Background Although the number of case reports on pulmonary infarction (PI) secondary to pulmonary embolism (PE) is increasing in recent years, its misdiagnosis remains common, mainly as pneumonia. In patients with intermediate and high-risk pulmonary embolism, delays in diagnosis and timely treatment would lead to poor prognosis.
Objective By analyzing the pneumonia-misdiagnosed cases of patients with PI, we summarized their clinical characteristics and related risk factors, and constructed a multivariate joint model to improve the accurate diagnosis rate at early stage.
Methods This retrospective study included the hospitalized patients with pulmonary embolism at the First Affiliated Hospital of USTC from January 2017 to December 2023. In the group of pneumonia-misdiagnosed patients with intermediate to high-risk PI, we analyzed the clinical characteristics and compared the differences between the misdiagnosed groups and control group. Furthermore, using a multivariate Logistic regression analysis, we explored the independent predictive factors of the delayed diagnosis, analyze the predictive value of various indicators for the misdiagnosis by ROC curves, and compared the AUC values using Delong test.
Results Among 101 cases of PI patients, 70 of them were misdiagnosed as pneumonia. From 2017 to 2023, the misdiagnosis rate gradually decreased in percentages of 100.0%, 83.3%, 74.1%, 71.4%, 63.2%, 66.7%, and 50.0%, respectively (χ2trend=6.672, P=0.010). Based on the results of multivariate Logistic regression analysis, the characteristics of over sixty-years-old age (OR=18.271, 95%CI=4.373-76.339, P<0.001), fever (OR=16.073, 95%CI=3.510-73.786, P<0.001), chest pain (OR=6.660, 95%CI=1.571-28.233, P=0.010) and non-dyspnea (OR=7.783, 95%CI=2.049-30.249, P=0.003) were independent predictive factors for the misdiagnosis. Therefore, a multivariate joint model was constructed as the following equation: Y=-6.624+0.095×A (factor of age) +2.510×F (factor of fever) +2.683×N (factor of non-dyspnea chest pain). The model indicated the PI misdiagnosis parameters as AUC under the curve (OR=0.880, 95%CI=0.802-0.959, P<0.001), sensitivity (0.871) and specificity (0.806). According to Delong's tests, the predictive values were superior to single-factor indicators of age (Z=2.771, P=0.006), fever (Z=4.653, P<0.001) and non-dyspnea chest pain (Z=4.014, P<0.001) .
Conclusion Although the misdiagnosis rate of pulmonary infarction has decreased in recent years, clinicians should keep alert to the differential diagnosis of pulmonary infarction and pneumonia in elderly PE patients with symptoms of fever and non-dyspnea chest pain.