Proton pump inhibitors (PPI) have no effect on non-acid reflux events which can continue to provoke gastro-oesophageal reflux disease (GERD) symptoms. Baclofen, a γ-aminobutyric acid agonist, can decrease non-acid reflux but its symptomatic benefit in refractory GERD symptoms is understudied. To assess the efficacy of baclofen 10mg t.i.d. vs placebo as add-on therapy in PPI-refractory GERD symptoms, in a randomised, double-blind, placebo-controlled study. Patients with persisting typical GERD symptoms on b.i.d. PPI therapy were randomised to 4 weeks of baclofen 10mg or placebo t.i.d. Before and after treatment, patients underwent 24 h impedance-pH monitoring on-PPI. Throughout the study, patients filled out ReQuest diaries. Data were analysed using mixed models. About 60 patients were included (age 47.5 years [range 19-73], 41f/19 m), 31 patients were randomised to baclofen. One patient withdrew consent and five in the baclofen group stopped treatment due to side effects. There was a trend towards a better response for general wellbeing in the baclofen-treated group compared to placebo (p= 0.06). When subdividing patients according to symptom association probability (SAP), only the SAP+ (n= 25) group improved significantly with baclofen (pcorr = 0.02), and worsened with placebo (pcorr = 0.008). The total number of reflux events decreased over time (p= 0.01), mainly due to the baclofen condition (pcorr = 0.1). The number of reflux events with a high proximal extent dropped significantly after baclofen (pcorr = 0.009), but not placebo. Baclofen decreases several reflux parameters in PPI refractory GERD symptoms, but pH-impedance monitoring is necessary before treatment as only SAP+ patients experience clinical benefit after 4 weeks.