Application of an adapted prognostic score in a case series of caustic gastrointestinal injuries treated in Spanish hospital emergency departments: a multicenter study

Objective.

Given the variability in injury severity following caustic ingestion, we evaluated the diagnostic performance of an adapted version of the Bergamo Score (BG-CSa) for identifying potentially more severe episodes.

Methods.

Retrospective, multicenter study of patients aged $ 15 years treated for caustic ingestion at 8 Spanish emergency departments (2015–2021). The ability of the BG-CSa (excluding 2 components of the original score and replacing neutrophilia with leukocytosis) to predict severe endoscopic lesions (Zargar classification $ IIb) was assessed, as was its concordance with the care provided.

Results.

A total of 199 patients were included (median age, 47 years; 52.8% women). A BG-CSa $ 4 showed good discriminative ability (AUC-ROC = 0.830), with a sensitivity rate of 88.1% and a negative predictive value of 95.1% for Zargar $ IIb endoscopic lesions. None of the 5 false-negative patients (esophagogastroduodenoscopy [EGD] performed with severe findings despite BG-CSa < 4) died. Patients categorized as true positives according to the BGCSa had worse clinical outcomes, with a mortality rate of 27%. EGD use in patients with BG-CSa < 4, observed in 52.3% of cases, was more common than omission of EGD in patients with BG-CSa $ 4, which occurred in only 5.5%. Dyspnea at presentation was the only variable that differed significantly from true negatives (40% vs. 4%; P = .026).

Conclusions.

Although the BG-CSa showed a high negative predictive value, the occurrence of false negatives precludes recommending its use in isolation to rule out severe lesions or avoid diagnostic testing.