In very carefully selected patients, surgery is better for medically unresponsive GERD

Référence

Spechler SJ, Hunter JG, Jones KM, et al. Randomized trial of medical versus surgical treatment for refractory heartburn. N Engl J Med 2019;381(16):1513-1523.

Sommaire

These researchers identified patients referred for gastroesophageal reflux disease (GERD) unresponsive to whatever proton pump inhibitor (PPI) they were taking. Patients with severe reflux esophagitis, non-GERD endoscopic abnormalities, achalasia, or absent contractility were excluded. All patients were taken off their current PPI and given omeprazole 20 mg twice daily. Those patients who had an improvement of less than 50% in their GERD-HRQL score (ie, treatment failure) completed a baseline general quality of life scale, a depression scale, and an anxiety scale, and underwent pH monitoring while taking the omeprazole. They were enrolled in the study only if they had an association between their acid reflux symptoms and a pH of less than 4.0, and were approved by the study surgeon. Of 366 patients who entered the gauntlet, only 78 met these criteria and were randomized into 1 of 3 groups: (1) surgical treatment with laparoscopic Nissen fundoplication; (2) omeprazole 20 mg twice daily plus baclofen, followed by desipramine if baclofen failed; and (3) omeprazole 20 mg twice daily plus a baclofen placebo, followed by a desipramine placebo if the baclofen placebo failed. The primary outcome was treatment success at one year defined as at least a 50% improvement in the GERD-HRQL score. Groups were fairly well-balanced and analysis was by intention to treat. At one year, treatment success occurred in 67% in the surgery group, 28% in the active medical therapy group, and 12% in the group who received omeprazole but placebo for baclofen and desipramine. The difference between the surgery and medical groups was statistically significant, with a number needed to treat of 3 for the comparison between surgery and active medical therapy. Major adverse events were similar between groups, although one surgical patient required repeat surgery that was complicated by postoperative pneumonia. This was a Veterans Affairs study, so the population may not be as representative as it could be.