Impaired Cognitive Domains in Surgical Patients: A Systematic Review and Meta-Analysis

BACKGROUND: Cognitive impairment is a highly prevalent but frequently overlooked issue among surgical patients preoperatively. This systematic review and meta-analysis aimed to (1) determine the perioperative prevalence of impaired cognitive domains in surgical patients, (2) explore perioperative changes in the different domains, and (3) examine postoperative outcomes associated with preoperatively impaired cognitive domains. METHODS: Five electronic databases were searched from inception to March 19, 2024. Inclusion criteria were (1) surgical patients ≥18 years of age; (2) preoperative cognitive assessments using a neuropsychological battery; (3) reported the prevalence of impairment in specific cognitive domains or changes perioperatively; and (4) sample size of ≥100 surgical patients. The exclusion criteria included studies involving neurological surgery; cross-sectional, case-control, and case series studies; non-English articles; and studies with overlapping data. RESULTS: In total, of the 12,082 articles identified from 5 databases, 21 studies (5725 patients, 11 non-cardiac surgery studies, and 10 cardiac surgery studies) were included. Among the 6 cognitive domains assessed preoperatively, the pooled prevalence of impairment was highest in executive function (18%; 95% CI, 13%–24%), visuospatial function (16%; 95% CI, 6%–26%), and attention/working memory/processing speed (14%; 95% CI, 9%–18%). Perceptual-motor control (13%; 95% CI, 9%–36%), language (13%; 95% CI, 8%–17%), and learning/memory (12%; 95% CI, 8%–16%) had lower pooled prevalence. The cognitive domains that were assessed postoperatively showed a high prevalence of impairment at 1 week, with 35% (95% CI, 4%–66%) in attention/working memory/processing speed, 34% (95% CI, 16%–51%) in executive function, and 28% (95% CI, 16%–40%) in learning/memory. The pooled prevalence subsequently decreased within 3 months to 16% (95% CI, 3%–35%) in attention/working memory/processing speed, 15% (95% CI, 6%–24%) in executive function, and 12% (95% CI, −2% to 25%) in learning/memory. CONCLUSIONS: The prevalence of preoperatively impaired cognitive domains was the highest in executive function, followed by visuospatial function and attention/working memory/processing speed. Identifying commonly impaired cognitive domains may help optimize cognitive assessments in the perioperative setting. Further research is needed to clarify the clinical utility of assessing specific cognitive domains in surgical populations to improve postoperative outcomes and reduce cognitive deterioration.