Studies have shown that controlling blood glucose closely and carefully improves survival and morbidity among intensive care patients, but the ideal target remains controversial.
The controversy was evident at the Endocrine Society's annual meeting in Toronto in June. At a featured debate during the meeting, attendees surveyed were evenly split between those who said the goal should be to keep glucose levels below 110 mg/dL and those who argued for a higher target (for example, 120 to 150 mg/dL).
Those who advocate for a lower target—and the intensive insulin therapy it requires—contend that it prevents serious complications such as bloodstream infections and organ loss, shortens ICU stays and cuts mortality for some patients, including those in surgical ICUs.
Supporters of a higher target worry about the risks of causing hypoglycemia in critically ill or septic patients, the difficulty—and sometimes impossibility—of having adequate nurses or hospital systems available to achieve low targets safely and the confounding factor of parenteral feeding, with its resulting high glucose and lipid levels.
Support for a lower target
“Why should [the target glucose level] be below 110 in the ICU?” Greet Van den Berghe, MD, PhD, asked during the debate. “Because that's what the two randomized, controlled trials have shown, and nobody has shown otherwise.”
There is no doubt that hyperglycemia can be present in critical illness and that its severity is associated with adverse outcomes in many patients. Therefore, Dr. Van den Berghe said, preventing it—by whatever means—is an important way to avoid complications that increase morbidity and mortality in the ICU.
Dr. Van den Berghe's team from the Catholic University of Leuven in Belgium has been studying intensive insulin therapy in the ICU for most of the decade. Her much-quoted study of surgical ICU patients on mechanical ventilation, which appeared in the Nov. 8, 2001, New England Journal of Medicine, found that strictly controlling blood glucose levels to at or below 110 mg/dL reduced “in-house mortality by 34 percent, bloodstream infections by 46 percent, acute renal failure requiring dialysis or hemofiltration by 41 percent, the median number of red-cell transfusions by 50%, and critical-illness polyneuropathy by 44 percent.” This benefit, Dr. Van den Berghe said, was mostly attributable to the large impact of glucose control among patients treated in the ICU for several days.
Dr. Van den Berghe and her colleagues followed that study with one of medical ICU patients, which was published in the Feb. 2, 2006, New England Journal of Medicine. In this population, intensive insulin therapy did not significantly reduce overall mortality. Dr. Van den Berghe pointed out, however, that mortality was reduced in the target population of patients treated for at least three days and that the intention-to-treat analysis showed significantly reduced morbidity. Overall, she added, the intensive insulin regimen particularly prevented kidney injury and permitted early weaning from mechanical ventilation, early discharge from the ICU and shorter hospital stays.
Potentially worrisome, however, and in need of further study, was that among patients who spent fewer than three days in the ICU, mortality appeared higher in the group randomly assigned to intensive insulin therapy than in those who received insulin only when their blood glucose levels exceeded 215 mg/dL (conventional care). The question of whether brief therapy causes harm is now being examined in a large, multinational study in Australia, New Zealand and Canada, Dr. Van den Berghe said.
Moreover, she said, it's possible that the only subgroup of patients who may not be served by a target of 110 mg/dL are those who had a history of diabetes before their ICU stay. These patients, she speculated, may not be used to living with such a low blood sugar level, so it may be better to keep them at the levels they are accustomed to, even if those levels are higher.
John M. Miles, MD, professor of medicine and endocrinology at the Mayo Clinic in Rochester, Minn., argued the other side of the debate, contending that the target of 110 mg/dL is not based on adequate scientific evidence. Controlling blood glucose levels to 120 to 150 mg/dL may be just as beneficial as maintaining glucose levels below 110 mg/dL, he said, with the higher target potentially reducing the risk of causing hypoglycemia among this already very sick population. The issue, he said, isn't whether to control glucose, but how ambitious the target should be.
The goal of 110 mg/dL or lower, which is now included in guidelines from the American College of Endocrinology and the American Diabetes Association, is problematic, Dr. Miles said, because it is based on findings in patients receiving parenteral nutrition. In general, he said, parenteral feeding provides critically ill patients with too many calories, including too much glucose and too many lipids. Intensive insulin therapy may only be offsetting problems caused by parenteral feeding, he said, and the lower glucose target therefore may not be relevant to all ICU populations, let alone other hospitalized patients.
“Because parenteral nutrition has considerable potential for exerting adverse effects, the optimal [blood glucose] level in other clinical settings is not known,” he said. “Therefore, the target should default to a less ambitious range, maybe 120 to 150 or 130 to 160.”
Dr. Van den Berghe's group has just started studying the effect of early parenteral nutrition on the outcome of intensive glucose control in the ICU. “In my view, parenteral nutrition has nothing to do with the impact of tight glucose control,” she said. She pointed out that parenteral nutrition may lead to hyperglycemia when given without controlling glucose levels, thereby reducing its benefit, but added that the question still needs to be resolved. In terms of whether the 110 mg/dL target is arbitrary, “It's the ‘normal’ value, and we thought it would be a reasonable target for our studies,” she said. “But we actually don't know if this level is the most optimal. No other levels have been studied.”