In the United States, we use twice as much insulin per capita as any other country. While this increased utilization may be partially due to the increased incidence of type 2 diabetes in minority populations, a greater frequency and severity of obesity, high fat intake, and a more sedentary lifestyle, the major reason is an historical lack of availability of oral therapy other than sulfonylureas.

Until May 1995, only sulfonylureas were available, and U.S. physicians and diabetic patients were confined to oral monotherapy. When sulfonylureas failed, insulin therapy had to be initiated. In some cases, control could be achieved with a sulfonylurea by day and either NPH insulin at bedtime or mixed short-acting and intermediate-acting insulin at the evening meal. But in most cases, achieving adequate control required twice daily mixed insulin.¹

Internationally, when monotherapy failed, combination therapy with metformin and a sulfonylurea was initiated. This combination therapy resulted in an average decrease in fasting glucose of 60 mg/dl and a 1.7% decline in the HbA1c and usually postponed the need for insulin for 3–5 years—a considerable portion of type 2 patients’ tenure with known diabetes.²

Utilization of combined metformin and sulfonylurea therapy was quickly adopted in this country when metformin became available. As a result, in the first year of metformin’s availability, insulin starts decreased by 50%, and total insulin sales decreased by 5–10%. With the availability of troglitazone and other thiazolidinediones, these figures are likely to decrease further.

Disadvantages of Insulin Utilization in Type 2 Patients
Insulin is not the ideal therapy for type 2 diabetes. Although glycemic control can be improved, because of the suppression of endogenous insulin production, hepatic insulin levels are low, and hepatic glucose production is not suppressed. To achieve normal or near-normal glycemic control, high dose insulin with increased serum insulin levels must be utilized. While high serum insulin levels derived from exogenous insulin may or may not be associated with acceleration of hypertension and atherosclerosis, the utilization of high dose exogenous insulin does result in weight gain. The need to increase caloric intake to either treat or avoid hypoglycemia rather than the hyperinsulinemia itself is the factor responsible for this weight gain.

Should other therapies fail, insulin utilization to decrease hyperglycemia with its resultant positive effects on the microangiopathic and macroangiopathic processes is necessary, and the weight gain has to be accepted. However, if insulin doses can be minimized by utilizing insulin in combination with oral hypoglycemics, then better glycemic control and fewer macrovascular complications might be expected.

When metformin became available in 1995, the initiation of insulin therapy in type 2 diabetes patients at our diabetes center came to a screeching halt. However, I wondered if those patients who had previously failed sulfonylurea therapy and had been started on twice daily mixed insulin could be converted back to oral therapy.

To avoid unnecessary clinic visits and expense and to avoid possible ketoacidosis, only those patients who had a random C-peptide above 0.8 ng/ml were included. Recognizing that these patients were unlikely to be controlled with monotherapy, I devised a protocol for conversion similar to a protocol I had used to convert type 2 patients to oral therapy after insulin utilization for a symptomatic onset of type 2 diabetes or for insulin utilization with an intercurrent illness.³

At the first visit, insulin was reduced by one-third, and metformin, 500 mg twice daily, was started. At the second visit 2 weeks later, if a random serum glucose was between 100 and 200 mg/dl, then insulin was reduced by an additional one-third of the initial dose, and a sulfonylurea (initially glyburide, 2.5 mg twice daily, but later glime-peride, 2 mg/day) was added to the metformin. At the third visit 2 weeks later, if the random serum glucose was between 100 and 200 mg/dl, insulin was discontinued, and the doses of metformin and sulfonylurea were adjusted if necessary. Similar adjustments were made at the fourth visit, which occurred 2 weeks after insulin had been discontinued. If at any visit the serum glucose was between 200 and 300 mg/dl, the oral hypoglycemic dose was increased, and insulin was not reduced. If at any time the serum glucose was in excess of 300 mg/dl, the patient was placed back on twice daily mixed insulin with or without metformin.³

Surprisingly, initial conversion was very successful. Of the first 75 patients, 60 (80%) were successfully converted. Those who were less obese (body mass index <30), less insulin-resistant (needing an average of 0.7 U/kg of insulin), and had the shortest duration of insulin utilization were more likely to be successfully converted. Overall, the primary failure rate was 23%, and the secondary failure rate was ~30% per year. The minority of primary failures and the majority of secondary failures could be controlled on combination metformin and a sulfonylurea plus premixed insulin given before supper. This indicates that secondary failures are at a less advanced stage of the disease than primary failures.³

Once Daily Insulin and Combination Oral Therapy
Reinitiation of insulin therapy was first attempted with premixed (70/30 human NPH and regular) insulin given at supper. This I found to be a better regimen than NPH given at bedtime. In type 2 diabetes, the highest glucose is usually before breakfast, and glucose levels drop during the day. Therefore, NPH given at bedtime can lower the fasting glucose, and glycemic control can be maintained with oral hypoglycemics for the rest of the day. However, most Americans eat sparingly and exercise most during the day, eat most at the evening meal, and are least active after the evening meal. Because of this, type 2 diabetes patients in my experience have better results with oral therapy by day and 70/30 insulin at supper rather than NPH insulin at bedtime.

