Background: We evaluated the association between admission serum glucose-to-potassium ratio (GPR) and injury severity as well as early transfusion requirements in patients with traumatic pelvic fractures.
Methods: This single-center, retrospective cohort study included 84 adult patients with isolated or predominantly pelvic fractures admitted between January 2020 and December 2024. Patients with concomitant non-pelvic skeletal fractures were excluded to isolate the metabolic response attributable to pelvic injury. GPR was calculated from admission serum glucose and potassium levels. Higher transfusion requirement (HT) was defined as ≥4 units of packed red blood cells within 24 h. Receiver operating characteristic (ROC) analysis identified the optimal GPR cut-off using the Youden index. Internal validation was performed using bootstrap resampling (1000 iterations), and model calibration was assessed with the Hosmer–Lemeshow test. The incremental discriminatory value of GPR beyond the Injury Severity Score (ISS) was evaluated by comparing AUC values using the DeLong test, and reclassification metrics including the category-free net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were calculated. Sensitivity analyses were conducted using alternative transfusion thresholds (≥6 and ≥10 units).
Results: The optimal GPR cut-off was 34 (area under the curve (AUC) = 0.730; 95% CI: 0.593–0.853; sensitivity 78.8%; specificity 59.0%). Patients with GPR ≥ 34 (
n = 43) had significantly higher ISS values (median 25 [IQR: 16–34] vs. 9 [5–17];
p < 0.001), greater transfusion volumes (median 3 [0–6] vs. 0 [0–1] units;
p < 0.001), and longer intensive care unit (ICU) stays (3 (0–6) vs. 0 (0–1) days;
p < 0.001). In univariable logistic regression, GPR was significantly associated with HT (OR = 1.059 per unit increase; 95% CI: 1.015–1.104;
p = 0.008); however, significance was not retained in the multivariable model after adjustment for ISS (
p = 0.194). ISS remained the sole independent predictor (OR = 1.128;
p < 0.001). The combined ISS + GPR model yielded an AUC of 0.857, representing a modest increment over ISS alone (AUC = 0.849; ΔAUC = 0.009; DeLong
p = 0.566). Bootstrap-corrected AUCs confirmed minimal optimism (GPR alone: 0.726; ISS + GPR: 0.847). The Hosmer–Lemeshow test indicated adequate calibration for all models (
p > 0.05). The category-free NRI was 0.627 (
p = 0.009), whereas the IDI did not reach significance (0.017;
p = 0.290). Sensitivity analysis at the ≥6-unit threshold yielded consistent results (GPR AUC = 0.709).
Conclusions: Admission GPR is significantly associated with injury severity, hemorrhagic burden, and transfusion requirements in patients with traumatic pelvic fractures. Although GPR does not independently predict transfusion needs beyond ISS, it yields significant reclassification improvement and may serve as a practical, rapidly obtainable adjunct for early risk stratification in the acute trauma setting. Level of Evidence: III (retrospective prognostic study).
Full article
