Intensive insulin therapy improves outcomes in children in the ICU

In pediatric patients with hyperglycemia in the ICU, intensive age-adjusted insulin therapy improved short-term outcomes and reduced mortality, according to findings from a recent study.

“Targeting blood glucose concentrations to an age-adjusted normal fasting concentration with insulin infusion throughout intensive care improved morbidity and reduced mortality, despite a substantial risk for biochemical hyperglycemia,” the researchers wrote in The Lancet.

The researchers conducted a prospective, randomized, controlled study to investigate the effect of targeting age-adjusted normoglycemia with insulin fusion. They enrolled 317 infants aged ≤1 year and 383 children aged ≥1 year who were admitted to the University Hospital of Leuven, in Belgium.

Researchers randomly assigned patients to blood glucose concentrations of 2.8 mmol/L to 4.4 mmol/L for infants and 3.9 mmol/L to 5.6 mmol/L for children with insulin infusion throughout their stay or to insulin infusion to prevent concentrations from exceeding 11.9 mmol/L.

Outcomes improved

Mean blood glucose concentrations were lower among patients who received intensive therapy compared with those who received conventional therapy, according to the study. For infants, mean blood glucose concentrations were 4.8 mmol/L in the intensive treatment group and 6.4 mmol/L in the conventional treatment group (P<.0001). In children, mean concentrations were 5.3 mmol/L in the intensive group and 8.2 mmol/L in the conventional group (P<.0001).

Hypoglycemia (blood glucose ≤2.2 mmol/L) occurred in 25% of patients in the intensive group (P<.0001) and in 1% of patients in the conventional group; hypoglycemia (blood glucose ≤1.7 mmol/L) occurred in 5% of patients in the intensive group compared with the conventional group (1%; P=.001), according to the study.

The duration of stay in the pediatric ICU was 5.51 days (95% CI, 4.65-6.37) in the intensive group and 6.15 days (95% CI, 5.25-7.05) in the conventional group (P=.017).

The inflammatory response was attenuated at day 5, as indicated by lower C-reactive protein in the intensive group (–9.75 mg/L; 95% CI, –19.93 to 0.43) than in the conventional group (8.97 mg/L; 95% CI, –0.9 to 18.84), according to the study.

In the intensive group, 38% of patients had extended stay in the pediatric ICU; 47% of patients had extended stay in the conventional group (P=.013). Three percent of patients died in the intensive group and 6% died in the conventional group, according to the study.

“For future studies, an accurate continuous blood glucose sensor for use in the pediatric ICU, which was not available when we undertook our study, would be preferable to keep the risk for hyperglycemia to a minimum,” the researchers wrote. In addition they are conducting a follow-up study to investigate the effect of intensive age-adjusted insulin therapy on long-term survival, morbidity and neurocognitive development. – by Christen Haigh

Lancet. 2009; 373:547-556.

This study is a pediatric equivalent of the recently published meta-analysis on the clinical outcomes influenced by attempting to normalize blood glucose in adult intensive care units (JAMA. 2008;300:933-944). This analysis showed that tight blood glucose control in adult ICU settings did not improve mortality (the primary outcome variable), but increased the risk for hypoglycemia. The current study confirms the increased hypoglycemia risk for intensive blood glucose control in the pediatric ICU setting. Because mortality is low in pediatric ICU settings, mortality was not the primary clinical endpoint variable. Instead, the investigators used length of stay as an index of morbidity. Using this measure, intensive glucose control reduced the average length of stay slightly (two-thirds of a day), but the fact that the 95% CI around the duration of pediatric ICU stay overlapped substantially in both groups of patients reduces one's confidence in the robustness of the conclusion that intensive glucose control has a truly meaningful effect on length of stay in the pediatric ICU for those subjects with short pediatric ICU stays. This observation is supported by the fact that the median lengths of stay in the pediatric ICU were the same in both groups (three days).

On the other hand, only 38% of the subjects whose blood glucose was controlled remained in the pediatric ICU longer than three days, whereas 47% of the children whose blood glucose values were not aggressively lowered stayed in the pediatric ICU longer than three days. This difference is, obviously, an important one in practice. Although not a primary endpoint variable, mortality was monitored as a secondary safety endpoint. Only 2.6% of the patients in the intensively treated group died versus 5.7% of the patients in the conventionally treated group (P=0.038), a clearly significant and important difference.

However, those that developed hypoglycemia had a death risk for 6.5% compared with the mortality risk for 3.8% in those who did not become hypoglycemic. This difference, nonetheless, was not statistically significant and, when hypoglycemic deaths were corrected for the other clinical variables that are independently associated with mortality, hypoglycemia per se, were not an independent risk factor for death. Taken together, these data provide support for the judicious control of blood glucose in the pediatric ICU setting. However, because approximately half of the patients were less than 1 year of age, because 75% were post-cardiac surgical patients, and because the long-term neurodevelopmental consequences of hypoglycemia were not measured, it is important to interpret these results very cautiously, especially in the case of generalizing them to older children and children who are admitted to the pediatric ICU for illness and injury other than that sustained by cardiac surgery. Additionally, this is a single pediatric study, and it is important to recognize that 'positive' studies are also among the studies evaluated in the adult ICU meta-analysis described above that concluded that intensive blood glucose control did not improve mortality.

– Dennis M. Bier, MD

Endocrine Today Editorial Board member

This study is an addition to the body of evidence that suggests hyperglycemia is associated with poor outcomes in a variety of clinical settings and that intensive glycemic control improves clinical outcome. It is also another reminder that tight glycemic control is associated with significant risk for hypoglycemia, and we need to work more judiciously to put systems in place in our hospitals to both improve the hyperglycemia while avoiding hypoglycemia. It is also possible that the commonly recommended target range of 80 mg/dL to 110 mg/dL for intensive control of glycemia is too low and needs to be modified.

– Etie Moghissi, MD, FACP, FACE

Associate Clinical Professor of Medicine
University of California, Los Angeles