The effect of different preventive strategies during total joint arthroplasty on periprosthetic joint infection: a...

Periprosthetic joint infection after total joint arthroplasty has a large incidence, and it may often require two or more stages of revision, placing an additional burden on clinicians and patients. The purpose of this network meta-analysis is to evaluate the effect of four different preventive strategies during total joint arthroplasty on the prevention of periprosthetic joint infection.

The study protocol was registered at PROSPERO (CRD: 42,023,448,868), and the literature search databases included Web of Science, PubMed, OVID Cochrane Central Register of Controlled Trials, OVID EMBASE, and OVID MEDLINE (R) ALL that met the requirements. The network meta-analysis included randomized controlled trials, retrospective cohort studies and prospective cohort studies with the outcome of periprosthetic joint infection. The gemtc R package was applied to perform the network meta-analysis to evaluate the relative results of different preventive strategies.

This network meta-analysis study included a total of 38 articles with 4 preventive strategies and negative controls. No improvement was observed in antibiotic-loaded bone cement compared with negative controls. Chlorhexidine showed the highest probability of delivering the best preventive effect, and povidone iodine had the second highest probability. Although vancomycin ranked after chlorhexidine and povidone iodine, it still showed a significant difference compared with negative controls. In addition, the incidence after applying chlorhexidine was significantly lower than that after applying negative controls and vancomycin. In the heterogeneity test between direct and indirect evidence, there was no apparent heterogeneity between them.

The study indicated that chlorhexidine, povidone iodine and vancomycin showed significant efficacy in preventing periprosthetic joint infection after total joint arthroplasty, while antibiotic-loaded bone cement did not. Therefore, more high-quality randomized controlled trials are needed to verify the results above.