Serum Glycoprotein VI in Acute Pulmonary Embolism: Association With Radiological Thrombus Burden and Clinical Risk Indicators.

Abstract

Background: Pulmonary embolism (PE) is a life-threatening condition requiring rapid diagnosis and accurate risk stratification. Biomarkers reflecting thromboinflammatory processes may improve early clinical assessment. Glycoprotein VI (GPVI), a key receptor involved in platelet activation, has been implicated in thrombus formation; however, its clinical relevance in PE remains unclear.

Objective: To examine the association of serum GPVI with radiological thrombus burden and clinical risk indicators in acute PE, while retaining the original PE-versus-control comparison only as a secondary exploratory case-control discrimination analysis, not as a diagnostic-accuracy analysis.

Methods: In this cross-sectional study, the analytic cohort comprised 50 patients with CTPA-confirmed acute PE and 25 control participants. Serum soluble GPVI (sGPVI) was measured by a colorimetric sandwich ELISA. The study-recorded CTPA thrombus-burden categories were analyzed as lower, intermediate, and higher radiological burden and separately from an exploratory ESC-aligned risk reconstruction. For ICU admission, parsimonious multivariable and incremental analyses evaluated GPVI beyond PESI with internal validation. A PE-versus-control ROC comparison with D-dimer was retained only as a secondary exploratory case-control discrimination analysis and was not treated as an estimate of diagnostic accuracy.

Results: Among patients with PE, GPVI differed across the study-recorded radiological burden categories (lower, n = 16: 192.95 [185.80-215.39]; intermediate, n = 13: 258.54 [238.60-276.29]; higher, n = 21: 307.90 [285.92-321.95] ng/L; p < 0.001). In the exploratory ESC-aligned analysis, GPVI differed across risk categories (Kruskal-Wallis p = 0.045) and showed an ordinal association with increasing risk category (Spearman ρ = 0.331, p = 0.02). GPVI was higher in patients admitted to the ICU than in those not admitted (298.24 [278.42-315.89] vs. 206.17 [191.15-238.33] ng/L, p < 0.001) and had an ICU-admission AUC of 0.969. In a parsimonious model adjusted for PESI, GPVI remained associated with ICU admission (OR per 10 ng/L, 2.31; 95% CI, 1.36-3.92; p = 0.002), although the magnitude should be interpreted cautiously given the small sample size, limited number of ICU events, and strong separation. These effect estimates are exploratory and should not be interpreted as stable clinical effect sizes. Adding GPVI to PESI increased the apparent AUC from 0.885 to 0.968 (ΔAUC, 0.083; bootstrap 95% CI, 0.013-0.200); the optimism-corrected AUC was approximately 0.954, and repeated stratified 5-fold cross-validation yielded a mean AUC of approximately 0.916. In the secondary PE-versus-control analysis, GPVI was higher in PE (268.12 [215.39-298.47] vs. 114.07 [89.10-136.80] ng/L, p < 0.001) and showed an AUC of 0.968 versus 0.884 for D-dimer (ΔAUC, 0.084; bootstrap 95% CI, 0.008-0.168; p = 0.03); this case-control estimate is susceptible to spectrum bias and is not an estimate of real-world diagnostic accuracy.

Conclusion: Serum GPVI was associated with radiological thrombus burden and severity-related clinical indicators in acute PE. These exploratory findings support further evaluation of GPVI as a potential risk-stratification biomarker but do not establish diagnostic utility. Confirmation in larger, clinically representative prospective cohorts with hard clinical outcomes is required.