Concordance between Observers for the Diagnosis and Endoscopic Classification of Esophageal Lesions Due to Caustic Ingestion in Children

Abstract

Guillermo Yanowsky-Reyes, Yamir F Monrroy-Martin, Sergio A Trujillo-Ponce, Jaime Orozco-Perez, Rafael Santana-Ortiz, Fernando Duque-Zepeda, J Jesus Perez-Molina, Christian L Breuillet-Barrera, L Gerardo Padilla-Meza

Introduction: The accidental ingestion of caustic materials among children continues to be a public health problem, with an incidence in international reports of 15.8 / 100, 000 per year. In our environment, at the Fray Antonio Alcalde Civil Hospital, this incidence has an average of 2.5 cases per month, around 30 cases per year. Objective. To evaluate the agreement between observers for the diagnosis and endoscopic classification of esophageal lesions due to the ingestion of caustics in children.

Material and methods: A historical cohort study was conducted in 2,190 days with patients diagnosed with esophageal stenosis secondary to caustic ingestion during a period between 2010 and 2016. Photographic records of patients who met the inclusion criteria were recovered. Diagnostic endoscopy at the time of caustic ingestion. These reports and the fixed images of the esophagus were classified according to Zagar’s Scoring Systems for Caustic Injuries. The following variables were collected: sex, age in months, date of ingestion of caustic substance, type of ingested caustic substance, presence of complications in the acute stage, degree of initial injury according to Zagar’s classification. We reviewed the results of the esophagogram and the diagnostic endoscopy at the time of caustic ingestion and the first revision endoscopy. Later these images were evaluated by the same endoscopist and by two others, the three reports were classified according to Cohen’s Kappa coefficient, to compare the agreement between endoscopists, as well as the Kappa coefficient of Feiss to compare the general concordance etre the three endoscopists, both coefficients with a confidence interval of 95% (Kappa +/- 1.96 standard error). Statistical analysis was performed through SPSS version 19.

Results: A total of 43 patients with diagnosis of caustic ingestion at the Civil Hospital of Guadalajara “Fray Antonio Alcalde” were reviewed. The characteristics of the patients in relation to sex, 13 (30%) were female and 30 (70%) of the male sex The average age was 34.5 months, with a range of 12 to 96 months. 42 patients for caustic soda (98%) and one patient with burn by muriatic acid (2%). First-time endoscopy was performed within the first 18-24 hrs after the injury, they did not present any complications in the acute stage, and they had the diagnosis of caustic burn according to the Scoring Systems for Caustic Lesions of Zagar. , identifying 2 patients with grade I burns (4.6%), 8 patients with grade IIa burn (18.6%), 14 patients with grade IIb burn (32.5%), 17 patients with burn IIIa (39.5%) and 2 patients with burn IIIb (4.6%). A total of 63 images were obtained, which were identified according to the predominant pattern of lesion due to caustic ingestion based on Zargar’s classification. Each image was displayed in a time of 20-30 seconds, with the total evaluation ending in an average of 30 minutes. The variability of concordance between pairs of endoscopists was compared when integrating the 3 endoscopists, a Kappa Feiss Coefficient of 0.3948 was obtained, with a standard error of 0.0402 and a confidence interval of 95% from 0.3160 to 0.4736. With this tool the kappa coefficient observed for the degree of esophageal burn injuries according to the Zargar Classification was 0.6585 (good concordance strength) for grade I, 0.4393 (moderate concordance force) for grade II a, 0.3000 (force low concordance) for grade II b, 0.4674 (moderate agreement strength) for grade III a, and 0.0950 (agreement strength without agreement) for grade III b.

Discussion and conclusion: The overall agreement (Kappa Feiss) regarding the decision of the presence / absence of mild / severe esophageal lesions (grade I, II a vs. grade II b, III a, III b) had a moderate reliability, but poor for leveling fine of each of the grades, with the strength of negligible agreement (kappa 30 0.0950) for grade IIIb, low (kappa 0.3000) for grade IIb, moderate for grade IIa (kappa 0.4393) and IIIa (kappa 0.4674) and good for grade I (kappa 0.6585). We believe that endoscopy is a mandatory and effective technique that should be performed to prevent unnecessary hospitalization and medication, as well as to plan initial treatment and to predict patients who are at risk of developing esophageal stenosis and / or obstruction of the exit gastric and can be performed without complications in expert hands.

PDF