Hyperglycemia is associated with increased mortality in ICUs, independent of the severity of illness, according to the largest and most definitive report on the subject to date. In a review of 216,775 consecutive patients who were admitted for the first time to ICUs at Veterans Affairs medical centers, the association between hyperglycemia and increased ICU mortality was strongest in patients with cardiovascular disorders, such as mycordial infarction, unstable angina, and stroke, and in those without diagnosed diabetes.
For cardiovascular disorders and many other diseases for which hyperglycemia was associated with increased ICU mortality, the risk of death increased in a stepwise fashion with increases in the level of mean blood glucose from 111-145 mg/dL to more than 300 mg/dL. Patients without diagnosed diabetes has an increased risk of death associated with hyperglycemia that ranged from 40% at the lowest level of hyperglycemia to a four-fold greater risk at the highest levels, while individuals with diagnosed diabetes did not have significantly increased risk of death unless their blood glucose level exceeded 146 mg/dL.
The study involved mostly men (97%) in 177 surgical, medical, and cardiac ICUs at 73 VA medical centers during 2002-2005. Two-thirds of the patients were older than 60 years of age, and 29% had diagnosed diabetes. Using a model that has been validated in determining the severity of illness; analyzed the likelihood of mortality associated with different glycemic levels (70-110 mg/dL, 111-145 mg/dL, 146-199 mg/Dl, 200-300 mg/Dl) in groups of patients in 78 diagnostic categories.
Even among those disorders where there was a significant relationship between hyperglycemia and mortality there was still variability in the magnitude of the mortality risk. For hyperglycemia in the lowest and highest ranges, patients with ischemic stroke were between 3 and 15 times more likely to die than were those with a normal mean blood glucose level (70-110mg/dL. But although hyperglycemia was significantly associated with ICU mortality in patient with pneumonia, the likelihood of death did not change in step with the severity of hyperglycemia.
Other diseases, such as chronic obstructive pulmonary disease and liver failure, showed no significant relationship between hyperglycemia and ICU mortality. Previous studies have reported hyperglycemia as a significant risk factor for mortality in the ICU in patients with an acute MI, coronary artery bypass graft, or stroke, and for patients in general staying in pediatric or adult ICUs. But many of these studies have been limited by inadequate adjustment for the severity of illness, small sample sizes, and measurement of blood glucose levels only at entry to the ICU. Randomized trials of critically ill patients in surgical and medical ICU have shown that tighter control of blood glucose level reduces morbidity and mortality. But a few recent trials involving patients with acute MI have reported mixed results, which have been attributed to insufficient power and the inability to establish glycemic control in intervention groups.
This led investigators to wonder if the benefits of tighter glycemic control are generalizable across all critically ill patients and all disease types.