Laparoscopic anti-reflux surgery for the treatment of idiopathic pulmonary fibrosis (WRAP-IPF): a multicentre, randomised, controlled phase 2 trial

Transcription

1 Laparoscopic anti-reflux surgery for the treatment of idiopathic pulmonary fibrosis (WRAP-IPF): a multicentre, randomised, controlled phase 2 trial Ganesh Raghu, Carlos A Pellegrini, Eric Yow, Kevin R Flaherty, Keith Meyer, Imre Noth, Mary Beth Scholand, John Cello, Lawrence A Ho, Sudhakar Pipavath, Joyce S Lee, Jules Lin, James Maloney, Fernando J Martinez, Ellen Morrow, Marco G Patti, Stan Rogers, Paul J Wolters, Robert Yates, Kevin J Anstrom, Harold R Collard Summary Background Abnormal acid gastro-oesophageal reflux (GER) is hypothesised to play a role in progression of idiopathic pulmonary fibrosis (IPF). We aimed to determine whether treatment of abnormal acid GER with laparoscopic anti-reflux surgery reduces the rate of disease progression. Methods The WRAP-IPF trial was a randomised controlled trial of laparoscopic anti-reflux surgery in patients with IPF and abnormal acid GER recruited from six academic centres in the USA. We enrolled patients with IPF, abnormal acid GER (DeMeester score of 14 7; measured by 24-h ph monitoring) and preserved forced vital capacity (FVC). We excluded patients with a FVC below 50% predicted, a FEV₁/FVC ratio of less than 0 65, a history of acute respiratory illness in the past 12 weeks, a body-mass index greater than 35, and known severe pulmonary hypertension. Concomitant therapy with nintedanib and pirfenidone was allowed. The primary endpoint was change in FVC from randomisation to week 48, in the intention-to-treat population with mixed-effects models for repeated measures. This trial is registered with ClinicalTrials.gov, number NCT Findings Between June 1, 2014, and Sept 30, 2016, we screened 72 patients and randomly assigned 58 patients to receive surgery (n=) or no surgery (n=). 27 patients in the surgery group and 20 patients in the no surgery group had an FVC measurement at 48 weeks (p=0 041). Intention-to-treat analysis adjusted for baseline anti-fibrotic use demonstrated the adjusted rate of change in FVC over 48 weeks was 0 05 L (95% CI 0 15 to 0 05) in the surgery group and 0 13 L ( 0 23 to 0 02) in the non-surgery group (p=0 28). Acute exacerbation, respiratory-related hospitalisation, and death was less common in the surgery group without statistical significance. Dysphagia (eight [%] of 28) and abdominal distention (four [14%] of 28) were the most common adverse events after surgery. There was one death in the surgery group and four deaths in the non-surgery group. Interpretation Laparoscopic anti-reflux surgery in patients with IPF and abnormal acid GER is safe and well tolerated. A larger, well powered, randomised controlled study of anti-reflux surgery is needed in this population. Funding US National Institutes of Health National Heart, Lung and Blood Institute. Copyright 2018 Elsevier Ltd. All rights reserved. Introduction Idiopathic pulmonary fibrosis (IPF) is an age-related chronic fibrotic lung disease that is characterised by physiological progression and episodes of acute respiratory worsening. 1 Abnormal acid gastro-oesophageal reflux (GER) has been well described in patients with IPF and is thought to play a role in pathogenesis and progression of disease through increased risk for microaspiration of gastric acid and other non-acid components of the gastric refluxate. 2,3 Evidence to date is mixed regarding the clinical impact of abnormal acid GER and treatment of GER on the progression of IPF. 4,5 Several retrospective cohort studies have described the prevalence of abnormal acid GER in patients with IPF 6 9 and assessed the association of antiacid therapies for GER (H2-receptor antagonists and proton-pump inhibitors) with clinical outcomes in IPF. 10,11 Two large secondary analyses 12,13 of data from clinical trials reached different conclusions; one reported a significant reduction in rate of physiological decline and acute exacerbation with anti-acid therapies, 12 the other reported no effect. 13 Microaspiration of both acid and non-acid components might be important to progression of IPF. A large retrospective cohort study 14 of patients with IPF showed improved survival in patients who had anti-reflux surgery compared with those taking anti-acid therapy. A larger case series 15 of 27 patients with progressive IPF who underwent laparoscopic anti-reflux surgery showed excellent safety outcomes and a trend toward stabilisation of physiology as measured by forced vital capacity (FVC). These data suggest that surgical treatment for GER might provide added benefit for patients with IPF and abnormal acid GER. Lancet Respir Med 2018 Published Online August 9, S (18) See Online/Comment S (18) Department of Medicine (G Raghu MD, L A Ho MD, S Pipavath MD) and Department of Surgery (C A Pellegrini MD, R Yates MD), University of Washington, Seattle, WA, USA; Duke Clinical Research Institute, Durham, NC, USA (E Yow MS, K J Anstrom PhD); Department of Medicine (K R Flaherty MD) and Department of Surgery (J Lin MD), University of Michigan, Ann Arbor, MI, USA; Department of Medicine (K Meyer MD) and Department of Surgery (J Maloney MD), University of Wisconsin, Madison, WI, USA; Department of Medicine, University of Chicago, Chicago, IL, USA (I Noth MD); Department of Medicine (M B Scholand MD) and Department of Surgery (E Morrow MD), University of Utah, Salt Lake City, UT, USA; Department of Medicine (J Cello MD, P J Wolters MD, H R Collard MD) and Department of Surgery (S Rogers MD), University of California San Francisco, San Francisco, CA, USA; Department of Medicine, University of Colorado Denver, Aurora, CO, USA (J S Lee MD); Department of Medicine, Weill Cornell School of Medicine, New York, NY, USA (F J Martinez MD); and Departments of Medicine and Surgery, University of North Carolina, Chapel Hill, NC (M G Patti MD) Published online August 9,

