Dressings for venous leg ulcers: systematic review and meta-analysis

Several errors in this paper by Palfreyman and colleagues persisted through the editorial process to publication (BMJ 2007;335:244-8, 4 Aug, doi: 10.1136/bmj.39248.634977.AE). In fig 2 of the full version (fig 1 of the abridged version), the results for Moffatt 1992 should be 13/30 healed in hydrocolloid arm and 7/30 in the low adherent arm (not the other way round), giving a pooled relative risk of 1.09 (0.89 to 1.34). In addition, the discussion of trial quality in both versions should have said that five [not three] trials stated the method of randomisation,^{w3 w4 w10 w11 w13} and the full version should have said that two studies [not one study] reported a priori sample size calculations.^{w9 w10} The discussion of external validity in both versions (and table 1 in the full version) should have stated that three trials [not just one trial] used life table analysis.^{w9 w17 w39} In both versions, the duration of the included trials should read four to 40 [not 48] weeks. Finally, the authors omitted to acknowledge the support and advice of the Cochrane Wounds Group in the course of the review and to thank Rona Lochiel for helping to develop the protocol.