Radiologic findings as a determinant and no effect of macrolide resistance on clinical course of Mycoplasma pneumoniae pneumonia

Subjects

Between September 2010 and December 2015, five hospitals participated in this study, including Seoul National University Children’s Hospital (Seoul), Seoul National University Bundang Hospital (Seongnam), Eulji Hospital (Seoul), Chungnam University Hospital (Daejeon), and Chonnam University Hospital (Gwangju). Cases from three hospitals were included from June 2014 to December 2015 to collect pneumonia cases from three different regions distant from Seoul, while those from Seoul National University Children’s Hospital and Eulji Hospital were enrolled throughout the study period.

This study enrolled the children who were hospitalized with a confirmed diagnosis of MP pneumonia and also were treated with macrolides without introducing or changing with tetracyclines or fluoroquinolones. Nasopharyngeal aspirates or swabs were obtained from children hospitalized with acute lower respiratory tract infections (LRTIs) as a strategy of routine patient care. Diagnosis of MP pneumonia was based on 1) the presence of infiltration in chest radiography together with clinical symptoms and signs such as fever (defined as ≥38 °C by axillary temperature), cough, or sputum and 2) MP detected by culture or positive test results for the MP P1 gene by PCR. MP-specific antibody was measured using a microparticle agglutination test Serodia-MycoII kit (Fujirebio, Tokyo, Japan). A 4-fold or greater rise in titers or a single titer of ≥1: 640 were considered to indicate current MP infection [5]. The following conditions were excluding criteria: underlying diseases that can modify the clinical course of MP pneumonia, hospital-acquired pneumonia, bacterial coinfection, and treatment with other anti-mycoplasmal antibiotics such as levofloxacin or tetracycline.

Nasopharyngeal samples were kept at 4 °C and sent to the laboratory of the Seoul National University Children’s Hospital for the culture of MP and analysis of macrolide resistance within 24 h from the collection of the samples.

Mycoplasma pneumoniae detection

Cultivation of MP was performed at the Seoul National University Children’s Hospital. Reference strain M129 (ATCC 29342) was cultured in parallel with the clinical samples using pleuropneumonia-like organism (PPLO) broth and agar. Two hundred microliters of the nasopharyngeal specimen were serially diluted 64-fold. The broth medium was composed of 70 mL of PPLO broth, 20 mL of horse serum, 10 mL of 25% yeast extract, 2.5 mL of 20% glucose, 200 μL of 1% phenol red, 1 mL of 2.5% thallium acetate, 0.5 mL of 200,000 units/mL penicillin G potassium, and 0.5 mL of 20,000 μg/mL cefotaxime. The agar was prepared with the same components as the broth medium except that cefotaxime was omitted and 1.2% agar powder was added instead of broth powder. The broth and the agar media were incubated aerobically at 37 °C for 6 weeks.

The plates were observed daily to identify color changes in the broth medium from red to transparent orange. Upon color change, 10 μL were sub-cultured onto agar plates. Spherical MP colonies were observed under a microscope at 100X magnification. In addition, P1 gene was amplified by PCR for the confirmation of MP. DNA was also extracted directly from the nasopharyngeal samples and subsequently amplified for P1 gene. Each reaction mixture was prepared with the following: 2 μL of template, 2 μL of 10X reaction buffer, 2.0 mM of MgCl₂, 0.2 mM of dNTP mix, 4 pmoL of each forward and reverse primer, 1 U of Taq DNA polymerase (Takara Bio Inc., Shiga, Japan), and distilled water to a final volume of 20 μL. Reaction conditions were set as follows: denaturation at 95 °C for 30 s, 35 cycles of annealing at 58 °C, and elongation at 72 °C for 40 s. Cases were included when cultures grew MP that was subsequently confirmed to be positive for the P1 gene by PCR. Cases with positive MP PCR of the nasopharyngeal aspirates were also included.

