Modified frailty index predicts programmed cell death activation and sepsis risk after total hip arthroplasty: a...

To investigate the relationship between mFI-defined frailty and the activation of programmed cell death (PCD) pathways-apoptosis, pyroptosis, necroptosis, and ferroptosis-in elderly patients who developed infection-related sepsis after total hip arthroplasty (THA), and to evaluate whether integrating mFI with circulating cell death biomarkers improves prognostic stratification of post-THA sepsis severity and clinical trajectory. The original framing of the combined model as a "predictive" tool was revised: the combined biomarker model now describes the ability to stratify severity and clinical trajectory once sepsis is diagnosed, not to predict its occurrence.

A retrospective cohort study was conducted on 188 elderly patients (≥65 years) who underwent primary THA between January 2020 and January 2024. Frailty was assessed using the mFI-11 tool. All patients were monitored for postoperative infection-related sepsis (Sepsis-3 criteria) within 30 days. In patients who developed sepsis (n = 30), circulating biomarkers of pyroptosis (gasdermin D, NLRP3, caspase-1), necroptosis (RIPK3), ferroptosis (GPX4, 4-HNE), apoptosis (caspase-3), and DAMPs (HMGB1, cell-free DNA) were measured at 48 h post-onset. Multivariate logistic regression, multiple linear regression, and ROC curve analyses were performed.

Post-THA sepsis occurred in 30 patients (16.0%), with incidence increasing across frailty strata: non-frail 3.3%, mild frailty 14.9%, moderate-to-severe frailty 29.7% (P = 0.006). mFI score was an independent predictor of post-THA sepsis (OR = 1.52, 95% CI: 1.28-1.81, P < 0.001). Patients with sepsis exhibited markedly elevated gasdermin D (386.4 ± 128.6 vs. 42.8 ± 18.4 pg/mL), RIPK3 (52.4 ± 21.8 vs. 6.4 ± 2.8 ng/mL), HMGB1 (62.4 ± 24.8 vs. 8.4 ± 3.6 ng/mL), and depleted GPX4 (8.6 ± 3.4 vs. 34.2 ± 10.6 ng/mL) compared with non-sepsis THA patients (all P < 0.001). Within the sepsis cohort, mFI positively correlated with gasdermin D (r = 0.74), RIPK3 (r = 0.71), and HMGB1 (r = 0.78), and inversely with GPX4 (r = -0.72; all P < 0.001). A combined model integrating mFI with HMGB1, gasdermin D, and GPX4 achieved AUC = 0.904 for stratifying sepsis severity and clinical trajectory once diagnosis is established, superior to mFI alone (AUC = 0.742; DeLong P < 0.001).

Frailty, as quantified by mFI, is strongly associated with the concurrent activation of pyroptosis, necroptosis, and ferroptosis pathways in post-THA infection-related sepsis. Integrating mFI with DAMP and cell death biomarkers substantially improves prognostic stratification of sepsis severity and clinical trajectory once diagnosis is established, supporting the clinical utility of frailty-guided perioperative cell death monitoring. A preoperative prediction model based on mFI, haemoglobin, and ASA class provides clinically accessible risk stratification, warranting prospective multicenter validation before clinical implementation.