BackgroundIn the diagnosis of pulmonary embolism (PE), the d-dimer threshold is based on studies conducted in Western countries, where the incidence rate is 5 times higher than that in Asian countries, including Japan. If we could elevate the d-dimer threshold based on the low pre-test probability in the Japanese population, we could omit the computed tomography pulmonary angiography (CTPA) which might lead to radiation exposure and contrast-induced nephropathy. Therefore, we aimed to determine a new d-dimer threshold specific to Japanese individuals.MethodsWe conducted a retrospective cohort study at an emergency department in Japan, using medical charts collected from January 2013 to July 2017. We included patients whose d-dimer were measured for suspicion of PE with low or intermediate probability of PE and CTPA were performed. The primary outcome was failure rate of the new d-dimer threshold, defined as the rate of PE detected by CTPA among patients with d-dimer under the new threshold ranging from 1000 to 1500 μg/L by 100. The new d-dimer threshold was appropriate if the upper limit of 95% confidence interval of the failure rate of PE was approximately 3%.ResultsIn 395 patients included, the number of patients with PE was 24 (the prevalence was 6.1%). If the d-dimer threshold was 1100 μg/L, the failure rate was 0% (0/119), the upper limit of the 95% confidence interval of the failure rate was 3.1%, and 30% (119/395) of the CTPA might be omitted.ConclusionThe new d-dimer threshold could safely exclude PE. This result can be generalized to other Asian populations with a lower incidence of PE. Further prospective studies will be needed.