2 Correspondence to: Dr Ganesh Raghu, Department of Medicine, University of Washington, Seattle, WA 98195, USA or Dr Harold R Collard, Department of Medicine, University of California San Francisco, San Francisco, CA 94143, USA hal.collard@ucsf.edu Research in context Evidence before this study We searched PubMed for reports of clinical trials investigating the role of abnormal acid gastro-oesophageal reflux in the progression of idiopathic pulmonary fibrosis (IPF), published between Jan 1, 1998, and Dec 31, We used the search terms idiopathic pulmonary fibrosis AND gastro-oesophageal reflux, and filtered for reports published in English, which yielded 143 results. Abnormal acid gastro-oesophageal reflux is thought to play a role in the progression of IPF. Two large secondary analyses of data from clinical trials reached different conclusions regarding the effect of anti-acid medications on disease progression. Additional data suggest that anti-reflux surgery might provide added benefit. Added value of this study This phase 2 randomised, controlled trial aimed to determine whether normalisation of abnormal acid gastro-oesophageal reflux with laparoscopic anti-reflux surgery reduced the rate of disease progression. Laparoscopic anti-reflux surgery was safe and well tolerated but disease progression over 48 weeks defined as change in forced vital capacity did not reduce significantly. Respiratory-related hospitalisation and death were less common in the surgical group without statistical significance. These results provide the first prospective controlled data addressing this hypothesis. Implications of all the available evidence Laparoscopic anti-reflux surgery is safe and well tolerated in patients with IPF and abnormal acid gastro-oesophageal reflux. Further investigation of the effect of anti-reflux therapies on disease progression is needed. See Online for appendix In this randomised controlled trial, we hypothesised that treatment of patients with IPF and abnormal acid GER with laparoscopic anti-reflux surgery would prevent further insult to the lung and slow the progression of disease (as measured by change in FVC), and reduce the rate of clinically meaningful events, such as acute exacerbation, respiratory-related hospitalisation, and death. To test this hypo thesis, we did laparoscopic anti-reflux surgery in patients with IPF and abnormal acid GER and compared outcomes with patients who did not receive such surgery. Methods Study design The WRAP-IPF trial was a randomised, unblinded, controlled trial of laparoscopic anti-reflux surgery performed at six academic centres in the USA, with a coordinating centre at the Duke Clinical Research Institute. Sites were chosen on the basis of expertise in the care of patients with IPF and experience with laparoscopic anti-reflux surgery in the IPF population. Five sites were involved in the design of the trial, including standardisation of the surgical intervention; a sixth site was added during the trial. An independent data monitoring committee approved the protocol and regularly reviewed the data, giving recommendations to the funder regarding study continuation. The protocol was approved by the institutional review boards at each site. The protocol is in the appendix. Participants Participants met the 2011 consensus criteria for the diagnosis of IPF as determined by the site investigator. 1 Concomitant therapy with US Food and Drug Administration-approved medications for IPF (ie, nintedanib and pirfenidone) was allowed. All potential participants underwent high-resolution oesophageal manometry and 24-h ph testing during screening and only those with evidence of abnormal acid GER (defined by a DeMeester score of 14 7) were eligible. 