• Research type

    Research Study

  • Full title

    A Randomized, Double-Blind, Placebo-Controlled, Phase 2 Study of Safety, Tolerability and Efficacy of Pirfenidone in Patients with Rheumatoid Arthritis Interstitial Lung Disease



  • Contact name

    Felix Woodhead

  • Contact email

  • Sponsor organisation

    University Hospitals of Leicester NHS Trust

  • Eudract number


  • Identifier


  • Duration of Study in the UK

    4 years, 0 months, 11 days

  • Research summary

    Summary of Research
    Rheumatoid arthritis (RA) is a chronic inflammatory disease that leads to joint damage, deformity, and a high rate of disability and unemployment. There is a lack of effective treatment options for RA. RA is seen more in women than men, it costs the economy millions of pounds a year and it can shorten a patients lifespan by approximately 10 years.

    Interstitial lung disease (ILD) is commonly seen in those patients with RA. ILD significantly contributes to increased mortality in patients with RA (a study has shown that patients with RA related ILD are at nearly 3 times the risk of premature death compared to patients without this complication).

    Current treatment for RA has achieved a great improvement in the control of articular disease, however, unfortunately these benefits have not extended to rheumatoid arthritis associated lung disease (RA-ILD). Pirfenidone treatment is safe and well tolerated in idiopathic pulmonary fibrosis (IPF). RA-ILD shares many characteristics with IPF and therefore pirfenidone may be of benefit in RA-ILD. This study is designed to evaluate whether pirfenidone 2403 mg a day reduces the decline in a patients lung function or mortality over 52 weeks, compared with placebo, in patients with RA-ILD.

    Summary of Results
    Scar tissue can build up in peoples’ lungs without an obvious cause. This leads to worsening breathlessness and usually, sadly, death in an average of three to five years. The commonest disease where this happens is called Idiopathic Pulmonary Fibrosis (IPF). The deterioration in IPF can be monitored by a reduction in the value of the volume of air a patient can breathe out called Forced Vital Capacity (FVC). Pirfenidone is from the class of drugs called antifibrotics which reduce the formation of scar tissue. Pirfenidone has been shown to slow the worsening of FVC in IPF.
    A similar condition to IPF can occur in roughly one in ten patients with rheumatoid arthritis, a condition where the body’s immune system attacks the joints and other organs. This is known as rheumatoid arthritis-associated interstitial lung disease (RA-ILD). When the TRAIL1 study (‘Treating RA ILD’) was planned, it was not clear if antifibrotic medications would be useful in RA-ILD, so the investigators set out to answer that question by giving study subjects who all had RA-ILD affecting at least 10% of their lungs on a CT scan (a 3-dimensional ‘body scan’ using Xrays) either pirfenidone, or an identical-looking inactive tablet called a placebo. This type of study is called a randomised, placebo-controlled trial, and is felt to be the best way of answering whether a drug works in a disease.
    It is thought that a 10% drop in FVC is particularly concerning and so the main single question on which we judge the significance of a trial (the ‘primary endpoint’) was chosen to be the proportion of subjects on either pirfenidone or placebo who either had a 10% FVC drop or died. This is known as a composite endpoint. When one compares the proportion of subjects reaching the primary endpoint based on whether they were given pirfenidone or placebo, any difference can be due to chance or due to the effect of the drug. The chances of detecting a difference due to the effect of the drug are increased by increasing the number of study participants. Statistical specialists can calculate how many subjects are need to show a difference in outcome that we can say is very likely to be due to the drug. This concept is known as the study power. When a study has fewer than this number of subjects, it can be possible that even if the drug does have some effect we cannot prove it.
    As well as looking at the primary endpoint, investigators record when they design a study a number of things that they want to compare between the groups. These are known as prespecified secondary endpoints. Although the overall success of a trial is usually assessed on a whether there is a difference in the primary endpoint that is unlikely to be due just to chance (‘statistically significant’), people reading the study report are often interested in these secondary endpoints too. Sometimes investigators see apparent differences between study groups in a different measurement that they had not planned to look at when they designed the study. One of these, so-called non-prespecified measurements was the effect of the drug depending on whether subjects had a CT appearance called Usual Interstitial Pneumonia (UIP). This may be important because UIP is the pattern seen in IPF, and may be associated with disease that gets worse quicker.
    The investigators had planned to recruit 270 RA-ILD patients to answer whether pifenidone reduced the proportion either dying or having a 10% FVC drop. Recruitment started in 2017 and took place in the USA, Canada, Australia and the UK. Unfortunately the trial was halted prematurely in March 2020 when the covid-19 pandemic meant that it was not safe for these clinically-vulnerable patients to attend study visits. At that time, only 123 subjects had been recruited. The fact that the number was much smaller than planned (the study was ‘underpowered’) meant that it would be very unlikely to be able to show a difference in the primary endpoint. We found that 11% of subjects taking pirfenidone either died or had a 10% FVC drop, compared to 15% of those taking the dummy tablet. It was not possible to say this was not just a chance difference.
    There was a difference, however, in the absolute size of the change in FVC over time between groups. After one year, the FVC of subjects taking pirfenidone had dropped by 66 ml, compared to 146 ml in those taking placebo. The worsening on drug was roughly half that seen on placebo, and statistical test show this is very unlikely just to be a chance finding. Another interesting observation relates to the comparison of subjects with the UIP CT appearance and those with other patterns. However, two factors make it harder to be sure there is a real difference. The first is that this analysis hadn’t been planned when the study was designed (it was not prespecified). The other is that, by chance, there was a difference between the proportion of patients allocated to pirfenidone with a UIP pattern (54%) and the proportion on placebo with UIP (78%). The annual worsening of FVC in patients with the UIP pattern on pirfenidone was 43 ml, whereas that of UIP patients on placebo was 169 ml. There was not a statistically significant difference between FVC worsening in subjects with a non-UIP pattern based on drug use.
    No difference in quality of life was seen in the groups. Roughly half of the patients on pifenidone had side effects which the local investigators felt could be related to their tablets, compared to about one fifth of those taking placebo. The most common side effects were nausea, fatigue and diarrhoea. About a quarter of subjects taking pirfenidone stopped treatment due to side effects, compared to one in ten on placebo. These differences were statistically significant.
    To conclude, the TRAIL1 study had to close prematurely due to covid, when far fewer subjects than planned had enrolled. It was not possible to prove that it reduced death or ‘significant’ lung function drop but it did appear to reduce the absolute decline in FVC loss by about half. This is similar to the effect of pirfenidone in IPF. In 2019, the INBUILD study showed that another antifibrotic, nintedanib, also reduced lung function decline in non-IPF lung fibrosis which had worsened despite conventional (not antifibrotic) treatment. 13% of the INBUILD population had RA-ILD. INBUILD met its primary endpoint, unlike TRAIL, and nintedanib is now being used in such patients, including some with RA-ILD. The TRAIL may provide some support to the case for antifibrotics in RA-ILD, although a significant proportion of patients had side effects causing them to stop treatment.

  • REC name

    East Midlands - Leicester South Research Ethics Committee

  • REC reference


  • Date of REC Opinion

    15 Sep 2017

  • REC opinion

    Further Information Favourable Opinion

Shape the future of the HRA website.