Impact of Frailty Upon Surgical Decision-Making for Left-Sided Colon Cancer

Background: Frailty is characterized by reduced physiologic reserve, and for patients with colon cancer, frailty is associated with increased morbidity after resection. One commonly cited reason for performing an end colostomy vs a primary anastomosis in left-sided colon cancer is the belief that frail patients do not have the physiologic reserve to withstand the morbidity associated with an anastomotic leak. We explored the impact of frailty on the type of operation performed in patients with left-sided colon cancer. Methods: We queried the American College of Surgeons National Surgical Quality Improvement Program for patients with colon cancer who underwent a left-sided colectomy from 2016 to 2018. Patients were categorized using the modified 5-item frailty index. Multivariate regression was used to identify independent predictors of complications and type of operation performed. Results: Of 17,461 patients, 20.7% were considered frail. Frail patients received an end colostomy more often than nonfrail patients (11.3% vs 9.6%, P=0.01). On multivariate analysis, frailty was a significant predictor for total medical complications (odds ratio [OR] 1.45, 95% CI 1.29-1.63) and readmission (OR 1.53, 95% CI 1.32-1.77) but was not independently associated with organ space surgical site infections or reoperation. Frailty was independently associated with receiving an end colostomy vs a primary anastomosis (OR 1.23, 95% CI 1.06-1.44), but an end colostomy did not decrease the risk of reoperation or organ space surgical site infections. Conclusion: Frail patients with left-sided colon cancer are more likely to receive an end colostomy, but an end colostomy does not lower the risk of reoperation or organ space surgical site infections. Based on these results, frailty alone should not prompt the decision to perform an end colostomy, but further studies are needed to guide surgical decision-making in this understudied population.


INTRODUCTION
Frailty is characterized by decreased physiologic reserve, decreased ability to maintain homeostasis, and decreased tolerance to stressors such as surgery. 1,2 The Canadian Study of Health and Aging developed a 70-item frailty index to predict mortality and morbidity in patients. 3 However, this index was difficult to apply in clinical research settings and was therefore adapted using the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) to create the modified frailty index (mFI). 4 The mFI was originally based on 16 variables mapped to 11 factors but was abbreviated to a validated 5-factor modified *Dr Riker is now affiliated with South Florida Surgical Oncology, GenesisCare USA, Fort Myers, FL. frailty index (mFI-5) because variables were removed from the NSQIP user files. 5 The mFI has been shown to predict morbidity and mortality after surgery across multiple surgical subspecialties. [6][7][8][9][10][11][12][13][14][15][16][17] In elective surgery for gastrointestinal cancer, stratifying at 2 or more mFI-11 characteristics has been found to optimally predict increased risk of postoperative morbidity and mortality. 18 Frailty is common among the elderly and has emerged as an important risk factor for poor outcomes after surgery. In the United States, 53 million people are over the age of 65 years and account for more than 40% of all inpatient surgical procedures. 19,20 In patients undergoing colorectal surgery for cancer, frailty is an independent predictor of 30-day postoperative outcomes and readmission. 14,15 Yet little is known about the impact of frailty on surgical decisionmaking in left-sided colon cancer. This study explored the

METHODS
Institutional review board approval was obtained for this retrospective analysis of the ACS NSQIP participant user files from 2016 to 2018. Current Procedural Terminology and International Statistical Classification of Diseases and Related Health Problems, 10th Revision, codes were used to identify patients undergoing colectomy for left-sided colon cancer. Patients with a diagnosis of rectal cancer and those undergoing a diverting ileostomy procedure were excluded.
The mFI-5 was used to classify patients into frail and nonfrail groups (Table 1). Each factor is assigned 1 point, and the frailty score is the total score divided by 5. Significant frailty is defined as mFI-5 ࣙ0.4 (2 or more positive factors).
The primary aim of this study was to explore the impact of frailty on the type of operation performed (primary anastomosis or end colostomy) in patients with left-sided colon cancer who underwent a left-sided colectomy. The secondary outcome of interest was to identify the impact of frailty on 30-day postoperative outcomes: wound and medical complications, reoperation, and hospital readmission. Wound complications were defined as superficial incisional surgical site infection (SSI), deep incisional SSI, organ space SSI, or wound disruption. Medical complications were pneumonia, unplanned intubation, pulmonary embolism, failure to wean off ventilator >48 hours, progressive renal insufficiency, acute renal failure, urinary tract infection, stroke/cerebrovascular accident, cardiac arrest requiring cardiopulmonary resuscitation, myocardial infarction, transfusion, deep vein thrombosis requiring therapy, and sepsis. Anastomotic leak is not a defined outcome variable in the main ACS NSQIP data user file. Consequently, we used surrogate measures to quantify anastomotic leak: organ space SSI and reoperation.
Descriptive statistics are presented to show the distribution of each outcome of interest. Pearson chi-squared and 2-sided Fisher exact tests were used to analyze differences in categorical variables between the frail and nonfrail groups, and t test was used to analyze differences in continuous variables. Univariate analysis was conducted between each outcome of interest and patient characteristic. Outcomes of interest were wound complications, medical complications, readmission, and reoperation. For the multivariate analysis, patient characteristics were age, body mass index (BMI), comorbidities, and frailty. Perioperative risk factors were sepsis within 48 hours prior to surgery, transfusion of ࣙ1 unit of whole/packed red blood cells within 72 hours prior to surgery, and preoperative serum albumin level <3.5 mg/dL. Emergency surgery and the type of procedure performed (primary anastomosis or end colostomy) were also included in the multivariate analysis. Multivariable logistic regression models were fit to predict each outcome of interest. Variables considered for selection into each model were based on clinical relevance and univariate significance. All analyses were performed using SPSS Statistics, v. 25 (IBM Corporation). An alpha of 0.05 was the threshold for statistical significance.