Determination of macrolide resistance

PCR to amplify domain V of the 23S rRNA gene was performed on cultured MP isolates or DNA extracted from nasopharyngeal samples. The primers used were MP23SV-F (5′-TAA CTA TAA CGG TCC TAA GG) and MP23SV-R (5′-ACA CTT AGA TGC TTT CAG CG). DNA from the reference strain M129 (ATCC 29342), was used as a positive control, and distilled water was used as a negative control. The 851- bp PCR products were purified using an AccuPrep® PCR Purification Kit (Bioneer, Inc., Daejeon, Korea), and samples were sequenced to identify the transitions in domain V of the 23S rRNA gene that have been associated with macrolide resistance [13]. The presence of 23S rRNA gene transitions were regarded as resistant to macrolides. For 56 cultured strains, the minimal inhibitory concentrations (MICs) were determined by use of a microdilution method. Briefly, frozen stocks were thawed and diluted in a broth media to a concentration of 10⁴ color change units (CCU)/mL. The MICs of MP strains were then measured using microdilution method in triplicate for the following 4 antimicrobial agents: azithromycin, clarithromycin, erythromycin, and roxithromycin. Each antimicrobial was serially diluted 2-fold, with a range from 0.0008 μg/mL to 128 μg/mL. MIC was defined as the lowest antibiotic concentration at which color did not change at the time when the color of positive control (media containing MP strains only) changed [14]. If negative control (media only) showed color change, the test was discarded. MIC of the reference strain M129 was determined as a control.

Clinical data collection and assessment of clinical outcome

The patients’ demographics, clinical manifestations, number of febrile days, duration of macrolide treatment, and the laboratory results were collected on a standardized form through review of the medical records by a pediatrician in each hospital. A febrile day was defined as a day on which the body temperature exceeded 38 °C at least once. Clinical parameters such as oxygen requirement, mechanical ventilator support were reviewed. Chest x-rays at the time of admission were reviewed independently by two radiologists. Chest x-ray findings were categorized into homogeneous dense lobar consolidation, patchy consolidation, nodular opacities, and bilateral parahilar infiltration (Fig. 1). Parapneumonic effusion, defined as ≥1 cm width on the decubitus view, was also included in the analysis. The pediatricians and radiologists were all blinded to the results of the macrolide susceptibility of MP. In the five participating hospitals, macrolide agents were initially chosen for the treatment of MP pneumonia. All physicians were unware of the results of 23S rRNA mutation during the treatment. The results from the susceptibility test did not have any impact on the choice of antibiotics because the susceptibility results take up to 6 weeks.

Statistical analyses

The data were analyzed using SPSS software version 19.0 (SPSS Inc., Chicago, IL, USA). For all of the statistical tests, a p-value of <0.05 was considered statistically significant. Differences between categorical variables were tested using the chi-squared or Fisher’s exact test with univariate and multivariate logistic regression analysis. The Mann-Whitney U test or t-test were used to compare the age and duration of febrile days between the groups as appropriate.

Demographics of patients with MP pneumonia

Between September 2010 and December 2015, a total of 2436 nasopharyngeal samples were obtained from children hospitalized with acute LRTIs. Of those, MP was detected in 250 samples (10.3%). A total of 116 children who were hospitalized with MP pneumonia were included after excluding cases that had any of the exclusion criteria such as underlying diseases, hospital-acquired pneumonia, bacterial coinfection, use of quinolone or tetracycline. Overall, 50.9% of the patients were male, and the mean age was 6.0 years; 9.5% were <2 years old, 31.9% were 2–4 years old, and 58.6% were ≥5 years old. The mean duration of fever was nine days (range 1–30 days), and the mean days of admission were 9 days. MP-specific antibody titers were ≥4-fold or greater in 52 (44.8%) patients and single antibody titers ≥1:640 in 49 (42.2%) patients. Remaining 15 patients did not have paired serum samples.

