The prevalence of unrecognized cognitive impairment in geriatric surgical patients: A systematic review and meta‐analysis

Abstract

AbstractBackgroundPre‐existing cognitive impairment is emerging as a predictor of poor postoperative outcomes in older surgical patients (Guenther, 2020). Due to a substantial under‐diagnosis of cognitive impairment in the elderly, the prevalence of undiagnosed cognitive impairment has not been well defined in those undergoing surgery (Elman, 2018). The objective of this systematic review and meta‐analysis is to determine the pooled prevalence of unrecognized cognitive impairment in the elderly undergoing elective noncardiac surgery.MethodAn expert librarian conducted the literature search, which included MEDLINE (Ovid), PubMed (non‐MEDLINE records only), Embase, Cochrane Central, Cochrane Database of Systematic Reviews, PsycINFO, and Emcare Nursing for relevant articles from 1946 to April 2021. Inclusion criteria were (1) patients ≥ 60 years old undergoing elective non‐cardiac surgeries; (2) preoperative cognitive impairment assessed by validated cognitive screening tests; (3) published in English language. Descriptive analysis was conducted for the cognitive impairment group. Data were extracted from each study and prevalence of cognitive impairment with 95% confidence interval (CI) was calculated for each study. We used the random effect model to calculate the pooled prevalence value with 95% CI. Comprehensive meta‐analysis software was used for statistical analysis (Borenstein, 2013).ResultOf 8,895 citations, 20 studies were included consisting of 4,159 patients undergoing a variety of elective non‐cardiac surgeries. All but one study were prospective cohort studies, with the exception being a cross‐sectional study. The mean age among patients with unrecognized cognitive impairment was 73.1 ± 7.2 years, and 43.6% were male. The pooled prevalence of cognitive impairment was 40.2% (95% CI: 31.0%, 50.1%; P=0.005; predictive interval: ‐1.47 to 2.27 (Table 1). The forest plot displays non‐overlapping CI's, indicating high heterogeneity within the data (I2: 97%). Influential analysis of any study showed that removal of each study did not significantly alter the pooled prevalence result. Meta‐regression analysis based on age, gender, and body mass index (BMI) did not change the final inference of our results.ConclusionThe prevalence of unrecognized cognitive impairment is very high at 40.2% in geriatric patients undergoing elective non‐cardiac surgery.