Prediction of hidden blood loss after intramedullary nail fixation of femoral shaft fractures in elderly patients: development and validation of a clinical nomogram based on Gamma regression

Intramedullary nailing (IMN) is the gold standard for treating femoral shaft fractures (FSFs) in the elderly, who are at high risk for complications due to age-related physiological changes. Hidden blood loss (HBL) remains an underrecognized complication that can compromise patient outcomes, particularly in elderly patients with reduced physiological reserve. This study aimed to identify risk factors for HBL following IMN fixation of FSF in elderly patients and develop a clinically applicable prediction nomogram. We retrospectively analyzed 113 patients aged ≥65 years who underwent closed reduction and IMN fixation for AO/OTA 32-type FSFs between July 2019 and June 2024. HBL was quantified using the formulas described by Nadler and Gross. We performed univariate analysis first, followed by LASSO regression for variable selection and Gamma regression to model the HBL data. A nomogram was then constructed and validated, and its predictive performance was compared with that of Random Forest and Support Vector Machine (SVM) models using the area under the receiver operating characteristic curve (AUC). The study cohort comprised 113 elderly patients (mean age 73.0 ± 7.9 years). The mean HBL was 1214.75±307.14 mL, accounting for 78.87% of total perioperative blood loss, with a 40% transfusion rate. Independent risk factors for HBL included male gender, elevated BMI, AO/OTA 32-C fractures, preoperative anemia, narrow femoral medullary cavity, ASA score≥3, and general anesthesia. The nomogram based on Gamma regression demonstrated acceptable discriminative ability (AUC=0.77), whereas the SVM model achieved superior predictive performance (AUC=0.848). HBL represents a major component of perioperative blood loss in elderly patients with FSF treated by IMN. The risk factors we identified provide valuable insights for targeted interventions in this vulnerable population, and the nomogram we developed serves as a practical tool for preoperative risk stratification in geriatric trauma care.