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


Drugs plus diet was better than diet alone for newly diagnosed diabetes

ACP J Club. 1995 July-Aug;123:5. doi:10.7326/ACPJC-1995-123-1-005

Source Citation

United Kingdom Prospective Diabetes Study Group. United Kingdom Prospective Diabetes Study (UKPDS) relative efficacy of randomly allocated diet, sulphonylurea, insulin, or metformin in patients with newly diagnosed non-insulin dependent diabetes followed for three years. BMJ. 1995 Jan 14;310:83-8.



To determine the 3-year effectiveness of diet plus drug therapy compared with diet alone in lowering plasma glucose levels in patients with newly diagnosed non-insulin-dependent diabetes mellitus (NIDDM).


Randomized controlled trial.


15 British diabetes clinics.


2769 patients aged 25 to 65 years with newly diagnosed NIDDM, fasting plasma glucose levels between 6.1 mmol/L and 14.9 mmol/L, and no hyperglycemic symptoms. Exclusion criteria were severe vascular disease, accelerated hypertension, proliferative or preproliferative retinopathy, renal failure with plasma creatinine level > 175 mmol/L, life-threatening disease, illness requiring steroids, occupation precluding insulin treatment, or ketonuria > 3 mmol/L. 2520 patients (mean age, 52 y; 1455 men; 81% white) completed 3-year follow-up.


Patients were stratified for obesity and allocated to diabetic diet alone (n = 664) or diet plus chlorpropamide (n = 446), glibenclamide (n = 472), or insulin (n = 676). Patients who were obese (> 120% of ideal body weight) were also allocated to metformin (n = 262). Doses were titrated according to the glycemic response.

Main Outcome Measures

Fasting plasma glucose level and insulin level, glycosylated hemoglobin (HbA1c) level, and body weight.

Main Results

At 3 years, the median fasting plasma glucose level was lower in nonobese patients receiving diet plus chlorpropamide (7.0 mmol/L), glibenclamide (7.6 mmol/L), or insulin (7.4 mmol/L) than in patients receiving diet alone (9.0 mmol/L) (P < 0.001). Results were similar in obese patients allocated to metformin (7.7 mmol/L). The mean HbA1c level was lower, and mean weight and fasting plasma insulin level were higher among patients receiving any of the 3 drugs than among those receiving diet alone (P < 0.001). In obese patients allocated to metformin, the mean fasting plasma insulin level was lower and body weight was not changed.


The addition of chlorpropamide, glibenclamide, insulin, and metformin to a diabetic diet achieved lower plasma glucose levels than diet alone in patients with newly diagnosed non-insulin-dependent diabetes mellitus.

Sources of funding: Medical Research Council; British Diabetic Association; National Institutes of Health (U.S.); British Heart Foundation; Health Promotion Research Trust; Becton Dickinson; Boehringer Mannheim; Bristol Myers Squibb; Hoechst; Lilly; Lipha; Novo Nordisk; Clothworkers Foundation.

For article reprint: U.K. Prospective Diabetes Study, Diabetes Research Laboratories, Radcliffe Infirmary, Oxford, United Kingdom OX2 6H3.


In 2 to 3 years, the United Kingdom Prospective Diabetes Study (UKPDS) may become as important to physicians who treat patients with NIDDM as the Diabetes Control and Complications Trial (DCCT) is for the treatment of insulin-dependent diabetes. This trial is testing 5 commonly used treatments for NIDDM and asks which is most effective at lowering glucose levels (intermediate outcome) and preventing microvascular and macrovascular disease (long-term outcomes) over 10 years. Other than the ill-fated University Group Diabetes Program (UGDP), no other large, long-term randomized trials in NIDDM have been done. Because > 95% of patients with diabetes have NIDDM and because they have more serious outcomes and die more frequently of macrovascular disease than of microvascular disease, the relevance and necessity of this trial is enormous (the DCCT primarily addresses microvascular disease).

While we await the long-term outcome data, the current study from the UKPDS gives us important clinical information. First, all active pharmacologic treatments lower glucose levels more effectively than diet alone. Second, chlorpropamide, glibenclamide (in the United States, glyburide), and insulin (starting with 1 shot of ultralente/d) are equally effective in patients who are not obese. In obese patients, chlorpropamide, insulin, and metformin are equally effective and glibenclamide is somewhat less effective. Third, only patients receiving metformin do not gain weight. (This will not be a surprise to most of us.) Fourth, 10% to 15% of patients require additional therapy to maintain a fasting glucose level < 15 mmol/L (290 mg/dL). Fifth, some patients (27% in the study) refuse to use insulin. Finally, none of the patient groups had a mean glucose level within the normal range.

We await the results of the long-term outcomes from this important study.

Jacqueline A. Pugh, MD
Audie L. Murphy Memorial Veterans Affairs Hospital San Antonio, Texas