ENG I DEU
Everybody wants them, few have them – the „hard endpoints“ in clinical studies. What are these “hard endpoints” and is being “hard” relevant for endpoints that are used for developing messages from studies?
In clinical studies, the term „endpoint“ describes the outcome of a treatment, e.g. hospital discharge, or a certain diagnosis, e.g. cardiac infarction. In an interventional study, the endpoints are specified prior to data collection. This reduces the risk of presenting only those results, which the researchers like. However, because studies consume considerable time and expenses, usually one primary and several secondary endpoints are defined in order to structure the results. For example, “death” could be a primary endpoint and “cardiac death” as well as “death related to other causes” could be secondary endpoints.
The most important endpoint is designated as “primary”, for severe, fast-progressing diseases this is usually death. Death is a classic “hard” endpoint because there is no room for interpretation in diagnosis and its relevance is undeniable. Death might even be the only real hard endpoint because it is hard to identify any other endpoint that describes such a clear “yes/no” phenomenon: it is impossible to be “a little death”. However, it is universally accepted that endpoints are the “harder”, the more important the specific event is during disease progression and the clearer any criteria for diagnosis are. For example, cardiac insufficiency is an important endpoint in cardiovascular diesease, but especially early stages are not easy to diagnose. This implies the risk that due to difficulties in diagnosing, an observed frequency of cardiac insufficiency is prone to bias.
Thus, a “hard” endpoint may indicate a high quality of a study. For example, the progression of renal disease in patients starting dialysis is evaluated. The incidence of loss of appetite, pruritus (itching) and weight loss are recorded as endpoints (all of these may be observed in dialysis patients). In this case “hardness” of the endpoints would increase from appetite and pruritus to weight loss because the first two are influenced by the patient’s perception, therefore leaving more room for interpretation. For loss of appetite, the only information available is based on the patient’s perception and cannot be tested. Pruritus possibly creates symptoms like scratch marks, therefore the doctors diagnosis does not only rely on the patient’s feelings. In contrast, weight loss can be measured (as long as calibrated scales are used). Additionally, weight loss may be the beginning of a generalized physical breakdown, which negatively influences disease progression. Therefore, this endpoint is the one with the highest clinical importance and the most reliable.
However, if a study about any supplement X shows that all of these three endpoints are positively affected, would reduced weight loss be the main message for the doctors? Regarding validity of the endpoints, this is correct. But further aspects need to be taken into account, e.g. validity of an endpoint may not reflect the importance perceived by the target group.
In our example, weight loss at the beginning of dialysis causes little suffering for patients and doctors. More doctors will probably take notice of the message regarding pruritus. This complication is rarer, but may cause considerable suffering – for both patient and doctor. Therefore, “hard” endpoints are important, but not the only factor that should be considered when communicating the results of a study.
Author: Dr. Christoph Messer