Article type
Year
Abstract
Introduction: The literature on the use of estrogens in the prevention of osteoporosis lacks RCTs reporting fracture as the main clinical outcome.
Objectives: To examine the quality of RCTs of the effects of estrogen treatment on measures of bone density.
Methods: Searching of Medline and Excerpta Medica databases, bibliographies of published reviews, and review of the reference list by experts and pharmaceutical manufacturers. Data extraction and quality assessment were performed in duplicate, with assistance of a manual. Raters were blinded as to authors, their affiliations and the publication details.
Results: Of 73 eligible RCTs published between 1977 and 1993, 6 had no extractable data, and 18 (25%) were published in duplicate. Total quality scores increased over time, but this was accounted for by improvements only in the measurement technologies used. There was no improvement in the quality of randomisation methods, the extent to which withdrawals were accounted for, or in the baseline comparability of treated and control patients. Neither sample sizes nor durations of follow-up increased over time, and the trials were far too small and of too short duration to detect changes in fracture rates. Effect size estimates (annual changes in BMD at the forearm and spine) were unrelated to measurement technique, publication date and study quality. Over time there was a shift, from unopposed to opposed estrogens, and from oral to transdermal estrogen delivery.
Discussion: This clinical trial literature has failed to develop in an appropriate manner. Investigators appear to have been more concerned with the technology of measuring a surrogate outcome, and delivering treatment, than in conducting trials that really inform clinical decision making.
Objectives: To examine the quality of RCTs of the effects of estrogen treatment on measures of bone density.
Methods: Searching of Medline and Excerpta Medica databases, bibliographies of published reviews, and review of the reference list by experts and pharmaceutical manufacturers. Data extraction and quality assessment were performed in duplicate, with assistance of a manual. Raters were blinded as to authors, their affiliations and the publication details.
Results: Of 73 eligible RCTs published between 1977 and 1993, 6 had no extractable data, and 18 (25%) were published in duplicate. Total quality scores increased over time, but this was accounted for by improvements only in the measurement technologies used. There was no improvement in the quality of randomisation methods, the extent to which withdrawals were accounted for, or in the baseline comparability of treated and control patients. Neither sample sizes nor durations of follow-up increased over time, and the trials were far too small and of too short duration to detect changes in fracture rates. Effect size estimates (annual changes in BMD at the forearm and spine) were unrelated to measurement technique, publication date and study quality. Over time there was a shift, from unopposed to opposed estrogens, and from oral to transdermal estrogen delivery.
Discussion: This clinical trial literature has failed to develop in an appropriate manner. Investigators appear to have been more concerned with the technology of measuring a surrogate outcome, and delivering treatment, than in conducting trials that really inform clinical decision making.