Title : Admission predictors of intraoperative blood loss in pelvic ring fractures: Comparison of statistical and clinically oriented models
Abstract:
Background: Intraoperative blood loss during definitive internal fixation of pelvic ring fractures remains insufficiently understood. While previous studies have evaluated bleeding risk based on fracture morphology and injury severity, prediction based on admission physiological parameters has not been well established. In our prior work, preoperative transfusion was associated with blood loss; however, this approach is subjective. Therefore, an objective model using admission data is required.
Purpose: To evaluate whether intraoperative blood loss can be predicted from admission factors and to compare a statistically derived model with a clinically constructed model.
Methods: We retrospectively analyzed 34 patients who underwent definitive internal fixation for pelvic ring fractures at a single center between November 2022 and May 2025. Admission variables included age, vital signs, laboratory data, Injury Severity Score (ISS), and Young–Burgess classification. First, correlation and non-parametric analyses were performed. Subsequently, stepwise multiple regression was used to construct a statistical model. In addition, a clinical model was manually constructed by selecting variables to reflect clinical reasoning, specifically incorporating injury severity and coagulation status.
Results: The statistical model included respiratory rate, Hemoglobin (Hb), Shock Index (SI), and age, demonstrating good performance (modified R² = 0.64). In contrast, the clinical model included age, ISS, SI, Hb, and Fibrinogen (Fib), resulting in a lower model fit (modified R² = 0.49). Notably, low Hb and low Fib were significant predictors of increased blood loss. Thus, variables reflecting systemic physiological status were consistently selected, whereas fracture-related factors were not independently predictive.
Discussion: Previous studies on pelvic fractures have primarily focused on early hemorrhage, in which fracture type and injury severity play central roles. However, intraoperative blood loss occurs during definitive internal fixation, where surgical exposure and dissection disrupt hematoma tamponade, leading to re-bleeding. Therefore, intraoperative bleeding may be more strongly influenced by systemic condition, including circulatory instability and coagulation impairment.
Conclusion: Intraoperative blood loss during pelvic ring fixation can be predicted using admission factors and appears to be more strongly influenced by systemic physiology than by fracture pattern or injury severity. Therefore, early assessment of physiological and coagulation status may be important for perioperative management.