16 Exclusion criteria included FVC below 50% predicted, FEV₁/FVC ratio of less than 0 65, a room air partial pressure of oxygen in the blood of less than 60 mm Hg, a 6-min walk distance of less than 50 m, history of acute respiratory illness in the past 12 weeks, inability to safely undergo laparoscopic anti-reflux surgery (as judged by the investigators), body-mass index greater than 35, known severe pulmonary hypertension (defined as mean pulmonary artery pressure >35 mm Hg on right heart catheterisation or right ventricular systolic pressure >50 mm Hg on echocardiography), and listing for lung transplantation. Full inclusion and exclusion criteria are given in the appendix. Behavioural and medical therapy (eg, H2-receptor antagonists and proton-pump inhibitors) for acid GER was allowed in both groups if the participant and their clinician felt it was necessary. Eligible patients gave written informed consent. Randomisation and masking We allocated participants (1:1) to surgery or no surgery using a computer-generated randomisation scheme. Randomisation was not stratified because of the unblinded study design. Study staff used a web-based trial management system to automatically allocate participants to treatment at the time of enrolment. Procedures The appendix contains a complete description of the protocol for laparoscopic anti-reflux surgery. All sites followed this standardised protocol. Briefly, the peritoneal cavity was accessed laparoscopically. After division of all short gastric vessels and the gastrohepatic ligament, the phreno-oesophageal membrane was divided. The gastric 2 Published online August 9,

3 fundus, gastroesophageal junction, and distal oesophagus were fully mobilised. Next, a retro-oesophageal window was made and the diaphragmatic crura were approximated securely around the oesophagus. The gastric fundus was wrapped around the distal oesophagus over a bougie, creating a floppy 360 fundoplication, which was secured to itself anteriorly with three permanent sutures at 1-cm intervals. Additional permanent collar sutures were used to fixate the wrap to the right and left pillars of the oesophageal crus. Participants were followed up from time of randomisation to 52 weeks. All participants had study visits at baseline, 12, 24, 36, and 48 weeks, during which spirometry, 6-min walk testing, and patient-related outcome assessments were done. For safety, spirometry was avoided for 8 weeks after surgery. Participants in the surgery group had clinical visits for preoperative evaluation, anti-acid surgery, and postoperative management as clinically indicated, and underwent repeat 24-h ph testing at 24 weeks to assess the efficacy of the surgery. Participants in the no surgery group had the option of receiving laparoscopic anti-reflux surgery from 24 weeks after randomisation if their clinician deemed it medically necessary for their abnormal acid GER. All participants were contacted by telephone for safety assessments at weeks 4, 8, 16, 20, 28, 32, 40, 44, and 52. All participants completed an exploratory questionnaire on reflux symptoms at baseline and 48 weeks. As part of an exploratory sub-study not reported, all participants were provided with in-home spirometers (PMD Healthcare, Allentown, PA, USA) for daily measurement of FVC, and underwent phlebotomy at study visits. Participants in the surgery group underwent bronchoalveolar lavage at the time of surgery. Outcomes The primary endpoint was change in FVC from randomisation (baseline) to 48 weeks measured on inoffice spirometry during study visits in accordance with ATS/ERS standards. 17 Secondary endpoints were acute exacerbation, non-elective hospitalisation (both all-cause and respiratory-related), death, change in cough severity (measured by a visual analogue scale), change in dyspnoea severity (measured by the University of California San Diego Shortness of Breath Questionnaire), 18 change in health-related quality of life (measured by the St George s Respiratory Questionnaire), 19 change in 6-min walk distance, and time to selected event-driven composite endpoints of disease progression (appendix). Additional secondary endpoints were time to disease progression, change in ICECAP-O score (continuous and categorical), change in EuroQOL-5 dimension scores (continuous and categorical), reduction in acid GER by serial ph testing and GER questionnaire, categorical change in FVC, and quantitative change in HRCT fibrosis score and honeycombing. These additional secondary endpoints will be reported at a later time. Acute exacerbation was defined allocated to surgery group 28 received surgery 1 refused surgery Figure 1: Trial profile Surgery (n=) No surgery (n=) Age, years 70 6 (6 0) 69 2 (7 4) Sex Male 28 (97%) 19 (66%) Female 1 (3%) 10 (34%) Race White 27 (93%) 27 (93%) Asian 1 (3%) 1 (3%) African 0 1 (3%) Other 1 (3%) 0 History of cigarette smoking 20 (69%) 15 (52%) History of gastro-oesophageal reflux disease 28 (97%) 28 (97%) History of hiatal hernia 12 (41%) 16 (55%) Medication at baseline Nintedanib 8 (28%) 3 (10%) Pirfenidone 9 (31%) 9 (31%) H2-receptor antagonist 1 (3%) 4 (14%) Proton-pump inhibitor 24 (83%) 26 (90%) FVC%pred 76 5 (16 0) 74 3 (13 3) DLCO%pred 41 6 (11 0) 43 4 (11 5) 24-h oesophageal ph measurement Total time ph <4, % 10 2 (12 6) 14 6 (17 4) Supine time ph <4, % 8 7 (15 3) 8 6 (8 8) DeMeester score 37 7 (42 3) 43 1 ( 8) 6-min walk distance, m 400 (81) 381 (96) Patent-reported outcome measures* UCSD Shortness of Breath Questionnaire score (0 120) 7 (20 7) 34 9 (22 9) Visual analogue scale for cough severity (0 100) 2 (20 7) 31 6 (26 7) St George s Respiratory Questionnaire total score (0 100) 35 8 (16 1) 38 8 (18 1) Data are n (%) or mean (SD). FVC%pred=percentage of predicted forced vital capacity. DLCO%pred=percentage of predicted diffusion capacity for carbon monoxide. UCSD=University of California San Diego. *Lower score equals better health (ie, less severe dyspnoea, less severe cough, and better quality of life). Table 1: Baseline characteristics 72 patients assessed for eligibility 58 randomly allocated to treatment 2 lost to follow-up 1 patient decision 1 death included in intention-to-treat analysis 14 excluded 9 DeMeester score < ineligible for surgery 3 other allocated to no surgery group 26 did not receive surgery 2 received surgery after 24 weeks 1 received surgery before 24 weeks 7 lost to follow-up 2 patient decision 1 lung transplantation 4 death included in intention-to-treat analysis Published online August 9,

