Mycoplasma pneumoniae triggers pneumonia epidemic in autumn and winter in Beijing: a multicentre, population-based epidemiological study between 2015 and 2020

ABSTRACT The objective of this paper is to explore the characteristics of Mycoplasma pneumoniae (MP) epidemics in Beijing, China. Patients with acute respiratory tract infection (ARTI) were enrolled from 35 sentinel hospitals in Beijing, 2015–2020. Their medical records were reviewed and respiratory specimens were collected for assay for nucleic acids of 24 respiratory pathogens, including MP. The genotypes of MP were analysed using a real-time PCR method. The domain V of 23s rRNA gene was sequenced to identify macrolide-resistant mutations. A total of 41,677 specimens of ARTI patients were included, with an MP positive rate of 6.16%. MP prevalence mainly occurred between August and January, and peaked in October. The increase in the MP detection rate was coincident with the elevation of the reported number of patients with pneumonia in the 35 sentinel hospitals. One or more respiratory pathogens were co-detected in 27.1% of the MP-positive patients. Type 1 MP remained predominant, and the macrolide-resistant rate of MP had exceeded over 90%. A2063G mutation accounted for 99.0% of macrolide-resistant MP infections. MP epidemic in Beijing mainly occurred between August and January with a remarkable high macrolide-resistant rate. MP is one of the important contributors to the pneumonia epidemic in autumn and winter in Beijing.


Introduction
Mycoplasma pneumoniae (MP) infection is often considered a self-limiting disease. Due to the relatively mild nature, the similarity in presentation to respiratory tract infections caused by other respiratory pathogens, and the lack of reliable point-of-care diagnostic tests to confirm the microbiological diagnosis, many sporadic MP infections and even outbreaks are likely to be undetected.
However, MP is one of the leading pathogens that cause human upper and lower respiratory tract infections. After world-spread use of pneumococcal conjugate vaccine (PCV13), MP has become the most common pathogen of community-acquired bacterial pneumonia in children [1] and accounts for 10-40% of cases of community-acquired pneumonia (CAP) [2,3]. Among European adults, the incidence of MP pneumonia is 4-8% during endemics and as high as 20-40% during epidemics [4,5]. In a report by the multinational Asian Surveillance Network for Drug-Resistant Pathogens (ANSORP), MP accounts for 11% of adults with CAP in eight Asian countries [6].
The positive detection rate of MP reaches 10-30% in CAP children in China [7]. In addition to respiratory tract infections, MP probably induces extrapulmonary damage, such as diarrhoea, inflammation of the central nervous system, acute myocardial injury, arthritis, or otitis media [2,[8][9][10]. These extrapulmonary illnesses are a result of MP infection itself and/or an excessive immune response or inflammation triggered by the "superior antigen" of MP, which can eventually develop into a life-threatening disease [11][12][13].
Macrolides have been used as first-line antibiotics for the treatment of MP infection for several decades [14]. However, since pneumonia caused by macrolide-resistant MP (MRMP) was first reported in Japan in 2000, the prevalence of MRMP has been increasing due to the extensive application of macrolide antibiotics [15][16][17][18]. There are significant differences in the prevalence of MRMP among different countries or regions in the world. Namely, the prevalence of MRMP remains under 30% [19,20] in the United States and Europe. In contrast, countries or regions in Asia have a significantly higher prevalence, which exceeds 80% in Japan [21] and even reaches 90% in China [14,22,23]. Considering the emergence and epidemic of MRMP, it is more challenging for the administration of MP infection, particularly for the infections among children and teenagers [24].
A worldwide organized surveillance program similar to influenza surveillance has not been initiated for MP infection. Presently, the few studies on the characteristics of MP epidemics in Beijing are mostly based on one or several clinical settings [22,23], and covered only a small subset of the population with MP infection [25][26][27]. Namely, these studies only included CAP and excluded upper respiratory tract infections, or included children and excluded adults, which is why it is difficult to assess the true impact of this pathogen on public health. This study was based on the Beijing Respiratory Pathogen Surveillance System (RPSS), which intended to cover a more complete illness profile of respiratory infections caused by MP (including acute upper respiratory tract infection, non-severe pneumonia, and severe pneumonia), and included patients of all age categories who presented at or were admitted to 35 sentinel hospitals in Beijing. Thus we conducted a systematic exploration of the prevalence of MP to obtain the comprehensive insight into the characteristics of MP epidemics in Beijing.

Ethics statement
The protocol of this study was approved by the Ethics Committee of the Beijing Center for Disease Prevention and Control (BJCDC). A written informed consent was obtained from each of the included patients or their legal guardians.

