Injection sclerotherapy-induced esophageal strictures: Risk factors and prognosis

Injection sclerotherapy (IS) has become an effective modality for the treatment of bleeding esophageal varices. Despite improvements in equipment, sclerosant solutions and operator technique, injection sclerotherapy-induced esophageal strictures (ISES) remain a significant cause of patient morbidity. To analyze the risk factors and prognosis of ISES, the records of 117 patients who underwent IS over a 6-year period at a single teaching institute were reviewed. The predictive value of multiple risk factors including the patient's age, Child's risk classification, previous bleeding episodes, etiology of varices, cumulative amount of sclerosant used, and the number of IS treatments were determined using ANOVA. A P value of < 0.05 was considered significant. In all cases, a free-hand injection technique, flexible endoscopes and sodium morrhuate were used. During a mean follow-up period of 228 days (1-1,469 days), 41 patients (35%) died and 24 patients (20.5%) developed symptomatic strictures. The cumulative amount of sclerosant used (81.4 ± 9.5 ml) and the number of IS treatments (6.5 ± 0.7) required in the stricture group was significantly greater than in the nonstricture group (49.1 ± 2.7 and 4.0 ± 0.3, respectively). The risk of stricture formation did not correlate with the volume of sclerosant injected per treatment, cause of varices, number of previous bleeds, or Child's hepatic risk class. A mean of 3.6 ± 4.5 dilations was required for treatment of established strictures and 18 patients (75%) required r 4 dilations. One esophageal perforation occurred following dilation. Mortality correlated with hepatic risk class as 30/41 (73%) of deaths occurred in Child's C patients. It is concluded that the risk of ISES can be predicted from the number of IS treatments and the cumulative volume of sclerosant required to obliterate esophageal varices. Management of strictures usually requires several dilations and can be performed with low risk to the patient. The high mortality rate observed in these patients is not a direct consequence of IS but appears to reflect the severity of their underlying liver disease.