4 Change in forced vital capacity (L) Surgery group No surgery group A p value at 48 weeks = Surgery group No surgery group B 54 % 51% 54% 46% 65% 58% 51% 81% 88% 75% 78% 78% 80% AEx-D 71% 0 Surgery 6 12 MLC Weeks after randomisation Figure 2: Change in forced vital capacity and deaths by treatment group (A) The mean change in forced vital capacity from randomisation to week 48. Error bars are standard error. (B) The disease course of the five patients who died during the study. Percentages are measured forced vital capacity in percentage of predicted and acute exacerbations. Each bar ends at the time of the patient s death. MLC=metastatic lung cancer. MI=myocardial infarction. AEx-D=definite acute exacerbations. AEx-S=suspected acute exacerbations. AEx-S AEx-D AEx-S AEx-S per protocol and was adjudicated (definitive vs suspected) by site investigators according to international working group perspectives. 20,21 Safety of laparoscopic anti-reflux surgery was assessed via key operative and postoperative outcomes, including participants symptoms, intraoperative compli cations, length of stay in intensive care unit and hospital, and postoperative complications or readmission to hospital within 30 days of surgery. Statistical analysis Based on a two-sample t test, our study had 80% power to detect an effect size of 0 8 SDs (about 0 20 L) for change in FVC at a significance level of 0 05, allowing for 10% missing data. We assumed that concomitant therapy for IPF with nintedanib or pirfenidone would affect the change in FVC over time. 22,23 We used mixed-effects models for repeated measures to compare differences in the primary endpoint across treatment groups at baseline, 12, 24, 36, and 48 weeks in the intention-to-treat population. 24 We estimated treatment effect via contrast estimates (and their 95% CIs) of differences between groups. We included an indicator variable in the regression model for baseline use of nintedanib or pirfenidone. We assessed the validity of the 36 MI 42 AEx-S 48 mixed-effects model using standard modelling diagnostics and goodness-of-fit measures. For participants who were randomised to the no surgery group and subsequently underwent surgery after 24 weeks (allowed per protocol), postoperative measurements of FVC were excluded from the primary endpoint analysis (all other endpoints used all 48 weeks of data). Because of concerns that missing primary endpoint data might be informative and therefore did not meet a core assumption of missingness at random for the mixed-effects model 25 we used Lachin s worst-rank analysis for post-hoc analyses. 26 This nonparametric approach assumes that missing data are informative and reflect poor outcomes. We used mixed-effects models to analyse secondary endpoints, Cox proportional hazards regression models for time-to-event endpoints, and χ² and Fisher s exact tests for binary and categorical outcomes. We used SAS version 9.4 (SAS Institute, Cary, NC, USA) for all statistical analyses. An independent data safety and monitoring committee oversaw the study. This study is registered with ClinicalTrials.gov, number NCT Role of the funding source The funder participated in study design and reviewed drafts of the protocol. The funder had no role in data collection, data analysis, data interpretation, or writing of the report. The corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Between June 1, 2014, and Sept 30, 2016, we screened 72 patients and randomly assigned treatment to 58 participants (figure 1). 33 (57%) patients were enrolled at one site (University of Washington, Seattle, WA, USA). Of the patients assigned to receive laparoscopic anti-reflux surgery, 28 (97%) had the procedure and 27 (93%) completed the study. The median time from randomisation to surgery was 6 7 weeks (SD 4 4). Of the patients assigned to not receive surgery, two (7%) had laparoscopic anti-reflux surgery after week 24, and 21 (72%) completed 52 weeks of follow-up. The differential dropout between the surgery and no surgery groups was informative, driven primarily by an imbalance in deaths (surgery 1 vs no surgery 4). Baseline characteristics were similar between the surgery and no surgery groups, although male sex and use of antifibrotic medication at baseline were more common in the surgery group (table 1). Mean age was 70 6 years (SD 6 0) in the surgery group and 69 2 years (7 4) in the no surgery group. Most participants were former smokers, had mild to moderate physiological impairment, and regularly used anti-acid medications. The mean baseline DeMeester scores were high in both groups, and all but two patients had already been diagnosed with GER disease. 4 Published online August 9,

