AbstractBACKGROUND & AIMS:
Elimination diets are effective treatments for eosinophilic esophagitis (EoE), but foods that activate esophagitis are identified empirically, via a process that involves multiple esophagogastroduodenoscopies (EGDs). No optimized approach has been developed to identify foods that activate EoE. We aimed to compare clinical strategies to provide data to guide treatment.METHODS:
We developed a computer-based simulation model to determine promising empiric elimination strategies based on reported prevalence values for foods that activate EoE. We conducted a review, searching PubMed through October 1, 2017, for prospective and retrospective studies of EoE and diet. Each patient in our simulated cohort was assigned a profile comprising as many as 12 foods known to induce EoE, including dairy, wheat, eggs, soy, nuts, seafood, beef, corn, chicken, potato, pork, and/or rice. To balance the strategy success rate with the number of EGDs required for food identification, we applied an efficiency frontier approach. Strategies on the frontier were the most efficient, requiring fewer EGDs for higher or equivalent success rates relative to their comparable, neighboring strategies.RESULTS:
In all simulations, we found the 1,4,8-food and 1,3-food strategies to be the most efficient in identifying foods that induce EoE, resulting in the highest rate of the correct identification of food triggers balanced by the number of EGDs required to complete the food elimination strategy. Both strategies begin with elimination of dairy; if EoE remission is not achieved, the 1,3 diet proceeds to eliminate wheat and eggs in addition to dairy, and the 1,4,8 strategy removes wheat, eggs, dairy, and soy. In the case of persistent EoE after the second round of food elimination, the 1,3-food strategy terminates, whereas the 1,4,8-food diet eliminates corn, chicken, beef, and pork. The 1,4,8-food diet resulted in correct identification of foods that activated esophagitis in 76.68% of patients, with a mean of 4.13 EGDs and a median of 6 EGDs. The 1,3-food strategy identified foods that activated esophagitis in 42.76% of patients, with a mean of 3.36 EGDs and a median of 2 EGDs required.CONCLUSIONS:
In this modeling analysis, we found the 1,4,8-food and 1,3-food elimination strategies to be the most efficient in detection of foods that induce EoE in patients. However, the ideal elimination strategy will vary based on clinical priorities. Additional research on specific foods that induce EoE are needed to confirm the predictions of this model.