Frailty and Operative versus Nonoperative Mortality in Common Emergency General Surgery Conditions
Author(s):
Caleb W. Curry; Marisa R. Imbroane; Andrew Tran; Esther Tseng; Christopher Towe; Jeffrey Claridge; Vanessa Ho
Background:
In the older adult, development of abdominal sepsis may be catastrophic. Frailty increases perioperative risk, but it is unknown if there is a frailty level for which nonoperative management would be favored over operative management for common emergency general surgery conditions.
Hypothesis:
The effect of frailty on operative and nonoperative mortality will differ by disease.
Methods:
We identified patients aged 65 and older with appendicitis, diverticulitis, cholecystitis, peptic ulcer, or ischemic bowel in January to September 2017 using the National Readmissions Database. Operative and nonoperative management were identified based on presence or absence of a major abdominal surgery on index admission. Time to death up to 90 days post-discharge was the primary outcome. Frailty quintiles (1-5; 5 indicating most frailty) were calculated using a deficit accumulation score derived from 38 possible diagnoses. Cox proportional hazards analysis, stratified by disease type, estimated how frailty affected time to death after operative and nonoperative management (reference=nonoperative quintile 1), adjusted for age, sex, and shock.
Results:
We included 304,573 patients (56% female, median age 77 years [IQR 71-84]). Mortality was highest for ischemic bowel (24%) and lowest for appendicitis (1%) (Figure 1). Mortality hazard tended to increase over frailty quintiles for each disease except ischemic bowel. The effect of frailty on mortality for operative and nonoperative patients differed by disease. In appendicitis and cholecystitis, operatively managed patients had a lower mortality hazard at all frailty levels. For diverticulitis, peptic ulcer disease, and ischemic bowel, nonoperatively managed patients had a lower mortality hazard over most frailty levels.
Conclusions:
Frailty is an important prognosticator, but its magnitude of impact on survival for operatively and nonoperatively managed patients differed by disease. In particular, operative management favored survival over all frailty quintiles in appendicitis and cholecystitis, suggesting frailty alone may not be a strong enough indicator to decline surgical intervention in those disease processes.