Association between perioperative acute kidney injury and in-hospital mortality in patients aged 80 years and older undergoing hemiarthroplasty for femoral neck fractures: a retrospective cohort study

This is the authors' abstract. We don't add key points for this paper.

Early mortality remains a major concern in very elderly patients undergoing surgery for femoral neck fractures. Although frailty and comorbidity indices are commonly used for perioperative risk assessment, the prognostic significance of perioperative renal dysfunction in this population remains incompletely understood. This study aimed to evaluate the association between perioperative renal deterioration classified using a study-specific operational definition of acute kidney injury (AKI), dynamic changes in renal function, and in-hospital mortality in patients aged 80 years and older undergoing hemiarthroplasty for femoral neck fractures. This retrospective cohort study analyzed 200 patients aged ≥ 80 years who underwent hemiarthroplasty for intracapsular femoral neck fractures between January 2019 and May 2026 and had the required perioperative clinical and laboratory data. Frailty was assessed using the Modified Frailty Index (mFI-11), and comorbidity burden was evaluated using the Charlson Comorbidity Index (CCI). Serum creatinine and blood urea nitrogen (BUN) levels were recorded at admission and on postoperative days 0, 1, and 3. AKI was classified using a study-specific operational definition based on KDIGO serum creatinine thresholds applied to measurements available through postoperative day 3, without uniform verification of the KDIGO temporal criteria or assessment of urine output. The primary outcome was in-hospital mortality. Exploratory linear mixed-effects models were used to evaluate longitudinal renal changes, while univariate logistic regression and Firth penalized multivariable logistic regression analyses were performed to evaluate associations with in-hospital mortality. The final analysis included 200 patients with the required perioperative laboratory measurements through postoperative day 3. The mean age was 86.6 ± 4.9 years, and 143 patients (71.5%) were female. A total of 50 patients (25.0%) met the study-specific operational AKI definition, and 12 patients (6.0%) died during hospitalization. The proportion meeting the operational AKI definition was significantly higher among non-survivors than survivors (66.7% vs. 22.3%, p = 0.002). Non-survivors had higher CCI scores (median [IQR], 7 [5.75–7.25] vs. 5 [4–6], p = 0.004) and significantly greater increases in serum creatinine levels during the postoperative period (Δcreatinine, p = 0.026). Exploratory linear mixed-effects analysis showed a significant mortality-group-by-time interaction for serum creatinine (Wald χ² [3] = 33.47, p < 0.001), which remained significant in the log-transformed sensitivity analysis. Although the corresponding interaction for BUN was significant in the primary raw-scale model ( p < 0.001), it was not robust to sensitivity analyses excluding the influential observation or using log-transformed values. In contrast, mFI-11 scores did not differ between groups ( p = 0.952). In the primary Firth penalized model, operationally defined AKI was associated with in-hospital mortality after adjustment for CCI (OR 5.47, 95% CI 1.67–20.28; p = 0.005), while CCI was also associated with mortality (OR 1.70, 95% CI 1.17–2.51; p = 0.006). Perioperative renal deterioration classified according to the study-specific operational AKI definition was associated with in-hospital mortality in patients aged 80 years and older undergoing hemiarthroplasty for femoral neck fractures in an exploratory model adjusted only for CCI and based on 12 mortality events. Exploratory longitudinal analyses supported different temporal patterns of serum creatinine between survivors and non-survivors, whereas the corresponding BUN finding was not robust in sensitivity analyses. No statistically significant association between mFI-11 and in-hospital mortality was demonstrated in this cohort. These findings support the clinical relevance of perioperative renal dysfunction but should be interpreted cautiously given the limited number of mortality events.