Increased risk for persistent intestinal metaplasia in patients with Barrett's esophagus and uncontrolled reflux exposure before radiofrequency ablation

Abstract

Background & aims: Radiofrequency ablation (RFA) is a safe alternative to esophagectomy for patients with dysplastic Barrett's esophagus (BE). Although some studies have indicated that RFA is effective at eradicating dysplasia, most have found that RFA is not as effective in eradicating intestinal metaplasia. We investigated whether uncontrolled reflux is associated with persistent intestinal metaplasia after RFA.

Methods: Thirty-seven patients with BE underwent RFA, high-resolution manometry, and 24-hour impedance-pH testing; they received proton pump inhibitors twice daily. Patients returned every 2 months for repeat treatment or standard surveillance. Patients were classified as complete responders (CRs) if all intestinal metaplasia was eradicated in fewer than 3 ablation sessions. We analyzed clinical parameters to identify factors associated with a CR or incomplete responder (ICR).

Results: Among the 37 patients, 22 had a CR and 15 had an ICR. Mann-Whitney U tests revealed that length of BE, size of hiatal hernia, and frequency of reflux, but not acid reflux, differed between CRs and ICRs. CRs had fewer weakly acidic events than ICRs (29.5 vs 52; P < .05) and total reflux events (33.5 vs 60; P < .05), and a trend toward fewer weakly alkaline events (1.0 vs 5.0; P = .06). No other clinical or manometric features differed between groups.

Conclusions: Uncontrolled, predominantly weakly acidic reflux despite twice-daily proton pump inhibitor therapy before RFA increases the incidence of persistent intestinal metaplasia after ablation in patients with BE. Length of BE and size of hiatal hernia also were associated with persistent intestinal metaplasia after RFA.

Figure 1

Impedance-pH recordings demonstrating acid reflux (A), WAR (B), and WAlkR (C). The green colorization on the impedance tracings illustrates the retrograde flow of refluxate to the most proximal recording site (17 cm). The corresponding pH tracing in red at the bottom demonstrates the nadir pH to be 2.3 in A, 5.8 in B, and 7.2 in C.

Figure 2

Reflux exposure was compared between ICR and CR after recalculating the data based on at least 3 ablations. (A) AR varied considerably, but there was no significant difference between ICR and CR. (B) Weakly alkaline reflux events (pH>7) were uncommon, however, they were significantly more frequent in ICR compared to CR. Weakly acidic reflux (C and D) events accounted for the bulk of total reflux events and were significantly more frequent in ICR compared to CR.