Traditionally, gastro-oesophageal reflux is deemed to have occurred when oesophageal pH falls below 4. Other 'non-traditional' pH changes that do not fall below pH 4, that fall below 4 for only brief intervals, or that occur when basal pH is less than 4 are usually disregarded. The aim of this study was to determine whether these non-traditional pH changes represent gastro-oesophageal reflux or are artefactual. The 3-h postprandial combined oesophageal pH and manometric records of 22 patients referred for investigation of suspected gastro-oesophageal reflux were reviewed. All pH falls of > or = 0.5 pH units were analysed for manometric evidence of reflux that was classified as definite, probable, or possible. In total, 196 traditional and 223 non-traditional pH events were scored and analysed. The majority of traditional (80%) and non-traditional (60%) events were associated with definite manometric evidence of reflux, although a greater proportion of non-traditional events were associated with only probable evidence of reflux (33%) compared with traditional events (18%). The proportions of possible reflux were similar in the two groups. Limiting pH events to only those satisfying traditional criteria excluded an additional 32% with definite manometric evidence of reflux and 49% with definite or probable evidence of reflux. Most pH falls that remained above 4 or fell across 4 for < 15 sec occurred in the 1st h postprandially, compared with traditional pH events, which occurred equally throughout the 3-h period. We conclude that traditional criteria for scoring pH episodes substantially underestimate the number of reflux episodes.