5 The adjusted rate of decline in FVC over 48 weeks was 0 05 L (95% CI 0 15 to 0 05) in the surgery group and 0 13 L ( 0 23 to 0 02) in the no surgery group (p=0 28, figure 2A). Comparisons between groups at interval time points are shown in the appendix. 27 patients in the surgery group and 20 patients in the no surgery group had an FVC measurement at 48 weeks (p=0 041). Worst-rank analysis showed that the between-group difference was statistically significant (p=0 017), favour ing the surgery group. The rank order of all participants is given in the appendix. Acute exacerbation, respiratory-related hospital isation, nonelective hospitalisation, and lung transplantation were less common in the surgery group, without statistical significance (table 2). One participant in the surgery group and four participants in the non-surgery group died. All deaths were preceded by definite or suspected acute exacerbations within 2 weeks of death (figure 2B). Time to event analysis showed that time to death or 10% decline in FVC (p=0 038, figure 3A) and time to death, 10% decline in FVC, or acute exacerbation (p=0 048) were longer in the surgery group, but differences in other measures of disease progression were not significant (figure 3B, table 2) Surgery did not significantly affect change in cough severity, dyspnoea severity, health-related quality of life, 6-min walk distance, or reflux symptom severity (table 2, appendix). The mean DeMeester score after laparoscopic anti-reflux surgery was 7 7 (SD 12 8), with a mean decrease from baseline of 30 6 points (36 8). Laparoscopic anti-reflux surgery was generally safe and well tolerated (appendix). There was one bleeding event in the surgery group (defined as >100 ml), which was successfully managed intraoperatively and had no sequelae. Postoperative dysphagia (eight [%] of 28) and abdominal distention (four [14%] of 28) were the most common symptom-based adverse events (appendix). The mean postoperative stay in an intensive care unit was 4 8 h (SD 23) and the mean length of hospital stay was 1 2 days (0 88; appendix). Two (7%) of patients had postoperative complications. The first had moderate dehydration 20 days after surgery and was managed as an outpatient. The second was readmitted to hospital 4 days after surgery for shock and respiratory failure attributed to metabolic acidosis. Dehydration and adrenal insufficiency were identified and treated successfully, leading to stabilisation and discharge after 7 days. Both patients had no sequelae of their complications and their FVC improved over the 48 weeks of follow-up (2% and 10%, respectively). There were no episodes of pneumonia or respiratory infection in either group. Discussion This phase 2 randomised controlled trial in patients with IPF and abnormal acid GER showed that laparoscopic anti-reflux surgery is safe and well tolerated but did not show that such surgery significantly slows the rate of FVC decline. This trial was underpowered for Surgery (n=) No surgery (n=) p value Clinical events* Acute exacerbation 1 (3%) 4 (16%) 0 19 Respiratory hospitalisation 2 (7%) 6 (21%) 0 25 Non-elective hospitalisation 5 (17%) 8 (28%) 0 35 Lung transplantation 0 1 (3%) >0 99 Disease progression Death 1 (3%) 4 (18%) % FVC decline or death 2 (9%) 7 (%) % FVC decline, acute exacerbation, 2 (9%) 7 (28%) or death Respiratory hospitalisation or death 2 (9%) 5 (19%) 0 16 Non-elective hospitalisation or death 5 (17%) 7 (26%) % FVC decline, 5 point UCSD Shortness of Breath Questionnaire increase, respiratory hospitalisation, or death 15 (57%) 15 (56%) 0 74 Change in symptoms and physical function Visual analogue scale for cough severity 4 74 ( 3 14 to 12 63) 7 15 ( 2 01 to 16 31) 0 69 UCSD Shortness of Breath Questionnaire 0 71 ( 5 62 to 4 21) 0 69 ( 6 24 to 4 86) >0 99 St George s Respiratory Questionnaire 1 04 ( 3 66 to 5 74) 3 18 ( 8 35 to 1 99) min walk distance, m ( 4 80 to 37 88) 9 13 ( to 33 02) 0 65 Data are n (%), n (Kaplan Meier estimates), or slope estimate (95% CI). FVC=forced vital capacity. UCSD=University of California San Diego. *We used the χ² and Fisher s exact tests to calculate p values. Rates are Kaplan Meier estimates at 48 weeks. We used the log-rank test at 48 weeks to calculate p values. Slope parameter estimates with 95% CIs from a mixed effects model with adjustment for baseline use of nintedanib or pirfenidone therapy. Table 2: Secondary endpoints the observed effect size, with about 400 patients required to achieve 90% power. We believe the non-significant numerical differences in the primary and secondary endpoints warrant further study of laparoscopic anti-reflux surgery as treatment for patients with IPF. These results suggest that patients with IPF and abnormal acid GER might be an important subgroup of this disease and further research is needed to discover whether they could benefit from anti-reflux surgery to prevent microaspiration and further injury to the lung. Our primary endpoint of change in FVC failed to reach statistical significance using our prespecified analytical method of mixed-effects model for repeated measures. This method assumes that missingness of data is random and non-informative (ie, rather than from clinically significant events, such as death). This assumption is not met by the results of this study because differential dropout was driven by an increased number of deaths in the nonsurgery group. Such informative missingness is often handled by non-parametric rank methodology and this approach has previously been used for research into IPF. 22 We did a post-hoc exploratory analysis of the primary endpoint using Lachin s worst-rank analysis, 25 an approach that assumes that missing values are informative and reflect poor outcomes. The difference between groups was significant (p=0 017) favouring the surgical group. Future studies of physiological decline in patients with IPF should carefully consider their analytical methods and how to handle potentially informative missingness of data. Published online August 9,

