The obesity paradox for postoperative delirium exists only in patients who are otherwise well

Design

Retrospective cohort with prespecified interaction analysis, propensity score matching and restricted cubic splines

Population

4320 patients aged 65 or over having non-cardiac, non-neurological surgery under general anaesthesia, 2015-2024

Primary outcome

Postoperative delirium within 7 days on the Confusion Assessment Method

Effect

Delirium 14.5 per cent; overweight aOR 0.81 (0.68-0.96) and class 1 obesity 0.76 (0.60-0.97) overall, but 0.69 and 0.65 in low comorbidity burden and no difference in high burden (interaction P = 0.040)

Four thousand three hundred and twenty patients aged 65 or over who had non-cardiac, non-neurological surgery under general anaesthesia between 2015 and 2024 were assessed daily for seven days with the Confusion Assessment Method. Postoperative delirium occurred in 625, 14.5 per cent.

The headline association looks like the familiar obesity paradox. Against normal weight, adjusted odds ratios were 0.81 (95 per cent CI 0.68 to 0.96) for overweight, 0.76 (0.60 to 0.97) for class 1 obesity, and 1.02 (0.78 to 1.33) for class 2 obesity or above. But the prespecified analysis was an interaction, and it was significant (P = 0.040). In patients with a Charlson Comorbidity Index of 0 to 2, overweight and class 1 obesity were clearly protective (0.69, 0.53 to 0.90 and 0.65, 0.46 to 0.92). In patients with an index of 3 or more, no body mass index category differed from normal weight at all. Propensity-matched estimates agreed.

What this changes is the preoperative risk conversation. A well-preserved older patient carrying some extra weight is at lower delirium risk than their thin equivalent; a multimorbid patient carrying the same extra weight is not, and reading their body mass index as reassurance is the error the interaction identifies. It also fits the usual explanation for the obesity paradox - that body mass index is a poor proxy for nutritional reserve once illness is present - and that explanation is easier to act on than the paradox itself. Take the comorbidity count first, and read weight in that light.

• Calculate a comorbidity index before interpreting body mass index as a delirium risk factor.
• Do not treat overweight as protective in a multimorbid older patient; the effect vanished in that group.
• Class 2 obesity or above showed no benefit in any group.
• Assess delirium daily with a validated tool rather than relying on nursing report.
• This is a retrospective single-centre cohort, so the size of the interaction may not transfer.

The statistics, in plain English

The interaction term is the finding, and it is the right way to test this question: rather than reporting subgroups and inviting the reader to compare them, it formally asks whether the two variables modify each other. A P of 0.040 is a modest signal, so the interaction is present but not strongly established. Odds ratios of 0.65 to 0.81 in a retrospective cohort cannot establish that weight protects against delirium; body mass index is standing in for nutritional reserve, frailty and much else that was not measured.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.