Outcomes of M. pneumoniae pneumonia with mixed infection

2.1 Study population

This retrospective research was conducted over a 48-month period (January 2014 to December 2017) and included children with severe MPP at a tertiary hospital in Fujian Province, P. R. China. M. pneumoniae infection was diagnosed based on positive serological results for total antibody ($\ge$ 1 : 640), IgM antibody and related clinical manifestations (dry cough, fever, and radiological pulmonary abnormalities). The severity of MPP was determined according to standard guidelines [11] (see the Supplementary material).

The exclusion criteria were abnormal development of the airway, presence of a congenital disease, presence of a genetically transmitted metabolic or autoimmune disease, presence of a neoplastic disease, malnutrition, human immunodeficiency virus infection, tuberculous infection, and incomplete clinical data.

2.2 Demographic and clinical data

The following information was collected from the medical records of the included patients: (1) basic demographic and clinical data (age, gender, hospitalization time, provision of supplemental oxygen, and duration of fever); (2) the data from laboratory tests performed within 24 h of admission (C-reactive protein [CRP] level, d-dimer levels, hemoglobin [Hb] level, lactate dehydrogenase [LDH] level, platelet [PLT] count, pre-albumin [PA] level, procalcitonin [PCT] level, red blood cell volume distribution width [RDW], and white blood cell [WBC] count), and the data from blood culture exam, chest radiography, and electrocardiography; (3) data on indicators of the degree of disease severity (systolic blood pressure, temperature, mental health status, status of the gastrointestinal system, heart beat rate, pupillary reflexes, glucose level, sodium level, potassium level, creatinine level, urea level, Hb level, PLT count, prothrombin time or partial thromboplastin time, WBC count, acidosis [pH], total CO${}_2$, PCO${}_2$, and arterial PaO${}_2$). Based on the worst values of the indicators of disease severity that were obtained within 24 h after hospital admission, the Pediatric Risk of Mortality (PRISM) III score (higher PRISM III score levels being associated with increased estimated mortality) and Pediatric Critical Illness Score (PCIS) were calculated. Within 72 h after admission, serum samples from the patients were examined for co-infection with the help of blood culture and deep sputum cultures for bacterial infection (a positive response was defined as two consecutive positive results for the same bacterial species), passive agglutination test for antibodies against Mpneumonia (FUJIREBIO INC, Japan), indirect fluorescence test for antibodies against nine noxaes that are commonly implicated in co-infection (adenovirus, Chlamydia pneumoniae, Coxiella burnetti, influenza viruses A and B, Legionella pneumophila, M. pneumoniae, parainfluenza viruses 1, 2, and 3, and respiratory syncytial virus) (VIRCELL.S.L PNEUMOSLIDE IgM,Spain), fluorometric PCR assay for Epstein-Barr virus and cytomegalovirus (Sun Yat-sen University Daan Gene C, Ltd . P. R. China), colloidal gold technique for enterovirus type 71 (Beijing Wantai Biological Pharmaceutical Co, Ltd. P. R. China), and ELISA for Herpes simplex virus (Virion-Serion R & D Co, Ltd. Germany).

2.3 Ethical compliance

The present research has received the approval of our hospital’s ethics committee (approval no. 2017-042). The patients, as well as the members of their family, provided their informed consent for participation in this study, and their data were anonymized.

2.4 Statistical analysis

All data were analyzed with the help of the software SPSS version 23.0 (IBM). The results obtained from descriptive analyses are presented as absolute frequencies or rates for categorical variables, as medians (Q${}_{25}$-to-Q${}_{75}$ values) for quantitative variables with non-parametric distribution, and as mean $\pm$ SD for quantitative variables with normal distribution. Between-group comparison of quantitative variables was performed by one-way analysis of variance or the Kruskal-Wallis test, as applicable. Comparison of categorical variables was done with the $\chi$${}^2$ test. Binary multivariate logistic regression analysis was used to evaluate the association of co-infection with the severity of MPP and the hospitalization period. Statistical significance was assumed at P values below 0.05.

3.1 Demographic data

Between January 2014 and December 2017, 205 children with severe MPP were admitted to our hospital, but 10 were excluded because of incomplete clinical data. Further, in the 11 other cases that were excluded, the reason for exclusion was malnutrition in 3 cases, congenital tracheal stenosis in 4 cases, methylmalonic acidemia in 2 cases, and acute lymphoblastic leukemia in 2 cases (Fig. 1). Thus, the final number of included patients was 184 (122 boys and 62 girls; male-to-female ratio, 1.97 : 1).