6 Event rate (%) Number at risk Surgery group No surgery group Event rate (%) Number at risk Surgery group No surgery group A 100 Surgery group No surgery group 90 Log rank p value at 48 weeks: B Log rank p value at 48 weeks: Figure 3: Kaplan-Meier estimates (A) Time to the composite endpoint of 10% decline in forced vital capacity or death. (B) Time to the composite endpoint of respiratory hospitalisation or death Time after randomisation (weeks) The effect of laparoscopic anti-reflux surgery on FVC decline and clinical events in patients with IPF and abnormal acid GER could have been reduced by the near universal use of anti-acid medications in the no surgery group. This possibility is supported by the slower rate of decline in FVC in our control group than that of historical clinical trial control groups. 22,23 Previous retrospective data have suggested that laparoscopic anti-reflux surgery is more beneficial than anti-acid therapy is, 14 perhaps because anti-acid medications do not fully control acid reflux. 7 In this study, abnormal acid GER was normalised after laparoscopic anti-reflux surgery in all patients. Another possible explanation is that abnormal non-acid reflux is an important driver of disease progression that requires treatment, and that treatment of abnormal acid GER alone is insufficient in these patients. Future studies of treatment for abnormal GER in IPF should directly address the effect of anti-acid medications on the progression of disease, and the distinct contributions of acid and non-acid GER The number of clinical events and death in the no surgery group are consistent with what is predicted on the basis of baseline demographics and physiology. 27 All deaths were preceded by definite or suspected acute exacerbations. This finding suggests that patients with IPF and abnormal acid GER might be at particularly high risk for developing an acute exacerbation, a hypothesis that is supported by retrospective data showing increased gastric pepsin in the bronchoalveolar lavage fluid of patients with acute exacerbation of IPF. 30 In such a population, composite clinical endpoints, such as time to acute exacerbation or death, might be more attractive efficacy outcomes than pulmonary physiology is, because they are more rapidly met and avoid the issue of informative missingness of data. Notably, two deaths (one in the surgery and one in the no surgery group) involved concomitant serious medical complications metastatic lung cancer and myocardial infarction), which affect their attribution to the progression of IPF disease. Laparoscopic anti-reflux surgery was generally safe and well tolerated, with only one participant requiring readmission to hospital for shock from postoperative dehydration and previously unrecognised adrenal insufficiency. Anti-reflux surgery was performed by highly experienced surgeons at established academic centres of excellence for IPF, with 17 (61%) of 28 surgeries done by one centre. It is unclear how generalisable these safety and tolerability findings are to the community population. Other limitations include the small number of participants and the informative missingness of data, both of which reduce statistical power. The trial was not designed to stratify the use of proton-pump inhibitors or H2-receptor antagonists, thus the potential effects of the specific antiacid treatment are unknown. Although the no surgery group included more women than the surgery group did, which could be a potential source of bias, historical data suggest that this imbalance would improve outcomes in the no surgery group (ie, bias against finding a benefit to surgery). 31,32 Findings from our analysis of change in percentage of predicted FVC (which corrects for sex differences in physiology) were similar to our primary FVC analysis (data not shown). Randomised trials of surgical interventions are difficult to do, and patients with IPF and abnormal acid reflux are a challenging subgroup to identify because of the invasive testing required. Future investigators and sponsors will need to anticipate these logistical difficulties and think creatively about study design and analytical methods. We suggest a few points for consideration. First, consider whether identification of elevated DeMeester scores is essential to predictive enrichment or could all patients with reflux, regardless of severity, be enrolled (perhaps stratified by symptoms or DeMeester scores)? Second, recognise that most patients who are not allocated to receive surgery might require anti-acid medications and other behavioural modifications, which will need to be incorporated into the design of the study and interpretation 6 Published online August 9,

