Knowledge of the D-dimer test result influences clinical probability assessment of pulmonary embolism

Background: In patients with suspected pulmonary embolism (PE), an unlikely or non-high probability assessment combined with a normal D-dimer test can safely exclude the diagnosis. We studied the influence of early D-dimer knowledge on clinical probability assessment. Methods: A questionnaire was sent to 150 randomly selected pulmonologists and internists in the Netherlands, presenting six hypothetical case-descriptions of patients with suspected PE. Physicians were randomized to receive one of three versions. The version contained a normal, an abnormal or no D-dimer result with each case-description. Each version contained two cases with an abnormal D-dimer result, two cases with a normal D-dimer result and two cases with no D-dimer result. Results: A total of 71 physicians (47%) returned the questionnaire; the three versions were equally represented. Compared to the control cases in which no D-dimer was given, knowledge of an abnormal D-dimer resulted in more "likely" clinical scores using the Wells' score (absolute increase in "likely" of 25-37%, p = 0.005, 0.111 and 0.144), while knowledge of a normal D-dimer resulted in more "unlikely" scores (absolute increase in "unlikely" of 27-44%, p = 0.001 and 0.070). D-dimer knowledge did not influence the probability assessment when the clinical suspicion was very high. Conclusion: Knowledge of the D-dimer test influences the physician in how the clinical probability for PE is scored. This will have direct clinical consequences, such as unnecessary imaging testing or inappropriate exclusion of the diagnosis. Physicians should therefore make sure that they examine the patient before they take notice of the D-dimer test result. (C) 2010 Elsevier Ltd. All rights reserved