A secondary endpoint is a predefined measurement objective in a clinical trial that is evaluated alongside the primary endpoint. While the primary endpoint addresses the main question of the study and is decisive for the statistical sample size calculation, secondary endpoints provide supplementary information on efficacy, safety, or quality of life. They are predefined in the study protocol and the Statistical Analysis Plan (SAP) and are arranged hierarchically. Secondary endpoints are an essential element of the overall evidence of a study: even if the primary endpoint is met, consistent results in secondary endpoints provide a more robust basis for regulatory decisions and HTA assessments. Conversely, inconsistent or negative secondary endpoints can call the overall benefit of an intervention into question.
Definition and Differentiation from the Primary Endpoint
The fundamental difference between primary and secondary endpoints lies in statistical priority and evidentiary weight. The primary endpoint determines the sample size (power calculation) and bears the main burden of statistical proof. Secondary endpoints serve to characterize the benefit of an intervention in further dimensions, without the study necessarily being formally sufficiently powered for them.
Typical secondary endpoints in clinical trials include: overall survival (OS) as a supplement to progression-free survival as the primary endpoint, patient-reported outcomes (PROs) such as quality of life, time to response, duration of response, biomarker changes, and safety-related metrics such as adverse events. In studies with an active comparator, non-inferiority analyses for secondary endpoints are frequently added. Pharmacokinetic parameters, responder analyses, and time-to-event variables also count as secondary endpoints depending on the study design. The exact definition—including measurement time point, measurement instrument, and statistical analysis method—must be pre-specified for each secondary endpoint in the protocol and the SAP before unblinding occurs.
Hierarchical Test Strategies and Alpha Control
Since the simultaneous statistical testing of multiple endpoints creates the problem of multiple testing, clinical trials with secondary endpoints must follow a predefined testing strategy. The most common approaches are the hierarchical (sequential) testing strategy and the Bonferroni correction. In a hierarchical strategy, secondary endpoints are only formally tested inferentially if the superior endpoint was statistically significant.
The European guideline ICH E9(R1) and the EMA guidance on multiple endpoints require an explicit description of the hierarchy and testing procedure in the SAP. Secondary endpoints that are not part of a controlled multiplicity strategy are declared as exploratory endpoints and may not support confirmatory conclusions. This distinction is crucial for regulatory assessment and HTA procedures.
Regulatory and HTA Relevance
Authorities such as the EMA and BfArM evaluate secondary endpoints as supporting evidence within the framework of marketing authorization procedures. In the EPAR (European Public Assessment Report), primary and secondary endpoints are evaluated and commented on individually. Weak or inconsistent results in secondary endpoints can limit the interpretation of the positive primary result or lead to conditions of marketing authorization. Within the framework of Scientific Advice, the EMA explicitly requests a discussion of the planned endpoint hierarchy so that sponsors can obtain early clarity on the regulatory utility of their secondary endpoints.
Within the framework of Health Technology Assessment (HTA) according to EU Regulation 2021/2282, secondary endpoints play a special role: endpoints with direct patient relevance—even if defined as secondary—are often weighted more heavily by the G-BA and other HTA authorities than intermediate primary endpoints such as biomarkers or surrogate parameters. Sponsors should therefore coordinate with HTA experts during the study planning phase to determine which secondary endpoints are particularly relevant for the subsequent benefit assessment and prioritize their measurement and evaluation accordingly.
Documentation and Reporting
The complete and transparent reporting of all predefined secondary endpoints is a regulatory requirement. In the Clinical Study Report (CSR) according to ICH E3, primary and all pre-specified secondary endpoints must be presented with full results—regardless of whether the results were significant or published. Selective reporting of individual endpoints (outcome reporting bias) is a serious scientific and regulatory violation.
In the Clinical Trials Information System (CTIS), sponsors are obliged to publish the results of all primary and secondary endpoints within the legally prescribed period. GCP inspections verify the consistency between the endpoints defined in the protocol, the SAP, the CSR, and the published results.
Frequently Asked Questions (FAQ)
How many secondary endpoints are appropriate in a clinical trial?
There is no fixed upper limit, but an excessive number of secondary endpoints increases the multiplicity problem and complicates regulatory assessment. The EMA recommends limiting secondary endpoints to what is scientifically necessary and clearly distinguishing between confirmatory (within the multiplicity strategy) and exploratory endpoints. In practice, Phase III trials typically have 3 to 10 secondary endpoints with a defined hierarchy.
Can a secondary endpoint be retroactively changed to a primary endpoint?
No—such a change after data review (post-hoc change) is not accepted by regulators and is considered a serious protocol violation. Changes to the endpoint hierarchy may only be made before data review (before unblinding) and by means of a formal protocol amendment, which must be notified to the ethics committee and the competent authority. Any subsequent shifting of endpoints is to be regarded as a source of bias.
What is the difference between secondary and exploratory endpoints?
Secondary endpoints are predefined and—if part of a multiplicity strategy—inferentially testable. Exploratory endpoints are also predefined but are not part of the confirmatory testing procedure. They serve to generate hypotheses for future studies and may not support confirmatory statements. The distinction must be clearly documented in the SAP.