by Antony Bayer

Reviewer: Henrik R. Wulff

The authors have reviewed a large number of protocols aubmitted to an ethics committee. Many of these had an upper age limit, which in many cases seemed unjustified. They regard that as ethical age discrimination. The paper is commendably short, and I only miss one piece of factual information. The authors only state that the upper age limit varied between 45 and 100 years, and they might have added one line telling the reader a little about the distribution of the limit (e.g. median and interquartile range). I take it that the authors see it as their limited aim to raise this issue, so that researchers in future will not include an age limit without explicit justification in the protocol. The paper serves this purpose well.

If I were the researcher I would consider two methodological issues:

a) It is usually crucial, e.g. in randomised clinical trials, to plan the study in such a way that the drop-out rate is minimised. What do I know or think about the drop-out rate in my sort of study among younger and older people? If I suspect it to be higher among elderly patients (e.g. a trial with frequent visits to the outpatient department over a longer period of time) I may conclude that an age limit is justified.

b) What will happen to the external validity of the trial, if I exclude elderly people? If I have no reason to believe that younger and older people will react differently to the treatment, an upper age limit will not matter very much, because then I have equally little reason to believe that I cannot extrapolate the results from one age group to another. If, on the other hand, I suspect that the two age groups may react differently, I should either do two trials (one of 'young' and one of 'old' people) or I should do one trial which prior stratification as regards age. I should not just pool old and young patients. The authors might have included a discussion along these lines, but I can also accept that they leave it to the readers to think for themselves.