Steroids for the treatment of corrosive esophageal injury: a statistical analysis of past studies.

Caustic esophageal injury causes substantial morbidity and mortality. However, the use of corticosteroids to treat this problem has been evaluated in a limited number of studies because adequate sample size is difficult to obtain. We analyzed 361 subjects with corrosive esophageal injury derived from 10 retrospective and three prospective publications. We divided cases into those treated with corticosteroids and antibiotics (T) and those that received neither modality (NT) based on inclusion and exclusion criteria. Forty-one percent of NT cases developed esophageal stricture and 19% of T cases developed this complication (P less than .01). There were no reported strictures among 72 first-degree esophageal burns (combined T and NT cases). The T group contained 54 strictures among 228 patients (24%) with either second- or third-degree burns. The NT group of 25 patients with the same burn severity suffered 13 strictures (52%) (P less than .01). Reports of death and gastrointestinal hemorrhage did not increase among steroid-treated patients. We do not recommend corticosteroid therapy for first-degree esophageal injuries. However, this therapy may be useful in preventing strictures among patients with second- or third-degree corrosive esophageal burns.