Background

D-dimer testing is known to have a high sensitivity at simultaneously low specificity, resulting in unspecific elevations in a multitude of conditions. 

Methods

The present study sought to assess diagnostic and prognostic features of D-dimers in cancer patients referred to the emergency department for suspected pulmonary embolism (PE) and deep vein thrombosis (DVT). We included a total of 526 cancer patients with a final adjudicated diagnosis of PE (n=83) and DVT (n=69), whereas 374 cancer patients served as the comparative group, in which venous thromboembolism (VTE) has been excluded.

Results

For the identification of VTE, D-dimers yielded the highest positive predictive value of 96% (95% confidence interval (CI), 85 to 99%) at concentrations of 9.9 mg/L and a negative predictive value of 100% at 0.6 mg/L (95% CI, 97 to 100%). At the established rule-out cut-off level of 0.5 mg/L, D-dimers were found to be very sensitive (100%) at a moderate specificity of nearly 65%. Using an optimized cut-off value of 4.9 mg/L increased the specificity to 95% for the detection of life-threatening VTE at the cost of moderate sensitivities (64%). During a median follow-up of 30 months, D-dimer concentrations positively correlated with mortality in both cancer patients with VTE (p<0.0001) and without VTE (p=0.0008).

Conclusions

Although D-dimer testing in cancer patients is discouraged by current guidelines, very high concentrations above the 10-fold upper reference limit contain diagnostic and prognostic information and might be helpful in risk assessment, while low concentrations remain useful for ruling out VTE.