Current issues of ACP Journal Club are published in Annals of Internal Medicine


Addition of insulin or metformin after secondary failure to glyburide enhanced glycemic control

ACP J Club. 1992 Sept-Oct;117:46. doi:10.7326/ACPJC-1992-117-2-046

Source Citation

Trischitta V, Italia S, Mazzarino S, et al. Comparison of combined therapies in treatment of secondary failure to glyburide. Diabetes Care. 1992 Apr;15:539-42.



To compare glycemic control after the addition of either metformin or insulin in patients with non-insulin-dependent diabetes with secondary failure to glyburide.


Randomized crossover trial.


University medical center in Italy.


20 patients (mean age 54 y, duration of diabetes 12 y) with non-insulin-dependent diabetes and no renal, hepatic, cardiovascular, or systemic diseases, and no body weight change in the last 3 to 6 months were recruited. All patients had previously been successfully treated with glyburide but were subsequently unable to achieve good metabolic control despite a 4-week regimen of a controlled diet and 15 mg/d of glyburide.


Participants were randomized to the addition of either bedtime NPH insulin, 0.15 to 0.20 U/kg body weight, or metformin, 0.5 g, 3 times daily given 20 minutes before meals. After 8 weeks of therapy, patients had a 3-week washout period of only diet and glyburide and then the alternate treatment for 8 weeks. 16 patients (80%) completed the study.

Main outcome measures

Fasting plasma glucose (FPG) and postprandial plasma glucose (PPPG) were measured at baseline and at 2, 4, and 8 weeks for each crossover period. Weight, HbA1c, cholesterol, and triglyceride levels were measured at baseline and at 8 weeks during each treatment period.

Main results

Both therapies improved glycemic control. Using the pooled data, patients in the insulin group had baseline FPG, PPPG, and HbA1c values of 14.1 mM, 16.9 mM, and 9.6%. After 8 weeks these had decreased to 9.5 mM, 13.3 mM, and 7.8%. Patients in the metformin group had initial FPG, PPPG, and HbA1c values of 13.5 mM, 16.2 mM, and 9.5%. After 8 weeks these were 9.7 mM, 11.3 mM, and 8.0%. At weeks 2, 4, and 8 the PPPG in the metformin group was less than the insulin group (P < 0.05). Neither treatment changed mean cholesterol or triglyceride values. Body weight remained unchanged for metformin but increased for insulin (64.1 vs 66.0 kg; P < 0.01). No patients reported symptomatic hypoglycemic episodes.


Addition of either a fixed low dose of insulin at bedtime or metformin 3 times per day for patients with non-insulin-dependent diabetes who had secondary failure to glyburide was effective in improving glycemic control. Decreases of postprandial plasma glucose were greater with metformin.

Source of funding: Not stated.

Address for article reprint: Dr. V. Trischitta, Cattedra di Endocrinologia dell'Universita di Catania, Ospedale Garibaldi, Piazza S.M. di GesĂș, 95123 Catania, Italy.


The optimal treatment for patients with typical non-insulin-dependent diabetes after sulfonylurea failure continues to be debated. Replacing the sulfonylurea by a once-daily insulin injection is often ineffective, necessitating inconvenient multiple injections of different insulin types. Other alternatives include the addition of a single dose of insulin in the evening or multiple doses of metformin. This study compares these 2 methods of combined therapy using a careful design. The main limitation of the study is the understandable lack of masking of subjects and investigators regarding whether the oral or the injected agent was used. The design is strengthened by efforts to exclude persons with late-onset insulin-dependent diabetes by documenting previous success with a sulfonylurea, normal C-peptide levels, and absence of anti-islet-cell antibodies. The results convincingly show that at the time of recent sulfonylurea failure, both of these simple regimens can greatly improve glycemic control.

Each regimen had advantages. Metformin-sulfonylurea combined therapy required only oral administration, gave better PPPG levels than insulin-sulfonylurea combination therapy (but comparable FPG and HbA1c values) and caused no weight gain. The addition of a single dose of insulin taken in the evening may be more convenient for some patients than adding 3 doses of metformin and might also have achieved lower mean glucose values had upward adjustments of insulin dosage been attempted. Increased insulin doses may, however, increase the risk for hypoglycemia and weight gain. The main practical implication of this study is that combined drug therapy may come to be accepted for diabetes as it is for many other conditions. Circumstances under which metformin-sulfonylurea might be preferred over insulin-sulfonylurea therapy, and vice versa, remain to be defined. Similarly, cases for which metformin alone or multiple injections of insulin alone are the best approach must be identified. We can expect more studies aimed at refining these clinical decisions.

Matthew C. Riddle, MD
Oregon Health Sciences UniversityPortland, Oregon, USA

Matthew C. Riddle, MD
Oregon Health Sciences University
Portland, Oregon, USA