JAMA Published online 5 January, 2021 | https://jamanetwork.com/journals/jama/article-abstract/2774681
Digest author: Emer Kelly, incoming ERS E-Learning Director / 12 March, 2021
Previous studies have reported a high frequency of pulmonary embolism in patients presenting with acute exacerbations of COPD: a meta-analysis showed a prevalence of 16%. This raises the question of how to screen these patients for PE at the time of their admission. The fact that clinical presentation of PE and exacerbations of COPD can be similar adds to the challenge.
This multicentre cross-sectional study with prospective follow-up took place in 7 French centres. Patients with COPD admitted with worsening of respiratory symptoms were screened for inclusion. The diagnosis of COPD was based on spirometry or a prior diagnosis of COPD by a pulmonologist. The main exclusion criteria were contraindications for CTPA, hospitalization for more than 48 hours before inclusion, pre-existing anticoagulation treatment, pregnancy, severe COPD or a life expectancy of less than 3 months. Patients were included within 48 hours of admission and were classified as having clinically suspected or non-suspected pulmonary embolism.
Clinical probability of PE was assessed using the revised Geneva score; high clinical probability patients directly proceeded to spiral CTPA and leg compression ultrasound. With low or intermediate clinical probability, a d-dimer was performed. In patients with a negative d-dimer PE was considered excluded with no further testing. In case of confirmed venous thromboembolism (VTE), anticoagulation was administered according to guidelines. All patients were followed for 3 months for symptoms or signs of VTE or bleeding. Primary outcome was PE diagnosed within 48 hours of admission. If CTPA was inconclusive a VQ scan or lower limb ultrasound was done. The main secondary outcome was PE during the 3 month follow up in those deemed not to have VTE in the 48 hours. Other secondary outcomes were VTE at admission, in the 3 months of follow up and all-cause mortality.
2268 patients with suspected COPD exacerbations were admitted to the 7 hospitals. 1518 were excluded, mainly due to long-term anticoagulation or inability to consent. 17 (2.3%) of the 740 patients had a high Geneva pretest clinical probability: PE was confirmed in 5 patients on CTPA and 2 had DVT. In the remaining 12 patients without PE, 4 had lower limb compression ultrasonography: 2 were negative and 2 had isolated DVT. US was not performed in 8 patients. In the 723 patients with low or moderate pretest probability, d-dimer was done in 712 patients. This was <500ug/ml in 212 patients and PE was considered excluded. 478 of the 500 patients with a positive D-dimer underwent CTPA: 36 had a PE, 13 had DVT. Among the 442 patients with a negative CTPA, 8 had isolated DVT, 339 had a negative lower limb US and 95 did not have the US. 22 patients with positive D-dimer did not have a CTPA and 6 of these had VQ scans that were negative. In the 11 patients in the low to moderate probability group d-dimer was not tested but CTPA was done demonstrating PE in 3 patients and no PE in the remaining 8.
Using this algorithm in 740 patients PE was confirmed in 44 patients (5.9%) and an additional 10 patients had an isolated DVT (1.4%) with an overall prevalence of VTE of 7.3%. In the 686 patients deemed not to have PE or DVT, 5 patients were shown to have PE in the 3 month follow up of the study. The overall 3 month mortality rate was 6.8%; 14 (25.9%) of 54 in the patients with VTE within 48 hours of admission and 36 (5.2%) of 686 without it. Among the 299 patients suspected of having PE at admission 30 (10%) were diagnosed versus 14 (3.2%) of the 441 not suspected. In those who had VTE ruled out at inclusion and were not anticoagulated only 0.7% developed PE in the 3 months follow up period.
In this study the overall prevalence of PE was lower than in previous meta-analysis. Limitations to this study were the lack of full access to lower limb compression ultrasound and also the use of non-age adjusted d-dimer (although this was addressed in post-hoc analysis). This is the largest prospective study and was consistent with previous studies in showing increased mortality in COPD exacerbation patients with co-existing PE.