The Better Regimen, and a Scheduled Window It Wasn't Studied Around
Basal-bolus wins the general inpatient argument decisively. The real question is narrower: does a scheduled surgical NPO window, with its own specific timing uncertainty, call for a different answer — or just a more careful version of the same one?
Frank D., a 63-year-old man, has coached his grandson's Little League team for the past three springs and is hoping to be back on his feet, cane-free, before the season starts again in a few months — the reason he finally agreed to the total knee replacement his orthopedic surgeon has been recommending for two years. He has had type 2 diabetes for twelve years, well controlled on metformin and a modest basal insulin dose, with an A1c of 6.8% at his preoperative visit. His surgery is scheduled for early tomorrow morning, and the team is finalizing his perioperative glycemic management plan the evening before admission.
The general inpatient evidence on this question is fairly settled: scheduled basal-bolus insulin has repeatedly outperformed sliding-scale-only coverage on both glycemic control and complication rates in hospitalized non-ICU patients, dating back to the RABBIT-2 trial and confirmed in the literature since. What that evidence doesn't directly speak to is his specific situation — a defined, scheduled NPO window around a single surgical event, where his usual basal insulin dose, calculated for a normal eating day, could plausibly run him low if his operating room start time shifts later than planned and he ends up NPO longer than expected.
The better regimen on paper, and a specific reason to worry about tomorrow
I'd keep him on basal-bolus insulin, with a reduced basal dose on surgery day rather than switching to sliding-scale for the perioperative window. RABBIT-2 and the trials since have consistently shown basal-bolus outperforming sliding-scale-only in hospitalized patients — I don't see a good reason to default to a demonstrably weaker regimen just because he has a scheduled procedure.
The trial evidence you're citing wasn't built around a scheduled NPO window with an uncertain OR start time. If his case gets pushed and he's NPO six or eight hours longer than planned, a basal dose calculated for a normal eating day is a real hypoglycemia risk — not a rare edge case, something I've seen happen on delayed first cases more than once.
I'm not arguing sliding-scale is generally better — I'm arguing this one specific window has a specific hazard the general inpatient data doesn't model, and simplicity during exactly that window has real safety value.
I don't think this actually has to be a choice between the two. Reduce his basal dose by about a third the night before, hold mealtime insulin entirely while he's NPO, and resume his normal regimen once he's eating again after surgery. That captures basal-bolus's real outcome advantage while directly addressing the specific hypoglycemia risk an OR delay would create.
This isn't splitting the difference for its own sake — a protocolized dose reduction is a more precise fix for the actual problem than abandoning the better regimen entirely would be. If his case does get delayed, we check a glucose before he goes back and adjust from there.
Basal insulin reduced by one-third for the night before surgery, mealtime insulin held during the NPO window, and a point-of-care glucose check ordered before OR transfer regardless of timing. All three voices agreed this addressed the actual risk directly rather than trading one general regimen for another.
Not agreed: how much further to reduce the basal dose if his OR start time is delayed beyond four hours past the scheduled time. The anesthesiologist wanted a pre-specified further reduction built in now; the endocrinologist preferred to decide in real time based on the actual pre-OR glucose check rather than pre-committing to a second adjustment that might not be needed. Left for the morning of surgery to resolve.