Prevalence of macrolide resistance

The annual distribution of 116 MP-positive samples over 6 years is shown in Fig. 2. During the study period, MP pneumonia was prevalent in 2011 and in 2015. Overall, 82 samples (70.7%) carried transitions in the 23S rRNA gene. The transition observed in the 23S rRNA gene was exclusively the A2063G transition. The macrolide resistance rates during the two outbreaks were 53.8% (21/39) in 2011 and 84.6% (44/52) in 2015. Of the 59 MP isolates that underwent MIC testing for four macrolides (azithromycin, clarithromycin, erythromycin, and roxithromycin), the MICs of 23 MP isolates that lacked the transition in the 23S rRNA gene ranged from 0.001 to 0.008 μg/mL. In contrast, the MICs of the 36 MP isolates that had a transition in the 23S rRNA gene were significantly higher, ranging from 2 to >128 μg/mL (Table 1).

Macrolides	MICs (μg/mL) for strains
	With 23S rRNA transition (n = 36)			Without 23S rRNA transition (n = 23)
	Range	MIC50	MIC90	Range	MIC50	MIC90
Azithromycin	1 ~ 64	8	16	0.001 ~ 0.001	0.001	0.001
Clarithromycin	8~ > 128	64	128	0.001 ~ 0.002	0.001	0.002
Erythromycin	2~ > 128	16	128	0.001 ~ 0.004	0.001	0.002
Roxithromycin	0.032 ~ 128	8	32	0.001 ~ 0.008	0.001	0.004

MIC₅₀ and MIC₉₀ are MICs at which 50% and 90% of the isolates were inhibited by the antimicrobials, respectively

Clinical features of MP pneumonia

Overall, the mean days of fever were nine days. There was no significant difference in the total duration of fever between the macrolide-susceptible and macrolide-resistant groups (9.4 days vs. 8.9 days, P = 0.322). Most patients were tested for the detection of MP no later than 7 days except 14 (12%) patients who presented after 7 days of fever. While the number of febrile days before macrolide treatment did not differ between the groups (4.9 days vs. 4.3 days, P = 0.315), the mean days to defervescence after macrolide initiation were longer in the macrolide-resistant group than in macrolide-susceptible group (5.7 days vs. 4.1 days, P = 0.021). After the initiation of macrolide treatment, more cases had a fever for more than 7 days in the macrolide-resistant group compared with those in the macrolide-susceptible group (37.8% vs. 17.6%, P = 0.034). There were no differences in oxygen requirement, mechanical ventilator support between the groups. Significant differences were not observed between the two groups in the presence of extrapulmonary signs.

Relationship between clinical course and findings of chest x-ray

In comparing the findings of chest x-rays at the time of admission, the patients exhibited bilateral parahilar infiltration (n = 16), nodular opacity (n = 19), patchy consolidation (n = 37), and homogeneous dense lobar consolidation (n = 44). Parapneumonic effusion was observed in 22 patients. Overall, there was no significant difference in the proportions of specific chest x-ray patterns according to the presence of macrolide resistance (Table 2).

Mean days to defervescence following macrolide initiation were longer in children with MP pneumonia when they showed homogeneous dense consolidation (7.27 days vs. 4.01 days, P < 0.001) and parapneumonic effusion (9.73 days vs. 4.20 days, P < 0.001) than those who did not. In contrast, children with parahilar infiltration showed shorter duration of fever (1.63 days vs. 5.83 days, P < 0.001). Parapneumonic effusion was frequently accompanied by homogeneous dense lobar consolidation (n = 18) and patchy consolidation (n = 4). It is also notable that patients with parapneumonic effusion tended to show extrapulmonary signs more frequently than those without (77.3% vs.11.7%, P < 0.001).

Risk factors associated with prolonged fever in MP pneumonia

Multivariate analysis showed that extrapulmonary signs (aOR; 3.037, 95% CI; 1.057–8.726), homogeneous lobar consolidation (aOR; 3.610, 95% CI; 1.519–8.580), and parapneumonic effusion (aOR; 9.705, 95% CI; 3.031–31.075) were responsible for prolonged fever ≥7 days following macrolide treatment regardless of macrolide resistance.