Sex-related differences in D-dimer levels for venous thromboembolism screening

Abstract

Background: D-dimer is generally considered positive above 0.5 mg/L irrespective of sex. However, women have been shown to be more likely to have a positive D-dimer after controlling for other factors. Thus, differences may exist between males and females for using D-dimer as a marker of venous thromboembolic (VTE) disease. We hypothesized that the accuracy of D-dimer tests may be enhanced by using appropriate cutoff values that reflect sex-related differences in D-dimer levels.

Methods: This research is a secondary analysis of a multicenter, international, prospective, observational study of adult (18+ years) patients suspected of VTE, with low-to-intermediate pretest probability based on Wells criteria ≤ 6 for pulmonary embolism (PE) and ≤ 2 for deep vein thrombosis (DVT). VTE diagnoses were based on computed tomography, ventilation perfusion scanning, or venous ultrasound. D-dimer levels were tested for statistical difference across groups stratified by sex and diagnosis. Multivariable regression was used to investigate sex as a predictor of diagnosis. Sex-specific optimal D-dimer thresholds for PE and DVT were calculated from receiver operating characteristic analyses. A Youden threshold (D-dimer level coinciding with the maximum of sensitivity plus specificity) and a cutoff corresponding to 95% sensitivity were calculated. Statistical difference for cutoffs was tested via 95% confidence intervals from 2,000 bootstrapped samples.

Results: We included 3,586 subjects for analysis, of whom 61% were female. Race demographics were 63% White, 27% Black/African American, and 6% Hispanic. In the suspected PE cohort, 6% were diagnosed with PE, while in the suspected DVT cohort, 11% were diagnosed with DVT. D-dimer levels were significantly higher in males than females for the PE-positive group and the DVT-negative group, but males had significantly lower D-dimer levels than females in the PE-negative group. Regression models showed male sex as a significant positive predictor of DVT diagnosis, controlling for D-dimer levels. The Youden thresholds for PE patients were 0.97 (95% CI = 0.64 to 1.79) mg/L and 1.45 (95% CI = 1.36 to 1.95) mg/L for females and males, respectively; 95% sensitivity cutoffs for this group were 0.64 (95% CI = 0.20 to 0.89) and 0.55 (95% CI = 0.29 to 1.61). For DVT, the Youden thresholds were 0.98 (95% CI = 0.84 to 1.56) mg/L for females and 1.25 (95% CI = 0.65 to 3.33) mg/L for males with 95% sensitivity cutoffs of 0.33 (95% CI = 0.2 to 0.61) and 0.32 (95% CI = 0.18 to 0.7), respectively.

Conclusion: Differences in D-dimer levels between males and females are diagnosis specific; however, there was no significant difference in optimal cutoff values for excluding PE and DVT between the sexes.

Keywords: D-dimer; age-adjusted; deep vein thrombosis; diagnostic testing; logistic regression; pulmonary embolism; receiver operating characteristic; sex-adjusted; venous thromboembolism.