Patients were started on 0.2 U/kg of 70/30 insulin at supper and increased the dose by 2 U every 3 days until home glucose monitoring revealed that fasting glucose was <120 mg/dl or until hypoglycemia occurred. If glycemic control was not obtained in this way, the patients were put back on twice daily mixed insulin and metformin.

Improved Glycemic Control on Oral Therapy
Glycosylated hemoglobin dropped by 1.3% over the first 6 weeks in those patients who were successfully converted from twice daily mixed insulin to combination oral therapy. That this was not a Hawthorn phenomenon (compliance and better control induced by frequent office visits) was proven when, after 6 months, this improvement was maintained in those on oral combination therapy alone, those on oral combination therapy and evening 70/30 insulin, and those on twice daily 70/30 insulin and metformin. Those who because of cost or side effects were transferred back to twice daily mixed insulin alone had no significant change in glycemic control.

The improved glycemic control on combination oral therapy with or without insulin, I believe, is due to higher hepatic insulin levels and suppression of glucose production by the combined effects of higher endogenous hepatic insulin levels induced with sulfonylureas and the suppressant action of metformin on glycogenolysis and/or gluconeogenesis.

Weight Loss on Combination Oral Therapy
Another surprising outcome of this study was the weight loss seen in patients successfully converted from insulin to combination oral therapy. Initial weight loss was undoubtedly due to discontinuation of insulin and, with it, the need to eat more. However, the weight loss continued, and I believe it was due to the utilization of metformin.

With better glycemic control, weight gain is to be expected because of restoration of fluid volume and muscle mass and reversal of glycosuria and urinary calorie loss. Weight gain occurs with sulfonylureas, insulin, and thiazolidinediones. However, in spite of better glycemic control, monotherapy with metformin results in weight loss or no weight gain. A prospective, randomized, double-blinded study has shown that metformin decreases appetite and calorie intake.⁴

Oral Monotherapy
The question of whether insulin therapy in type 2 diabetes patients could be discontinued and replaced with oral monotherapy is often asked. The natural history of type 2 diabetes, in which glycemic control is obtained initially with monotherapy before combination therapy, makes this unlikely. While troglitazone reduces insulin needs by as much as 54%, only 15% of insulin-requiring patients are able to discontinue insulin. Sulfonylureas, such as glimeperide, reduce insulin needs by 38%, and metformin alone reduces insulin needs by 35%. Therefore, discontinuation of insulin therapy in type 2 diabetes requires the effects of at least two hypoglycemics with different mechanisms of action.

Other Potential Combinations
Could triple or quadruple therapy utilizing the different actions of thiazolidinediones, alpha-glucosidase inhibitors, metformin, and sulfonylureas avoid the use of insulin in patients with type 2 diabetes? The answer is probably yes, although no data on triple or quadruple therapy is available. However, the cost and potential multiplicity of side effects makes this impractical at present.

Could the same results be achieved using other combinations? The answer is unquestionably yes, with the combination of metformin’s hepatic action and troglitazone’s action on muscle looking promising.⁵ In addition, a sulfonylurea-thiazolidinedione combination could be as effective as a meformin-sulfonylurea combination.

Summary
With the availability of more than one group of oral hypoglycemics, conversion from insulin to combination oral therapy is possible in C-peptide-positive, insulin-utilizing patients with type 2 diabetes. Conversion from insulin results in better glycemic control and weight loss. Combination therapy with other oral hypoglycemics will probably also result in the avoidance or decreased use of insulin in type 2 diabetes patients.

References

¹Clements RS Jr., Bell DSH, Benbarka A, Capper SA: Rapid insulin initiation in non-insulin-dependent diabetes mellitus. Am J Med 82:415-20, 1987.

²DeFronzo RA, Goodman AM: Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus. N Engl J Med 333:541-49, 1995.

³Bell DSH, Mayo MS: Outcome on metformin-facilitated reinitiation of oral diabetic therapy in insulin-treated patients with non-insulin-dependent diabetes mellitus. Endocrine Pract 3:73-76, 1997.

⁴Lee A, Bray GA: Metformin decreases food consumption in obese non-insulin-dependent (NIDDM) diabetes. Diabetes 45 (Suppl 2): 170A, 1996.

⁵Inzucchi SE, Maggs DG, Spollett GR, Paige SL, Rife FS, Shulman GI: Efficacy and metabolic effects of troglitazone and metformin in NIDDM. Diabetes 46 (Suppl 1): 34A, 1997.

David S.H. Bell, MB, is a professor of medicine and director of the endocrine clinic at the University of Alabama at Birmingham School of Medicine.