The expression of gastric acidity as pH or antilog transformation of pH and the use of the mean or the median as the better summary variable are among the major problems concerning the assessment of antisecretory drugs' effects by means of continuous intraluminal pH monitoring. In this study, we reviewed data deriving from 23 normal subjects, 69 asymptomatic duodenal ulcer patients treated with bedtime (52) or twice daily (17) doses of H2-antagonists, and 10 H2-receptor blocker nonresponders. The whole group underwent continuous 24-h acidity monitoring, and the arithmetic mean and median were calculated on each individual pH and pH antilog transformed profile. One-way analysis of variance and Kruskall-Wallis test were used for statistical comparison between the four subgroups studied. The greatest discrimination between them was obtained when using the mean of pHs [p(F) approximately 0, p[KW] = 2.10(-12)), instead of the median of pHs (p[F] = 8.10(-8), p[KW] = 7.10(-8)). The antilog transformation of pH did not improve the results. It can be concluded that the use of pH to express gastric acidity and the choice of the mean to summarize the drug-related profiles provides a better characterization of the acidity behavior than would the antilog transformation of pHs and the median.