Evaluating Laboratory Diagnoses of Children’s Mycoplasma Pneumoniae Infection Reveal Effective Clinical Approaches


 This study aims to evaluate laboratory methods for the diagnosis of Mycoplasma pneumoniae . Ninety-three children infected with Mycoplasma pneumoniae were tested, and large sample size makes the results representative. The pharyngeal swabs and serum samples were mostly collected from the first week of the disease, or from the second to third week of the disease in recovery children. The pharyngeal swabs from healthy children were cultured and two of them were positive for the bacterium. The accuracy of laboratory diagnostic methods for Mycoplasma pneumoniae were evaluated against serum test with 4-fold increase or decrease in the titers of antibody in two serum samples. The specificity of culture method was high, which suggested that the method was suitable for clinical research. In acute phase, the sensitivities of single serological test or simultaneous amplification and testing of Mycoplasma pneumoniae (SAT) were 71.7% and 85%, respectively.


Introduction
Mycoplasma pneumonia (MP) is an important pathogen of community acquired pneumonia (CAP) in children. The incidence of MP accounts for about 10-40% of CAP. The severity of the disease varies greatly and may develop into severe and refractory MP that endangers patient's life or leads to serious sequelae 2 . It is believed that early diagnosis and timely as well as effective treatment are the key to reduce severe infections 3 . The clinical symptoms and signs of MP infection are often unspeci c, and laboratory diagnosis is becoming increasingly crucial.
At present, the commonly used diagnostic methods in laboratory include bacterial culture method, serological diagnosis, and molecular diagnosis. However, these methods have not been extensively veri ed for their clinical values. The sensitivity of bacterial culture method is low (60%, 4) and the culture needs special culture medium as well as other speci c laboratory conditions. The method requires a long growth cycle, which is not suitable for rapid clinical diagnosis. However, the method has high speci city, which is employed in either academic or clinical researches. Molecular diagnosis method is sensitive and fast, which is less affected by the immune status of the body. It is speci cally essential for infants 5 and is a preferable laboratory diagnosis method. Evaluation shows that the uorescence quantitative PCR method developed in our laboratory is more rapid and accurate than the nested PCR as well as PCR with commercial kits. Using dye method, the speci city of this method is the same as the probe method of commercial kits, but it is more cost-effective and more sensitive than other alternatives 6  included. To date, and to our knowledge, this is the largest MP infection target sample size so far. They were clinically diagnosed to have MP infection with symptoms such as fever (uncertain fever type or no fever), cough (persistent severe cough as the main manifestation, no or less phlegm), or with dyspnea and wheezing. They had rough pulmonary auscultation or dry and wet rales, who were positive for one of the three tests: serum MP-IgM antibody test, MP-PCR, or MP culture. They also had signi cant pulmonary X-ray signs, such as enhanced image of hilum of lung, bronchitis, interstitial pneumonia, homogeneous consolidation, atelectasis 15 . Patients were excluded if they had one of the following symptoms: multiple system and organ damage; severe liver, kidney, cardiovascular and hematopoietic diseases, immune diseases and psychosis.

Sample collection
Pharyngeal swab samples were collected from the healthy and MP children. Double blood samples were collected at acute and recovery phase of the MP children.
The bacteria were inoculated in pleuropneumonia-like organism (PPLO) culture medium, and were incubated in incubator. The indicator color change was used to judge the MP growth 16

Serological test
Passive agglutination assay was used to measure the titer of MP-speci c serum antibodies via SERODIA MYCO II (cat no. YZB/JAP8033-2013, Fuji Rebi, Japan) following the manufacturer's instructions. Fourfold increase or decrease in the titer was used as a diagnosis cut-off 9, 10, 11 .

Statistical analysis
The measurement data were expressed as means ± standard error, which were analyzed using SPSS15.0 statistical software. Data in percentage were compared using Chi-square or Fisher test. P<0.05 was considered to be statistically signi cant.

MP infection in asymptomatic carriers
Fluorescence quantitative PCR (RT-PCR) and bacterial culture showed that out of 349 samples that tested, 66 (18.9%) and 2 (0.06%) samples were positive, respectively.

MP infection in diseased patients
Serological test showed that out of 42 (45.2%) boys and 51 (54.5%) girls (aged 3-14 years), 59.5% and 68.6% of them were MP positive, respectively. The infection rates were similar between genders. When the results were analyzed based on age (under 3 years, between 3 and 5 years and older than 5 years), no difference in the infection rates were found between different age groups (p = 0.612, 0.444 and 0.913, respectively), and the infection rates were 23.08%, 60.98% and 82.05%, respectively (Figure 1a). The difference of infection rates between the three groups was statistically signi cant (p < 0.05, P = 0.017, P = 0.037) by Chi-square or Fisher test, suggesting that the infection rate of MP increases with age (Table 1).
60 (64.5%), 39 (41.9%), 68 (73%) and 59 (63.4%) children were diagnosed to have MP infection based on the Serological test, bacterial culture, RT-PCR and SAT, respectively. Compared with the gold standard, the sensitivity and speci city of RT-PCR were the highest (93.3%) and low (63.6%), respectively, suggesting a higher rate of false positive. The speci city and sensitivity of bacterial culture were the highest (100%) and low (65%), respectively. Since the culture period was about 14-21 days or longer, the method could effectively help improve early diagnosis precision. SAT had high sensitivity and speci city (85% and 75.8%, respectively). The combined diagnosis of single acute serum serological test with RT-PCR generated a sensitivity of 100% with the unchanged speci city. The combined diagnosis of single acute serum serological test with SAT increased the sensitivity to 95% with unchanged speci city (Figure 1b). This method had the highest coincidence rate, Yoden index and Kappa value. The ROC curves of various methods compared with gold standard are shown in Figure 2, and the sensitivity and speci city are shown in Table 2 and Figure 1c. These data suggest that combined SAT with single acute serum serological test is the most suitable diagnostic method.

