A Diagnostic Dilemma: PCR or Serology to Detect Mycoplasma Pneumoniae Pneumonia in Children

Cristina E. Figallo; Liz Bayes; Mariana Lanata; Veronica Etinger

doi:10.1542/peds.140.1_MeetingAbstract.39

Purpose

Mycoplasma pneumoniae (MP) is a common cause of community-acquired pneumonia (CAP) in children. The Infectious Diseases Society of America recommends testing children hospitalized with signs and symptoms suspicious for MP. The standard laboratory diagnosis relies mainly on PCR and/or serology. The objective of this study was to: (1) assess the value of each test and (2) determine which one should be performed based on duration of symptoms.

Methods

We conducted a chart review of patients between 1-18 years of age who had been hospitalized at Miami Children's Hospital between December 2012 and July 2014 with a diagnosis of MP pneumonia by ICD-9 code tested using PCR or IgM. Samples for respiratory panel PCR (RPP) were collected by nasopharyngeal swabs and processed for 20 respiratory pathogens. ELISA was used for IgM serology. We compared the frequency of positive Mycoplasma PCR with the frequency of positive IgM and the agreement between both tests using Kappa score. Fisher exact and Pearson Chi square test were used to analyze the association between Mycoplasma PCR and duration of symptoms.

Results

95 patients met inclusion criteria. 40 patients (42.1%) were tested for Mycoplasma by PCR and IgM serology. The overall percent of agreement between both tests was 34.2% and Kappa test showed only slight agreement (k=0.03). 71 patients had IgM serology from which 67 (94.4%) had positive IgM. Of all IgM positive, 45 patients (63.4%) had positive IgM in early disease (less than 7 days of symptoms). Only 26 out of these 45 patients were also tested by PCR resulting in 9 patients positive for a virus, 9 for Mycoplasma and 8 patients had negative PCR.

Conclusion

We observed a lack of consensus regarding laboratory testing. Serology was the test of choice as it had been used widely in clinical practice, while RPP became recently available at our institution. A poor level of agreement was observed between IgM and Mycoplasma PCR results. When IgM and PCR were analyzed by onset of symptoms 65.4% of the children who were IgM positive in early disease had a negative Mycoplasma PCR. Usually IgM takes 7-10 days to appear, so performing this test in the first week of disease should yield negative results. In conclusion, in order to prevent overuse of macrolides, IgM should not be obtained in early disease, as it most likely reflects an old infection and PCR should be considered instead.