D-dimer is a useful biomarker with a high negative predictive value among patients suspected of having deep vein thrombosis (DVT) and pulmonary embolism (PE). Among cancer patients, d-dimer is associated with DVT, PE and adverse outcomes including death. We have previously shown that patients with PE identified incidentally on routine cancer staging CT scans are frequently symptomatic, especially with fatigue and shortness of breath, and have adverse outcomes similar to cancer patients with suspected PE. It is not known if d-dimer levels among cancer patients with unsuspected PE (UPE) are similar to those in cancer patients with suspected PE or whether d-dimer levels correlate with symptoms or PE location.
We analyzed data from a prospective, single arm Phase II, IRB-approved clinical trial in which serial high sensitive D-dimer assays were performed in cancer patients diagnosed with DVT/PE and treated with a low molecular weight heparin (tinzaparin 175 units/kg). PE location was determined based on the most proximal segment of the pulmonary vasculature affected (main, lobar, segmental and subsegmental from most to least proximal). We also recorded signs and symptoms reported to the physician or nurse at the time of the PE diagnosis including tachycardia, hypoxia, fatigue, shortness of breath, dyspnea on exertion, cough, and chest pain. We used analysis of variance and linear regression models to evaluate the association between baseline clinical characteristics and baseline, pre-treatment d-dimer levels. We included log transformed d-dimer levels in all the models and report geometric means and 95% confidence intervals (CIs).
Among 91 treated patients, 62 had PE, of which 47 (75.8%) were UPE. The majority of patients (67%) had genitourinary tract or gastrointestinal tract malignancies, with the remainder having lung (n=11), pancreatic (n=7), or other (n=12) malignancies. Mean d-dimer levels at baseline were not significantly different between patients with UPE (mean 2377 ng/mL; 95% CI 1775–3182) and patients with suspected PE (mean 3694 ng/mL; 95% CI 2448–5574; p=0.11). Among signs and symptoms, only fatigue was significantly associated with higher d-dimer levels (d-dimer without fatigue: mean 1827 ng/mL, 95% CI 1370–2437 vs. with fatigue: mean 4727 ng/mL, 95% CI 2434–9177, p=0.0002); this association was also true for the subgroup of patients with UPE (p=0.0024). There was no significant association between fatigue and the location of PE. Among all patients with PE, there was a significant association between d-dimer and location of PE, with higher d-dimer levels associating with more proximal clots (p for trend=0.0023). Mean d-dimer levels based on the most proximal pulmonary artery segment involved were as follows: main pulmonary artery, 5198 ng/mL (95% CI 2858–9453); lobar artery, 2241 ng/mL (95% CI 1516–3314); segmental artery, 1816 ng/mL (95% CI 875–3769); and subsegmental, 931 ng/mL (95% CI 444–1951). This trend was also statistically significant when we restricted the analysis to UPE patients (p for trend=0.0005).
Cancer patients with UPE have elevated d-dimers similar to those found among cancer patients with suspected PE. Moreover, fatigue and more proximal clot location are significantly associated with higher d-dimer levels among cancer patients with UPE. Elevated d-dimers and presence of fatigue should prompt suspicion for PE among cancer patients.
Liebman:Sanofi Aventis: Honoraria; Glaxo Smith Kline: Honoraria; Esai: Honoraria; Pfizer: Honoraria. O'Connell:Astex Pharmaceuticals: Honoraria; Celgene: Speakers Bureau; Incyte: Honoraria.
Asterisk with author names denotes non-ASH members.