ABSTRACT
Importance: Ruling out pulmonary embolism (PE) among patients presenting to the Emergency Department (ED) with suspected or confirmed SARS-COV-2 is challenging due to symptom overlap, known increased pro-thrombotic risk, and unclear D-dimer test interpretation. Objective: Our primary objective was to assess the diagnostic accuracy of standard and age-adjusted D-dimer test thresholds for predicting 30-day pulmonary embolism (PE) diagnosis in patients with suspected SARS-COV-2 infection. Design, Setting, and Participants: This was a retrospective observational study using data from 50 sites enrolling patients into the Canadian COVID-19 ED Rapid Response Network (CCEDRRN) registry between March 1, 2020 to July 2, 2021. Adults ([≥]18 years) with SARS-COV-2 testing performed at index ED visit were included if they had any of the following presenting complaints: chest pain, shortness of breath, hypoxia, syncope/presyncope, or hemoptysis. We excluded patients with duplicate records or no valid provincial healthcare number. Main Outcomes and Measures: Our primary end point was 30-day PE diagnosis based on a positive computed tomography pulmonary angiogram (CTPA) or hospital discharge diagnosis code of PE. The outcome measure was the diagnostic accuracy of an age adjusted D-dimer strategy as compared to absolute D-dimer thresholds (500 - 5000 ng/mL). Results: 52,038 patients met inclusion criteria. Age-adjusted D-dimer had a sensitivity (SN) of 96% (95% CI 93-98%) and a specificity (SP) of 48% (95% CI 48-49%) which was comparable to the most sensitive absolute threshold of 500 ng/mL (SN 98%, 95% CI 96-99%; SP 41%, 95% CI 40-42%). Other absolute D-dimer thresholds did not perform well enough for clinical reliability (SN <90%). Both age-adjusted and absolute D-dimer performed better in SARS-COV-2 negative patients as compared to SARS-COV-2 positive patients for predicting 30-day PE diagnosis (c-statistic 0.88 vs 0.80). Conclusions and Relevance: In this large Canadian cohort of ED patients with suspected SARS-COV-2 infection, an age-adjusted D-dimer strategy had similar sensitivity and superior specificity to the most sensitive D-dimer threshold of 500 ng/mL for predicting 30-day PE diagnosis irrespective of SARS-COV-2 infection status. Adopting an age-adjusted D-dimer strategy in patients with suspected SARS-COV-2 may help avoid unnecessary CTPA testing without compromising safety.