U.S. Regulatory Perspective on the Minimal Clinically Important Difference in Chronic Obstructive Pulmonary Disease

Abstract

This paper outlines the regulatory issues surrounding the determination and use of minimally clinically important differences (MCID) in assessing measures of outcomes from treatments of chronic obstructive pulmonary disease (COPD). To place this discussion in context, it is important to understand the current state of approved therapies for COPD, as well as newer directions in therapy. This paper discusses the currently available, approved drug therapies for COPD in the United States and how they were approved. This is followed by an overview on the use of MCID for assessing outcomes in therapies for COPD, as well the more general experience with MCID from the U.S. regulatory perspective.

• COPD
• Regulation
• Drug
• Clinical trials
• Endpoints
• Viewpoint

FDA Perspective on Current COPD therapy

COPD is an important disease, affecting a large and growing percentage of the U.S. population [[1]]. Despite the morbidity and mortality burden from COPD in the United States, there are limited numbers of approved drug therapies for this indication. To date, most of the therapies specifically approved for use in COPD in the United States have been indicated as bronchodilators. These drugs are intended to treat the bronchospasm or increased airway tone that often, though not invariably, affects patients with COPD. These drugs include the anticholinergics (ipratropium bromide and tiotropium) and the beta-agonists (salmeterol xinafoate and fomoterol fumurate), as well as combination drugs containing both classes (i.e., ipratropium bromide and albuterol). Most recently, a single dose level of a fixed-combination product containing salmeterol xinafoate with the corticosteroid fluticasone propionate was also approved for treatment of COPD. Since all of these agents either partly or wholly exert their effects through bronchodilation, the primary assessments of efficacy were measures of airflow obstruction—namely the FEV₁. This results for most of these agents in fairly specific, limited claims such as for the treatment of “bronchospasm associated with COPD (including emphysema and chronic bronchitis) [[1]].” This indication reflects both the primary mode of action of the drug along with the assessments conducted by the drug manufacturer.

While the meaning of degrees of change in FEV₁resulting from a therapeutic intervention may be a matter of some debate, the FEV₁is a well accepted clinical measure with which physicians and other caregivers who care for COPD patients are well familiar. Many physicians, including those assessing the clinical trials data at the FDA, have a long history of assessing COPD patients over time with serial spirometry and a clinical comfort with the meaning and performance of that measure as a clinical assessment. Despite the FEV₁not always being well-validated in a strict sense for use as it is employed in these clinical trials, it is “validated” by long years of use in the clinic and through multiple clinical trials over many years—both for drug development, as well as for other scientific purposes. For a drug such as the combination product containing a long-acting beta-2 agonist and a corticosteroid, the assessment of FEV₁ is done bothacutely (i.e., how FEV₁ changes acutely following a single-dose) and chronically (i.e., how basal FEV₁ changes at trough drug levels over time). The former is intended to assess directly the bronchodilation effect of the drug, the latter is intended to assess changes in the chronic narrowing of the airways that may better reflect the “anti-inflammatory” actions of the corticosteroid on the airways.

It is important to point out that in addition to assessments of FEV₁ and other spirometric metrics (e.g., forced vital capacity or midexpiratory flow rates) in clinical trials for bronchodilator treatments, other secondary endpoints are employed, many of those quite commonly, such as symptom scores (as assessed by various symptom scoring instruments), rescue bronchodilator use (either beta agonist or anticholinergic), functional assessments (such as 6-minute walks) and exacerbation rates. Interestingly, for the drugs approved to date, these secondary assessments have often shown little or no additional evidence of efficacy above and beyond that demonstrated on the primary endpoint. Therefore, while the bronchodilators approved for COPD are proven to lead to improvement in airflow in patients with a reversible component to their COPD, this bronchodilation does not always lead to improvements in other assessments in how patients are affected by the disease in ways that are easily determined and documented in clinical trials. This, in part, is why the indication section of the labeling for these approved products is reasonably narrow and concise, making clear the drug's claim is for bronchodilation and not more broadly for other aspects of COPD.

A number of novel drugs are being developed for COPD and there are certainly many other molecular targets and disease pathways being explored for their potential to lead to new treatment modalities. Since these novel therapeutic approaches are often aimed at disease mechanisms other than bronchoconstriction, such as the underlying inflammation that is well documented to play a role in COPD, assessments of airflow may, in many cases, not be the best solitary assessment of outcome. Other clinical assessments, such as measures of lung volumes and residual capacity, exacerbation rates, disease-specific health-related quality of life instruments, symptoms, assessments of functional performance, and even changes over time in the lung parenchyma (as assessed by high-resolution CT scanning) are being proposed to the FDA and other regulatory agencies as either co-primary assessments along with serial FEV₁ or as sole primary assessments. FDA is certainly open to discussing the use of additional, novel endpoints in clinical trials for new treatments for COPD. Even where FEV₁ is retained as a primary or co-primary endpoint, the use of one or more of these additional assessments in clinical trials of novel agents may well add important information to the assessment of efficacy.

MCID and COPD Drug Trials

As new endpoints are developed and employed in clinical trials as primary assessments of outcomes from therapeutics, it is important that these novel endpoints be validated for that use. From a regulatory point of view, this means not only understanding how the endpoints perform in clinical trials in defined populations, but also gaining insight into the clinical meaning of results obtained with the endpoint. This latter part of accepting an endpoint for regulatory purposes is not always straightforward, particularly if the endpoint itself is not derived from or commonly used in clinical practice. While a physician may feel that she or he has some innate sense of what a particular change in spirometric measure might mean to their patient, it may be much harder for a clinical physician or, indeed, a regulatory physician to make sense of a certain, defined amount of change in a novel assessment tool (for instance, a newly created multi-domain assessment of patient outcome).

While the FDA does not generally require that an MCID be established for endpoints (whether established or new), for novel endpoints, defining a level of change that would have clinical import to the patient or their caregiver is useful for placing the findings of trials utilizing that endpoint into perspective. That kind of information on clinical relevance of a change is very helpful in helping to assess the risk-benefit balance for a novel therapeutic.

There are many ways one can approach exploring and establishing an MCID for a new assessment tool. To date, FDA has seen the use of correlations to other accepted measures (e.g., changes in FEV₁), anchoring to changes that matter to patients in other established or global ratings, the use of statistical-based approaches (e.g., based on the standard error of the mean) and other means. From the perspective of the FDA, there is no absolute “right” way to establish an MCID. Many different techniques can be used and each method may offer complementary value in elucidating an MCID. When a newly created and validated instrument is proposed for use in a clinical trial, particularly when it is proposed as a primary, co-primary, or important secondary endpoint, it is helpful to have the MCID pre-specified in the clinical protocol. This can help the FDA in assessing the design of the trial (e.g., in interpreting whether the power analysis for the trial is appropriate for the instrument in question) and assessing the planned analyses (e.g., whether the MCID may be used to define “responders” in a responder analysis).

There is neither a consensus within FDA nor any FDA policy as to how the MCID is most appropriately applied in interpreting study results, either for use in COPD or in other disease states. Across the agency, experience has varied. While the MCID has at times been used as a firm cut-off applied to the difference between the point estimates of the means of two comparative treatment groups, it has also been used to define “responders” where the comparison then is the difference in response rates between the test treatment and its comparator (e.g., placebo). Lastly, there have been instances where MCIDs have not been used in the statistical interpretation of the results, but primarily for giving context to changes seen in an endpoint. Since no policy exists in thisregard, sponsors interested in utilizing a new measure and in establishing and applying an MCID for that measure in COPD patients should speak to the relevant review division(s) prior to embarking on pivotal clinical trials utilizing that measure.

Conclusions

In summary, the FDA has seen renewed interest in new therapeutics for COPD and, given the consequences of this important disease and the current limitations in drug therapeutics for addressing it, we are heartened by this renewed interest. As new therapeutics have been proposed for development, we have seen a number of novel endpoints proposed for use in trials for drugs that are not expected to have a primary bronchodilatory effect. In interpreting the design, analysis, and results of clinical trials of drugs of novel mechanisms that may be assessed by novel endpoints, establishing a level of change that is important to patients (i.e., the MCID) is very helpful for placing the results into a clinical and regulatory perspective and thereby informing the risk-benefit analysis of that new therapeutic.