Abstract Background Incisional prophylactic intra-operative wound irrigation (pIOWI) with an aqueous antiseptic solution effectively reduces surgical site infections (SSIs); however, data on the use of aqueous chlorhexidine gluconate are limited and no direct comparisons to aqueous povidone iodine exist. Recognising the efficacy of chlorhexidine gluconate in skin antisepsis, we investigated its use as an aqueous solution for pIOWI, hypothesising it might offer similar advantages. We therefore compare the efficacy of aqueous chlorhexidine gluconate with aqueous povidone iodine for incisional pIOWI in prevention SSI in elective abdominal surgery. Methods A post hoc analysis of data from the randomised controlled EPO2CH trial was conducted including 699 patients to assess the effect of aqueous chlorhexidine gluconate compared to aqueous povidone iodine for incisional pIOWI on the incidence of SSI. Multiple imputation with chained equations were used to impute missing values. The association between type of irrigation and SSI was assessed using inverse probability of treatment weighted logistic regression. Findings Weighted regression analysis showed a lower SSI rate −3.03%, 95% c.i. −7.04 to 0.98 for pIOWI with aqueous chlorhexidine gluconate (4.05%) versus aqueous povidone iodine (7.08%) (number needed to benefit 33.0 (NNTH 102.0 to ∞ to NNTB 14.2). There was a lower SSI rate −3.65%, 95% c.i. −7.47 to 0.17 for superficial and a higher SSI rate 0.62%, 95% c.i. −0.61 to 1.85 for deep SSIs for aqueous chlorhexidine gluconate compared to aqueous povidone iodine. Interpretation While SSIs are lower for incisional pIOWI with aqueous chlorhexidine gluconate than with aqueous povidone iodine, the results are not statistically significant, with wide confidence intervals, indicating uncertainty. Based on these results, we recommend the use of aqueous povidone iodine for pIOWI and advise caution when considering aqueous chlorhexidine gluconate, given concerns about its potential to induce resistance and its limited clinical data.
Groenen et al. (2025) studied this question.