Study population and specimen collection
Based on the RPSS, patients with acute respiratory tract infection (ARTI) were enrolled from 35 representative sentinel hospitals of different kinds and levels in Beijing in the period from January 1, 2015, to December 31, 2020 (see Supplementary table 1 in  appendix supplementary). Approximately 5 patients with upper respiratory tract infection (URTI) and 15 patients with CAP were required to be recruited from each sentinel hospital per month.
URTI patients were included if they presented with fever and/or respiratory symptoms, e.g. cough, sputum production, and sore throat. CAP patients were included if they had evidence of CAP according to the guidelines for the diagnosis and treatment of adult community-acquired pneumonia in China (2016 edition) and the guidelines for the management of community-acquired pneumonia in children (the revised edition of 2013) [28][29][30]. They were further classified as severe CAP (SCAP) and non-severe CAP (NSCAP) based on illness severity according to the criteria described previously [31]. The inclusion and exclusion criteria of the study subjects were described in detail in appendix supplementary.
All of the cases were investigated by their physicians after enrolment using a uniform questionnaire, which mainly included demographic data, epidemiological data, clinical manifestations, radiographical imaging data, and laboratory data. Respiratory specimens were collected as early as possible (at the time of presentation for outpatients or within 3 days since admission for inpatients) and included at least one of the following types of specimens: pharyngeal swabs, nasopharyngeal swabs, nasopharyngeal aspirates, sputum, pleural effusion, tracheal aspirates, or bronchoalveolar lavage fluid.
In addition, the numbers of URTI patients and CAP patients were reported weekly by the 35 sentinel hospitals from January 1, 2017, to December 31, 2020.

Laboratory assay
Nucleic acid test was performed on the collected clinical specimens for nine respiratory viruses, MP and Chlamydia pneumoniae (CP) (see Appendix supplementary). Moreover, the specimens obtained from lower respiratory tract, a subset of the total specimens, were tested for a panel of 13 respiratory bacteria (see Appendix supplementary).
MP-positive specimens were further genotyped into types 1 and 2 with a real-time PCR method described by Zhao et al. [32]. The domain V of their 23s rRNA gene was sequenced to identify the macrolide-resistance mutation [16].

Statistical analysis
Continuous variables were presented as mean (95% CI) or median (interquartile range, IQR), and categorical variables were presented as numbers (%). Comparisons between different groups were conducted by the χ² test or Fisher's exact test. The restricted cubic spline (RCS) method was used to analyse the relationship between the probability of MP infection and age using R language 4.0.5 and RMS software package. Interaction between multiple respiratory pathogens was analysed by multiple-stage logistic regression. A two-sided Pvalue lower than 0.05 was considered statistically significant. Statistical analysis, except for RCS analysis, was done using SPSS 19.0 (SPSS Inc., Chicago, IL, USA).

Study population
From January 1, 2015, to December 31, 2020, a total of 41,677 specimens of patients with ARTI were included in this study, including 14,099 specimens of patients with URTI, 22,125 specimens of patients with SNCAP, and 5453 specimens of patients with SCAP. The mean number of specimens enrolled per month was 624 (95% CI, 605-644) between 2015 and 2019, which was reduced to 351 cases per month due to the COVID-19 pandemic in 2020. Male patients accounted for 55.9%, which is in line with the population structure in China. The median age was 37 years (P25, 9 years; P75, 67 years). Among the included patients, 2566 (6.16%) were confirmed as MP infections.

Seasonality
From 2015 to 2020, the prevalence of MP in patients with ARTI in Beijing exhibited a distinct one-peakone-year epidemic pattern. The yearly positive rates (from 2015 to 2020) of MP among patients with ARTI were 3.97%, 6.12%, 6.46%, 7.12%, 9.30%, and 1.52%. Hence, after an increasing trend from 2015 to 2019, the rate fell to the lowest level in 2020 ( Figure 1A).
The positive rates of MP among patients with ARTI per month increased significantly from August, peaked in October, and fell down in January of next year. The epidemics of MP lasted for about 6 months ( Figure 1B).
In this study, the MP epidemic season was defined as the consecutive months in which the PCR positive rate of MP was continuously over 5%, and the cases in the epidemic season accounted for 74% of the total cases ( Figure 1C).
In some of the examined years, the onset of the epidemic season was advanced or delayed by 1-2 months.

Positive rates of MP by age group and severity of illness
The positive rate of MP among patients aged 5-44 years was significantly higher than that in other age groups (χ 2 = 1655.765, P < 0.001), with a peak in 10-14-year-old group (16.06%) (Figure 2A). The median age of 2566 patients with confirmed MP infection was 24 years (P25, 8 years; P75, 37 years).
There was a nonlinear correlation between the probability of MP infection and age. For patients under 24 years of age, the probability of MP infection increased with age, while for patients aged 24 years and older, the probability of MP infection decreased as age increased ( Figure 2B). Regarding patients with SCAP, there was also a nonlinear correlation between the probability of MP infection and age, but the changing trend was almost opposite to that in overall ARTI patients. Namely, the probability of MP infection decreased with age among the SCAP patients under 26 years of age, then increased gradually as age increased until 68 years, and finally fell down again ( Figure 2C).

