Clinical Endpoints

Di Davide De Bortoli - Biostatistician & Data Manager

Introduction

One of the most important (if not the most important) steps toward a successful clinical trial is choosing the most appropriate primary endpoint. Choosing a wrong endpoint may, indeed, lead to unclear and/or inconclusive results and thus to a failure of the clinical trial, wasting resources and exposing patients to pointless and unnecessary risks.
The definition of the endpoints should take place during the design stage of a clinical trial and requires a close cooperation between all the stakeholders involved (e.g., clinicians, biostatisticians, sponsors, etc).
The two aspects are crucial because a clear definition and an adequate understanding of the endpoints, together with a correct design of a clinical trial, make sure that the information necessary to answer the clinical question are always collected and managed properly.
The aim of this article is to clarify the concept of Endpoint, particularly in comparison to the similar one of Objective, and to present the principal types and features.

Endpoint vs Objective

Outside clinical research, the meaning of "Endpoint" and "Objective" is similar and may be even used interchangeably in some cases. In Clinical Research, however, they represent two distinct concepts, although interconnected.
The objective of a clinical trial is the aim or the purpose of the study, usually defined as the clinical research question that the researchers are willing to investigate.
The study Endpoints, on the other hand, are the specific response variables that are chosen and used to measure and evaluate the objective and, ultimately, to answer the research question.
In some way, they both describe what the trial is trying to demonstrate or investigate, but while the objective is more general, comprehensible, and discursive, the endpoint is more detailed, defining the specific variables to be collected and used.
For example, in the case of a trial trying to demonstrate the efficacy of an antihypertensive drug after 6 months of treatment, compared to placebo, the primary objective may be: "To evaluate the efficacy of IP in improving the hypertensive status of patients after 6 months of treatment", while the corresponding endpoint may be "Difference between treatment groups in terms of reduction of the diastolic blood pressure as measured at baseline and after 6 months of treatment".

Definition of the endpoint: Primary, Secondary or Exploratory?

Depending on the relevance of the objective, endpoints may be distinguished in primary, secondary and exploratory.
The primary endpoint is usually corresponding to the primary objective and it is used to answer the primary research question. It is the starting point for the sample size calculation and the statistical results related to the primary endpoint determinates the success, or not, of the entire clinical trial; therefore it should be chosen carefully.
Secondary endpoints are either supportive measurements related to the primary objective or effect measurements related to secondary objectives. Finally, exploratory endpoints are used to generate some initial evidence on additional hypotheses that may be investigated in future trials.
Additionally, safety endpoints, if not akin to the primary one, are also considered when conducting a clinical trial. Safety endpoints are used to evaluate all potential threats to patients' well-being.

Characteristics and most common types of endpoints

The selection of endpoints is a complex process requiring a combination of clinical relevance and statistical reasoning. The most crucial part of this process is determining a suitable variable to measure the effect of the studied intervention. Appropriate endpoint selection ensures correct interpretation of results by providing a reliable answer to the study objective.
An optimal endpoint should be clinically relevant, precise, accurate, reliable, analyzable and interpretable; it should be capable of capturing the outcome of interest, affected by the intervention.

The most important characteristics of endpoints regard the objectivity (or not) and the nature of the measurement.
An endpoint can either be objective or subjective. Objective endpoints are variables that can be measured without conditioning or influences (e.g., death, blood pression, weight). Subjective endpoints, on the other hand, are those variables that represent an individual interpretation (e.g., pain, satisfaction, depression). Objective endpoints are less susceptible to bias and should always be preferred, if possible.
As regards to the nature of measurement, endpoints can be classified as follows:

  • Continuous (e.g., blood pression, BMI, heart rate)
  • Categorical (e.g., responder vs non-responder, improved vs stable vs worsened)
  • Time-to-event (e.g., time to event or survival time)

Finally, the main types of endpoints are: 

  • Direct: endpoints corresponding to variables that directly measure patient’s clinical condition, function or survival, representing by themselves the clinical outcome of interest
  • Composite: endpoints that provide a unique evaluation by combining multiple variables; these endpoints are useful when evaluating interventions that can improve patient’s condition in different ways or in case individual components of the endpoint are rare. Their interpretation may be challenging when the effect is not homogeneous in the individual components.
  • Surrogate: the measured variables are indicators of biological activity closely correlated with clinically meaningful endpoints that can act as substitutes for those. They could be predictive of a clinical benefit in a shorter observation period.
  • Patient reported: variables that report patient's status as perceived by him/herself, being provided directly from the patient, without interpretation of the response by a doctor or anyone else.
  • Digital: variables generated by a sensor, typically outside of a clinical setting during daily activities. The sensor could be worn, ingested, or even be part of a remote sensor system.

Conclusion

Unfortunately, there is no golden rule always applicable to choose the most appropriate endpoint: different scenarios and trial objectives require specific endpoint characteristics.
The selection of the best endpoints is crucial and complex process, that should be carefully evaluated during the design phase of the trial, weighting the knowledge and the requirements of all the different experts involved. A constructive exchange of views between clinicians, researchers and methodologists can increase the knowledge and understanding of the research objective and thus improve the likelihood of a trial’s success.


References

  • Center for Drug Evaluation and Research (CDER) and Center for Biologics Evaluation and Research (CBER): Multiple Endpoints for Clinical Trials: Guidance for Industry (2017)
  • Evans SR. Fundamentals of clinical trial design. J Exp Stroke Transl Med. 2010 Jan 1;3(1):19-27. doi: 10.6030/1939-067x-3.1.19. PMID: 21533012; PMCID: PMC3083073.
  • EUnetHTA – European network for Health Technology Assessment: Endpoints used for relative effectiveness assessment of pharmaceuticals: CLINICAL ENDPOINTS (February 2013)
  • ICH (International Council for Harmonisation of Technical Requirements for Human Use) Topic E8 General Consideration for Clinical Studies. Chapter 5.4.
  • ICH (International Council for Harmonisation of Technical Requirements for Human Use) Topic E9 Statistical Principles for Clinical Trials. Chapter 2.2.
  • ICH (International Council for Harmonisation of Technical Requirements for Human Use) Topic E9 (R1) Statistical Principles for Clinical Trials: Addendum: Estimands and Sensitivity Analysis in Clinical Trials.
  • Landers M, Dorsey R, Saria S. Digital Endpoints: Definition, Benefits, and Current Barriers in Accelerating Development and Adoption. Digit Biomark 2021 Sep 13;5(3):216-223. doi: 10.1159/000517885. PMID: 34703976; PMCID: PMC8490914.
  • McLeod C, Norman R, Litton E, Saville BR, Webb S, Snelling TL. Choosing primary endpoints for clinical trials of health care interventions. Contemp Clin Trials Commun. 2019 Nov 12;16:100486. doi: 10.1016/j. conctc.2019.100486. PMID: 31799474; PMCID: PMC6881606.