Systematic review and meta-regressions: management of eosinophilic esophagitis requires histologic assessment

Treatment endpoints in eosinophilic esophagitis (EoE) are response of eosinophilic inflammation and of symptoms. Steroids and diet therapy are effective in inducing histologic response in EoE, but there may be poor correlation between histologic and symptomatic response. Despite this, we find that in clinical practice symptoms are commonly used to guide management without assessing histologic response. We hypothesized that symptom response alone is not reliable in assessing response to therapy and is confounded by endoscopic dilation. We conducted a systematic review and meta-regressions to estimate the association of histologic and symptomatic response, stratified by whether concurrent dilation was permitted. We performed a systematic search of PubMed, EMBASE, and Web of Science for studies describing both histologic and symptomatic responses to dilation, steroid, and diet therapies. We abstracted the proportion of histologic response and symptom response. Studies were stratified by whether dilation was permitted. We performed meta-regressions of the association between the proportions with histologic and symptomatic response, stratified by whether dilation was permitted. We identified 1359 articles, of which 62 articles were assessed for eligibility, and 23 were included providing data on 1202 patients with EoE. Unstratified meta-regression of histologic versus symptomatic response showed moderate association and large heterogeneity (inconsistency index [I2] = 89%). In adult studies in which dilation was allowed, there was weak association between symptomatic and histologic response (β1 = 0.21), minimal symptomatic response of 67% and the heterogeneity persisted, I2 = 77%. In studies that prohibited dilation, maximal symptomatic response was 72% and was moderately associated with histologic response (β1 = 0.39) with less heterogeneity, I2 = 59%. Studies of EoE that permit dilation obscure the relation between histologic and symptomatic response and have a high floor effect for symptomatic response. Studies that prohibit dilation demonstrate moderate association between histologic and symptomatic response, but have a ceiling effect for symptomatic response. Our results demonstrate that success of dietary or medical management for EoE cannot be judged by symptoms alone, and require histologic assessment, particularly if dilation has been performed.