Guidance Recap Podcast | Multiple Endpoints in Clinical Trials
Thank you for joining us for another episode of the Guidance Recap Podcast. The Guidance Recap Podcast provides highlights for FDA guidance documents straight from the authors. My name is Kylie Haskins, and I am the host for today’s podcast. In today’s episode, I am excited to be talking with Dr. John Lawrence, a statistician in CDER’s Office of Biostatistics. Dr. Lawrence will be sharing some thoughts with us on the newly published final guidance titled, “Multiple Endpoints in Clinical Trials.” Welcome, Dr. Lawrence! Thank you for speaking with us today.
For listeners who are less familiar with multiple endpoints, can you provide background on the topic?
Sure, let me begin by explaining the general type of endpoint discussed in this guidance called an efficacy endpoint. In a standard clinical trial used to evaluate the effectiveness of a new drug or biological candidate, half the people enrolled in the trial are given the treatment and the other half are given a placebo. Then specific outcomes of what happens to people in the trial are measured, these are called efficacy endpoints, and the efficacy endpoints are compared between the two groups to assess the effects of the candidate treatment compared with placebo. In short, efficacy endpoints are measures designed to assess the intended effects of a drug. For example, in a clinical trial for a potential drug to treat the flu, a reduction in fever is a common efficacy endpoint. Because most diseases can potentially cause more than one altered health function, assessing multiple endpoints in a clinical trial is common.
A good example of a disease where multiple endpoints are necessary to evaluate the efficacy of a drug is treatment of migraine headaches. One symptom of migraine headaches is pain, but it is often accompanied by nausea, photophobia (which is sensitivity to light), and/or phonophobia (which is sensitivity to sound). Treatments must show an effect on both pain and the accompanying symptoms. One approach is to have two co-primary endpoints. The two co-primary endpoints are (1) having no headache pain at 2 hours after dosing and (2) a demonstrated effect on the most bothersome migraine-associated symptom. Patients are asked (before treatment) to identify their most bothersome migraine-associated symptom in addition to pain and then asked 2 hours after taking the drug if the symptom was relieved.
A second example is evaluation of drugs for treating heart disease. There can be a high mortality rate and also a high rate of hospitalizations for heart failure. An endpoint used in a trial to demonstrate efficacy for treating heart disease can include both of these endpoints. For example, a comparison (between the drug group and control group) of the time to the first occurrence of either endpoint, or the total number of all hospitalization and deaths.
The inclusion of multiple endpoints is a valuable clinical trial design option. However, assessing more than one endpoint in a clinical trial can lead to something called multiplicity, which is an increased likelihood of making false conclusions about a drug’s effects when multiple comparisons occur in a clinical trial. Multiplicity can occur because there is a chance of making a mistake during the assessment of each endpoint in a trial, and these chances of making a mistake can add up when assessing multiple endpoints if the appropriate statistical adjustments are not made. If there is a 10% chance of rain today and a 10% chance of rain tomorrow, then the chance of rain on either day is close to 20 percent. The chance of making a false conclusion with multiple endpoints adds up in a similar way. This guidance provides statistical methods for managing multiplicity within a study with multiple endpoints to control the chance of making erroneous conclusions about a drug’s effects.
This guidance discusses grouping endpoints in clinical trials. Can you explain more about this?
The decision on grouping endpoints is related to either establishing effectiveness to support approval or demonstrating additional meaningful effects. There is a bit of a hierarchy when grouping endpoints based on the endpoint’s clinical importance. The highest group in the hierarchy are endpoints that are absolutely required to get approval, and those are what we call primary endpoints. If a candidate fails to show an effect on those, it won’t get approved
Next in the hierarchy are secondary endpoints, which can support the primary endpoint(s) or demonstrate additional clinically important effects. Failure to show an effect on a secondary endpoint does not doom a candidate treatment. However, if an effect is demonstrated, it would be valuable for a sponsor to include that information in labeling.
The third category in the hierarchy includes all other endpoints, which are referred to as exploratory or tertiary. Sponsors may not have confidence that they will see an effect, and these endpoints may be studied for research purposes or for new hypotheses generation. They aren’t likely to put the resulting analysis in the labeling.
Can you review the circumstances for using different types of multiple primary endpoints in a clinical trial?
There are different situations for using multiple primary endpoints, which are the basis for concluding that the study has met its objectives. For some candidates, an effect would be required for both primary endpoints for a potential approval. With these co-primary endpoints, the chance of making an erroneous conclusion of efficacy is reduced compared with the use of any single endpoint. However, there is an increased chance of failing to detect beneficial effects if the appropriate statistical adjustments are not made. Then there are instances with two or more endpoints in a study when each endpoint by itself would provide sufficient justification for FDA to approve the candidate. These are called multiple primary endpoints. An example would be a cardiovascular drug candidate that could reduce the risk of a heart attack and lower the risk of being hospitalized for heart problems.
There are also instances when important clinical outcomes are combined into a single primary endpoint called a composite endpoint. In the example of heart disease, you could make up a composite endpoint that combines the effect on heart attacks and the effect on hospitalizations. You could assess how many heart attacks and hospitalizations each patient had and add those together to come up with a composite. There might be some uncertainty at the conclusion of the study whether the effect was on both the endpoints together or only on one of the endpoints, but it was of sufficient magnitude that it showed an effect on the entire composite.
Lastly, there are instances where within-subject clinical outcomes are combined into a single primary endpoint called a multi-component endpoint. A multi-component endpoint is similar to a composite endpoint except the multiple effects combined in the endpoint occur within each patient to give each individual an overall rating. For example, the primary endpoint in clinical trials of allogeneic pancreatic islet cells for Type 1 diabetes mellitus can be a response rate in which patients are considered responders only if they meet two dichotomous response criteria: normal range of HbA1c and elimination of hypoglycemia.
What are some of the reasons FDA issued this guidance?
From our point of view, the FDA thought it would be imperative to offer consistent advice regarding the use of multiple endpoints to industry and have a resource that FDA reviewers can provide to sponsors when questions about multiple endpoints arise. This guidance builds on the International Council for Harmonisation guidance for industry E9 Statistical Principles for Clinical Trials, which was released in September 1998. We aimed to provide greater detail on multiple endpoints than what was discussed in that document. It also corresponds with an FDA commitment under the Food and Drug Administration Amendments Act of 2007.
The draft version of this guidance was published in 2017. We received public comments from 20 groups and individuals on the document, which were mostly professors and individuals or groups from the pharmaceutical industry. One interesting debate was about the role of secondary endpoints. The FDA allows additional claims based on the evaluation of secondary endpoints whereas the European Medicines Agency does not. There was discussion about the pros and cons of controlling the familywise error rate for all endpoints versus other ways of defining the error rate. Another issue discussed in the comments was that the document did not say much about the impact of other sources of multiple comparisons in clinical trials such as multiple doses, multiple treatments, multiple time points of interim analyses prior to study completion. In order to keep the document a manageable size, the writing group decided to focus only on multiple endpoints. For the final version of the guidance, FDA reviewed each comment received on the draft guidance and thoroughly addressed each item, pared down the document to provide more concise recommendations, and moved some items to the appendix.
Can you mention a couple of key items that you would especially like the audience to remember?
Clinical trials are complicated, with sponsors commonly using multiple endpoints to measure many different things to assess the effectiveness of a drug candidate and for potential inclusion of important information in the label. When more than one endpoint is analyzed in a single trial, the likelihood of making false conclusions about a drug’s effects can increase due to an effect called multiplicity. FDA issued this guidance to describe various strategies for grouping and ordering multiple endpoints and for applying some well-recognized statistical methods to adjust for multiplicity to make sure the chance of erroneously saying that a drug is effective is close to zero.
Dr. Lawrence, thank you for taking the time to share your thoughts on the multiple endpoints final guidance. We all have learned so much from your experience and insights in this area, and we appreciate the hard work that you have invested to inform safe and effective use of the drugs and biologics we regulate. We would also like to thank the guidance working group for writing and publishing this guidance.
To the listeners, we hope you found this podcast useful. We encourage you to take a look at the snapshot and to read the guidance.