Discussion
MP is an important pathogen of community-acquired pneumonia in children. Its incidence is high and clinical manifestations are not speci c. Laboratory diagnosis has become an important complementary approach to determine the pathogens for medication. Diagnostic methods that provide early, rapid and accurate results are highly demanded though they are hardly available. This study evaluates several commonly used laboratory methods in order to provide guidance for clinical diagnosis.
Bacterial culture is highly speci c, which is used as the golden standard for MP diagnosis. In our study, its speci city reaches 100%. However, the sensitivity is relatively low (65%). It is noteworthy that culture of pharyngeal swab samples from 349 healthy children generated two positive samples, suggesting that they might be asymptomatic carriers of MP. Partly due to the amount of bacteria carried and the immune status of the body, the children were not affected. Asymptomatic MP carriers have also been reported previously. For instance, Jiang found that the MP infection might be related to the bacterial load. It is believed that high fever is more likely to occur when the MP load is greater than 10 7 copies/µl. However, the relationship between low MP load and pathogenesis is not clear 17 . In present study, the growth cycle of MP was 14 to 21 days, which is too long to meet the needs of early and rapid clinical diagnosis. Report showed that adding catalase to culture medium could accelerate MP growth 18 , but the results still need to be further veri ed.
Serological methods are widely used in clinic. However, there is still controversy about the diagnosis value based on single serum sample from acute phase. It was reported that the sensitivity of single serum sample from acute phase is 32%~35% 19 24 . According to the ROC curve drawn in this study, the threshold value is slightly higher than the best diagnostic point and thus may result in missed diagnosis. Considering the number of samples used in this study, larger study is needed to con rm the conclusion.
Molecular methods are rapid and sensitive. The sensitivity of RT-PCR protocol developed in our laboratory is 93.3%, but the speci city is low. The reasons might be due to asymptotic MP carriers. The detection rate of MP in swab samples of healthy children was 18.9% and were identi ed as MP by sequencing. It is reported that the carrying rate of MP in the respiratory tract of healthy people is 0.1%-56%. In addition, MP DNA can be detected continuously for 7 weeks to 7 months after infection 25 , although the detection of DNA cannot be diagnosed as MP infection. Li et al showed that SAT has high sensitivity and speci city in detecting MP infection, and has a high coincidence rate with RT-PCR (Kappa = 0.97) 26 . In this study, the sensitivity of SAT was slightly lower than RT-PCR, probably because RNA is easier to degrade than DNA, and is more susceptible to the in uence of sampling time and storage methods. In this study, samples were collected from children hospitalized for about a week, and most of them had already be treated with antibiotics. These may affect the detection rate of MP. In this study, the speci city of RNA is higher than that of DNA, which may be related to the fact that RNA can better re ect the presence of pathogens than DNA. However, total RNA, not messenger RNA were detected by the kit used. Therefore, the results cannot be used to distinguish between asymptomatic infection and healthy carriers. Based on our data, it is clear that at early disease phase, molecular detection is sensitive, rapid and suitable for the clinical needs. However, due to asymptomatic carrier of MP, it may lead to over diagnosis if it is solely based on molecular method. MP infection still needs to be con rmed with clinical data.
Joint diagnosis of serological and molecular methods for detection of MP infection is highly recommended 27,28 . Our RT-PCR protocol has a high sensitivity for detecting MP. Once combining with serological method, the sensitivity was further improved (up to 100%), while the speci city remained unchanged, leading to less missed diagnosis. However, this approach cannot solve the problem of low speci city caused by asymptomatic carrying. In the future, we will continue to investigate the relationship between DNA load and pathogenesis in order to address this problem. The sensitivity of detecting MP RNA in acute phase single serum samples combined with SAT was greatly improved (up to 95%). Because RNA is easy to degrade, its speci city is greater than DNA in MP detection, the accuracy of diagnosis is better than that of a single method. This combined method is the best diagnostic method tested in this study.

Conclusion
The current laboratory testing methods have their own limitations, their results should be veri ed based on clinical data for MP diagnose. In our study, the sensitivity of bacterial culture which required 14-21 days to complete is 65%, therefore, it is not appropriate for rapid clinical diagnosis. In the future, it is important to investigate how to improve the sensitivity and shorten the culture time.