Leukocyte glucose index for early risk stratification in acute pulmonary embolism

The leukocyte glucose index (LGI), derived from leukocyte count and admission glucose, may reflect inflammatory and metabolic stress in acute pulmonary embolism (APE). We examined its association with adverse outcomes and compared its discrimination with admission glucose and the simplified Pulmonary Embolism Severity Index (sPESI).

This retrospective study included 205 patients with computed tomography pulmonary angiography-confirmed APE. Endpoints were in-hospital mortality and composite adverse outcomes. Multivariable logistic regression assessed associations. Diabetes mellitus (DM) was defined as documented DM, glycated hemoglobin (HbA1c) ≥6.5%, or admission glucose ≥200 mg/dL. Strict DM required documented DM or HbA1c ≥6.5%. LGI × DM interactions were tested. Paired DeLong tests compared areas under the receiver operating characteristic curve (AUCs). Internal validation used 2000 bootstrap resamples. Sensitivity analysis evaluated in-hospital death or admission systolic blood pressure < 90 mmHg.

LGI was independently associated with mortality (adjusted odds ratio [OR] 1.64, 95% confidence interval [CI] 1.25-2.16, P < 0.001) and composite adverse outcomes (adjusted OR 3.50, 95% CI 2.10-5.83, P < 0.001). LGI showed better discrimination than admission glucose for mortality (AUC 0.909 vs. 0.859, P = 0.033) and composite adverse outcomes (0.889 vs. 0.845, P = 0.025). Adding LGI to sPESI improved discrimination for mortality (AUC 0.757 to 0.868, P = 0.0045) and composite adverse outcomes (0.779 to 0.886, P < 0.001). No significant LGI × DM interaction was observed. LGI remained associated with the objective endpoint.

Admission LGI was independently associated with adverse outcomes, showed better discrimination than admission glucose, and improved sPESI discrimination. External validation is required.