Malnutrition Linked to Higher Postoperative Delirium Odds

Evidence

Malnutrition and postoperative delirium. Older surgical patients classified as malnourished had more delirium—an acute disturbance of attention and awareness—during hospitalization than matched patients without malnutrition: 25.5% versus 11.2%. After statistical adjustment, malnutrition was associated with 2.36 times the odds of delirium, with a 95% confidence interval of 1.36–4.10. This is an odds comparison, not a 2.36-fold increase in probability.

The malnutrition and postoperative delirium study, published October 5, retrospectively examined 1,125 patients aged 65 or older undergoing elective or emergency surgery under general anesthesia at one institution. Researchers classified preoperative malnutrition using Global Leadership Initiative on Malnutrition criteria. Matching patients on measured characteristics produced a cohort of 165 patients with malnutrition and 330 without it; delirium occurred in 42 and 37 patients, respectively. Delirium was identified through routine clinical assessments or explicit physician diagnoses.

Predicting sitting during nusinersen treatment. The sitting-prediction study, published October 20, included 124 participants with spinal muscular atrophy who could not sit and received nusinersen. Participants had a median of seven assessments each; 62, or 50%, achieved sitting. Researchers randomly split participants 80/20 between model-training and testing datasets, using motor-function, respiratory and swallowing measures as candidate predictors.

The selected model combined multiple decision trees and had a testing concordance index of 0.74. This measures how well a model ranks time-to-sitting outcomes; it does not mean 74% prediction accuracy. Tube feeding at treatment initiation was associated with a predicted median time to sitting four months later than no tube feeding—not an observed causal delay.

Skin-barrier measurements during dupilumab treatment. The dupilumab skin-barrier study, published October 25, enrolled 24 Chinese patients with moderate-to-severe atopic dermatitis and 20 healthy volunteers. This prospective, non-randomized study followed patients receiving dupilumab every two weeks for 16 weeks. It was open-label, meaning treatment was known to participants and researchers.

Researchers measured transepidermal water loss—water escaping through the skin—alongside surface sampling with adhesive tape. Water loss measured after five tape-stripping procedures was lower at week 2 than before treatment and remained lower through week 16.

Analysis

Interpretation: These studies address different populations, measurements and questions, rather than corroborating a common biological mechanism. The malnutrition study provides the clearest quantified association. The sitting model offers exploratory prognosis within an already-treated population, not a test of nusinersen’s effectiveness. The skin study documents changes during dupilumab treatment, without establishing why those changes occurred.

Limitations

The malnutrition study’s retrospective, single-center design leaves room for unmeasured differences between groups despite matching. Cognitive and frailty assessments were sparse or inconsistent, and standardized muscle-mass measurements were unavailable. Routine documentation may also have missed delirium episodes.

For the sitting model, missing values were filled in before splitting data for evaluation. That creates a risk that information from testing data influenced model development, making performance look better than it would elsewhere. External validation remains necessary; feeding support may reflect underlying disease severity rather than cause later sitting.

The skin study was small, single-site and non-randomized. Changes in water loss do not by themselves establish clinical importance, and its descriptive molecular findings cannot establish a mechanism.

What to watch

Independent testing of the sitting model, with missing-data processing kept separate from the testing dataset, would show whether its reported discrimination holds up. A randomized patient comparison reporting the absolute water-loss change would strengthen assessment of the dupilumab findings.

For the lead finding, confidence would rise if a prospective surgical cohort retained a similar malnutrition–delirium association after standardized cognition, frailty and muscle-mass measurements and scheduled delirium assessments; substantial weakening after those adjustments would lower confidence.