Therapeutic Efficacy and Safety of Prolonged Macrolide, Corticosteroid, Doxycycline, and Levofloxacin against Macrolide-Unresponsive Mycoplasma pneumoniae Pneumonia in Children

Ha, Seok Gyun; Oh, Kyung Jin; Ko, Kwang-Pil; Sun, Yong Han; Ryoo, Eell; Tchah, Hann; Jeon, In Sang; Kim, Hyo Jeong; Ahn, Jung Min; Cho, Hye-Kyung

We compared the TTD after initial macrolide treatment between PMC group and each secondary treatment group through PS matching to reduce selection bias and to control potential confounding factors. The estimated PS for being assigned to each group was calculated for each patient using multiple logistic regression models with the following covariates: age, sex, radiographic findings, mixed infection with other pathogens. Patients were matched at a ratio of PMC:CST = 2:1, PMC:DXC = 1:1, and PMC:LFX = 2:1, based on a greedy 8-1 digits matching algorithm. This algorithm attempted to match the PMC subjects and the secondary treatment subjects on the first 8 digits of the PS. The PMC subjects that did not match were then matched to secondary treatment subjects on 7 digits of the PS. We processed through the algorithm sequentially to the 1-digit match on the PS. The patients with no corresponding match were excluded. PS matching was performed using SAS 9.1.3 (SAS Institute Inc., Cary, NC, USA). P values of < 0.05 were considered to be statistically significant.