Explanation of Primary and Secondary Endpoints

Clinical trials are designed to provide clinically important and meaningful information on particular measures of efficacy and / or safety. These are called primary endpoints. Analyses of primary endpoints can therefore generally be considered robust (i.e., they typically have greater scientific validity than analyses of secondary endpoints).

Other measures of efficacy and safety are often assessed in clinical trials and referred to as secondary endpoints. Although clinical trials are typically not designed to provide robust information on secondary endpoints the information they provide can be important e.g., for hypothesis setting or to support the information obtained from primary endpoints. However it should be noted that:

1. The number of patients enrolled into clinical trials is usually based solely on providing clinically and statistically meaningful information for primary endpoints not secondary endpoints.

   Therefore, a lack of statistical significance between treatments for secondary endpoints may be because the trial did not include enough patients (the trial is referred to as being underpowered for these endpoints). Conversely where there are statistically significant differences between treatments for secondary endpoints, it is important to recognise that this may be because the trial is oversensitive for these endpoints rather than the difference being clinically important (the trial is referred to as being overpowered for these endpoints).

2. The more secondary endpoints that are analysed the more difficult it can be to interpret statistical tests comparing treatments. This is because the more statistical tests that are conducted the more chance there is of finding a statistical difference which is not real or of missing a real difference.

We therefore recommend that qualified statisticians help in the interpretation of the results provided for secondary endpoints.

Uncontrolled Open Label Extension Studies

Some clinical trials are extended to give additional information on the longer term safety and efficacy of a medicine. In some cases these are not randomised or blinded and thus called uncontrolled open label extension studies.

Safety information from uncontrolled open label extension studies should be interpreted with caution as patients in these extension studies are likely to have tolerated the medicine during the randomised phase of the study and those patients who experience adverse events may withdraw from the study early making longer term assessments of safety difficult. Moreover, the adverse events reported may be related to the long term consequences of the disease under investigation, or a separate disease or condition, or concomitant therapies rather than the medicine under study. Efficacy endpoints for uncontrolled open label extension studies in the Register are summarised descriptively as formal statistical analyses are often not appropriate for data collected in an open label extension study. This is because patients in these trials are not randomised to different treatments and typically will have responded to the medicine in the randomised phase of the trial.