Clinically Useful Insights From the Early Results of the UKPDS

Reviewed by Matthew C. Riddle, MD

STUDY
The United Kingdom Prospective Diabetes Study (UKPDS) Group: United Kingdom Prospective Diabetes Study 24: A 6-year, randomized, controlled trial comparing sulfonylurea, insulin, and metformin therapy in patients with newly diagnosed type 2 diabetes that could not be controlled with diet therapy. Ann Intern Med 128:165-75, 1998.

SUMMARY
Objective. This trial was designed to show whether early treatment of hyperglycemia in type 2 diabetes can reduce morbidity and increase life expectancy and to compare the results of several alternative treatments.

Design and subjects. After screening out people with ketosis and therefore severe insulin deficiency, the UKPDS enrolled 4,075 people with newly diagnosed type 2 diabetes. Of these, 746 (18%) had excellent responses to initial lifestyle efforts. The main group of 2,769 (68%) had moderate responses and were randomized to sulfonylurea, metformin, or insulin treatment or to continued diet and exercise. Preliminary findings from this main randomization protocol have been published.¹

The study reviewed here deals with the remaining 560 subjects (14%), who continued to have either symptoms of hyperglycemia or fasting plasma glucose >15 mMol/l (270 mg/dl) despite efforts to improve lifestyle. This "primary diet failure" group was younger and less obese than the main group and had more retinopathy and impotence.

Methods. These subjects were randomized to treatment with sulfonylurea (glyburide [DiaBeta, Glynase, or Micronase] or chlorpropamide [Diabinese]), metformin (Glucophage), or ultralente insulin. If glycemic control became inadequate on therapy with one oral agent, another agent was added, and if oral combined therapy failed, insulin alone was started. If necessary, regular insulin could be added to the ultralente.

The main measurements were fasting plasma glucose, HbA_1c, weight, and frequency of hypoglycemia. These clinical measures and the evolution of treatments were followed for 6 years.

Results. Glycemic control improved greatly and equally with all treatments, with HbA_1c about 2% lower after 1 year. However, despite medication adjustments, there was a trend toward worse glycemic control over time. After 6 years, only half of the subjects had HbA_1c <8%, and only about 40% of the subjects were still on the single treatment initially assigned.

The group starting with insulin gained more weight (9.9 kg, 22 lb) than the sulfonylurea group (5.3 kg, 12 lb), while the metformin group had no weight change. The group starting with insulin had the most hypoglycemia, including some serious events.

Conclusion. The authors conclude that type 2 diabetic patients with primary diet failure frequently respond well to oral agents initially, and that to minimize hypoglycemia and weight gain, it may be wise to delay the use of insulin until combined oral therapies fail.

COMMENTARY
The UKPDS is the big diabetes intervention trial for this decade, as the Diabetes Control and Complications Trial was for the last one. Its long-awaited 10-year results will be presented this summer. Even if the main questions are not conclusively answered, as may be the case, this study will have greatly expanded our understanding of the natural history of type 2 diabetes. This report of 6 years’ experience with one subgroup of the study illustrates several clinically useful points.

First, the primary diet failure subgroup contained the more severely affected newly diagnosed people. These people differed from the rest in that they were younger, leaner, and had more impaired b-cell function.

Within the group, youth and leanness predicted the need to very rapidly advance to combination therapy or insulin. Sixty percent of those under 46 years of age and 50% of those with body mass index <23 needed additional treatment within a year of randomization. Many of them must have had late-onset (autoimmune-mediated) type 1 diabetes masquerading as type 2 diabetes. We need ways to identify and treat such people with insulin earlier than in the past. Another report from the UKPDS suggests that testing for antibodies to glutamic acid decarboxylase (GAD) may help us do this.2

In addition, the primary diet failure subgroup as a whole experienced secondary failure of the initial treatment. At 6 years, at least half of the subjects had advanced beyond their initial treatment to combination regimens, whether sulfonylurea-metformin or ultralente with multiple doses of regular insulin.

The same process, although less dramatic, has been evident in the main randomization group of the UKPDS1 and clearly reflects the natural history of type 2 diabetes. No single treatment can be expected to remain successful for more than a few years, not because of a limitation specific to the drug itself or noncompliance by its user, but because -cell function declines over time. With this experience and the availability of other new agents, such as acarbose and troglitazone, it is clear that multiple-drug regimens will routinely be used for type 2 diabetes.

Finally, some limitations of treatment with insulin alone were evident in this group of subjects. Switching to insulin after failure of oral combinations was often helpful, but only half of the subjects initially assigned to insulin maintained HbA_1c values <8% at 6 years. Moreover, the amount of weight gain and the frequency of hypoglycemia in the insulin group were worrisome.

This difficulty in maintaining control even with insulin is surely related to the natural progression of the abnormalities underlying type 2 diabetes. It also probably reflects the methods for using insulin in this trial. More individualized use of long-acting and rapid-acting insulins, including lispro, might increase the proportion of people with HbA_1c values <8% and reduce weight gain and hypoglycemia.

Even so, as the authors suggest, a trial of oral therapy seems desirable before starting insulin, even for people with quite high glucose levels at the outset. Also, and not mentioned in the authors’ discussion, continuing one or more oral agents after starting insulin may enhance the effectiveness of insulin. Overlapping a sulfonylurea with insulin should prevent escape of glycemic control when insulin is started,3 and combining metformin with insulin may limit weight gain.4 We await studies better defining the possible benefits of combining acarbose or troglitazone with insulin.

We look forward to many more reports from the UKPDS, a true landmark trial. Its insights have improved our treatment strategies already.

REFERENCES

¹UKPDS Study Group: U.K. Prospective Diabetes Study 16. Overview of 6 years’ therapy of type II diabetes: a progressive disease. Diabetes 44:1249-58, 1995.

²UKPDS Study Group: UKPDS 25: Autoantibodies to islet cell cytoplasm and glutamic acid decarboxylase for prediction of insulin requirement in type 2 diabetes. Lancet 350:1288-93, 1997.

³Riddle MC, Schneider J, and the Glimepiride Combination Group: Beginning insulin treatment of obese patients with evening 70/30 insulin plus glimepiride versus insulin alone.    21:1052-57, 1998.

⁴Yki-Jarvinen H, Nikkila K, Tulokas T, Vanamo R, Heikkila M: New thoughts on insulin therapy in NIDDM. 16th International Diabetes Federation Meeting Program, Helsinki, Finland, July 1997.