Optimal Modified Frailty Index Cutoff in Older Gastrointestinal Cancer Patients

Abstract

The newly characterized modified frailty index (mFI) is a robust predictor of postoperative outcomes for surgical patients. The present study investigates the optimal cutoff for mFI specifically in older gastrointestinal (GI) cancer patients undergoing surgery. All patients more than 60 years old who underwent surgery for a GI malignancy (esophagus, stomach, colon, rectum, pancreas, liver, and bile duct) were identified in the 2005 to 2012 National Surgical Quality Improvement Program, Participant Use Data File (NSQIP PUF). Patients undergoing emergency procedures, of American Society of Anesthesiologists (ASA) five status, or diagnosed with preoperative sepsis were excluded. Logistic regression modeling and 10-fold cross validation were used to identify an optimal mFI cutoff. A total of 41,455 patients (mean age 72, 47.4% female) met the eligibility criteria. Among them, 19.0 per cent (n = 7891) developed a major postoperative complication and 2.8 per cent (n = 1150) died within 30 days. A random sampling by a cancer site was performed to create 90 per cent training and 10 per cent test sample datasets. Using 10-fold cross validation, logistical regression models evaluated the association between mFI and endpoints of 30-day mortality and major morbidity at various cutoffs. Optimal cutoffs for 30-day mortality and major morbidity were mFI ≥ 0.1 and ≥0.2, respectively. After adjusting for age, sex, ASA, albumin ≥3g/dl, and body mass index ≥ 30 kg/m2, mFI ≥ 0.1 was associated with increased mortality (odds ratio (OR) 1.49, 1.30–1.71 95% confidence interval (CI), P < 0.001) and mFI ≥ 0.2 was associated with increased morbidity (OR 1.52, 1.39–1.65 95% CI, P < 0.001). For older GI cancer patients, a very low mFI was a predictor of poor postoperative outcomes with an optimal cutoff of two or more mFI characteristics.