Treating IPF: An Analysis of 3 Clinical Trials

Andrew F. Shorr, MD, MPH


August 29, 2012

This feature requires the newest version of Flash. You can download it here.

This is Andrew Shorr from Washington, DC, with a pulmonary and critical care literature update. I would like to talk about 3 recent studies in idiopathic pulmonary fibrosis (IPF) that I think are crucial for pulmonologists who care for these patients to appreciate. The first 2 are randomized controlled trials of intervention therapies, and the third is a severity-of-illness or staging tool for IPF.

The first clinical trial is the PANTHER-IPF trial, published in the New England Journal of Medicine,[1] and the second is the ACE-IPF trial, published in July in the American Journal of Respiratory and Critical Care Medicine.[2] Both were done under the auspices of the Idiopathic Pulmonary Fibrosis Clinical Research Network, funded by the National Institutes of Health. This was a multicenter collaborative effort to improve our clinical trial experience and clinical trial effort as it relates to IPF because, for a long time, much of the evidence related to IPF was from very low-quality glorified case series and observational studies.

PANTHER is a large randomized controlled trial that has 3 arms, 2 of which are ongoing. One arm is looking at N-acetylcysteine alone, which has been used in IPF based on an earlier study in the New England Journal of Medicine[3]; the second is a placebo arm; and the third is an arm that combined N-acetylcysteine with azathioprine and prednisone. The combination arm reflects what, anecdotally, a number of people have been using for this disease, because there has been some small, randomized controlled trial data supporting azathioprine and prednisone.[4]

It is important to realize that in this trial with approximately 150 patients in the arm comparing combination therapy with placebo, there was an increased risk for death with combination therapy. When you look at the study, which only reports outcomes for N-acetylcysteine with prednisone and azathioprine vs placebo, you see a population that kind of has moderate lung disease impairment. Forced vital capacities (FVCs) were in the 50%-60% range, and diffusing capacities of the lung for carbon monoxide (DLCOs) were in the 40%-45% range. Clear mortality imbalances led the Data and Safety Monitoring Board to stop this study early.

Rates of death in patients receiving combination therapy with those 3 drugs were substantially higher, as were rates of hospitalization. In addition, no signal or potential signal of physiologic benefit with combination therapy vs placebo was found. This clearly indicates that at this point, pulmonologists should not be using triple combination therapy for this disease state. In fact, it is harmful. The placebo arm is continuing to be followed, and enrollment and observation are continuing in the arm with only N-acetylcysteine.

The second randomized controlled trial, also from the IPF network, is ACE-IPF.[2] It looked at anticoagulation with warfarin in IPF. There is a strong physiologic basis for thinking about anticoagulation as important in fibrosis because activation of the clotting cascade can serve as a template for letting down a fibrosis. We know that the clotting system is active in a number of interstitial lung diseases, ranging from idiopathic usual interstitial pneumonia or IPF to sarcoidosis. Therefore, anticoagulation has been used. In addition, data suggest that the risk for venous thromboembolism is higher in patients with IPF, particularly following transplant. Therefore, there are sound physiologic and biologic reasons for looking at anticoagulation in this disease state.

In this study of approximately 150 patients, half were randomly assigned to placebo and half were randomly assigned to warfarin. What the investigators saw were increased risks for death and hospitalization in patients assigned to warfarin. Most of those deaths were categorized as respiratory events. The population in ACE-IPF was slightly sicker than the population in PANTHER. FVCs and DLCOs were a little bit lower. Again, a mortality disadvantage or signal of harm was seen in the intervention arm.

There was no signal of any potential physiologic benefit in terms of the mechanism of action for warfarin, and it was not evident in any of the prespecified subgroups that the authors investigated, whether it was based on DLCO or the need for supplemental oxygen. Now we have 2 well-done randomized controlled trials -- for what was thought to be a disease state that we could never do a randomized controlled trial for -- that show that what was considered an acceptable standard or what was potentially evaluated as something to consider in our most recent IPF guideline document is actually harmful and should not be utilized.

One of the things that has hampered efforts to do clinical trials for this disease state is a tool for ensuring that patients are equally severely ill. We don't have one global tool that captures the risk for mortality or severity of disease in IPF. FVC is a poor marker and DLCO is a poor marker. They certainly capture aspects of things; a 6-minute walk captures aspects of the risk for death in this disease state as well. The best predictors tend to be physiologic ones from looking at exertional activity by using a cardiopulmonary exercise test or something similar. However, that is cumbersome.

That led to a study by Ley and colleagues[5] that was published in the Annals of Internal Medicine several months ago. The investigators tried to develop and then validate a severity-of-illness staging system for idiopathic pulmonary fibrosis. They looked at an inception cohort of patients with IPF, developed their severity-of-illness scoring tool looking at mortality, and then validated it on a separate cohort of patients. The severity-of-illness scoring tool, or the GAP tool, looked at 3 variables: gender, with male gender being associated with worse outcomes; age, with age > 65 years being associated with worse outcomes; and physiology variables in terms of FVC and DLCO, with more points added for worse FVC or DLCO. If you had a relatively normal DLCO and FVC, you were young, and you were female, your 1-year mortality was relatively low: 6% vs something 5 times higher if you had the worst potential score.

When the investigators validated the severity-of-illness scoring tool in a separate population, it seemed to perform adequately and segregated groups into low risk for short-term mortality, higher risk for short-term mortality, and even intermediate risk for mortality. One problem with this study was that it did not incorporate ongoing physiologic measures such as the 6-minute walk or the response to therapy over time. However, there are changes over time that seem to herald what is going on in the future of this disease. Patients seem to have distinct phenotypes when they have a trajectory that is more rapid than others who have a more flat trajectory.

The C-statistic, or the assessment of how well the tool fit the data or how well it explained the variability, was about 0.7. For most scoring tools for severity of illness or prediction rules, we like the C-statistic to be in the 0.80-0.85 range. That is almost like the area under the curve in terms of how well it explains the data. There is a lot that this mortality prediction rule, although it is really simple, still does not explain.

I think this is a crucial and important first step. It kind of tells pulmonologists what we already know: that the worse you look, the worse you are in terms of what we know of physiology. However, I think there is still a lot of work to do.

This is Andy Shorr from the Washington Hospital Center.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.