Comprehensive assessment of risk factors and development of novel predictive tools for perioperative hidden blood loss in intertrochanteric femoral fractures: a multivariate retrospective analysis

Linbing Lou, Lei Xu, Xiaofei Wang, Cunyi Xia, Jihang Dai, Le HuEuropean 2024 · DOI

Abstract

The nomogram model facilitates an intuitive and quantitative assessment of the risk of perioperative hidden blood loss in patients with ITFs, providing robust support for clinical decision-making.

Introduction

Intertrochanteric femoral fractures (ITFs) represent a significant challenge in geriatric orthopedics, particularly within aging populations in Asia. While surgical intervention using Proximal Femoral Nail Antirotation (PFNA) is the standard of care, it is frequently complicated by "Hidden Blood Loss" (HBL). Unlike overt blood loss observed during surgery, HBL occurs through tissue infiltration and hemolysis. In ITF cases, HBL can constitute up to 82% of total perioperative blood loss, often exceeding one liter. This study aims to develop a predictive model to identify patients at high risk for significant HBL, enabling proactive clinical management.

Methods

This retrospective study analyzed 231 patients treated at Northern Jiangsu People’s Hospital between 2021 and 2023. The cohort was divided into a training set (n=185) and a validation set (n=46). Total blood volume was estimated using the OSTEO formula, with HBL calculated based on changes in hematocrit levels and transfusion volumes.

Variable selection was performed using LASSO (Least Absolute Shrinkage and Selection Operator) regression to identify the most significant predictors while minimizing overfitting.

Selection of optimal lambda via cross-validation in the LASSO model.
LASSO coefficient profiles for the clinical variables.

Following LASSO selection, multivariate logistic regression was employed to construct a clinical nomogram. Eight key predictors were finalized: gender, BMI, fracture classification (Evans), hypertension status, preoperative hemoglobin, serum calcium, and other laboratory markers.

Results

The developed model demonstrated high predictive accuracy. In the training set, the Area Under the Curve (AUC) was 0.947, while the validation set yielded an AUC of 0.902. These results indicate excellent discriminative ability. Decision Curve Analysis (DCA) further supported the clinical utility of the nomogram, showing a higher net benefit compared to standard "treat-all" or "treat-none" approaches. Key individual predictors included preoperative hemoglobin levels and the stability of the fracture.

Discussion

Several clinical insights emerged from the multivariate analysis:

  • Demographics and Morphology: Females and patients with a BMI ≥ 24 showed higher risk, likely due to hormonal influences on bone density and increased surgical trauma in patients with higher adipose tissue.
  • Biochemical Markers: Lower admission serum calcium was a significant risk factor, reflecting its essential role in the coagulation cascade.
  • The Hypertension Paradox: Patients without hypertension were at higher risk. This may be attributed to the relative elasticity of their vessels compared to the stiffened, atherosclerotic vessels found in chronic hypertensive patients, which may be more resistant to mechanical rupture during surgery.
  • Fracture Severity: Unstable fractures (Evans classification) correlated with higher HBL due to increased internal tissue damage from mobile bone fragments.

Conclusion

The proposed nomogram provides a reliable, quantitative tool for predicting HBL in ITF patients. While limited by its retrospective, single-center design, the model offers a practical means for surgeons to identify high-risk individuals and implement early interventions. Future research will focus on multi-center validation and the inclusion of additional variables such as dynamic bone density to further refine predictive accuracy.

Summary updated

Abstract

The nomogram model facilitates an intuitive and quantitative assessment of the risk of perioperative hidden blood loss in patients with ITFs, providing robust support for clinical decision-making.
BibTeX
@inproceedings{lou2024comprehensive,
  author = {Linbing Lou and Lei Xu and Xiaofei Wang and Cunyi Xia and Jihang Dai and Le Hu},
  title = {{Comprehensive assessment of risk factors and development of novel predictive tools for perioperative hidden blood loss in intertrochanteric femoral fractures: a multivariate retrospective analysis}},
  booktitle = {European},
  year = {2024},
  doi = {10.1186/s40001-024-02244-1},
}
Figure