Could the Pittsburgh Severity Score guide the treatment of esophageal perforation? Experience of a single referral center.

Background: Esophageal perforation (EP) is characterized by high morbidity and mortality. The Pittsburgh Severity Score (PSS) is a scoring system based on clinical factors at the time of EP presentation, intended to guide treatment. The aim of the study is to verify PSS usefulness in stratifying EP severity and in guiding clinical decisions.
Methods: All patients referred to our unit for EP between January 2005 and January 2020 were enrolled. Patients were stratified according to their PSS into three groups (PSS ≤ 2, 3-5, and >5): the postoperative outcomes were compared. The predictive value of the PSS was evaluated by simple linear and logistic regression for the following outcomes: need for surgery, complications, in-hospital mortality, intensive care unit (ICU) and hospital stay, time to refeeding, and need for reintervention.
Results: Seventy-three patients were referred for EP (male/female, 46/27). Perforations were more frequently iatrogenic (41.1%) or spontaneous (38.3%). The median PSS was 4 (interquartile range, 2-6). Surgery was required in 60.3% of cases. Pittsburgh Severity Score was associated with ICU admission, hospital stay, need for surgery and reintervention, postperforation complications and mortality. After regression analysis, PSS was significantly predictive of postperforation complications (p < 0.01), in-hospital mortality (p = 0.01), ICU admission (p < 0.01), need for surgical treatment (p < 0.01), and need for reintervention (p = 0.02).
Conclusion: Pittsburgh Severity Score is useful in stratifying patients in risk groups with different morbidity and mortality. It is also useful in guiding the therapeutic conduct, selecting patients for nonoperative management. Prospective studies are needed to confirm the role of the PSS in the treatment of esophageal perforation.
Level of Evidence: Management, Therapeutic/Care; level IV.
(Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.)