Cognitive Deficits in Executive and Language Functions Predict Postoperative Delirium.

OBJECTIVES

Postoperative delirium (POD) remains the most common complication in older adults, with cognitive impairment being the main risk factor. Patients with mild cognitive impairment, in particular, have much to lose from delirium; despite this, their cognitive impairment might be clinically overlooked. Understanding which cognitive domains are particularly predictive in this regard may improve the sensitivity of preoperative testing and allow for a more targeted application of resource-intensive measures to prevent delirium in the perioperative period. The authors conducted this study with the aim of identifying the most indicative cognitive domains.

DESIGN

A secondary analysis of a randomized controlled trial.

SETTING

At a single center, the University Medical Centre Hamburg in Hamburg, Germany.

PARTICIPANTS

Patients ≥60 years without major neurocognitive disorders (dementia, Mini-Mental State Examination score ≤23) scheduled for cardiovascular surgery.

MEASUREMENTS AND MAIN

RESULTS

Preoperative neuropsychologic testing and delirium screening were performed twice daily until postoperative day 5. A multiple logistic regression model was applied to determine the predictive ability of test performances for the development of delirium.

RESULTS

A total of 541 patients were included in the analysis; the delirium rate was 15.6%. After controlling for confounders, only low performance within the Trail Making Test B/A (odds ratio [OR] = 1.32; 95% CI: 1.05-1.66) and letter fluency (OR = 0.66; 95% CI: 0.45-0.96) predicted a particularly high risk for delirium development. The discriminative ability of the final multiple logistic regression model to predict POD had an area under the curve of 0.786.

CONCLUSIONS

Impairment in the cognitive domains of executive function and language skills associated with memory, inhibition, and access speed seem to be particularly associated with the development of delirium after surgery in adults ≥65 years of age without apparent preoperative neurocognitive impairment.