Secondary end point

The collection of all secondary endpoints is called the secondary endpoint family. Secondary endpoints are those that may provide supportive information about a drug’s effect on the primary endpoint or demonstrate additional effects on the disease or condition. Secondary endpoints might include a pharmacodynamic effect that would not be considered an acceptable primary efficacy endpoint but is closely related to the primary endpoint, (e.g., an effect consistent with the drug’s purported mechanism of action). A secondary endpoint could be a clinical effect related to the primary endpoint that extends the understanding of that effect (e.g., an effect on survival when a cardiovascular drug has shown an effect on the primary endpoint of heart failure-related hospitalizations) or provide evidence of a clinical benefit distinct from the effect shown by the primary endpoint (e.g., a disability endpoint in a multiple sclerosis treatment trial in which relapse rate is the primary endpoint). In all cases, when an effect on the primary endpoint is shown, the secondary endpoints can be examined and may contribute important supportive information about a drug’s effectiveness. Positive results on the secondary endpoints can be interpreted only if there is first a demonstration of a treatment effect on the primary endpoint family. The Type I error rate should be controlled for the entire trial, defined in section II.C as strong control. This includes controlling the Type I error rate within and between the primary and secondary endpoint families. Moreover, the Type I error rate should be controlled for any preplanned analysis of pooled results across studies; pooled analyses are rarely conducted for the planned primary endpoint, but are sometimes used to assess lower frequency events, such as cardiovascular deaths, where the individual trials used a composite endpoint, such as death plus hospitalization. Statistical testing strategies to accomplish this are discussed in section IV. Control of the Type I error rate for all endpoints depends upon the prospective designation of all primary and secondary endpoints. Generally, the endpoints and analytical plan should be provided at the time the trial protocol is finalized. The statistical analysis plan should not be changed after unmasking of treatment assignments, including unmasking for any interim analyses. Because study sample size is often determined based only on the amount of information needed to adequately assess the primary hypothesis, many studies lack sufficient power to demonstrate effects on secondary endpoints. If success on the secondary endpoints is important, the secondary endpoints should be considered when determining study design (e.g., sample size). An example of a secondary endpoint used to further characterize the drug’s effect is a measurement of the primary outcome variable at 30 days in a trial whose primary endpoint is the same outcome measured at 6 months. Another example is a secondary endpoint of the percentage of patients whose symptoms are “very improved,” when the primary endpoint is the percentage of patients with any amount of improvement for the same symptoms. Adjustment for multiplicity is necessary to demonstrate these additional effects. It is recommended that the list of secondary endpoints be short, because the chance of demonstrating an effect on any secondary endpoint after appropriate correction for multiplicity becomes increasingly small as the number of endpoints increases. Endpoints intended to serve the purpose of hypothesis generation should not be included in the secondary endpoint family. These should be considered exploratory endpoints.