Objective: Recent guidelines propose that both proton pump inhibitor (PPI) responders and nonresponders are included in eosinophilic esophagitis (EoE). Although multiple biopsies should be required to diagnose EoE because of patchy distribution of esophageal eosinophils, it is unclear whether multiple biopsies are required to evaluate histological effectiveness of PPI therapy. This study aimed to determine the optimal biopsy protocol after PPI therapy in patients with EoE. Methods: Of 110 EoE patients, 22 PPI nonresponders were enrolled. Intraepithelial eosinophils were counted in areas of high density in multiple biopsy specimens after PPI therapy. The prevalence of esophageal eosinophilia and peak eosinophil counts after PPI therapy was analyzed according to the biopsy sites and endoscopic findings. Positive predictive value (PPV) was calculated according to the number of biopsies. Results: Of 124 biopsies, 59 (47.6%) specimens showed esophageal eosinophilia (≥15 per high-power field). Eosinophil counts were significantly higher in specimens from the lower esophagus than in those from the upper esophagus but not in those from the middle esophagus. Prevalence of esophageal eosinophilia was 76.2, 40.9, and 24.3% in the lower, middle, and upper esophagus respectively. PPI nonresponders were diagnosed in all cases with 4 biopsy specimens obtained from the lower and middle esophagus, showing that PPV for non-effectiveness of PPI therapy was 0.910 (95% CI 0.773–1.000). The prevalence of esophageal eosinophilia and peak eosinophil counts was higher in cases with white plaques and linear furrows. Conclusion: Multiple biopsies should be required to evaluate histological effectiveness of PPI therapy in patients with EoE. Four biopsies from the lower and middle esophagus may be sufficient.