RESULTS
A total of 17,461 patients were included in the analysis, of whom 20.7% had an mFI-5 score ࣙ0.4 (frail) and 79.3% had an mFI-5 score ࣘ0.2 (nonfrail). Frail patients were older, had a higher BMI, and tended to have more comorbidities than nonfrail patients ( Table 2). Smoking and disseminated cancer were more prevalent in the nonfrail group.
Frail patients were more often received as transfers from outside hospital facilities and were more likely to have preoperative hypoalbuminemia, to require preoperative blood transfusion, to require emergency surgery, and to receive an end colostomy instead of a primary anastomosis compared to nonfrail patients (Table 3). Mean hospital stay was longer for frail patients than nonfrail patients (7.7 vs 6.1 days, respectively, P<0.01), and frail patients were more likely to be discharged to skilled nursing or rehabilitation facilities than nonfrail patients (16.4% vs 6.7%, respectively, P<0.01).
Multivariate analysis showed that frailty was independently associated with receiving an end colostomy vs a primary anastomosis (Table 6).
In the multivariate analysis, frailty was associated with a 45% greater risk for total medical complications (odds ratio

DISCUSSION
This analysis demonstrates that frail patients undergoing surgery for left-sided colon cancer had worse outcomes than nonfrail patients, with a greater risk for medical complications and readmission. Interestingly, frail patients were more likely to receive an end colostomy than nonfrail patients. In the frail population, receiving an end colostomy did not decrease the risk for organ space SSI, reoperation, or readmission compared to frail patients with a primary anastomosis. This finding suggests that frailty alone should not prompt the decision to perform an end colostomy. A primary anastomosis may be considered in the frail patient who does not have any other significant risk factors for anastomotic leak. Our analysis additionally showed that increasing frailty is associated with increasing rates of postoperative complications, longer hospital stays, and a greater percentage of patients requiring skilled nursing facilities at discharge, in concordance with other published studies. [14][15][16][17] An estimated 10% to 14% of colorectal cancer surgeries result in a permanent ostomy. 21 Patients with an ostomy report worse quality of life and social well-being. 21 Ostomyrelated problems include depression, dissatisfaction with appearance and sexual activity, gas, constipation, difficulty with clothing and travel, fatigue, and worry about ostomy noises. Ostomy-related postoperative complications, including peristomal dermatitis, parastomal hernia, stenosis, and prolapse, occur in 21% to 70% of cases. 22 A commonly cited reason for performing an end colostomy vs a primary anastomosis is the belief that frail patients do not have the physiologic reserve to withstand the morbidity associated with an anastomotic leak. Older age, higher ASA physical status classification scores, and early (0 to 3 days) anastomotic leaks have been found to be reliable predictors of failure to rescue anastomotic leak after colectomy for colon cancer. 23,24 The mFI-5 is a poor predictor of failure to rescue, 23 but frail patients have been found to be more vulnerable to failure to rescue once a leak occurs. 25    While these concerns are important, the data presented here suggest that frail patients do not have a higher rate of reoperation when controlling for other factors and may not need to receive end colostomies more frequently than nonfrail patients simply on the basis of frailty. Indeed, an end colostomy may subject frail patients to other complications, such as hernia and prolapse that would also be poorly tolerated in the setting of their decreased functional reserve. It is unlikely that all end colostomies can be avoided, as common emergency surgery indications to perform an end colostomy include perforated diverticulitis with fecal peritonitis, sepsis requiring pressor support, and large bowel obstruction. Our multivariate analysis showed that the following patient factors should also be taken into consideration when deciding whether to perform an end colostomy: smoking status, dialysis dependence, and preoperative serum albumin levels.
A 2022 ACS NSQIP study showed an increased risk of anastomotic leak in frail patients undergoing elective colectomy for colon cancer. 25 Methodologic differences between our investigations likely account for this difference in findings. Specifically, we focused our investigation on leftsided colon cancers to focus on colocolonic and colorectal anastomoses, while the other study included all sites of colon cancer. A single-center Japanese study published in 2020 concurred with our findings and showed no association between frailty and anastomotic leak. 26 While we were not able to specifically examine anastomotic leak with this dataset, our study used the related surrogate markers of organ space SSI and reoperation to help answer the question of whether frail patients undergoing colon surgery would benefit from an end colostomy vs primary anastomosis. We believe this study can help inform surgical decision-making for frail patients undergoing colectomy for left-sided colon cancer. However, as with other retrospective studies, additional study in the form of prospective randomized trials can help confirm the role that frailty may play in the surgical decision of end colostomy vs primary anastomosis in left-sided colon cancer. Our study has limitations. First, we used surrogate measures to quantify anastomotic leak rates. Second, the association between procedure choice and cancer stage could not be elicited, as cancer stage is not a variable in the ACS NSQIP data user file. The frail and nonfrail groups may have had differences in cancer staging, and staging differences may have been present in the frail patient population that received an end colostomy vs primary anastomosis. Additionally, patients who receive an end colostomy usually do so for circumstances that increase the risk for anastomotic leak (ie, fecal peritonitis, perioperative sepsis, large bowel obstruction); therefore, the increased risk for medical complications and readmission among frail patients with an end colostomy may be related to other factors not accounted for in this study and not to the end colostomy itself. Finally, any database study is limited by coding and recording errors that may be present, as well as the retrospective nature of the study.

CONCLUSION
Frail patients with left-sided colon cancer are more likely to receive an end colostomy, yet an end colostomy was not associated with a lower risk of reoperation or organ space SSI. Frailty alone should not prompt the decision to perform an end colostomy in left-sided colon cancer. Further studies are needed to guide surgical decision-making in this understudied population.