The diagnosis of deep venous thrombosis (DVT) is mostly dependent upon duplex ultrasound (DUS) examination of the affected extremity. Data from our laboratory implicates P-Selectin and microparticles (MPs) in the pathogenesis of venous thrombosis. A prospective study was performed to determine the sensitivity and specificity of plasma assays for D-Dimer, soluble P-selectin (P-selectin), and total MPs in patients with documented DVT by DUS. Three groups of individuals were examined: Group 1, 30 normal volunteers; Group 2, 22 patients positive for DVT on DUS (DVT); and Group 3, 21 patients with symptoms of leg pain but negative DUS for DVT (SMP). D-Dimer was measured by the Advanced D-Dimer assay (Dade-Behring, Deerfield, IL), P-Selectin was measured by ELISA (R&D, Minneapolis, MN.) and MPs were assayed by flow cytometry. Total MPs are cell fragments detected by anti-CD41 and anti-CD11b that are < 1 micron, as seen on flow cytometry. Group 1 individuals have D-Dimer levels of 1.53±0.12 mg/dl, P-Selectin of 0.34±0.05 ng/mg total protein (TP), and total MP of 370,103±41,910 particles/200 microliters of platelet-poor plasma. No differences in age (mean=48 vs. 51 yrs, respectively), weight, BMI, use of OCP/HRT, smoking, family history of DVT, or trauma history were noted between DVT and SMP patients. Patients with DVT were more likely to have traveled recently or have a malignancy present. 100% of DVT patients were at highest risk (score≥5) for thrombosis by thrombosis risk assessment (

Caprini, JA,
J. Thromb. Thrombolysis
,
9
:
253
,
2000
) while only 62% of SMP patients were at highest risk. Group 2 individuals (DVT) have D-Dimer values of 8.20±2.03 mg/dl, P-Selectin of 0.98±0.11 ng/mg TP, and MPs 129±17% of control. Group 3 individuals (SMP) have D-Dimer values of 3.12±0.79 mg/dl, P-Selectin of 0.55±0.08 ng/mg TP, and MPs 99±12% of control. Differences were statistically significant between groups 2 and 3 for D-Dimer (p=0.01) and P-selectin (p<0.01), while not significantly different for MP (p=0.18). Using a logistic regression model with dichotomous variables, we determined a sensitivity of 73%, a specificity of 81%, and an accuracy of 77% when combining D-Dimer, soluble P-selectin, and total MPs to differentiate DVT from SMP patients. The single variable most predictive for thrombosis was soluble P-Selectin.

Logistic Regression Analysis Using Dichotomous Variables.

VariablesThreshold ValueSensitivitySpecificityAccuracy
D-Dimer 3 mg/dl 64% 76% 70% 
Soluble P-Selectin 0.68 ng/ml TP 68% 81% 74% 
Total Microparticles 125% (compared to controls) 50% 67% 58% 
Combined Variables  73% 81% 77% 
VariablesThreshold ValueSensitivitySpecificityAccuracy
D-Dimer 3 mg/dl 64% 76% 70% 
Soluble P-Selectin 0.68 ng/ml TP 68% 81% 74% 
Total Microparticles 125% (compared to controls) 50% 67% 58% 
Combined Variables  73% 81% 77% 

A logistic regression model using continuous variables yielded similar results and was statistically significant (p=0.05) to distinguish DVT from SMP patients. This study demonstrates that combined plasma assays for venous thrombosis markers achieve moderate sensitivity and specificity in differentiating DVT from SMP patients. It also suggests that P-Selectin participates in the pathophysiology of DVT. More studies with larger numbers of patients are planned to confirm this assessment.

Supported by NIH 70766 (TWW).

Author notes

Corresponding author

Sign in via your Institution