When Insulin Is Started: What to Do With Oral Agents?

July 1, 2006

Which medications should be stopped first when a patient starts using insulin? Are there any that should be continued?

Which oral agents should be stopped first when a patient starts using insulin? Which-if any-should be continued?

- MD

No guidelines address this issue. Some clinicians stop oral agents when insulin therapy is started. They may reason that since oral agents are used to obviate the need for insulin, they are no longer necessary when insulin is required because of persistent hyperglycemia. This approach is usually the most cost-effective.

However, patients whose treatment regimen includes oral agents-particularly insulin-sensitizing drugs, such as metformin and the thiazolidinediones (TZDs)-experience better glycemic control with a lower insulin dose than patients who receive insulin alone. Thus, many diabetologists prefer to continue a patient's oral agents and simply add insulin to the regimen. There are 2 problems with this approach: first, it can be costly; and second, it can burden the patient with an overly complex regimen.

I prefer an intermediate approach, which consists of stopping any insulin secretagogue that is being used (eg, a sulfonylurea or meglitinide) but continuing TZDs and metformin. Since insulin therapy must be initiated presumably because of insufficient endogenous insulin production, it seems logical to discontinue use of secretogogues.

There are 2 caveats to keep in mind when considering this intermediate approach, however. The first is that a secretagogue is sometimes needed in patients who are initially given a single injection of insulin glargine-a popular way of starting insulin therapy. This approach may control fasting blood glucose levels, but because it does not address the problem of prandial insulin production, many patients experience significant increases in postprandial glucose levels. When that occurs, a secretagogue may be indicated.

A second caveat is that the use of insulin in combination with a TZD is associated with an increased risk of weight gain and edema. These adverse effects may outweigh the benefits of such a regimen in some patients.

What is needed is a study that randomly assigns patients already receiving 3 oral agents (eg, metformin, a TZD, and a secretagogue) either to a treatment arm that consists of complete insulin replacement with discontinuation of oral therapy or to an arm in which insulin is added at a lower dose and the oral agents are continued. The treatment goal would be an HbA1c of less than 7%; additional clinical outcomes, such as fasting glucose levels, postprandial glucose levels, weight, and macrovascular events, would be regularly assessed. I suspect that such a study would show that fewer cardiovascular complications develop in those patients who continue to receive sensitizing medications-but at this point, this is largely conjecture.

- Silvio E. Inzucchi, MD
    Professor of Medicine
    Section of Endocrinology
    Yale University School of Medicine
    New Haven, Conn