Comment on: Interventions to control MRSA: high time for time-series analysis?

Sir, I enjoyed Harbarth and Samore's thoughtful leader 1 on interven- tions to control MRSA and agree with their sentiments and rec- ommendations that more well-designed studies are urgently needed and that multiple interventions are most likely to work. I would like, however, to emphasize that there are many more published intervention studies assessing the effects of anti- biotic stewardship on MRSA than they quote, certainly more than the 'surprisingly few' that they state. 2-1 2 These studies (and there may be more as I have not performed a systematic review of the literature) consistently (11 of 11) demonstrate reduction in MRSA, usually after reduction in b-lactam and/or quinolone use. Our own work also shows macrolide use to be consistently important in causing increased MRSA rates. 13,14 While the Cochrane review 15 referred to by Harbarth and Samore did not show any evidence that antibiotic stewardship could reduce MRSA, three of the studies 8,9,12 not quoted by them are reasonable quality interrupted time series that have been published since the Cochrane Review. Moreover, the con- sistency of the findings and the scientific plausibility of all 11 studies counter their uncontrolled nature and add considerable weight to the argument that much more emphasis should be put on antibiotic stewardship to control MRSA, particularly in the face of failure of the current MRSA control strategies. The real problem is how to initiate and maintain major changes in our antibiotic prescribing habits. Major effort is required to be put into answering the difficult questions sur- rounding this need. It is not easy to make significant and persist- ent inroads into a hospital's reliance on b-lactams, quinolones and macrolides, although the advent of widespread MRSA admission screening should make it easier to avoid these anti- biotics in patients colonized with MRSA. 16,17