Background Persistent throat symptoms are commonly attributed to ‘laryngopharyngeal reflux’. Despite a limited evidence base, these symptoms are increasingly being treated in primary care with proton pump inhibitors. Objective To assess the value of proton pump inhibitor therapy in patients with persistent throat symptoms. Design This was a double-blind, placebo-controlled, randomised Phase III trial. Setting This was a multicentre UK trial in eight UK ear, nose and throat departments. Participants A total of 346 participants aged ≥ 18 years with persistent throat symptoms and a Reflux Symptom Index score of ≥ 10, exclusive of the dyspepsia item, were recruited. Intervention Random allocation (1 : 1 ratio) to either 30 mg of lansoprazole twice daily or matched placebo for 16 weeks. Main outcome measure Symptomatic response (i.e. total Reflux Symptom Index score after 16 weeks of therapy). Results A total of 1427 patients were screened and 346 were randomised. The mean age was 52 years (standard deviation 13.7 years, range 20–84 years); 150 (43%) participants were male and 196 (57%) were female; 184 (53%) participants had a mild Reflux Symptom Index minus the heartburn/dyspepsia item and 162 (47%) had a severe Reflux Symptom Index minus the heartburn/dyspepsia item. A total of 172 patients were randomised to lansoprazole and 174 were randomised to placebo. Main outcomes A total of 267 participants completed the primary end-point visit (lansoprazole, n = 127; placebo, n = 140), of whom 220 did so between 14 and 20 weeks post randomisation (‘compliant’ group); 102 received lansoprazole and 118 received placebo. The mean Reflux Symptom Index scores at baseline were similar [lansoprazole 22.0 (standard deviation 8.0), placebo 21.7 (standard deviation 7.1), overall 21.9 (standard deviation 7.5)]. The mean Reflux Symptom Index scores at 16 weeks reduced from baseline in both groups [overall 17.4 (standard deviation 9.9), lansoprazole 17.4 (standard deviation 9.9), placebo 15.6 (standard deviation 9.8)]. Lansoprazole participants had estimated Reflux Symptom Index scores at 16 weeks that were 1.9 points higher (worse) than those of placebo participants (95% confidence interval –0.3 to 4.2; p _adj = 0.096), adjusted for site and baseline severity. Secondary outcomes Ninety-five (43%) participants achieved a Reflux Symptom Index score in the normal range (< 12) at 16 weeks: 42 (41%) in the lansoprazole group and 53 (45%) in the placebo group. A total of 226 participants completed the end-of-trial follow-up visit (lansoprazole, n = 109; placebo, n = 117), of whom 181 were ‘compliant’. The mean Reflux Symptom Index scores at 12 months reduced from baseline in both groups [lansoprazole 16.0 (standard deviation 10.8), placebo 13.6 (standard deviation 9.6), overall 14.7 (standard deviation 10.2)]. A total of 87 (48%) participants achieved a Reflux Symptom Index score in the normal range at 12 months: 33 (40%) in the lansoprazole group and 54 (55%) in the placebo group. Likewise, the Comprehensive Reflux Symptom Score and Laryngopharyngeal Reflux – Health Related Quality of Life total scores and subscales all showed very similar changes in the lansoprazole and placebo cohorts at both 16 weeks and 12 months. Limitations Drop-out rate and compliance are issues in pragmatic clinical trials. The Trial Of Proton Pump Inhibitors in Throat Symptoms (TOPPITS) aimed to detect clinically relevant difference with 90% power. The 346 randomised participants reduced to 283 at the primary end point; 267 completed the primary outcome measure, 220 within the protocol time scale. Despite this, the powers to detect the clinically relevant difference in Reflux Symptom Index score at 16 weeks were 82% (compliant comparison) and 89% (pragmatic comparison). The lack of difference between lansoprazole and placebo is generalisable across NHS clinics. Conclusions Participants on lansoprazole did not report significantly better outcomes than participants on placebo on any of the three patient-reported outcome tools (Reflux Symptom Index, Comprehensive Reflux Symptom Score and Laryngopharyngeal Reflux – Health Related Quality of Life). This multicentre, pragmatic, powered, definitive Phase III trial found no evidence of benefit for patients by treating persistent throat symptoms with lansoprazole. Trial registration Current Controlled Trials ISRCTN38578686 and EudraCT number 2013-004249-17. Funding This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 25, No. 3. See the NIHR Journals Library website for further project information.