Wednesday, December 26, 2007

Top ten issues in hospital medicine for 2007, issue 9: more questions than answers concerning inpatient glycemic control

Just about the only thing everyone agreed on this year was that hyperglycemia should not be ignored in hospitalized patients. The unresolved questions relate to what target for what indication. For many of us those targets seem to be moving.

Most of the discussion has centered around glycemic control of critically ill patients in the ICU, particularly septic patients. A landmark 2001 study from Leuven, Belgium demonstrated a decreased mortality with intensive glycemic control in a population of surgical patients. Despite questions about how these results might be extrapolated to critically ill medical patients this study was met with general enthusiasm as insulin drip protocols sprung forth in many hospitals.

Evidence suggested that the greatest benefits of intensive glycemic control occurred with near normal glucose levels at the price of an increased incidence of hypoglycemia. Due to concerns about hypoglycemia the Surviving Sepsis Guidelines compromised and recommended less than 150mg/dl as the target.

In 2004 a “before and after” study (related editorial here) looked at the effects of institution of a glycemic control protocol in a “real world” ICU and found a reduced mortality following the intervention.

Enthusiasm was dampened by publication in 2006 of another study by the Leuven group (sometimes referred to as Leuven-2) in patients with medical illness. There was mortality benefit in patients who remained in the ICU for 3 or more days, harm in patients with less than 3 day stays and no significant mortality benefit overall. Inability to predict patients’ lengths of ICU stay made it difficult to apply conclusions from this study.

Two other recent studies which failed to show a mortality benefit were troubled by inadequate statistical power and a troubling incidence of severe hypoglycemia. Some critics regard these studies as failed trials rather than negative trials.

Questions also remain concerning non-critical patients on the ward and patients with specific diagnoses such as myocardial infarction (the DIGAMI studies yielded differing results and confusion as to whether intensive glycemic control or GIK treatment confers benefit) and stroke.

Thus the clinician faces a daunting task when confronted with a hospitalized patient with hyperglycemia: a glycemic control target must be planned. Selection of the target may depend on the patient’s diagnosis and, if in the ICU, the anticipated length of ICU stay. Moreover, the decision must be made with less than clear evidence from clinical studies. For patients in the medical ICU a possible strategy would be to start with the compromise target of 150mg/dl as recommended by the Surviving Sepsis Campaign, then, if/when it becomes clear that the patient’s ICU stay will be prolonged, shift to the more aggressive targets used in the Leuven studies.

The best synthesis of all the evidence came out this year in the July issue of Chest in the form of a review article and accompanying editorial. The authors seem to favor intensive glycemic control in critical illness provided the ICU is sufficiently organized and equipped to achieve recommended glycemic targets while minimizing hypoglycemia. They point out the ongoing NICE SUGAR trial which may give more definitive answers.

No comments: