The Efficacy and Mechanism Evaluation of Treating Idiopathic Pulmonary fibrosis with the Addition of Co-trimoxazole (EME-TIPAC): study protocol for a randomised controlled trial

Patients of either gender, aged 40 years or older, are being entered into the trial. Patients are considered to have IPF following a multidisciplinary team (MDT) consensus decision undertaken at a specialist centre (or MDT otherwise meeting the criteria of a specialist centre) following a review of an appropriate clinical history, characteristic features of usual interstitial pneumonia (UIP) on thoracic, high-resolution computed tomography (HRCT) and/or UIP histology confirmed by surgical lung biopsy according to the latest international guidelines [6]. There is no maximum time limit between the diagnosis of IPF and enrolment. Patients may receive orally administered prednisolone at a dose of up to 10 mg per day (in keeping with previous studies [11]), anti-oxidant therapy, pirfenidone, nintedanib or other licensed medication for IPF. If taking licensed medication for IPF, patients should be on a stable treatment regimen for at least 4 weeks prior to the screening visit to ensure that baseline values are representative. Patients must also have a Medical Research Council (MRC) Dyspnoea Score [16] of greater than 1, with this being defined as ‘Not troubled by breathlessness except on strenuous exercise’, at screening in order to exclude asymptomatic patients.

Patients with a FVC of more than 75%, as predicted based on the algorithm suggested by Crapo et al. [17] are not recruited into the trial; this is intended to exclude patients with mild disease and, instead, to identify patients more likely to meet the primary endpoint (see ‘Sample size’ section). Other reasons for exclusion include (1) a recognised significant co-existing respiratory disease, e.g. obstructive airways disease (defined as forced expiratory volume in 1 s (FEV₁)/FVC < 60% [18]), (2) a significant medical, surgical or psychiatric condition that would affect subject safety or influence the trial outcome or (3) untreated folate or vitamin B₁₂ deficiency, to ensure that no bone marrow or neurological adverse effects occur with folate therapy to vitamin B₁₂-deficient individuals. Sulphonamides are recognised to increase the risk of haemolysis in individuals with glucose-6-phosphate dehydrogenase (G6PD) deficiency although this risk remains low [17]. The prevalence of G6PD deficiency is higher in men of African, Asian or Mediterranean descent and, therefore, patients in these demographics are screened for the condition and, if found to have G6PD deficiency, are excluded from the trial. Patients with a self-reported respiratory tract infection within 4 weeks of screening (defined as two or more of cough, sputum or breathlessness and requiring antimicrobial therapy) are not eligible due to the difficulty of obtaining reliable baseline lung function. Those patients receiving a short course of antibiotics (for any indication) within 4 weeks of screening or long-term prophylactic antibiotic treatment (defined as more than 1 month of therapy) within 3 months of screening are also ineligible as this may have an impact on lung microbiota. A risk of sudden death in patients treated with co-trimoxazole and angiotensin-converting enzyme (ACE) inhibitors was identified [19] and is thought to be due to an increase in serum potassium due to an amiloride-like action of trimethoprim on sodium channels in the distal nephron [20]. For this reason, patients with a serum potassium greater than 5.0 mmol/L are excluded and patients with a baseline serum potassium of between 4.7 and 5.0 mmol/L, who are 66 years old or over and taking potassium-sparing diuretics (including ACE inhibitors or angiotensin-receptor blockers), are required to have an extra blood test to measure potassium 1 week after starting trial treatment.

The main method of patient identification is by review of ILD MDT meeting minutes or summaries, but is also via screening patient registries, hospital medical records and databases of research-interested patients. Potential recruits are being approached by local clinic teams and provided with a patient information sheet and given at least 24 h to read this prior to consent. Consent is taken by appropriately-trained clinicians or delegated members of staff. All patients are invited to provide additional consent to provide a sample of blood for deoxyribonucleic acid (DNA) analysis; however, this is optional and patients can refuse whilst remaining eligible for the main trial.

Following consent, patients meeting all inclusion criteria and none of the exclusion criteria (after review of their screening bloods) may be randomised without a subsequent visit. Randomisation is performed centrally according to a computer-generated randomisation code with the treatment group allocation sent to research pharmacists only. Minimisation factors are: research site and the use of baseline licensed medication for IPF.

Participants receive co-trimoxazole (generic) 960 mg as two tablets of 480 mg twice a day plus folic acid 5 mg once a day or placebo tablets (manufactured to appear identical to co-trimoxazole 480 mg) two oral tablets twice a day plus folic acid 5 mg once a day over a median treatment period of 27 (range 12–42) months. This is dispensed 3-monthly for the first 6 months then 6-monthly and supplied in bottles providing 1 month’s supply either by hospital pharmacy or via a courier. Participants are advised to store their medication below 25 °C but there will be no temperature monitoring after dispatch to the patient. Folic acid must be taken by participants whilst on the trial drug to reduce the risk of bone marrow depression associated with the long-term use of co-trimoxazole.

The trial drug may be reduced to two tablets once a day three times a week plus 5 mg folic acid once a day three times a week if patients develop gastrointestinal adverse effects or rash, have grade 1 hyperkalaemia (serum potassium > 5.0 mmol/L and < 5.5 mmol/L) or any other adverse event that, in the opinion of the local investigator, requires a dose reduction. In such cases, the dosing interval is to ensure that the dosing is spread throughout the week (e.g. Monday, Wednesday and Friday or equivalent). Once a patient has had their dose reduced, no re-escalation will be permitted, even if the adverse event leading to the reduction resolves.

The primary outcome is the time to death (all causes), transplant or first non-elective hospital admission. Individuals who withdraw will be censored on the date of withdrawal.

The individual components of the primary outcome: time to death (all causes) or transplant and time to first non-elective hospital admission will be analysed separately as secondary outcomes. In addition, respiratory-related events will be analysed separately from non-respiratory-related events. The following measurements will also be undertaken at baseline, 3 and 6 months post randomisation, then 6-monthly for the duration of the trial plus at the end of trial/hospitalisation.

This is being assessed in a number of ways: (1) the King’s Brief Interstitial Lung Disease (K-BILD) health-related quality of life questionnaire [21] which is the only validated ILD-specific instrument. This 15-question, self-completed patient questionnaire has a mean score of 53 (standard deviation 26) units in IPF and a minimum clinical significant difference of 8 units; (2) the MRC Breathlessness Score [16] which is the most widely utilised dyspnoea score for patients with ILD; (3) the EuroQol 5-dimension (EQ5D) questionnaire [22] will be used to determine quality-adjusted life years (QALYs) but a cost-utility analysis will not be undertaken. The K-BILD does not specifically capture cough so we are assessing cough using a Visual Analogue Scale. Overall, quality of life is being captured using a 6-point Likert global rating of concept scale.

Spirometry [23] and the total lung-diffusing capacity of carbon monoxide (DLCO) [24] are being measured according to current American Thoracic Society/European Respiratory Society guidelines. FVC and DLCO are both components of prognostic modelling algorithms [25], are frequently utilised in clinical trials [9] and are part of routine care.

Peripheral blood is being taken at baseline, 3, 6 and 12 months and at the end of the trial and stored for analysis of measures of (1) infection/inflammation, including C-reactive protein (CRP), which is an acute-phase serum protein that is increased in concentration in patients with inflammation and is a significant prognostic indicator for survival in patients with IPF [26]; (2) alveolar epithelial injury including surfactant protein (SP)-D which is elevated in serum in patients with IPF [27] and predicts mortality [28] and matrix metalloproteinase (MMP)-7 which is related to disease severity [29, 30] and is an independent predictor of mortality [31] and (3) neutrophil activity, including myeloperoxidase, which is almost exclusively expressed in neutrophils and is a marker of their activation and degranulation.

Sputum is being obtained, where possible, and sent for local microbiological culture and susceptibility testing; and a nasal swab will be sent for viral culture, if clinically indicated, in all patients.

Alveolar epithelial cell injury markers (SP-D, MMP-7) will be measured as for peripheral blood samples. Neutrophil elastase is indicative of neutrophil activation and is related to disease activity in IPF [32]. Neutrophil elastase inhibitors have been shown to reduce the development of pulmonary fibrosis in animal models [33]. Elastase activity will be measured using the EnzChek elastase assay (Molecular Probes). Pro-collagen III N-terminal peptide (PIIINP) is a collagen turnover peptide and a surrogate marker of fibrosis. It is elevated in patients with IPF [34] and is higher in patients with progressive and non-progressive IPF [35].

Patients are being asked to carry a card detailing their involvement in the trial and their local investigator contact details. This is for safety reasons and in order to maximise ascertainment of follow-up during hospital admission. Questionnaires, blood samples and routine microbiology are being collected in the same manner as undertaken at the routine visits.

Blood for full and differential white cell counts, urea and electrolyte assays and liver function assessment is being taken at baseline, 6 weeks, 3, 6, 9 and 12 months then 6-monthly for the duration of the trial and at the end of trial/hospitalisation. The 6-week and 9-month blood tests may be taken in primary care, with assessment of adverse events being captured by phone.

This trial complies with UK NHS Research Ethics Service guidelines on reporting of adverse events (National Research Ethics Service Safety and progress reports (Clinical Trials of Investigational Medicinal Products (CTIMPs)); http://​www.​hra.​nhs.​uk/​research-community/​during-your-research-project/​safety-reporting/​). As hospitalisation and death are outcome variables they will only be regarded as serious adverse events if they are drug related.

With 264 patients predicted to be randomised over a period of 30 months and an additional 12-month follow-up after the last patient is recruited (a total of 42 months after the first patient is enrolled, median patient trial duration of 27 months) we expect 96 primary endpoint events to occur uniformly over the duration of the trial. This will have 80% power (two-sided test, significance level of 5%) to show a change in hospitalisation-free survival from a median value of 28.8 months in the control arm to 51.1 months in the co-trimoxazole arm (hazard ratio of 0.56) using a log-rank test. This is based on a sensitivity analysis of all patients from TIPAC (including IIP and/or IPF) with reduced lung function (FVC < 75% predicted) using an intention-to-treat analysis.

All analyses will be conducted according to a detailed statistical analysis plan. Analyses will be adjusted for site and the use of baseline, licensed medication for IPF. The analysis populations are defined as intention-to-treat (all randomised individuals regardless of adherence), per-protocol (all randomised individuals who adhere to the trial medication to within 80% (based on pill counts)), modified-per-protocol (all randomised individuals who adhere to the high-dose regimen) and safety population (all patients randomised who received at least one dose of the trial treatment). Non-compliance will be dealt with using a Compliance-Adjusted-Causal-Effect (CACE) analysis, using compliance data from returned medication (pill counts).

The safety analysis will be based on the pre-defined population (as above). Summary tables will be presented for incidence rates (number of patients with at least one incidence) of adverse events and serious adverse events coded according to the Medical Dictionary for Regulatory Activities (MedDRA). Tables of change from baseline will be presented for the blood and other clinical laboratory assessments.

Requests for access to trial data and stored samples will be considered, and approved in writing where appropriate, after formal application to the TMG and TSC.

Full details of the analysis will be finalised in an analysis plan before database lock.

Analysis of the biomarkers from stored blood will be performed using the same linear mixed model as for the analysis of K-BILD. Analysis of routine microbiology and 16S ribosomal ribonucleic acid (rRNA) sequence data will be descriptive by tabulating the different types of microbiological cultures and their relative prevalence or signal strength.

The results of the trial will be published regardless of the direction of effect.

All patients participating in the trial will, at the end of the trial, be provided with a letter detailing their treatment allocation with a lay summary of the trial outcomes. General practitioners of patients participating in the trial will also be given a copy of the trial results.

We modified the protocol in February 2015 as we became aware of new data regarding the risks of hyperkalaemia with co-trimoxazole [19]. We started to exclude participants with a serum potassium greater than 5.0 mmol/L and increase monitoring of participants with a baseline serum potassium of between 4.7 and 5.0 mmol/L who are 66 years old or over and taking potassium-sparing diuretics.

We modified the protocol in May 2016 to widen recruitment criteria to make the trial more generalisable, taking in to account feedback from sites, the TMG and TSC. We removed a requirement for participants to be diagnosed within 2 years of entry into the trial. This exclusion criterion was originally included to prevent stable patients entering into the trial; however, the time of diagnosis was sometimes difficult to determine and also patients with stable disease are unlikely to have significantly reduced lung function (FVC). We increased the maximum FVC value from 70% predicted to 75% predicted as this permitted increased recruitment without reducing the anticipated event rate. We also permitted the 6-week and 9-month visits to be undertaken via the phone with local assessment of safety bloods in order to reduce the patient burden and increase acceptability.

The protocol also included a bronchoscopy sub-study at a small number of selected sites, at which bronchoalveolar lavage fluid would be obtained from a sub-set of 50 participants at baseline and then at 3 months post commencement of trial treatment. However, following recommendations by the DMC and TSC, recruitment to the sub-study was suspended due to a lack of recruitment. As a result it has not been included in this paper.