Wednesday, August 01, 2007

The diabetes mafia needs to refine its arguments

We get a lot of preaching these days about glycemic control in hospitalized patients and the evils of the sliding scale. A recent article in the American Journal of Medicine, Sliding Scale Insulin Use: Myth or Insanity?, illustrates what’s wrong with this preaching.

The authors make two important points: 1) hyperglycemia is under treated in hospitalized patients and 2) inpatients with uncontrolled hyperglycemia need basal coverage, usually with long acting insulin. Unfortunately, their arguments suffer from confusion about the term “sliding scale” and from assertions that go beyond what the evidence supports. As is the case with much of the preaching about glycemic control heard today, these defects in their arguments detract from the message.

Although the authors condemn sliding scale insulin, their examples of adverse effects of sliding scale treatment are primarily those in which sliding scale short acting insulin was the sole modality of treatment of hyperglycemia. In the real world, when patients enter the hospital with pre-existing diabetes on pharmacologic treatment, sliding scale insulin is usually given as an adjunct to such treatment.

While experts (including the authors) roundly condemn sliding scale insulin the basal-bolus regimens they recommend are in fact modified sliding scales. This is because the pre-prandial insulin doses, like traditional sliding scales, are subject to modification based on the prevailing capillary blood glucose. Unlike traditional sliding scales a portion of each pre-prandial dose is fixed. Moreover, currently recommended insulin drip protocols for critically ill patients are, conceptually, sliding scales since infusion rates are adjusted up and down based on the capillary blood glucose concentration. In this case the sliding scale, though modified by route of administration (intravenous), blood glucose target (more aggressive) and frequency of checkpoints (hourly) is a sliding scale none the less.

This tends to be confusing to clinicians. When experts categorically denounce sliding scales while recommending modified sliding scales what do they really mean? Physician buy in might improve if experts were more explicit about how current practice should change.

A second weakness of the article is that its recommendations go beyond current evidence. The authors state:

In view of this evidence, a recent position statement by the American Association of Clinical Endocrinologists recommended glycemic targets for hospitalized patients in the intensive care unit between 80 and 110 mg/dL, and in noncritical care settings a preprandial glucose goal less than 110 mg/dL and a random glucose less than 180 mg/dL. The Joint Commission on Accreditation of Healthcare Organization recently proposed tight glucose control for the critically ill as a core quality of care measure for all US hospitals that participate in the Medicare program.

These generalizations are not warranted by current evidence. Preliminary evidence suggests certain aggressive glycemic targets benefit specific subsets of patients, but these data do not warrant broad recommendations. The specific circumstances in which aggressive glycemic control is beneficial, as well as the optimal targets for various subsets of patients are far from settled.

Background: Review from Diabetes Care on inpatient management of hyperglycemia. Although somewhat dated, this review is timely and thorough in its coverage of multiple practical aspects of management.

No comments: