Randomized Clinical Trials > Primary Endpoint
A primary endpoint is the main measurement a trial is trying to assess. It answers the most important question in the trial. For example:
- In a weight loss study: What is the average weight loss after six months?
- In a trial for a new cancer drug: What is the average increase in survival time? Or, What is the time to death?
- In a migraine medicine trial: What is the number of migraines experienced by the patient?
Primary endpoints usually fall into one of three categories:
- Proportion of “successes” (i.e. cure or other successful treatment).
- Time to a “failure.” The failure event might be death, development of disease, or recurrence of disease.
- Measurable items, like weight loss in pounds, glucose levels or white blood cell counts.
Endpoints are defined before the trial starts. If the endpoint is decided at a later date, the results are likely biased (i.e. not accurate or not meaningful).
Importance of Primary Endpoints
Well-defined endpoints are important to ensure there is an adequate sample size (i.e. number of participants). If the sample size is too small, the trial will lack statistical power (i.e. the results will be weak and not generalizable). Fairly rare occurrences, like death, require much larger sample sizes than common occurrences, like getting a migraine.
Composite Primary Endpoint
A composite primary endpoint is used when a study has a fairly rare primary endpoint. Rare events may require a very large, expensive trial in order to get statistically significant results. Let’s say a study is investigating a new drug for anaphylaxis (a severe allergic reaction which can lead to death). If “death” is the endpoint, a trial could need about a million people to record one death(1). A composite primary endpoint might be “any non-fatal or fatal allergic reaction” — which would require much smaller numbers (thousands, or possibly hundreds).
A surrogate endpoint is a measure of success in a clinical trial. It is used in place of a primary endpoint to speed up approval process or prove that a drug appears to be working. For example, let’s say a new drug is tested to see if it reduces risk of death from diabetes. The drug appears to reduce high blood sugar, but it will take several years to gather mortality data. Reduced blood sugar is then used as a surrogate endpoint instead of the primary endpoint of reduction in number of deaths from diabetes — a figure which could take many years to calculate. Surrogates are often used because they are cheaper or easier to measure, but are weaker than stronger indicators like survival time.
When clinical trials have shown that a certain surrogate endpoint is a reliable predictor of some health benefit, it is called a validated surrogate endpoint. The FDA usually only accepts validated surrogates during a drug approval process; It may accept a “reasonably likely” alternative in the Accelerated Approval program, which gives patients with serious diseases faster access to promising treatments.
Benefits of a Surrogate Endpoint
- Smaller sample sizes and shortened time to study. For example, a small sample of about 150 patients might be needed to study a drug’s effect on cholesterol levels and the study could be completed in a year or two. If the goal was to measure reduction in the number of deaths from myocardial infarction (heart attack), a trial of 1000+ patients might be needed and the trial would likely last several years.
- Ethically sound: Surrogate endpoints can be used when measuring primary endpoints could be unethical. For example, a placebo-controlled trial is usually considered unethical if the investigative treatment decreases the risk of death or serious complications.
Other types of endpoints include:
A Secondary endpoint has secondary objectives. For example, a drug designed to prevent allergy-related deaths might also have a measure of whether quality of life is improved. A Secondary endpoint is therefore always paired with a primary one.
Secondary endpoints are usually not taken into consideration from a power or randomization standpoint when designing a trial; they are usually analyzed post-hoc. As they are not taken into consideration when designing the trial, secondary endpoints may not be statistically sound and should be interpreted with caution.
Binary Endpoints are endpoints with two choices, like success/failure, cure/no cure or remission/no remission.
Soft and Hard Endpoints
- Soft Endpoints: Subjective measurements and observations, like “patient reports feeling better” or “skin has fewer rashes.”
- Hard Endpoints: Objective endpoints that are well-defined, like “pounds lost at 6 months”.
American Academy of Allergy, Asthma and Immunology. Dying from Allergies: Fatal Anaphylaxis in the United States. Retrieved March 2, 2017 from: https://www.aaaai.org/global/latest-research-summaries/Current-JACI-Research/fatal-anaphylaxis
FDA Facts: Biomarkers and Surrogate Endpoints. Retrieved July 22, 2017 from: https://www.fda.gov/aboutfda/innovation/ucm512503.htm
If you prefer an online interactive environment to learn R and statistics, this free R Tutorial by Datacamp is a great way to get started. If you're are somewhat comfortable with R and are interested in going deeper into Statistics, try this Statistics with R track.Comments? Need to post a correction? Please post on our Facebook page.