Editorial
May 19, 2024
When the Third Time Is Not the Charm—Trial Outcomes in Idiopathic Pulmonary Fibrosis
Ana C. Zamora, Victor E. Ortega, Eva M. Carmona
JAMA. Published online May 19, 2024. doi:10.1001/jama.2024.8776
Idiopathic pulmonary fibrosis (IPF) is a progressive fibrotic lung disease that inevitably results in respiratory failure and death within 3 to 5 years of diagnosis.1 No effective treatment for IPF existed until the approval of 2 antifibrotic drugs (pirfenidone and nintedanib) more than 10 years ago. These drugs were approved by the US Food and Drug Administration based on demonstration of significant drug-associated differences in slowing the decline in lung function in 2 randomized clinical trials2,3; lung function was measured by forced vital capacity (FVC), which remains the main clinically relevant outcome for IPF.
Longitudinal decline in FVC is an indicator of disease progression in patients with IPF, and a decrease by 5% to 10% in the decline of FVC per year is associated with increased mortality.4 Issues with poor tolerability (nausea, profound fatigue, diarrhea, and rash) and limited efficacy (mostly by reducing the rate of progression without significant changes on how the patient feels) for the antifibrotic drugs pirfenidone and nintedanib leave patients with IPF with a similarly poor prognosis as there is no definite disease-altering or curative treatment apart from a lung transplant.
New therapeutics for IPF have been widely anticipated, but the phase 3 randomized clinical trials5,6 of 2 different therapeutic agents were prematurely halted due to safety concerns, lack of efficacy, or both. Both ziritaxestat and zinpentraxin alfa did not demonstrate beneficial effects in altering the decline of FVC compared with placebo.
In this issue of JAMA, Raghu and colleagues7 reported the findings from the ZEPHYRUS-1 trial, which is a phase 3, double-blind, randomized clinical trial including 356 patients with IPF (181 patients were treated with pamrevlumab, which is an inhibitor of the connective tissue growth factor, vs 175 patients in the placebo group). Pamrevlumab was initially developed as an anticancer drug and the rationale for its use in patients with IPF was based on its ability to target abnormal wound healing processes and inhibit fibrogenesis. Trial participants7 were required to be free of antifibrotic therapies prior to enrollment but, in contrast to the phase 2 trial, they could initiate additional treatment after randomization with either of the antifibrotic therapies (alone or both drugs plus the randomized drug).
The current trial7 did not demonstrate a significant difference between the pamrevlumab group and the placebo group for the primary outcome of absolute change in FVC from baseline to week 48 or differences for any of the secondary outcomes. The secondary outcomes were focused on individual and composite health care–related outcomes or death, machine learning–based quantitative pulmonary fibrosis measures derived from computed tomographic scans of the lungs, and patient-reported symptoms. The current findings7 contrast sharply with those from the phase 2 trials8,9 of pamrevlumab in which significant between-group differences were observed. The subgroup analyses in the current trial7 also were negative, including those additionally treated vs not treated after randomization with the 2 currently available antifibrotic therapies. These negative findings resulted in the termination of a parallel trial, ZEPHYRUS-2 (NCT04419558), and an open-labeled extension study.
Overall, the treatment groups were well matched for age, sex, baseline disease severity, and the proportion of patients who initiated concomitant treatment with 1 of the 2 antifibrotic therapies (or both) during the treatment period.7 There were several important innovative aspects of this trial,7 including evidence for the feasibility of home infusions and the integration of machine learning–based radiomic phenotypes to quantify interstitial fibrosis. The integration of patient-reported outcomes in the clinical trial design was well aligned with recommendations from the US Food and Drug Administration and recent calls4 for the evaluation of more meaningful patient-centered outcomes. Despite these strengths, the current trial7 is the latest example of a phase 3 trial of an IPF therapeutic agent that failed to show evidence for efficacy after more promising phase 2 studies.8,9
The failure of different investigational drugs to successfully progress from phase 2 to phase 3 is a fundamental problem limiting the development of much-needed effective treatments for IPF. One can surmise different contributing or causative factors leading to this recurring problem. First, attention must be paid to the primary outcome of slowing the decline in lung function measured by FVC, which is a complex and highly variable physiological phenotype influenced by extrinsic restriction, interstitial edema, and air trapping in individuals with airways disease—factors that are usually independent of IPF and its progression. The high interindividual variability in FVC is further exaggerated by the unpredictable course of IPF progression, the time necessary to detect significant differences, and the poor correlation with patient-reported outcomes and other functional parameters.4 Insufficiently powered phase 2 studies with outliers driving big changes in FVC also may have contributed to the misleading promising results.5,10 The continued use of FVC as a primary outcome in randomized clinical trials in IPF remains a topic of debate among the scientific community, industry partners, and funding agencies.4 Although it may seem practical to reduce the target difference in FVC from 10% to 5% to facilitate the detection of a significant effect (especially when antifibrotic therapies are introduced while patients are receiving a randomized treatment), this would be a step in the wrong direction. More effective therapeutics are needed beyond what the standard of care can provide. Idiopathic pulmonary fibrosis is a complex, systemic disease that may best be characterized with outcomes that integrate physiological variables, exacerbations, hospitalizations, and mortality outcomes with physical function, patient-reported outcomes, and the quantitative computed tomographic scan–based measures of lung texture and fibrosis (eg, Computer-Aided Lung Informatics for Pathology Evaluation and Ratings11 and blood biomarkers). This comprehensive approach considers outcomes that matter for the individual patient while enabling the detection of therapeutic responsiveness and the identification of responder subgroups.