Fig. 1.Flow chart showing the selection and exclusion of the patient population in this study.

The patients were classified into the following three groups: single M. pneumoniae infection (n = 66), mixed infection with a single noxa other than M. pneumoniae (n = 58), and mixed infection with two or more noxae other than M. pneumoniae (n = 60). These three groups were matched for gender and age (P $>$ 0.05; Table 1). For the analysis, the study population was segmented into two severity levels based on the PRISM III scores: the group with score $>$ 5 (n = 48) and the group with score $\le$ 5 (n = 136) [12].

Mixed infection with viral or bacterial pathogens was detected in 118 children (64.1%), and 200 noxae were detected from 184 encounters (Table 2). In 102 check-ups (55.4%), the mixed infection involved viruses, while in 29 check-ups (15.8%), the mixed infection involved bacteria. Eleven patients tested positive for three different noxae other than M. pneumoniae; two patients tested positive for four different noxae other than M. pneumoniae; one patient tested positive for five different noxae other than M. pneumoniae; and one patient tested positive for six different noxae other than M. pneumoniae. Cytomegalovirus was isolated in 36 of the 102 (35.3%) patients with viral co-detection; Epstein-Barr virus was isolated in 30 of the 102 patients (29.4%); Chlamydia pneumoniae, in 19 patients (18.6%); influenza B virus, in 12 patients (11.8%); and respiratory syncytial virus, in 9 patients (8.8%).

Klebsiella pneumoniae, Escherichia coli, and Staphylococcus aureus were isolated in 5 patients each of the 29 (17.3%) patients with bacterial co-infection, and Pseudomonas aeruginosa was isolated in 4 of the 29 cases (13.8%). The percentage of patients with bacterial mixed infection was not significantly different between the three groups. There was no significant difference in the co-detection rates of noxae based on patient age or gender, or based on prematurity, respiratory comorbidity, cardiovascular complications, or adverse outcomes.

3.2 Clinical data

Fever was the most frequent symptom, and the duration of fever was not significantly different between the three groups. The proportion of infants who required invasive intubation significantly rose as the mixed infection rate increased (P$<$ 0.05; Table 1). Extrapulmonary complications were also common among the patients. The proportion of patients with heart failure and extrapulmonary complications that involved two or more systems significantly differed between the three groups; however, the rates of prematurity, non-invasive ventilation, pulmonary complications, and cardiovascular complications were not significantly different. With regard to laboratory examinations, increase in the number of pathogens was associated with a progressive increase in the proportion of patients with total length of stay in hospital (TLSH) $>$ 14 days, PCIS $<$ 80, and mixed infection of bacteria (P$<$ 0.05). A higher number of pathogens was also associated with a significant increase in TLSH, d-dimer levels, PRISM III score, and the estimated mortality based on PRISM III (P$<$ 0.05). However, the WBC count, Hb level, PLT count, and RDW, as well as the CRP, PCT, PA, and LDH levels were not significantly different based on the number of pathogens (Table 1).

3.3 Correlation between mixed infection and severity of severe MPP

The distribution of the PRISM III scores was skewed; therefore, the scores were transformed with the Blom formula so that their distribution was normal. For the purpose of statistical analysis, the study population was segmented into two severity levels: the severity level was considered as 0 when the PRISM III score was $\le$ 5, and it was considered as 1 when the PRISM III score was $>$ 5. Binary logistic regression analysis was used to identify the variables that showed an independent association with the PRISM III score and PCIS. We found that mixed infection with more than one pathogen (other than M. pneumoniae) was positively associated with the PRISM III score ( β = 0.760, odds ratio [OR] = 2.139, 95% confidence interval [CI] = 1.391-2.390, P = 0.001) and PCIS ( β = 1.203, OR = 3.328, 95% CI = 1.723-6.731, P = 0.000).

3.4 Correlation between mixed infection and TLSH

The distribution of the TLSH data was skewed; therefore, the data were transformed with the Blom formula to ensure the distribution was normal. For the purpose of statistical analysis, the study population was segmented into two severity levels: the severity level was considered as 0 when TLSH was $\le$ 14 days, and it was considered as 1 when TLSH was $>$ 14 days. Binary logistic regression analysis was conducted to identify variables that showed an independent association with TLSH. We found that mixed infection with more than one pathogen (other than M. pneumoniae) was positively correlated with TLSH ( β = 0.730, OR = 2.075, 95% CI = 1.404-3.066, P = 0.000).