Objective  To investigate the associated factors of acute exacerbation of chronic obstructive pulmonary disease (AECOPD) complicated by pulmonary thromboembolism/pulmonary thrombus formation (PTE/PTS).Methods  A retrospective analysis was conducted on 121 patients diagnosed with AECOPD at the First Affiliated Hospital of Fujian Medical University between January 2013 and December 2023, who underwent complete computed tomography pulmonary angiography (CTPA) or radionuclide ventilation/perfusion (V/Q) imaging. Based on imaging findings, 44 patients diagnosed with PTE/PTS were assigned to the PTE/PTS group, while the remaining 77 patients formed the non-PTE/PTS group. Baseline characteristics, comorbidities, clinical manifestations, and auxiliary examination results between the two groups were compared. Binary logistic regression models were employed to identify independent risk factors for AECOPD complicated by PTE/PTS. Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic performance of relevant indicators for AECOPD complicated by PTE/PTS.

Results  Significant differences existed between groups in female gender, age, hypertension, asymmetric lower limb edema, D-dimer, troponin I (cTnI), brain natriuretic peptide (BNP), hemoglobin (HGB), lower limb deep vein thrombosis (DVT), right heart enlargement, pulmonary hypertension, and pleural effusion (P ＜ 0.05). Multivariate logistic regression analysis revealed significant associations between female gender (OR = 5.777, 95%CI: 1.667 to 20.027), D-dimer (OR = 1.158, 95%CI: 1.015 to 1.322), BNP (OR = 1.001, 95%CI: 1.000 to 1.002), and AECOPD with PTE/PTS (P ＜ 0.05). ROC curve analysis results indicated that the optimal cutoff value for D-dimer was 0.565 mg/L (sensitivity 86.4%, specificity 49.4%), with an area under the curve (AUC) of 0.717 (95%CI: 0.626 to 0.808, P ＜ 0.001); The optimal cutoff for BNP was 222 ng/L (sensitivity 56.8%, specificity 20.8%), with an AUC of 0.702 (95%CI: 0.605 to 0.800, P ＜ 0.001). Combining D-dimer and BNP improved predictive performance (AUC = 0.753, 95%CI: 0.662 to 0.843), however, both sensitivity (65.9%) and specificity (27.3%) were limited.

Conclusion  Female gender, elevated D-dimer, and elevated BNP may be independent predictors of AECOPD complicated by PTE/PTS and hold diagnostic value. In patients with AECOPD, D-dimer serves as a sensitive screening marker for PTE/PTS, while BNP acts as an independent complementary predictor. Combined assessment of both markers may improve predictive performance, however, the combined model exhibited low specificity and therefore cannot yet be directely applied for clinical stratification.