7 of the results to determine if anti-acid therapy is associated with disease progression in IPF. An understanding of this relationship will be important to the proper interpretation of any differences between groups in a follow-up surgical study. Third, consider approaches that minimise the chances of crossover or dropout by patients randomly allocated to not receive surgery. Such approaches could include sham surgery to blind participants to the intervention received, or an incentive that encourages continued participation in the no surgery group. Fourth, anticipate meaningful missingness in the dataset and use analytical methods (eg, rank analyses) that are appropriate to the data structure. Finally, consider alternative endpoints to FVC and how they might help to improve power and clinical interpretability. It could be that surgical treat ment of reflux has a major effect on acute exacerbation, respiratory-related hospitalisation, and death, and that a com posite endpoint that includes these events would provide reasonable power and improve clinical interpretability. In summary, the results of this trial show that laparoscopic anti-reflux surgery in patients with IPF and abnormal acid GER is safe and well tolerated. The potential therapeutic value of laparoscopic anti-reflux surgery in this disease subgroup should be confirmed in larger, more generalisable trials. Contributors GR, CAP, EY, KRF, KM, IN, JC, LAH, JSL, SP, JL, JM, FJM, EM, MGP, SR, PJW, KJA, and HRC designed the study. EY and KJA analysed the data. All authors contributed to study conduct, data interpretation, and manuscript preparation. Declaration of interests All authors are coinvestigators of GR and HRC and received a research grant to do this study from the US National Institutes of Health, National Heart, Lung and Blood Institute as co-principal investigators. GR is a consultant to Bellerophan, Boerhinger-Ingelheim, BMS, Biogen, Fibrogen, Nitto, Patara, Promedior, Sanofi, and Veracyte for other IPFrelated studies, outside of the submitted work. HRC reports personal fees from Bayer, Boehringer Ingelheim, Global Blood Therapeutics, Genoa, ImmuneWorks, Navitor, Parexel, PharmAkea, Prometic, Toray, and Veracyte, outside of the submitted work. EY reports grants from the US National Institutes of Health during the conduct of this study. KRF reports grants and personal fees from Boehringer Ingelheim and Roche/Genentech and personal fees from Veracyte, Fibrogen, and Sanofi-Genzyme, outside of the submitted work. KM reports grants from the US National Institutes of Health, Genentech/Roche, InterMune, Parion, Nivalis, and Promedior, and personal fees from University of Nebraska & Practice Point Communications, National Jewish Health, and PILOT CME Speaker, outside of the submitted work. IN reports, personal fees from Genentech/Hoffmann La Roche, Sanofi, and Global Blood Therapeutics, and personal fees and non-financial support from Boehringer Ingelheim, outside of the submitted work. IN also has a patent pending for Toll-like interacting protein and its pharmacogenetics. MBS reports other support from Boehringer Ingelheim and Genetech during the conduct of this study. LAH reports personal fees from Genentech and Boehringer Ingelheim, outside of the submitted work. JSL reports grants from the US National Institutes of Health and personal fees from Celgene, Genentech, and Boehringer Ingelheim, outside of the submitted work. JL is a proctor and surgical site mentor for Intuitive Surgical. FJM reports grants, personal fees, non-financial support, and other from AstraZeneca, Boehringer Ingelheim, PeerView, and GlaxoSmithKline; personal fees and non-financial support from the American College of Chest Physicians, ConCert, Inova Fairfax Health System, Novartis, Prime, Chiesi, Sunovion, and Theravance; personal fees from Columbia University, Haymarket Communications, Integritas, Inthought Research, MD Magazine, Methodist Hospital Brooklyn, New York University, Pearl Pharmaceuticals, Unity, UpToDate, WebMD/MedScape, Western Connecticut Health Network, Academic CME, Clarion, Patara, France Foundation, and PlatformIQ; grants and other support from Afferent and Gilead; grants, personal fees, and non-financial support from Miller Communications, National Association for Continuing Education, Potomac, Puerto Rican Respiratory Society, Stromedix/Biogen, University of Alabama Birmingham, and Genentech/Roche; nonfinancial support and other from Nitto; other support from ProterrixBio, Veracyte, and Bayer; and non-financial support from Zambon, outside of the submitted work. PJW reports grants from Medimmune, grants from Genentech, and personal fees from Roche, outside of the submitted work. KJA reports grants from the US National Institutes of Health and is on the Data Safety and Monitoring Board for Boehringer Ingelheim and Promedior. KJA is also a consultant for Boehringer Ingelheim and Promedior. CAP, JC, SP, JM, EM, MGP, SR, and RY declare no competing interests. Acknowledgments The US National