Multiple white flat lesions on upper endoscopy: a systematic review and meta-analysis of the association with proton pump inhibitor exposure.

Abstract

Background: Recognition of multiple white flat lesions (MWFLs), or white flat elevated mucosa (WFEM), is increasing. Their tendency to mimic gastric intestinal metaplasia often leads to unwarranted biopsies. To address this, we reviewed existing evidence covering their endoscopic characteristics, relationship with H. pylori, detection frequencies, and longitudinal course, including the observational association with proton pump inhibitor (PPI) therapy.

Methods: From 1 January 2010 to 30 November 2025, we searched PubMed, Embase, Web of Science, Scopus, CENTRAL, and East Asian databases without language restrictions. We screened the reference lists of included studies. We pooled study-reported multivariable-adjusted odds ratios (ORs) for PPI exposure using a random-effects model wherever possible. Detection frequencies were reported descriptively, as pooling was not appropriate given the ascertainment heterogeneity.

Results: We retained five observational studies (N = 5,065). Pooling was feasible for three studies contributing cross-sectional exposure-outcome estimates (k = 3; Japan n = 2, China n = 1; N = 2,907) reporting multivariable-adjusted ORs. In this limited exploratory synthesis, PPI exposure, however defined, was associated with higher odds of MWFLs (pooled adjusted OR 2.91, 95% CI 1.98-4.29; I²=16.9%). As all studies were at moderate risk of bias, residual confounding and detection bias cannot be excluded. Study-level detection frequencies in four unselected cohorts (N = 4,902) ranged from 3.0% to 10.4% (these are detection frequencies, not population prevalence estimates). Regardless of how variables were defined, MWFLs were typically associated with H. pylori-negative or post-eradication status rather than active infection. Longitudinal evidence was inadequate to determine malignant potential or whether biopsy can be safely deferred.

Conclusions: Uncertainty regarding the malignant potential of typical MWFLs precludes any conclusion on the safety of biopsy omission. In the limited available cross-sectional evidence, PPI exposure was associated with higher MWFL odds, but this estimate should be viewed as hypothesis-generating rather than causal. Careful description of suspected MWFLs, best achieved with image-enhanced endoscopy, is essential while prospective, protocol-driven investigations resolve questions regarding their natural history, malignant potential, and the resolution of diagnostic uncertainty.

Registration: PROSPERO CRD420251248816.