Prolonged Wireless pH Monitoring in Patients With Persistent Reflux Symptoms Despite Proton Pump Inhibitor Therapy.

ABSTRACT

BACKGROUND & AIMS: Wireless pH monitoring measures esophageal acid exposure time (AET) for up to 96 hours. We evaluated competing methods of analysis of wireless pH data. METHODS: Adult patients with persisting reflux symptoms despite acid suppression (n = 322, 48.5 ± 0.9 years, 61.7% women) from 2 tertiary centers were evaluated using symptom questionnaires and wireless pH monitoring off therapy, from November 2013 through September 2017; 30 healthy adults (control subjects; 26.9 ± 1.5 years; 60.0% women) were similarly evaluated. Concordance of daily AET (physiologic <4%, borderline 4%-6%, pathologic>6%) for 2 or more days constituted the predominant AET pattern. Each predominant pattern (physiologic, borderline, or pathologic) in relation to data from the first day, and total averaged AET, were compared with other interpretation paradigms (first 2 days, best day, or worst day) and with symptoms. RESULTS: At least 2 days of AET data were available from 96.9% of patients, 3 days from 90.7%, and 4 days from 72.7%. A higher proportion of patients had a predominant pathologic pattern (31.4%) than control subjects (11.1%; P = .03). When 3 or more days of data were available, 90.4% of patients had a predominant AET pattern; when 2 days of data were available, 64.1% had a predominant AET pattern (P < .001). Day 1 AET was discordant with the predominant pattern in 22.4% of patients and was less strongly associated with the predominant pattern compared with 48 hour AET (P = .059) or total averaged AET (P = .02). Baseline symptom burden was higher in patients with a predominant pathologic pattern compared with a predominant physiologic pattern (P = .02). CONCLUSIONS: The predominant AET pattern on prolonged wireless pH monitoring can identify patients at risk for reflux symptoms and provides gains over 24 hours and 48 hours recording, especially when results from the first 2 days are discordant or borderline.