A network meta-analysis of the timing of wound dressing removal.

Introduction: Wounds are invariably dressed postoperatively but the evidence for the timing of dressing changes is limited. This meta-analysis evaluated whether the relative risk of wound infection varies depending on when dressings are changed.

Methods: A frequentist random-effects network meta-analysis was conducted on the results of a systematic review of the MEDLINE^®, Ovid^®, Scopus^®, Web of Science™ and PubMed^® databases and the Cochrane Central Register of Controlled Trials performed in May 2023. Evidence quality was graded using the Confidence In Network Meta-Analysis tool.

Results: A total of 4 studies were included with 878 patients. A significant increase in the relative risk of wound infection was found when dressings were left in situ for more than 4.5 days when compared with 48 hours (3.18, 95% confidence interval: 1.22-8.33). There were no significant differences in the relative risk of infection between the other groups. Model heterogeneity and inconsistency were insignificant (Cochran's Q: 0.44, p=0.51). The quality of the evidence was graded as generally very low and risk of bias evaluations showed it to be of high concern for bias.

Conclusions: Late dressing changes significantly increase the risks of wound infection and changes at 48 hours minimise these risks. There was no advantage demonstrated for earlier dressing changes. Ensuring that dressings are changed appropriately can minimise patient harm and health service costs.