Does early surgery improve outcomes for periprosthetic fractures of the hip and knee? A systematic review and meta-analysis.

INTRODUCTION

Previous evidence has established that early surgery is beneficial to improve outcomes for individuals with native hip fractures in the elderly population. Patients who sustain a periprosthetic fracture have been demonstrated to have similar demographics and outcomes as those with native fractures around the hip and knee. We therefore set out to determine if there is a similar difference in perioperative outcomes between early and delayed surgery for periprosthetic fractures of the hip and knee through a systematic review and meta-analysis.

METHODS

Literature search outputs were screened for studies meeting the inclusion criteria. The groups of early surgery and delayed surgery were defined by study authors. The primary outcome measure was 30 day mortality. Where there was sufficient study homogeneity, a random-effects meta-analysis was performed. Individual study risk of bias was assessed using the ROBINS-I criteria, with the GRADE criteria used for independent outcome evaluation. The review protocol was registered on PROSPERO prior to commencement (Registration number CRD42019149360).

RESULTS

The inclusion criteria was met in 11 studies (n = 3006). Mean time to surgery from admission for reporting studies was 64 h. 59.6% patients underwent early surgery as defined by the study authors. We identified a significantly lower risk of 30 day mortality for those with early surgery versus delayed surgery (RR 0.21; 95% CI 0.05, 0.90; p = 0.04, n = 2022). There were also significantly better outcomes for early versus delayed surgery regarding: medical complications, length of stay, transfusion risk, and reoperation. The quality of evidence for all the individual outcomes was low or very low.

CONCLUSIONS

There is evidence that delaying surgery in those with periprosthetic fractures of the hip and knee has a deleterious impact on mortality and other important patient outcomes. There are, however, notable limitations to the existing available literature, with further appropriately designed large-scale studies required to confirm these findings.