Institutes of Health National Heart, Lung and Blood Institute funded this work (grant HL119089). We are grateful to all the patients, coordinators, coinvestigators, and subinvestigators who participated in this study at all sites. References 1 Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med 2011; 183: Raghu G, Meyer KC. Silent gastro-oesophageal reflux and microaspiration in IPF: mounting evidence for anti-reflux therapy? Eur Respir J 2012; 39: Lee JS, Collard HR, Raghu G, et al. Does chronic microaspiration cause idiopathic pulmonary fibrosis? Am J Med 2010; 123: Johannson KA, Strambu I, Ravaglia C, et al. Antacid therapy in idiopathic pulmonary fibrosis: more questions than answers? Lancet Respir Med 2017; 5: Ghebre YT, Raghu G. Idiopathic pulmonary fibrosis: novel concepts of proton pump inhibitors as antifibrotic drugs. Am J Respir Crit Care Med 2016; 193: Patti MG, Tedesco P, Golden J, et al. Idiopathic pulmonary fibrosis: how often is it really idiopathic? J Gastrointest Surgery 2005; 9: Raghu G, Freudenberger TD, Yang S, et al. High prevalence of abnormal acid gastro-oesophageal reflux in idiopathic pulmonary fibrosis. Eur Respir J 2006; 27: Sweet MP, Patti MG, Leard LE, et al. Gastroesophageal reflux in patients with idiopathic pulmonary fibrosis referred for lung transplantation. J Thorac Cardiovasc Surg 2007; 133: Tobin RW, Pope CE 2nd, Pellegrini CA, Emond MJ, Sillery J, Raghu G. Increased prevalence of gastroesophageal reflux in patients with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998; 158: Raghu G, Yang ST, Spada C, Hayes J, Pellegrini CA. Sole treatment of acid gastroesophageal reflux in idiopathic pulmonary fibrosis: a case series. Chest 2006; 1: Linden PA, Gilbert RJ, Yeap BY, et al. Laparoscopic fundoplication in patients with end-stage lung disease awaiting transplantation. J Thorac Cardiovasc Surg 2006; 131: Lee JS, Collard HR, Anstrom KJ, et al. Anti-acid treatment and disease progression in idiopathic pulmonary fibrosis: an analysis of data from three randomised controlled trials. Lancet Respir Med 2013; 1: Kreuter M, Spagnolo P, Wuyts W, et al. Antacid therapy and disease progression in patients with idiopathic pulmonary fibrosis who received pirfenidone. Respiration 2017; 93: Lee JS, Ryu JH, Elicker BM, et al. Gastroesophageal reflux therapy is associated with longer survival in patients with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2011; 184: Raghu G, Morrow E, Collins BF, et al. Laparoscopic anti-reflux surgery for idiopathic pulmonary fibrosis at a single centre. Eur Respir J 2016; 48: Published online August 9,

8 16 Johnson LF, Demeester TR. Twenty-four-hour ph monitoring of the distal esophagus. A quantitative measure of gastroesophageal reflux. Am J Gastroenterol 1974; 62: Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Respir J 2005; 26: Eakin EG, Resnikoff PM, Prewitt LM, Ries AL, Kaplan RM. Validation of a new dyspnea measure: the UCSD Shortness of Breath Questionnaire. University of California, San Diego. Chest 1998; 113: Jones PW, Quirk FH, Baveystock CM. The St George s Respiratory Questionnaire. Respir Med 1991; 85 (suppl B): Collard HR, Ryerson CJ, Corte TJ, et al. Acute exacerbation of idiopathic pulmonary fibrosis. An international working group report. Am J Respir Crit Care Med 2016; 194: Collard HR, Moore BB, Flaherty KR, et al. Acute exacerbations of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2007; 176: King TE Jr, Bradford WZ, Castro-Bernardini S, et al. A phase 3 trial of pirfenidone in patients with idiopathic pulmonary fibrosis. N Engl J Med 2014; 370: Richeldi L, du Bois RM, Raghu G, et al. Efficacy and safety of nintedanib in idiopathic pulmonary fibrosis. N Engl J Med 2014; 370: Mallinckrodt CH, Kaiser CJ, Watkin JG, Molenberghs G, Carroll RJ. The effect of correlation structure on treatment contrasts estimated from incomplete clinical trial data with likelihood-based repeated measures compared with last observation carried forward ANOVA. Clin Trials 2004; 1: Little RJA, Rubin DB. Statistical analysis with missing data. 2nd edn. London: John Wiley & Sons, Lachin JM. Worst-rank score analysis with informatively missing observations in clinical trials. Control Clin Trials 1999; 20: Collard HR, Brown KK, Martinez FJ, Raghu G, Roberts RS, Anstrom KJ. Study design implications of death and hospitalization as end points in idiopathic pulmonary fibrosis. Chest 2014; 146: Collard HR, Brown KK, Martinez FJ, Raghu G, Roberts RS, Anstrom KJ. Study design implications of death and hospitalization as end points in idiopathic pulmonary fibrosis. Chest 2014; 146: Collard HR, Yow E, Richeldi L, Anstrom KJ, Glazer C. Suspected acute exacerbation of idiopathic pulmonary fibrosis as an outcome measure in clinical trials. Respir Res 2013; 14: Lee JS, Song JW, Wolters PJ, et al. Bronchoalveolar lavage pepsin in acute exacerbation of idiopathic pulmonary fibrosis. Eur Respir J 2012; 39: Han MK, Murray S, Fell CD, et al. Sex differences in physiological progression of idiopathic pulmonary fibrosis. Eur Respir J 2008; 31: Ley B, Ryerson CJ, Vittinghoff E, et al. A multidimensional index and staging system for idiopathic pulmonary fibrosis. Ann Intern Med 2012; 156: Published online August 9,