Discharge to a Skilled Nursing Facility After Hip Fracture Results in Higher Rates of Periprosthetic Joint Infection
Emily A Treu,Jeffrey J Frandsen,Graham J DeKeyser,Brenna E Blackburn,Michael J Archibeck,Lucas A Anderson,Arthroplasty for Hip Fracture Consortium,Jeremy M Gililland,Elie Ghanem,Gerald McGwin,Kyle H Cichos,Brandon Boyd,Brett Crist,Jim Keeney,Michael Ewing,Madeline A Sauer,Simon C Mears,Benjamin M Stronach,Jeffrey B Stambough,Joshua W Mueller,Michael O'Malley,Stuti Patel,Antonia F Chen,Christopher M Melnic,Eric Jordan,Amanda Cao,Erik N Hansen,David Sing,Patrick F Bergin,Eldrin Bhanat,George W Stayer,Mariegene E Almand,Ugur Yener
DOI: https://doi.org/10.1016/j.arth.2024.04.002
Abstract:Background: Femoral neck fractures (FNFs) in elderly patients are associated with major morbidity and mortality. The influence of postoperative discharge location on recovery and outcomes after arthroplasty for hip fractures is not well understood. Methods: A multisite retrospective cohort from 9 academic centers identified patients who had FNF treated with hemiarthroplasty or total hip arthroplasty between 2010 and 2019. Patients who had diagnoses of dementia, stroke, age > 80 years, or high energy fracture were excluded. Discharge location was identified, including home-based health services (HHS), inpatient rehabilitation (IPR), or a skilled nursing facility (SNF). Rates of reoperation, periprosthetic joint infection (PJI), and mortality were compared between cohorts. Multivariate logistic regressions were performed, adjusting for age, American Society of Anesthesiologists (ASA) score, body mass index, sex, and tobacco use. Statistical significance was defined as P < .05. Results: A total of 672 patients (315 HHS, 144 IPR, and 213 SNF) were included in this study. The average follow-up was 30 months. The SNF cohort was significantly older (P < .0001) with higher ASA scores (P < .0001) than the HHS cohort. In a logistic regression model adjusting for age, ASA score, and body mass index, the SNF cohort had higher mortality rates than the HHS cohort (P = .0296) and were more likely to have PJI within 90 days (odds ratio = 4.55, 95% confidence interval = 1.40, 4.74) and within 1 year (odds ratio = 3.08, 95% confidence interval = 1.08, 8.78). Time to PJI was significantly shorter in the SNF cohort (SNF 38 versus HHS 231 days, P = .0155). No differences were seen in dislocation or reoperation rates between the SNF and HHS cohorts. No differences were seen in complication rates between the IPR and HHS cohorts. Conclusions: Discharge to a SNF after arthroplasty for FNF is associated with increased mortality and higher rates of PJI. Hip fracture care pathways that uniformly discharge patients to SNFs may need to be re-evaluated, and surgeons should consider discharge to home with HHS when possible.