The trusted source for
healthcare information and
Abstract & Commentary
This study was conducted to test the hypothesis that hyperglycemia or relative insulin deficiency (or both) predisposes to complications during critical illness. At ICU admission, 1548 consecutive patients who required mechanical ventilation and admission to a surgical ICU were randomly assigned to receive either conventional or intensive insulin therapy. In the intensive therapy group, a continuous infusion of insulin was started when the blood glucose level exceeded 110 mg/dL and then adjusted to maintain normoglycemia (80-100 mg/dL). In the conventional group, a continuous infusion of insulin was not started until the blood glucose level exceeded 215 mg/dL and then adjusted to maintain the value between 180 and 200 mg/dL. At ICU discharge, a conventional approach was adopted for both groups (blood glucose 180-200 mg/dL).
At study entry, 75% of the subjects had a blood glucose level at the upper limit of normal after an overnight fast (> 110 mg/dL), but only 12% had a level in the nonfasting diabetic range (> 200 mg/dL). A minority (13%) had a history of diabetes and 5% were receiving insulin. In the intensive-treatment group, almost all subjects received insulin compared to 39% in the conventional group. As anticipated, the mean morning blood glucose level was lower (103 ± 19 mg/dL) in the intensive insulin group compared to the conventional group (153 ± 33 mg/dL).
The study was prematurely terminated when interim analysis indicated conventional treatment was inferior. During the study, 4.6% of patients randomized to intensive treatment died compared to 8.0% in the conventional group (P < 0.04). The reduction in mortality occurred exclusively in patients who remained in the ICU for > 5 days (10.6% intensive insulin vs 20.2% conventional; P = 0.005). Patients randomized to intensive-insulin therapy also reduced overall in-hospital mortality by 34%, septicemia by 46%, acute renal failure requiring dialysis or hemofiltration by 41%, the median number of red-cell transfusions by 50%, and critical illness polyneuropathy by 44%. Those cases of critical care neuropathy that developed resolved more rapidly in the intensive insulin group. In both groups, there was a positive linear correlation between the risk of polyneuropathy and the mean blood glucose level. Further, markers of inflammation (C-reactive protein level, white-cell count, body temperature) were less often abnormal (P < 0.02) in the intensive insulin group. (Van den Berghe G, et al. Intensive insulin therapy in the surgical intensive care unit. N Engl J Med. 2001;345:1359-1367.)
The major finding of this study was the striking decrease in mortality and morbidity in patients randomized to receive strict glycemic control (blood glucose £ 110 mg/dL). Van den Berghe and colleagues hypothesize that effectiveness of this therapy resulted from its ability to decrease the risk of infection and associated complications. Markers of inflammation were less frequently abnormal in patients randomized to strict glycemic control. When antibiotics were indicated, they were used for a shorter interval. The greatest reduction in mortality involved deaths due to multiple organ failure with a proven septic focus. Both groups had a similar number of patients who remained in the ICU > 5 days (P = 0.1), and the reduction in mortality occurred exclusively in this long-stay cohort. Longer ICU stays are known to increase risk of nosocomial infection, and hyperglycemia would logically increase this risk.
Findings of this study suggest that critically ill patients may gain substantial benefit from metabolic therapy designed to normalize blood glucose levels. Hyperglycemia has known deleterious effects on macrophage and neutrophil function. Sepsis and related syndromes are the chief cause of multiple-organ failure and death in critically ill patients. Some benefits are less easily explained, eg, fewer transfusions and the lower incidence of acute renal failure and critical-illness polyneuropathy but could result from a quicker recovery with fewer complications.
The study was carefully designed and there was no evidence of confounding factors. Subjects were enrolled consecutively at admission to the ICU. Of those eligible, only 14 were excluded, 5 because they were enrolled in another trial and 9 because they had a do-not-resuscitate order. Group assignments were made with sealed envelopes, balanced by permuted blocks, and stratified according to the type of critical illness. The groups were similar at entry in regard to demographic variables, including presence of renal failure, preexisting diabetes, hyperglycemia, severity of illness (APACHE II), and intervention needs (TISS). There were no between group differences in nonprotein calorie (P = 0.2) or nitrogen (P = 0.3) intake. Insulin doses were adjusted according to a strict algorithm by a team of ICU nurses, assisted by a physician not involved in care of the patients. Strict glycemic control appears to be a preventive approach that merits consideration for use in critically ill patients.
Publication of this study took my colleagues and me by surprise, and the paper has generated a great deal of discussion. Although the rationale for tight glycemic control is clear, the strategy of routinely and aggressively managing all ICU patients to prevent hyperglycemia is something new for most clinicians. This was a well-done study. In-depth examination of Van den Berghe et al’s methods, data reporting, and interpretation fails to reveal major defects, and I think we must take it seriously. By show of hands at the journal club at which this paper was discussed, the majority of my faculty colleagues and our fellows indicated that they would modify their ICU practice as a result of its findings. This is remarkable for a group that is usually skeptical and tends not to jump at the latest "breakthrough" reported in the critical care literature.
Some caution is warranted if insulin drips are to be instituted across the board in ICU patients if their blood glucose levels exceed the modest threshold used in this study. Van den Berghe et al took care to monitor their patients’glucose levels carefully and to prevent episodes of serious hypoglycemia. If the same care is not used in everyday practice, morbidity and even mortality could actually be increased if severe and prolonged hypoglycemia were to be induced, despite any beneficial effects of higher insulin levels on inflammation and infection.
It would be reassuring if other groups confirmed these results, something that ought not be too difficult if the benefits of tight glucose control are as substantial as suggested by this study. Assessment of the frequency and severity of hypoglycemia when this regimen is used should be carried out as well.