Tight Glucose Control Nearly Won This Argument Once Before It Was Reversed
A 63-year-old man with diabetes is about to start his CABG, and his glucose target for the case is still unsettled. The trial that once made tight control the standard for cardiac surgery patients specifically was later contradicted by a larger one — and the room hasn't fully agreed on which one should govern an anesthetized patient who can't report his own hypoglycemia.
Vincent D., a 63-year-old man, has run a small print shop since taking it over from his father, and has managed type 2 diabetes for the past eleven years, currently on metformin and a basal insulin regimen with an A1c of 7.4% — reasonably controlled, not tightly so. His admission fasting glucose was 142mg/dL and he has no documented history of severe hypoglycemia on his current regimen, which is worth establishing before the discussion starts, because the argument that follows turns almost entirely on how much hypoglycemia risk he can be exposed to and on how anyone would detect it.
Progressive angina over the past two months, culminating in a positive stress test and catheterization showing severe three-vessel disease, brought him to CABG today — his first cardiac surgery and his first general anesthetic of any kind. His diabetes puts intraoperative glucose management squarely in front of the team before the case begins, and the target range remains a genuinely contested question in cardiac surgery specifically, not settled the way it might appear from either trial alone. Van den Berghe's Leuven surgical ICU trial, which first popularized a target of 80 to 110mg/dL, was conducted in a population dominated by cardiac surgical patients much like him and found real reductions in complications. NICE-SUGAR, larger and later, found the opposite — increased mortality on the same tight target, driven substantially by severe hypoglycemia. He resembles the population of the trial that found benefit and the population of the trial that found harm about equally well, which is why the deciding argument is unlikely to come from either one: under general anesthesia he cannot report the tremor, sweating or confusion that would normally be the first warning that his glucose is falling.
Setting the glucose target before the first incision
I'd target tight control, 80 to 110, through the case. The original surgical ICU trial that established this range was conducted in a population overwhelmingly like his — cardiac surgery patients — and found real reductions in infection and other complications. That's the trial most directly analogous to what we're actually doing today.
I'd push back toward a moderate target, 140 to 180, instead. NICE-SUGAR was a much larger, multi-center trial conducted after that original one, and it found tight control significantly increased mortality — substantially driven by severe hypoglycemia. That's not a smaller or weaker trial we're setting aside; it's the larger, more rigorous one.
And I'd specifically flag why hypoglycemia risk matters even more for him than for a typical NICE-SUGAR patient — he's about to be under general anesthesia, unable to report the tremor, sweating, or confusion that would normally be the first warning sign of a falling glucose. We'd be relying entirely on scheduled lab draws to catch it, which is a real detection gap tight control doesn't have in an awake patient.
That's a genuinely important distinction I hadn't weighted heavily enough — you're right that an anesthetized patient loses the early-warning system a tight target depends on to be caught safely. I don't have a way to argue around that detection gap specifically.
I'll go with the moderate target. I still think the infection-reduction signal from the original trial is real, and I'd want us disciplined about getting him to a moderate range promptly rather than letting hyperglycemia drift, but not at the cost of a hypoglycemia risk we can't monitor for clinically while he's under.
Agreed: moderate glycemic target of 140-180mg/dL maintained through the case via insulin infusion, with frequent glucose checks and explicit discipline against letting hyperglycemia drift toward the upper end of that range.
Not agreed as a fully settled question: how much of NICE-SUGAR's hypoglycemia signal reflects the target range itself versus differences in feeding protocols and glucose monitoring technique between the two trials, a distinction the surgeon raised but that neither voice treated as definitively resolved. The moderate target was adopted as the safer choice for today's intraoperative setting specifically, not as a claim that the underlying scientific debate between the two trials is fully closed.