Strict glucose control may not reduce mortalities in the ICU

Recent findings challenge the current recommendations for glucose control in the critically ill.

Compared with usual care, strict glucose control may be associated with an increased risk for hypoglycemia in the critically ill, according to recent data published in JAMA.

Researchers from Dartmouth-Hitchcock Medical Center and other sites in New Hampshire and Vermont conducted a meta-analysis to examine the risks and benefits of tight glucose control vs. usual care in surgical and medical ICU patients.

Affects on mortality

The researchers collected data for 8,432 intensive care patients from 29 controlled, randomized studies that assessed strict glucose control (<150 mg/dL) vs. usual care.

The number of hospital mortalities did not differ significantly between tight control and usual care: 21.6% vs. 23.3% (RR=0.93; 95% CI, 0.85-1.03).

Similarly, the researchers reported no difference in mortality when they stratified strict control vs. usual care by glucose goal or ICU setting. Stratifying by glucose goal, they reported that patients with the very tight glucose control (≤110 mg/dL) had mortality rates of 23.2% in the strict control group vs. 25.2% in usual care (RR=0.90). Patients who achieved moderately tight control (<150 mg/dL) in the strict control group had a 17.3% mortality rate, and those in the usual care group had an 18% mortality rate (RR=0.99).

When stratified by ICU setting, surgical patients who received tight control had an 8.8% mortality rate vs. 10.8% in usual care (RR=0.88). Medical patients who underwent tight control had a 26.9% mortality rate, compared with a 29.7% mortality rate for those who underwent usual care (RR=0.92).

Increased risks

According to the researchers, patients who received strict glucose control had a lower risk for septicemia, compared with usual care (10.9% vs. 13.4%; RR=0.76; 95% CI, 0.59-0.97). Patients who underwent tight control had an almost fivefold increased risk of hypoglycemia, compared with those in usual care (13.7% vs. 2.5%; RR=5.13; 95% CI, 4.09-6.43).

In an accompanying editorial, Simon Finfer, MBBS, FJFICM, of the George Institute for International Health, and Anthony Delaney, MBBS, FJFICM, of the Royal North Shore Hospital in Sydney wrote, “There is no simple or clear answer to the complex problem of glycemic control in critically ill adults; at present, targeting tight glycemic control cannot be said to be either right or wrong.”

Based on their findings, the researchers suggested that the current recommendations for glucose control in critically ill patients be reevaluated until additional data are available. – by Stacey L. Adams

JAMA. 2008;300:933-944.

The physiological regulatory systems of healthy individuals are capable of accurately and precisely controlling blood sugar under almost all circumstances of daily life. Individuals with impaired counter-regulatory systems due to injury, acute illness or chronic illness (such as is the case with diabetes), lose their ability to precisely control blood glucose levels. Physicians attempting to restore tight control over blood sugar excursions in these instances have learned repeatedly that they are, unfortunately, not as good as Mother Nature at this task. Thus, it is not surprising to learn that, in this study, critically ill adults subjected to tight glucose control (≤110 mg/dL) had a fivefold increased risk of developing hypoglycemia (glucose ≤40 mg/dL). This result will come as no surprise to endocrinologists who take care of patients with diabetes. Nor will they be surprised by the fact that trying to tightly control blood sugar was no easy task, with 21% of the reported studies unable to achieve an average blood sugar within 5 mg/dL of their stated target. Somewhat surprisingly, there appeared to be no less hypoglycemia in studies where the subjects’ average blood sugar was ≤150 mg/dL.

Tight glucose control was also not associated with preventing the need for dialysis or with improving mortality (the primary outcome variable); although the study was not adequately powered to confidently detect small differences in mortality such as the 1.7% difference found between the subjects whose glucose was tightly controlled and those who received usual care. Perhaps this too should not have been unexpected, since the blood sugar of the usual care patients were not “uncontrolled.” In fact, the average blood sugar levels in the usual control group were about 160 mg/dL and, in almost all studies, the average was below the renal threshold for glucose.

Importantly, the study did demonstrate a 24% lower relative risk of septicemia (absolute risk difference = 2.5%), although this improvement was limited to patients in surgical ICUs and was not found in medical or in medical-surgical ICUs. Moreover, since no overall reduction in mortality was detected, the authors speculate that the septicemia reduction may have taken place primarily in less severe cases.

What’s a doctor to do with these results? Until much larger, adequately powered, blinded studies are performed, the answer, at least in the medical ICU setting, seems clear: “primum non nocere.” Severe hyperglycemia should be avoided, and judicious efforts should be taken to keep the majority of blood glucose values below the renal threshold. In surgical ICUs the operating principle is almost certainly the same: since there is significant heterogeneity among studies, mortality rate is not unambiguously improved by tight blood glucose control and the absolute difference in the risk of developing sepsis is relatively small — although, clearly, of clinical significance.

– Dennis M. Bier, MD

Endocrine Today Editorial Board member