Zachary T. Bloomgarden, MD


August 30, 2001


When should a patient with type 2 diabetes start treatment with insulin?

Loakim Tsigeridis, MD

Response from Zachary T. Bloomgarden, MD

It is clear that insulin deficiency is as much a part of the development of type 2 diabetes as is insulin resistance.[1] However, in type 2 diabetes insulin has been relegated to being a "therapy of last resort." This is a far cry from the recommendations that followed the release of the University Group Diabetes Program (UGDP) study results in the 1970s. In this 9-year study, 1027 patients were treated with placebo, a constant-insulin dose, variable-insulin doses, phenformin, or tolbutamide. In the UGDP, plasma glucose levels were maintained at 121 mg/dL in the variable-dose-insulin group, 30-40 mg/dL lower than in the placebo or fixed-insulin treatment group, a difference similar to that observed in the United Kingdom Prospective Diabetes Study (UKPDS). Phenformin and tolbutamide were reported to increase mortality. Despite leading to better glycemic control, insulin failed to improve mortality over that seen with placebo, although there was a suggestion from post-hoc analysis that patients in the insulin treatment groups with good glucose control had fewer cardiovascular (CV) events than did those with fair or poor control.[2]

The UKPDS showed that treatment with insulin led to a 10-year mean HbA1c level of 7.1%, in comparison with 7.9% in the conventional treatment group. Hypoglycemia, however, occurred in 37% of patients annually with insulin, exceeding the 11% and 18% frequency with the sulfonylureas (SU) chlorpropamide and glyburide, respectively. Mean weight gain was 6.5 kg over 10 years with insulin, but 4.2 kg with chlorpropamide and 5.1 kg with glyburide. Outcome measures did not differ between insulin and SU, but there was a suggestion that patients treated with metformin experienced superior CV outcomes after a 6-year lag period. However, patients receiving both metformin and SU appeared to have worse outcome than those receiving SU alone. The UKPDS also showed that long-term monotherapy is unlikely to lead to maintenance of good glycemic control. At 3, 6, and 9 years, HbA1c levels < 8% were seen in 48%-55%, 35%-38%, and 16%-21% of patients, respectively, suggesting that over time, patients with type 2 diabetes require combination therapy and/or insulin.[3]

Insulin may indeed allow better maintenance of control than oral agents. A comparison of insulin treatment resulted in HbA1c levels that were 1.6% lower compared with SU after 2 years in a group of patients initially treated with SUs.[4] In particular, this seems to be the case with multiple-dose insulin (MDI) regimens, which may be as useful in type 2 as in type 1 diabetes.[5] A comparison of conventional treatment with MDI in type 2 diabetes has indeed shown a convincing decrease in microvascular endpoints.[6] Such improvement may also be seen with combination therapy of insulin and oral agents.[7] Both SU[8,9] and metformin[10] appear to provide benefit in combination with insulin, with bedtime insulin being a convenient and physiologically advantageous approach in this setting. Innovative approaches involving devices such as insulin pens, new long-acting insulins such as insulin glargine, and future new insulin formulations such as inhaled insulin may further expand opportunities for insulin use in patients with type 2 diabetes.

At present, it seems logical to use insulin as initial therapy in the clinical setting of patients with type 2 diabetes whose glycemic control is unstable and who appear unlikely to show a sufficiently prompt response to an oral agent to reduce the risk to the patient of continual serious hyperglycemia. When hepatic insufficiency, congestive heart failure, or renal insufficiency are present, metformin is contraindicated and thiazolidinedione administration may be problematic. In these circumstances in particular, insulin therapy should be considered. Although insulin may cause more hypoglycemia in large patient groups, selected patients at risk for hypoglycemia may be deemed better candidates for insulin than for SU treatment, particularly with use of insulin lispro. The rapid-acting insulin secretagogues repaglinide and nateglinide are also of use in this setting.