Pathogens co-detected with MP
Nucleic acids of respiratory pathogens other than MP were detected in specimens of 27.1% (695/2566) of the patients infected with MP, including lower respiratory tract specimens (sputum, tracheal aspirates, bronchoalveolar lavage fluid) of 564 patients and upper respiratory tract specimens (pharyngeal swabs) of 131 patients( Table 1). The most common co-detected pathogen was S. pneumoniae (240 cases) followed by H. influenzae (137 cases), S. maltophilia (80 cases), and P. aeruginosa (77 cases).
The most frequent co-detected pattern was observed between MP and S. pneumoniae (2.01%), followed by that between MP and H. influenzae (1.15%), MP and S. maltophilia (0.67%), and MP and P. aeruginosa (0.64%) ( Figure 3A). Most of the interactions referring to MP were negative, predominantly occurring between MP and all 10 viruses, and between MP and 6 bacteria, while positive interaction was identified only between MP and 2 bacteria (S. pneumoniae, S. pyogenes) ( Figure 3B).

Contribution of MP infection to the prevalence of CAP in Beijing
A total of 2,290,510 patients with ARTI were reported in Beijing in this study, including 987,045 cases of URTI and 1,303,465 CAP cases. The median number of the reported cases per month was 9265 (P25, 7848 cases; P75, 12,476 cases).
The number of the reported cases per month of CAP always surged earlier by about 1 or 2 months than that of URTI cases. Namely, the number of CAP cases began to rise from October, reached the peak in December, and fell down in January, while the number of URTI cases rose dramatically in December, peaked in January, and then decreased in February (Figure 4).
Referring to the positive detection rates of 11 common respiratory pathogens, the number of the reported URTI cases per month approximately rose and fell at the same time as the positive detection rate of influenza virus among URTI cases ( Figure 4A-C). The elevation in the reported number of the reported CAP cases per month initiated in the period during which the positive detection rate of MP among CAP cases increased dramatically. However, the detection of the other respiratory pathogens (including influenza virus) remained at a low level in the same period. As the positive detection rate of influenza virus among CAP cases rose, this increasing trend in the number of CAP cases was further enhanced. Generally speaking, the total detection of MP and influenza virus fluctuated approximately with the number of CAP case reported, while the total detection of the other respiratory pathogens independently remained at an almost constant level in the same period ( Figure 4D-F).
During the pneumonia epidemic season between October and January, MP and influenza virus accounted for 27.78% and 26.79% of CAP, respectively ( Figure 4G).

The alteration of genotype
A total of 866 specimens of MP-positive patients from 2018 to 2020 were used for the analysis of the P1 gene type in this study. Type 1 accounted for 62.0% cases in 2018, 60% in 2019, and increased significantly to 83.7% in 2020 (Table 2).

