Chasing a Normal Blood Sugar After a Stroke, and Why That Instinct Lost Its Trial
Hyperglycemia after a stroke looks like an obviously fixable problem — until the trial built to fix it found nothing to show for the effort except more dangerously low blood sugars, and the team has to decide how tightly to actually control a number that looked worth controlling.
R.P., a 69-year-old man who spent most of his career running a small auto-repair shop before handing it to his son two years ago, still comes in most mornings "just to make sure nobody's slacking," his son says with an eye-roll that has clearly played out before. He has type 2 diabetes managed on metformin and glipizide, with an A1c of 8.4% on his last check six months ago — adequately controlled by his own account, though his son mentions he's been "bad about the diet lately." He developed left-sided weakness and slurred speech at 8:20am while having coffee at the shop, and was brought in by his son within thirty minutes.
His NIHSS on arrival is 8, and imaging confirms a small-to-moderate right MCA territory infarct without hemorrhage; he was outside the thrombolysis window by the time of clear symptom onset and is being managed medically with close monitoring for early deterioration. His admission fingerstick glucose is 268mg/dL, consistent with his known diabetes and further elevated, likely, by the acute physiologic stress of the stroke itself. The question in front of the team is how tightly to control that number over the next several days. Hyperglycemia has a real, mechanistically plausible relationship to worse ischemic outcomes — observational cohorts have linked it to larger final infarct volumes, plausibly through favoring anaerobic metabolism in tissue that's already oxygen-starved — but SHINE, the trial built specifically to test whether correcting it aggressively actually helps, randomized patients exactly like him to intensive insulin-infusion control (target 80-130mg/dL) against standard sliding-scale management (target under 180mg/dL) and found no improvement in 90-day functional outcome from the intensive strategy, with significantly more episodes of severe hypoglycemia in that arm — a real, measurable harm traded for a benefit the trial never found.
On admission, deciding how tightly to chase the glucose
His glucose is 268 and he's had a stroke — those two facts sitting together makes me want to start an insulin infusion and bring that number down toward normal, not just watch it. Hyperglycemia pushes ischemic tissue toward anaerobic metabolism, and cohort studies have tied it to bigger final infarcts. That's a mechanism I'd rather correct than leave alone in a patient who's already had part of his brain go without oxygen.
That's exactly the hypothesis SHINE was built to test, and it came back neutral. Intensive control, target 80-130 via insulin infusion, against standard sliding-scale management under 180 — no improvement in 90-day functional outcome for the intensive arm, and significantly more severe hypoglycemia. We'd be trading a real, measured harm for a benefit the trial specifically looked for and didn't find.
I'd run standard sliding-scale, target 140-180, and leave the infusion pump in the supply closet.
You're right that the mechanism sounds compelling, and I don't think it's wrong exactly — I just think the causal arrow runs partly the other direction than the intuition suggests. A bigger stroke produces a bigger physiologic stress response, and that stress response is what's driving some of the glucose number, not purely the reverse. Hyperglycemia is partly a marker of how sick he already is, not a fully independent lever we can pull to make him less sick.
Which is a different route to the same place the intensivist landed: standard sliding-scale, target 140-180, avoid both extremes. I'd just add that we shouldn't be surprised SHINE came back neutral — the biology gives a real reason to expect that result, not just the trial's own numbers.
Agreed: standard sliding-scale insulin, target 140-180mg/dL, avoiding both severe hyperglycemia and hypoglycemia, rather than starting an intensive IV insulin infusion. Home metformin and glipizide held during the acute stay, with diabetes management and adherence addressed as part of discharge planning.
Not agreed, and named explicitly rather than smoothed over: how to manage a genuinely refractory hyperglycemia case, well above 300mg/dL despite sliding-scale dosing, which sits outside SHINE's clean binary framing of "intensive versus standard" and would need its own judgment call if it arose. Not this patient's problem tonight, but flagged as a real gap in what the trial data actually covers.