Delirium risk profiles in a population-based study of United States older adults undergoing common noncardiac surgeries.

Background: Older adults often require surgical care and are at elevated risk of delirium. We explored delirium risk profiles across the population of U.S. older adults who underwent one of 10 common noncardiac surgeries.

Methods: We analyzed Health and Retirement Study (HRS) participants linked with Medicare billing data who underwent the following 10 noncardiac surgeries from 2000 to 2018 at age 65 or more: total knee arthroplasty (TKA), total hip arthroplasty (THA), spine surgery, cholecystectomy, colorectal surgery, hernia repair (ventral, umbilical, or incisional), endarterectomy, prostatectomy, transurethral resection of the prostate (TURP), and hysterectomy. Demographic and health covariates were obtained from the HRS dataset. Latent cognitive ability was calculated from cognitive testing, proxy reports, and demographics at the preoperative HRS interview. We compared standardized differences for delirium risk factors across the 10 surgeries and qualitatively clustered them into phenotypical subgroups.

Results: We analyzed 7424 older adults (mean age 76 ± 6 years, 45% male). Endarterectomy patients presented with the highest burden of nearly all health and cognitive factors, implying higher delirium risk (e.g., stroke, 22%; depressive symptoms, 30%; high school or less education, 73%; frailty, 42%; lowest latent cognitive ability). A second "general surgery" phenotype, including cholecystectomy, colorectal, and hernia surgery patients, experienced more frailty (29%-32%) and depressive symptoms (24%-26%), with moderate comorbidity burden. A third "pain" phenotype, which included TKA, THA, and spine surgery patients, commonly reported moderate or severe pain (47%-53%) and impairment in activities of daily living (ADL, 23%-30%), but fewer comorbid medical conditions. The remaining surgery types (hysterectomy, prostatectomy, TURP) were not phenotypically grouped and generally had lower risk features for delirium.

Conclusion: In an epidemiological cohort of US older adults, we identified clinically meaningful heterogeneity in delirium risk profiles across different surgical types, which may have implications for delirium risk stratification and delirium prevention or treatment.