Postoperative Delirium: Preventing It Before the First Confused Night
An 81-year-old is about to have his hip fixed. Everyone agrees delirium prophylaxis matters — the disagreement is whether a scheduled sedative, a scheduled antipsychotic, or simply doing the unglamorous non-drug things well is what actually prevents it in someone his age.
H.B., an 81-year-old man, lives alone, still drives to church twice a week, and fractured his hip stepping off a curb he'd navigated without incident for a decade — the kind of fall his daughter describes as "completely unlike him," a detail the team notes because unremarkable falls in otherwise-steady patients his age sometimes signal an evolving cognitive or sensory change nobody has caught yet. He is scheduled for hip fixation within the day, mildly hypoactive on baseline exam but oriented and conversant, and carries three of the strongest known risk factors for postoperative delirium at once: his age, a hip fracture requiring urgent surgery, and a Mini-Cog score suggesting mild, previously undiagnosed baseline cognitive impairment.
Postoperative delirium in a patient carrying this risk profile isn't a rare complication to watch for reactively — incidence in hip-fracture patients over 80 with baseline cognitive impairment runs high enough that prevention, not early detection, is the actual target, and the three plausible strategies rest on genuinely different evidence bases. Dexmedetomidine's alpha-2 agonism produces a sedation state closer to natural sleep architecture than benzodiazepine or high-dose opioid sedation, and Su's randomized trial found low-dose nocturnal infusion cut delirium roughly in half over the first seven days — in patients over 65 admitted to intensive care after non-cardiac surgery, which is the detail that has to be checked against H.B. rather than assumed for him, since a hip fixation does not automatically route an 81-year-old to an ICU bed. Prophylactic antipsychotics have not cleared the same bar: REDUCE randomized nearly 1,800 critically ill adults at high delirium risk to haloperidol or placebo and found no survival benefit and no reduction in delirium incidence, and Schrijver's trial in acutely hospitalized older patients reached the same place. The multicomponent non-pharmacologic bundle sits apart from both: it treats delirium as substantially environmental and physiologic rather than purely pharmacologic, and its own trial evidence is real, though it asks more of the ward's actual staffing and routine than either drug does.
His daughter, sitting in on the pre-operative discussion, mentions that he refused his hearing aids at the emergency department overnight because he "didn't want to lose them," a small detail that turns out to matter more than it sounds — uncorrected hearing loss is itself an independent, modifiable delirium risk factor, and a patient who can't hear the team clearly is more likely to become disoriented regardless of which drug is or isn't running. It's the kind of finding that never shows up in a trial's inclusion criteria but sits directly in the path of whatever strategy the team ultimately chooses.
Pre-operative planning, the afternoon before surgery
I'd start a low-dose dexmedetomidine infusion perioperatively — his alpha-2 sedation profile more closely resembles natural sleep than the benzodiazepine- or opioid-heavy sedation he'd otherwise get, and trials in high-risk elderly surgical patients like him have shown real delirium-incidence reduction on that basis, not just severity reduction.
I'd stay away from prophylactic antipsychotics specifically — worth naming since it's often reached for reflexively in patients like him. REDUCE randomized close to 1,800 high-risk patients to prophylactic haloperidol and found neither a survival benefit nor a reduction in delirium incidence; Schrijver's trial in acutely hospitalized older patients found the same. The literature isn't uniform — Wang's earlier trial in elderly post-operative patients did show a lower incidence — but the larger and more recent trials are the negative ones, and that asymmetry is the whole basis for not reaching for it here.
Before either drug, get the non-pharmacologic bundle right — his hearing aids and glasses back on him the moment he's out of the OR, early mobilization tomorrow rather than three days from now, protected sleep hours overnight, frequent reorientation, and no restraints or unnecessary lines. Inouye's Hospital Elder Life Program is the trial that established this, and it reduced delirium incidence on exactly these unglamorous levers — none of which carry drug risk in a man already facing a general anesthetic. And it speaks directly to his hearing aids, which is not a footnote to the drug question: sensory impairment was one of the targeted risk factors, and he is currently without correction he normally uses.
I'm not arguing against the dexmedetomidine — his risk profile is real enough that layering pharmacologic prophylaxis on top of the bundle is reasonable. I'm arguing the bundle isn't optional scaffolding around the drug decision; it's the foundation the drug decision sits on top of.
Agreed: low-dose dexmedetomidine infusion started perioperatively, layered on top of the full non-pharmacologic bundle (hearing aids and glasses restored immediately post-op, early mobilization begun the same evening, protected overnight sleep hours, frequent reorientation, no physical restraints). Hip fixation proceeded without complication.
He remained oriented through post-operative day two with no CAM-ICU-positive episodes recorded, though the team explicitly noted this as a single reassuring outcome, not evidence resolving which piece of the combined strategy carried the actual protective weight.