Background: D-dimer is routinely measured as part of the clinical diagnosis algorithms for venous thromboembolism (VTE). In these algorithms, low D- dimer cut-off values are used to generate a dichotomous test result that is sensitive, but very non-specific for VTE. A consequence of any test dichotomisation is loss of information that is hidden in the continuous spectrum of results. For D-dimer, the information conveyed by extremely elevated results may be particularly relevant. Our aim was to assess the differential diagnosis of extremely elevated D-dimer levels in a hospital setting.
Methods: Retrospective cohort study of patients > 18 years with an extremely elevated (> 5000 μg/l; > 10x cut-off to exclude VTE) D-dimer test result. Electronic medical records were reviewed for diagnoses.
Results: A total of 759 extremely elevated D-dimer results were identified. After exclusion of 120 duplicate cases, 53 patients undergoing cardiopulmonary resuscitation, and 5 cases without diagnostic information, 581 cases were analysed. Their D-dimer ranged between 5030 and 239,000 μg/l, with a mean of 17,598 μg/l (SD 22,972 μg/l). Altogether, 89% of these patients had a diagnosis of VTE, sepsis and/or cancer. The prevalence was highest for pulmonary embolism (183 patients; 32%), followed by cancer (168 patients; 29%), sepsis (142 patients; 24%), trauma/surgery (142 patients; 24%), and deep vein thrombosis (73 patients; 13%).
Conclusion: Although D-dimer testing has a reputation for being very non-specific, an extremely elevated D-dimer is uniquely associated with severe disease, mainly including VTE, sepsis and/or cancer. These results suggest that, even if sharply elevated D-dimers are a seemingly solitary finding, clinical suspicion of severe underlying disease should be maintained.