A methodological evaluation of 14 controlled clinical trials in myelomatosis.

A methodological evaluation of the 14 randomized clinical controlled trials so far published on myelomatosis was performed. A quite large heterogeneity was observed among the trials, especially in their presentation of patients included and in the criteria applied for therapeutic response. Additionally, none of the studies accounted for the randomization procedure - and particularly whether the randomizations were performed consecutively. 19 comparisons of the effectiveness of various drugs or drug combinations were performed and in 12 of these an insignificant result was obtained. 2 of these with more than 25% probability have overlooked a beneficial effect of one of the regimens of more than 50% and 6 trials of more than 25%. It is stressed that these situations will arise when sample sizes that are too small are applied in clinical trials. It is concluded that internationally accepted criteria for diagnosis and therapy response are needed. Additionally, a more thorough statistical planning, prior to initiation of a trial, and a more extensive analysis of the results would be helpful.