Identifying potential predictive indicators for reimplantation timing in two-stage revision: a meta-analysis and system review

Purpose: The two-stage exchange revision represents a pivotal strategy in the management of prosthetic joint infections, wherein the judicious timing of reimplantation serves as a crucial determinant for therapeutic success. At present, attempts have been made to utilize predictive models to establish the optimal timing for reimplantation; however, their predictive accuracy remains unsatisfactory. This inadequacy primarily arises from the lack of dependable predictive indicators, which demonstrate inconsistent effectiveness across various studies and occasionally yield contradictory outcomes. Therefore, identifying solid predictive indicators is in desperate need.

Methods: Studies reporting outcomes of the two-stage exchange revision till June, 2023 were systematically retrieved, screened and subject to quality analysis. Basic characteristics of these studies were firstly summarized. Subsequently, factors of interest regarding clinical information, blood and body fluid test results, pathogen test results of the recurrent and recurrent-free cohorts were extracted and submitted to a fixed or random effects model. Meanwhile, evaluation of publication bias and sensitivity was performed.

Results: After filtering, a total of 45 studies were finally involved. Pooled analysis suggested that the recurrent cohort exhibited elevated incidences of body mass index (BMI) > 30, smoking and alcohol habits, diabetes, sinus tract, positive culture, and G^-, polymicrobial and drug-resistant infections. Additionally, higher levels of C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), synovial white blood cell (WBC) count, and polymorphonuclear percentage (PMN%) were observed in the recurrent cohort. On the other hand, the results of D-dimer and fibrinogen were ambiguous, and no difference regarding peripheral WBC count was observed. Forest plots suggested a low risk of publication bias. Besides, sensitivity analysis indicated good stability of the aforementioned indicators, except D-dimer and fibrinogen.

Conclusion: To sum up, BMI > 30, smoking and alcohol habits, diabetes, sinus tract, positive culture, and G-, polymicrobial and drug-resistant infections, CRP, ESR, synovial WBC and PMN% exhibited significant differences between recurrent and recurrent-free cohorts. Therefore, these indicators may be considered as potential predictive factors for the further development of a prognostic model that aids the determination of reimplantation timing. Nevertheless, the efficacy of these indicators remains to be further confirmed.

Registration number: Prospero ID: CRD42022296568.