Retrospective analysis of multiple factors combination for early identification of pulmonary tuberculosis complicated with pulmonary embolism / 中华传染病杂志
The univariate analysis showed that patients with pulmonary tuberculosis complicated with pulmonary embolism had higher ratio of chest tightness (67.6%(23/34) vs 22.1%(30/136)), syncope (23.5%(8/34) vs 0.7%(1/136)), fever (55.9%(19/34) vs 36.0%(49/136)), hemostaticdrug use (100.0%(34/34) vs 13.2%(18/136)), history of venous thrombosis (8.8%(3/34) vs 0.7%(1/136)), atrial fibrillation (11.8%(4/34) vs 2.2%(3/136)) and D-dimer levels (4.090 0(1.035 0, 10.790 0) mg/L vs 0.850 0(0.432 5, 2.145 0) mg/L) than those of simple pulmonary tuberculosispatients, and the differences were all statistically significant ( χ2=26.35, 28.19, 4.47, 96.44, 7.75, 6.30 and Z=-4.65, respectively; all P<0.050). The arterial partial pressure of oxygen (PaO 2)(61.90(52.95, 73.00) mmHg vs 82.00 (75.00, 87.00) mmHg, 1 mmHg=0.133 kPa) and albumin ((28.83±4.98) g/L vs (32.76±5.65) g/L) of patients with pulmonary tuberculosis complicated with pulmonary embolism were lower than those of simple pulmonary tuberculosispatients, and the differences were both statistically significant ( Z=-5.21 and t=3.71, respectively, both P<0.001). Binary regression analysis showed that chest tightness (odds ratio ( OR)=3.494, 95%confidence interval ( CI) 1.208 to 10.100, P=0.021), D-dimer ( OR=1.285, 95% CI 1.079 to 1.530, P=0.005) and PaO 2( OR=0.931, 95% CI 0.895 to 0.970, P=0.001) were the independent diagnostic indicators for pulmonary embolism in pulmonary tuberculosispatients. The areas under the ROC curve of chest tightness, D-dimer, PaO 2, and the combination of the three indicators (the combination factor) were 0.728, 0.758, 0.834, and 0.890, respectively. The optimal cut-off value of the combination factor was -3.1, with the sensitivity of 0.824 and the specificity of 0.824.