Comparison Of Five Frailty Assessment Tools In Predicting Postoperative Adverse Outcomes In Elderly Patients With Colorectal Cancer

ObjectiveTo compare the preoperative frailty status,measurement consistency and prediction efficiency of postoperative adverse outcomes of elderly patients with colorectal cancer assessed by Frailty Phenotype(FP),FRAIL Scale(FS),Clinical Frailty Scale(CFS),Edmonton Frailty Scale(EFS)and Groningen Frailty Indicator(GFI),simultaneously considering its clinical applicability,which provides a reference for selecting appropriate frailty assessment tools for the target population before clinical surgery.MethodsDuring December 2020 to October 2021,a total of 207 elderly patients who underwent radical resection of colorectal cancer in the First Affiliated Hospital of Soochow University were selected by convenience sampling method;FP,FS,CFS,EFS and GFI were used to evaluate the frailty and Kappa was used to test the consistency in pairs;The subjects were followed up from operation to 3 months after discharge,and the adverse postoperative outcomes were recorded.Controlling covariates,a logistic regression analysis was used to explore the effects of different frailty evaluation tools on postoperative complications,unplanned readmission,postoperative functional decline or death(combined endpoint);Receiver operating characteristic(ROC)curve was used to describe the prediction efficiency of five frailty assessment tools for each adverse postoperative outcome,and to test whether there were differences in area under the ROC curve(AUC)between different frailty assessment tools;The database was established by Excel 2016,and statistical analysis was carried out by SPSS 21.0 and medcalc 20.0 software.Results1.The prevalence of preoperative frailty in elderly colorectal cancer measured by different frailty assessment tools were 19.8%(FP),11.6%(FS),22.2%(CFS),10.1%(EFS)and 33.3%(GFI),respectively;The results showed that the consistency fluctuation range between different tools was large.The kappa value was between 0.338 and 0.607(all P<0.001),and the consistency of FP and CFS was the highest.2.207 elderly patients with colorectal cancer were followed up from operation to 3months after discharge,and the subjects did not fall off during hospitalization,of which 57 cases had postoperative complications(27.5%);After discharge,they were followed up by telephone for 3 months,during which 6 cases(2.9%)were lost,and 201 cases were finally collected with complete data.Among the 201 subjects,9 died(4.5%),38 had postoperative functional decline(18.9%),and the incidence of postoperative functional decline or death(combined endpoint)was 23.4%;Among the 192 surviving subjects,22(11.5%)had unplanned readmission.3.Controlling the covariates of age,Charlson Comobidity Index(CCI),Body Mass Index(BMI),preoperative hypoproteinemia,preoperative anemia,American Society of anesthesiologists(ASA)classification,operation method and whether intraoperative blood transfusion,only the frailty defined by FP or CFS was an independent predictor of postoperative complications,with Odds Ratio(OR)values of 3.495 and 2.653 respectively,all P<0.05;Both CFS and FP had moderate predictive efficacy for the risk of postoperative complications.The AUC were 0.714(95%CI: 0.647～0.774)and 0.737(95%CI:0.671～0.795),respectively.There was no significant difference in AUC(P>0.05).The AUC of FS,EFS and GFI was 0.643(95%CI: 0.574～0.708),0.665(95%CI: 0.596～0.729)and0.635(95%CI: 0.565～0.700),respectively.4.Controlling the covariates of age,CCI,BMI,preoperative hypoproteinemia,preoperative anemia,ASA classification and operation method,frailty defined by FP or GFI was an independent predictor of functional decline or death(combined endpoint),with OR values of 3.643 and 4.938 respectively,all P<0.05;The predictive efficacy of FP,FS,CFS,EFS and GFI on postoperative functional decline or death was moderate,with AUC of 0.742(95%CI: 0.675～0.801),0.722(95%CI: 0.655～0.783),0.759(95%CI: 0.694～0.817),0.748(95%CI: 0.683～0.807)and 0.791(95%CI: 0.728～0.845)respectively;The AUC difference between the five tools was not statistically significant(all P>0.05).5.Controlling the covariate of pathological stage,the frailty defined by FP or CFS was an independent predictor of unplanned readmission,with OR values of 6.448 and 2.869,respectively,all P <0.05;The predictive efficacy of FP,FS,CFS,EFS and GFI for unplanned readmission in elderly patients with colorectal cancer was low,with AUC of 0.580(95%CI:0.507～0.651),0.543(95%CI: 0.470～0.615),0.625(95%CI: 0.553～0.694),0.559(95%CI:0.486～0.630)and 0.590(95%CI: 0.517～0.660),respectively,The AUC difference between the five tools was not statistically significant(all P>0.05).Conclusion1.The incidence of preoperative frailty in elderly colorectal patients assessed by different tools varies widely.Except for the high consistency of FP and CFS,there is only moderate to low consistency between other tools.2.The five frailty assessment tools all have moderate predictive value for postoperative functional decline or death in elderly patients with colorectal cancer,and there is no difference in predictive efficacy;Both FP and CFS have moderate predictive value for postoperative complications,and there is no difference in predictive efficacy.Both can also identify elderly patients with colorectal cancer surgery who have a higher risk of unplanned readmission,but their predictive ability is insufficient;FS,EFS and GFI could not effectively predict postoperative complications and unplanned readmission 3 months after discharge.3.In terms of the predictive efficacy of the five frailty assessment tools for the three postoperative adverse outcomes concerned in this study,FP and CFS have the best predictive efficacy.Considering the clinical applicability,CFS is more suitable for the evaluation of preoperative frailty in elderly patients undergoing colorectal cancer surgery.