Prevalence of Frailty and Reliability of Established Frailty Instruments in Adult Elective Colorectal Surgical Patients: a Prospective Cohort Study

Background: Frailty is associated with adverse outcome in older surgical patients but the strength of the association varies widely between studies. This may be due in part to the wide range of instruments used to assess frailty. Large population studies now demonstrate that frailty is prevalent in all adult age groups. However, there are limited data on the association between frailty and surgical outcome in younger patients. We studied the agreement between three established frailty instruments and their predictive value for postoperative complications in an adult colorectal surgery population. Methods: We studied patients aged 18 years or over presenting for elective colorectal surgery at a major teaching hospital. Preoperative frailty assessments were performed using the clinical frailty scale (CFS), accumulation deficit (AD) and frailty phenotype (FP) instruments. Data were collected on postoperative complications, length of stay, and readmission. The correlation and agreement between the frailty measures was examined in pairwise comparisons. Measures of association and prediction for postoperative complications were calculated and compared for the instruments. Results: Frailty scores were recorded in 200 patients (91 male), mean(range) age 57(18-92) years. The prevalence of pre-frailty was 52-67% and that of frailty 2-32% depending on the instrument used. Agreement between the instruments was poor, kappa 0.08 to 0.17 in pairwise comparisons. Outcome data were available on 160 patients. Only the frailty phenotype was significantly associated with adverse outcomes, odds ratio 6.1(1.5-24.5) for postoperative complications. The three frailty scoring instruments studies had good sensitivity (CFS-90%, AD-96%, FP-97%) but poor specificity (CFS-12%, AD-13%, FP-18%) for the prediction of postoperative complications. All three instruments were poorly predictive of adverse outcomes with likelihood ratios of CFS-1.03, AD-1.03 and FP-1.12. Conclusions: This study showed a significant prevalence of pre-frailty and frailty in adult colorectal surgical patients of all ages. There was poor agreement between three established frailty scoring instruments. Our data do not support the use of current frailty scoring instruments in all adult colorectal surgical patients. However, the significant prevalence of pre-frailty and frailty across all age groups of adult surgical patient justifies further research to refine frailty scoring in surgical patients.