Establishment of a Predictive Diagnostic Model for Acute Mycoplasma Pneumoniae Infection in Elderly Patients with Community-acquired Pneumonia

We established a diagnostic model to predict acute Mycoplasma pneumoniae (M. pneumonia) infection in elderly Community-acquired pneumonia (CAP) patients. We divided 456 patients into acute and non-acute M. pneumoniae infection groups. Binary logistic regression and receiver operating characteristic (ROC) curves were used to establish a predictive model. The following independent factors were identified: age ≥ 70 years; serum cTNT level ≥ 0.05 ng/mL; lobar consolidation; mediastinal lymphadenopathy; and antibody titer in the acute phase ≥ 1:40. The area under the ROC curve of the model was 0.923 and a score of ≥ 7 score predicted acute M. pneumoniae infection in elderly patients with CAP. The predictive model developed in this study has high diagnostic accuracy for the identification of elderly acute M. pneumoniae infection.