Integrated SIRI and Lipid Profile for Early Prediction of Bloodstream Infection in AML During Induction Chemotherapy

Bloodstream infections (BSIs), a frequent and life-threatening complication during acute myeloid leukemia (AML) induction chemotherapy, carry high mortality; however, current predictive models lack robust combined inflammatory-metabolic biomarkers.

We conducted a retrospective analysis of 225 AML patients (2020-2024). The systemic inflammation response index (SIRI) and lipids measured at baseline. BSIs were confirmed according to Centers for Disease Control and Prevention/National Healthcare Safety Network (CDC/NHSN) criteria during neutropenia. Predictors selected via univariate analysis (P<0.05) and multivariable logistic regression using backward selection based on the Akaike information criterion (AIC). A nomogram was constructed. Model validation included receiver operating characteristic curve analysis and area under the curve (ROC-AUC), calibration curves (1,000× bootstrap), and decision curve analysis (DCA).

Among 225 AML patients, BSIs incidence was 24% (54/225). Patients with BSIs exhibited significantly elevated systemic inflammation (SIRI: 2.52 ± 0.38 vs 1.57 ± 0.29; P<0.001) and atherogenic dyslipidemia, characterized by higher low-density lipoprotein cholesterol (LDL-C: 3.43 ± 0.91 vs 2.56 ± 0.72 mmol/L; P<0.001) and lower high-density lipoprotein cholesterol (HDL-C: 0.61 ± 0.19 vs 0.92 ± 0.25 mmol/L; P<0.001). The SIRI-lipid nomogram incorporated six independent predictors, including SIRI (OR=3.36, 95% CI 2.00-6.07), LDL-C (OR=5.98, 95% CI 2.84-14.13) and HDL-C (OR=0.06, 95% CI 0.01-0.64). The nomogram achieved an AUC of 0.926 (95% CI 0.879-0.973) and demonstrated excellent calibration, with a mean absolute calibration error of 0.014 based on 1000 bootstrap samples. DCA showed clinical utility across decision thresholds. SIRI remained an independent predictor of BSIs after multivariable adjustment (OR=3.28) and correlated with prolonged hospitalization (P=0.007).

The SIRI-lipid integrated nomogram provides clinically applicable prediction of BSIs risk in AML induction therapy, with validated clinical utility. Elevated SIRI combined with atherogenic dyslipidemia, characterized by high LDL-C and low HDL-C, represents actionable risk indicators enabling early clinical interventions.