Optimizing the information obtained from continuous 24-hour gastric pH monitoring

Our objective in performing this study was to determine the best way to summarize experimental data obtained from 24-h continuous intragastric pH-metry. For this purpose, we reviewed 605 circadian intragastric pH recordings performed under different clinical conditions. Included were 82 normal subjects, 126 patients with active duodenal ulcer, 272 ulcer patients in clinical remission treated with a single bedtime dose (at 10 PM) of various H2 antagonists and 125 patients with active ulcer treated with a single morning dose (8 AM) of omeprazole. For each group, we calculated the time spent at different pH levels scaled with 0.1 incremental steps in three different time intervals: 24 h (5 PM to 4:59 PM), daytime (8 AM to 7:59 PM), and nighttime (8 PM to 7:59 AM). Then we calculated the mean and the median pH, as well as the time spent in min greater than 3.0 pH units over a mainly drug-related time window (8 PM to 7:59 AM) for each treatment and its respective placebo group. The mean and the median pH, as well as the time spent in minutes greater than 1.2 pH units over 24 h were also calculated for the untreated groups of normal subjects and duodenal ulcer patients. Considering the individual pH recordings, bi- or plurimodal density distributions of pHs were found in more than 90% of the 605 subjects, no matter the time window and the clinical condition examined, and log-normal distributions were found in only three cases. Considering the groups instead of single pH recordings, conversely, normal subjects and duodenal ulcer patients showed average left-skewed unimodal density distributions, whereas bimodal patterns were the rule in the groups of patients treated with H2 antagonists and omeprazole. The time spent greater than 3.0 pH units was the most powerful acidity index in assessing drug effectiveness in the various subsets of antisecretory compounds, in that the Student's t test values obtained with this variable were almost invariably higher than those obtained with the mean and median pH levels. Also, the time spent greater than 1.2 pH units was the most powerful summary variable in differentiating normal subjects from duodenal ulcer patients (p less than 0.01, compared with the mean and median pH). We conclude that the time spent above a given pH threshold of interest is a more accurate index than the mean and median pH in summarizing gastric acidity changes under all clinical conditions, and this is because it better represents the experimental pH measurements from which it comes.