbA1c levels did not differ between SMBG and usual-care groups. Second, does SMBG improve quality of life? Not really, and it might cause harm. It is argued that SMBG empowers patients and enables them to take control of their disease. Alternately, someone with an HbA1c level of 7.8% taking metformin, diligently recording daily glucose excursions (or worse, being chastised in the clinic for not testing enough), and not being able to do anything about the data might be distressed. In a cohort of 2300 patients that did not use insulin, SMBG increased depression, stress, and worry compared with no testing (5). Furthermore, there was a dose response, with more testing leading to worse quality of life (5). In the ESMON trial, the SMBG group had more symptoms of depression (P =0.01) and anxiety (P =0.07) than did the no-monitoring group. Third, is SMBG cost-effective? No. Supplies for SMBG are expensive and account for 60% of drug and device expenditures for type 2 diabetes. SMBG supplies are often covered by prescription benefits, and in most jurisdictions, more is spent on SMBG supplies than on oral hypoglycemic drugs. For those who do not use insulin, improvements in glycemic control are small (2). Any benefits related to such small changes in HbA1c levels are mostly estimated from observational studies, whereas previous cost-effectiveness analyses have markedly underestimated harms and disutility of SMBG (2). The rigorous health economic analysis of the DiGEM trial showed that, from a health-care-payor perspective, at 1 year SMBG only costs an additional US $175/patient/y compared with usual care. However, SMBG did not lead to clinically worthwhile changes in glycemic control or change in resource use, and increased depression and anxiety. In summary, there is now robust evidence that recommending routine SMBG for patients with type 2 diabetes who do not use insulin does not improve glycemic control, impairs quality of life, and is not cost-effective. Although people with diabetes should be taught SMBG to better deal with intercurrent illness or symptoms attributed to dysglycemia, the resources expended on routine SMBG in most patients with type 2 diabetes are better spent elsewhere. Of course, this does not apply to the minority of patients who use insulin or who have troublesome hypoglycemia where SMBG plays an important role and where clinical experience and observational evidence probably suffice to guide current practice.