Early Prediction of Length of Stay After Pancreaticoduodenectomy

Background: Enhanced recovery after surgery (ERAS) is an evidence-based clinical pathway designed to standardize and optimize care. We studied the impact of ERAS and sought to identify the most important recommendations to predict shorter length of stay (LOS) after pancreaticoduodenectomy (PD).

Methods: We retrospectively reviewed all patients undergoing PD at our institution between January 2014 and June 2018. We compared clinicopathologic outcomes for patients before and after ERAS implementation. We defined "A-recommendations" as those that were graded "strong" and had "moderate" or "high" levels of evidence. We then compared outcomes of the ERAS group with adherence to "A-recommendations" and performed a subset analysis of "A-recommendations" over the first 72 h after surgery, which we termed "early factors".

Results: A total of 191 patients underwent PD during the study period. We excluded 87 patients who had minimally invasive PD (22), vascular reconstruction (53), or both (12). Of the 104 patients studied, 56 (54%) were pre-ERAS and 48 (46%) were ERAS. There were no differences in comorbidities or demographics between these groups, and morbidity, mortality, and readmission rates were also similar (P > 0.6). Median LOS was 3.5 d shorter in the ERAS group (7 versus 10.5 d, P < 0.001). Adherence to "A-recommendations" within ERAS was associated with a decreased LOS (r = -0.52 P = 0.0001). Patients with >5 "early factors" had a median LOS of 6 d, whereas patients with <5 "early factors" had a median LOS of 9 d (P = 0.008).

Conclusions: ERAS is an effective protocol that standardizes care and reduces LOS after PD. Implementation of ERAS resulted in a 3.5-day reduction in our LOS with no change in morbidity, mortality, or readmissions. Adherence to ERAS protocol "A-recommendations" and ≥5 "early factors" may be predictive of shortened LOS.