Hospitalized patients: simple algorithms best for glucose control.(Clinical Rounds)

BOSTON -- Hospitals should stop using "sliding scale" insulin orders and replace them with simple algorithms for determining dosages of intravenous insulin, Dr. Irl B. Hirsch told physicians at the annual meeting of the American Association of Clinical Endocrinologists.

"This is the problem on the floor with sliding scale insulin--especially when you're dealing with severe insulin deficiency," Dr. Hirsch said, showing a picture of a roller coaster ride. "It doesn't work."

Keeping the algorithms simple is important because many hospitals are flummoxed by how to implement new, tighter glycemic controls for diabetic patients, said Dr. Hirsch, a professor at the University of Washington, Seattle, Several physicians in the audience told him that nurses opposed introducing intravenous insulin in their units.

In December, the American Association of Clinical Endocrinologists and the American College of Endocrinology recommended upper limits for glycemic targets of 100 mg/dL during labor and delivery, 110 mg/dL in intensive care units, and 110 mg/dL (preprandial) to 180 mg/dL (maximal) in non-intensive critical care units. Although the American Diabetes Association and numerous other groups have endorsed the new standard. Dr. Hirsch suggested the Joint Commission on Accreditation of Healthcare Organizations may have to make it a requirement for accreditation before some centers will act.

"We have to convince the nurses and the administrators that we need to put the resources into this at the level of the hospital because we get better outcomes for our patients," he said, saying the data are too overwhelming too ignore. Among the outcomes he cited were 43% reduction in intensive care unit mortality and 34% reduction in hospital mortality for critically ill patients given intravenous insulin (N. Engl. J. Med. 345[19]:1359-67, 2001).

Whether intravenous insulin or improved glycemic control caused the better outcomes is still unknown, said Dr. Hirsch, who suggested the answer was both.

At his hospital, a new protocol for insulin infusion was introduced in July 2002. A case-control study of 78 patients conducted before and after the change found the proportion of hyperglycemic patients (defined as glucose levels ...