Discussion
This study showed that MP infection in Beijing occurred throughout the year. A higher prevalence was between August and January, showing that MP infection exhibited a one-peak-one-year epidemic pattern. Several studies have reported that MP epidemic surges every 3-7 years probably due to antigenic shift or decreased herd immunity [33]. Among five consecutive MP seasons, we observed that MP epidemic in Beijing peaked in 2016, followed by that in 2019. However, a similar interval between two adjacent epidemics was not observed in this study due to a short surveillance period.
In this study, we observed that the positive rates of P1 adhesin gene-based nucleic acid test for MP were always over 6.16% during the five consecutive MP seasons, and we assumed that the epidemic season of MP occurred if the PCR positive rate of MP among ARTI patients exceeded 5% per month in Beijing. This threshold contributed to the timely identification of MP epidemic season and the comparability of different studies. Of course, the MP prevalence intensity varies even between different regions of China, so a local value of the threshold of MP detection probably needs an adjustment according to the characteristics of MP prevalence in a particular region.
Public health measures against COVID-19, such as wearing masks in public places, keeping social distance, prohibiting large assemblies, and isolating and quarantining patients and close contacts as early as possible, were implemented in Beijing since January 23, 2020. These measures were confirmed to effectively control or even exterminate COVID-19 epidemic in Beijing. Meanwhile, they also dramatically reduced the prevalence of multiple respiratory pathogens, including MP [34]. Although the role that each measure played in the course of preventing epidemics of these respiratory pathogens is unclear, they, as a whole, were shown to effectively stop or slow down the spread of these respiratory pathogens, including MP, in the population. This is also the first time that the several-decade epidemic pattern of MP in Beijing was altered completely by artificial interventions.
We observed a nonlinear correlation between the risk of MP infection and age of patients with ARTI. Among individuals aged under 24 years, the risk of MP infection increased with age, while among those aged 24 years and older, the risk of MP infection decreased as age increased. We compared the MP positive rates among different age groups and found that the MP positive rate was higher in the 5-44year-old group and peaked in the 10-14-year-old group, which is approximately consistent with a previous study that showed that MP infection occurred more often among children and teenagers aged 5-14 years [35]. The distribution of the MP detection rate by age was lower at both ends and higher in the middle, which was significantly different from many other respiratory pathogens (e.g. enterovirus, respiratory syncytial virus). The reason underlying this specific age distribution is not completely clear and may be associated with the expression of the receptor of MP on epithelial cells of the respiratory tract, and/or with host immune defence. Cilia of the epithelial cells in the respiratory tract are considered the initial attachment site for P1 cytadhesin in the course of MP infection [36]. Community-acquired respiratory distress syndrome (CARDS) toxin of MP is considered the most important virulence factor that presumably triggers cell pyroptosis and inflammatory injury via the NLRP3 signalling pathway [37]. According to Bajantri et al., the "common antigen" between MP and the host presumably induces excessive immune response and causes refractory MP infections [5]. The population aged 5-44 years possesses more mature cilia of epithelial cells, which facilitates the MP attachment and thereby increases the risk of MP infection. These individuals also have a more mature immune system, which contributes to maintaining the balance between immune protection and immune injury, thus preventing MP-infected host from developing a severe illness. In this study, we observed that the changing trend of MP infection probability with age among SCAP patients was almost opposite to that among total ARTI patients, which suggested that young children have a higher risk to develop severe illness even though they have lower probability of MP infection.
Clinical manifestations of pneumonia caused by MP are indistinguishable from those of pneumonia caused by other respiratory pathogens. In this study, in respiratory specimens of 27.1% (695/2566) of MP-positive patients, one or more other respiratory pathogens were co-detected. These co-detected pathogens may have been only planted in the respiratory tract or may have participated in the pathogenic process together with MP. Currently, as SARS-CoV-2 spreads around the world, a co-infection of MP and SARS-CoV-2 is of great concern. Xing et al. [38] reported that, among 68 COVID-19 cases, the MP detection rate among the confirmed COVID-19 patients was 23.3%  1  2  3  4  5  Obtained from URT  121  8  2  --131  1244  10.5  Pharyngeal swabs  121  8  2  --131  1244  10.5  Obtained from LRT  368  128  42  17  9  564  1322  42.7  BALF  54  15  1  70  217  32.3  Tracheal aspirates  1  1  2  4  50  Sputum  313  113  42  16  8  492  1101  44.7  Total  489  136  44  17  9  695  2566  27.1 URT, upper respiratory tract; LRT, lower respiratory tract; BALF, bronchoalveolar lavage fluid. For pathogens "X" and "Y", the numerator was the number of patients in whom "X" and "Y" were co-detected, and the denominator was the total number of patients who were both tested "X" and "Y". Bigger size and darker colour of the circles indicate higher co-detected rates between two pathogens. in Qingdao city and 2.63% in Wuhan city. Given the COVID-19 pandemic, this kind of co-infection occurs more frequently and probably leads to more severe clinical outcomes; thus, we should pay more attention to it, particularly during the MP epidemic season. Before the COVID-19 pandemic, influenza was the leading cause of acute respiratory infections. Influenza epidemics have also been considered the main cause for the dramatic elevation in the incidence of pneumonia in autumn and winter in Beijing. However, we observed that the onset time of pneumonia epidemic in Beijing was about 1 or 2 months earlier than that of the influenza epidemic, and was approximately coincident with that of MP epidemic, with no observed epidemics of other common respiratory pathogens in the same period (October and November). Moreover, MP accounted for 27.78% of CAPs in pneumonia epidemic season, and our previous study also reported that MP, after influenza virus, was the second most common pathogen of CAP in Beijing [29]. Taken together, the high prevalence of pneumonia in autumn and winter in Beijing was initiated by the MP epidemic, and then further increased and prolonged by the subsequent influenza epidemic. Thus influenza vaccination alone is not enough to significantly reduce the high prevalence of pneumonia in autumn and winter in Beijing, and some of the abovementioned measures confirmed effective against MP should also be taken into consideration.
In conclusion, MP prevalence exhibited a onepeak-one-year pattern in Beijing with an epidemic season that usually occurred between August and January with a peak in October, which was altered completely by public health measures against COVID 19 in 2020. Type 1 MP remained dominant while the prevalence of MRMP exceeded over 90%. Notably, it was MP, other than the influenza virus, that first triggered pneumonia epidemics in autumn and winter in Beijing each year. These facts should be considered when public health policies are made.