Characteristics associated with effectiveness in postoperative delirium research: a systematic review of randomised...

Postoperative delirium remains prevalent despite extensive research through randomised trials aimed at reducing its incidence. Understanding trial characteristics associated with interventions' effectiveness facilitates data interpretation.

Trial characteristics were extracted from eligible trials identified through two systematic literature searches. Multivariable meta-regression was used to investigate trial characteristics associated with effectiveness estimated using odds ratios. Meta-analysis was used to investigate pooled effectiveness.

We identified 201 eligible trials. Compared with China, trials from the USA/Canada (ratio of odds ratio, 1.89; 95% confidence interval, 1.45-2.45) and Europe/Australia/New Zealand (1.67; 1.29-2.18) had an 89% and 67% higher odds ratio, respectively, suggesting reduced effectiveness. The effectiveness was enhanced when the incidence of postoperative delirium increased (0.85; 0.79-0.92, per 10% increase). Trials with concerns related to deviations from intended interventions reported increased effectiveness compared with those at low risk (0.69; 0.53-0.90). Compared with usual care, certain interventions appeared to have reduced the incidence of postoperative delirium in low-risk trials with low-to-moderate certainty of evidence. However, these findings should be considered inconclusive because of challenges in grouping heterogeneous interventions, the limited number of eligible trials, the prevalence of small-scale studies, and potential publication bias.

The effectiveness of postoperative delirium trials varied based on the region of trial origin, the incidence of delirium, and the risk of bias. The limitations caution against drawing definitive conclusions from different bodies of evidence. These findings highlight the imperative need to improve the quality of research on a global scale.

PROSPERO (CRD42023413984).