Tight glycemic control in critically ill kids: benefits, risks still unclear

Intensive insulin therapy for critically ill adults was rapidly adopted as standard care after 2001 when an apparent benefit was established after cardiac surgery, then medical ICU patients. Eleven years later, after a wave of minor harm signals, the NICE-SUGAR study confirmed for most intensivists that the excess hypoglycemia from intensive glucose control was potentially lethal in adults, and that the risks outweighed the benefits for most patients. The patients thought most likely to benefit from intensive glucose control -- those having major surgery or acute strokes -- also seem to be harmed by the strategy, especially if they are in ICUs. The pendulum of standard care swung back to a relaxed approach to glucose control (a goal of <140-180 mg/dL is often used in U.S. ICUs today).

In children, owing to a series of smaller conflicting studies, the debate as to the relative benefits and harms of intensive glucose control during critical illness rages on.

What They Did

At 13 ICUs in England, investigators randomized 1,369 critically ill nondiabetic children (newborns to 16 year olds, mechanically ventilated, on pressors) to receive tight glucose control (72-126 mg/dL) or liberal blood sugar control (< 216 mg/dL triggered an insulin drip, which was turned off when glucose fell <180 mg/dL). 60% of the children had undergone cardiac surgery. A computer titrated all the insulin according to a centralized algorithm.

The primary outcome was ventilator-free days at 30 days. Secondary outcomes included complications and intensity of therapy in the ICU (days of pressors, presence of infections, organ failures, hypoglycemia, etc.) and length of stay in the ICU and hospital.

What They Found

There was no difference in the primary outcome of days alive and free of mechanical ventilation at 30 days (23.6 vs. 23.2), nor in 30-day survival (95% in both groups) or 12-month survival. Nor were there any significant differences in almost any other measured outcome (ICU/hospital stays, transfusions, days on pressors, infections, etc.), although the tight glycemic control group had numerically better numbers in most of these.

However, kids receiving tight glycemic control had less need for renal replacement therapy (9% vs. 13.5% in the liberal-glycemic-control group), almost 30 fewer patients, a number needed to treat of about 20.

16% of patients getting intensive insulin therapy had moderate to severe hypoglycemia, 4 times the rate of those on a liberal strategy. Only 3% of tightly-controlled patients had seizures requiring medication, and only 2% of those liberally controlled. There were 70 cases of severe hypoglycemia (blood glucose < 36 mg/dL) in the tight-control children, and only 11 in the liberally-controlled kids. Most kids having severe hypoglycemia had multiple episodes.

The prespecified subgroup of children in the ICU for reasons other than cardiac surgery (40% of the total) did have a shorter total average hospital stay, by almost 2 weeks, at the end of a year of follow-up.

What It Means

For children, the jury is still out on just how tightly glucose should be controlled during critical illness:

• Vlasselaers et al found intensive insulin therapy reduced PICU stay and mortality in 700 critically ill children (Lancet 2009). Although intensively-treated kids had more hypoglycemia, after 3 years of followup they did not have measurable neurocognitive developmental delays, as compared to the liberally-treated children.

• Agus et al found no benefit of intensive glucose control in kids after cardiac surgery (NEJM 2012).

So, to all the kids out there reading PulmCCM, but not undergoing heart surgery: Would you take a 1 in 20 chance of getting to avoid dialysis and go home from the hospital sooner (with tight glucose control), if it costs you a 1 in 10 increased chance of having multiple episodes of severe hypoglycemia? What do you mean, you're not sure? How are you going to get into Harvard if you don't keep up with your statistics homework? And your mother and I also want to know, how is it that you can get straight As in social studies, but you keep failing your spontaneous breathing trials?

Intensive glucose control is an intuitively irresistible idea whose actual utility, when subjected to the scientific method, has been continually shrinking with almost every new published study. If we accept the findings of this trial and its forebears, tight glucose control is now a niche therapy -- a reasonable option for critically ill children not undergoing heart surgery. It might very well prevent renal failure and have other unmeasured benefits that lead to faster recovery after critical illness in kids. Or, the benefits seen here may be due to lack of blinding (kids getting intensive insulin get better care in other unmeasured ways) or (for renal failure) chance positive results after measurement of multiple outcomes. Also, for me, multiple episodes of severe hypoglycemia would be tough to watch kids go through as the cost of doing business to achieve these benefits (although I don't take care of kids). And we can expect severe hypoglycemia to occur more often in community practice, outside the confines of a clinical trial.

You can almost hear the clicking of the keyboard keys tapping out the next grant proposal with this excellent but unsatisfying manuscript's equivocal parting words: As with any trial, further studies would be required to assess whether these findings apply to routine clinical practice in other settings. I couldn't have said it better myself.

Duncan Macrae et al. A Randomized Trial of Hyperglycemic Control in Pediatric Intensive Care. N Engl J Med 2014; 370:107-118. DOI: 10.1056/NEJMoa1302564