Second, the heterogeneity of IPF in the current cohort,7 and in trials for IPF in general, may be a contributing factor in these negative trials. The diagnosis of IPF is based on clinical, radiographic, and histological criteria that do not consider disease heterogeneity.12 Idiopathic pulmonary fibrosis can result from different pathogenic mechanisms, including mendelian, heritable diseases that are identified in up to 30% of patients with IPF, most commonly from telomere or surfactant protein pathway genes or pathogenic variation associated with pleuroparenchymal fibroelastosis.13-15 Different forms of heritable IPF, common loci from genome-wide association studies (such as the MUC5B locus) and biomarkers (such as telomere length) are usually not evaluated in phase 2 or 3 trials. The ongoing PRECISIONS trial (NCT04300920) is randomizing patients with IPF to N-acetylcysteine or placebo based on the TOLLIP variant genotype. Stratifying for telomere length, genetic variations, polymorphisms (ie, MUC5B and TOLLIP), or even for individuals with different quantiles of an IPF polygenic risk score could identify subgroups that might be more likely to respond.16 For example, there is evidence that patients with MUC5B (rs35705950) variant T allele genotypes experience increased survival during treatment with antifibrotic agents compared with those with the homozygous GG genotype.17 Genetics is only one example of a biomarker of disease heterogeneity that can identify individuals who will respond to drugs. In patients with severe asthma, drug responder subgroups for biological agents are determined by peripheral blood eosinophil count.18
Third, the concurrent use of available antifibrotic therapies during randomized treatment must also be considered, which was a fundamental difference between the phase 2 trial of pamrevlumab8 and the current phase 3 trial of pamrevlumab.7 The introduction of antifibrotic agents in combination with pamrevlumab or placebo during the course of the current trial7 may have obscured detectable changes in FVC. The decision to permit the initiation of antifibrotic agents after randomization was likely due to ethical considerations that must be acknowledged in future trials because prior phase 2 trials were conducted when antifibrotic use was not a widely implemented standard-of-care therapy for patients with IPF.
In addition, one must consider how mechanistic, drug-discovery studies identify novel molecular targets and drugs and whether these approaches are identifying drugs relevant to critical IPF pathogenic mechanisms in humans. Pamrevlumab exhibited efficacy in preclinical tests, similar to other drugs that have advanced to phase 2 trials based on bleomycin- and radiation-induced murine models of fibrosis.19,20 Should the reliance on findings from animal model systems continue? Or should researchers look to human-based approaches (such as human organoid lung-in-a-chip technology) or multiomic approaches in lung tissue (such as proteomics and single-cell transcriptomics)?
There is a critical need to reconsider how phase 2 clinical trials are designed and conducted to inform phase 3 trials starting with molecular target conception to the testing of the most relevant outcomes for the individual patient. Phase 2 trials should be representative of the general population from which phase 3 trials should be drawn and test similar related outcomes.
More adaptive approaches should be used that consider diverse and composite outcomes that integrate physiological, functional, radiographic, and molecular biomarkers while considering heterogeneity and predictive biomarkers for responder subgroups to inform phase 3 trials. An example of this is the ongoing Precision Interventions for Severe and/or Exacerbation-Prone (PreCISE) Asthma Network by the National, Heart, Lung, and Blood Institute21 that integrates deep phenotyping and biomarker identification in an innovative adaptive platform across 5 simultaneous phase 2 and proof of concept trials. Adaptive trial designs that consider complex factors will enhance the understanding of IPF and its subtypes while accelerating the development of much-needed therapeutics for appropriate patients.
