Reframing Surgical Risk in Frailty: Causal Mediation Analysis of Mortality and Readmission after Pancreatectomy.
Key claim: After pancreatectomy, frailty-associated 30-day mortality is largely mediated by postoperative complications (~81%), whereas frailty-associated readmission is driven mainly by recovery vulnerability beyond complications.
Abstract
BACKGROUND: Frailty predicts poor outcomes after pancreatectomy, but whether excess mortality and readmission reflect more postoperative complications, worse rescue after complications, or broader recovery vulnerability remains unclear. STUDY DESIGN: We performed a retrospective study of the American College of Surgeons NSQIP Pancreatectomy database (2014 to 2023) including 71,104 patients undergoing pancreatic resection. Frailty was defined as modified 5-item frailty index score 2 or more. Multivariable logistic regression evaluated associations between frailty and 30-day mortality and readmission. Causal mediation analysis quantified total, direct, and complication-mediated indirect effects. Interaction modeling compared failure-to-rescue and nonprecedented deaths by frailty status. RESULTS: Of 71,104 patients, 15,779 (22.2%) were frail. Frailty was independently associated with 30-day mortality (adjusted odds ratio 1.29, 95% CI 1.13 to 1.46; p < 0.001) and readmission (adjusted odds ratio 1.11, 95% CI 1.06 to 1.16; p < 0.001). For mortality, the total effect of frailty was significant (risk ratio [RR] 1.32, 95% CI 1.23 to 1.40; p < 0.001), the indirect effect through complications was significant (RR 1.25, 95% CI 1.22 to 1.27; p < 0.001), and the direct effect was not (RR 1.07, 95% CI 0.99 to 1.14; p = 0.400); complications mediated 80.72% of the frailty-mortality association. The largest contributors were unplanned reintubation, bleeding requiring transfusion, septic shock, acute kidney injury requiring dialysis, MI, and organ or space surgical site infection. Major complications increased mortality similarly in frail and nonfrail patients, without meaningful frailty-related differences in failure to rescue. For readmission, complications explained only 17% of the frailty association. CONCLUSIONS: After pancreatectomy, frailty-associated mortality is largely explained by postoperative complications, whereas frailty-associated readmission is driven predominantly by vulnerability beyond complications. These findings support complication prevention, prehabilitation, and intensified postdischarge follow-up for frail patients.
Why this matters
Frailty scores are routinely used as preoperative risk stratifiers, but the causal pathway from a high score to a bad outcome has been largely a black box. This mediation analysis separates two mechanisms — complication occurrence (dominant for mortality) versus post-discharge recovery vulnerability (dominant for readmission) — which matters for how a frailty signal should be acted on at the measurement layer, since the same score points to different downstream vulnerabilities depending on the endpoint.