Mycoplasma pneumoniae detections before and during the COVID-19 pandemic: results of a global survey, 2017 to 2021

Background Mycoplasma pneumoniae respiratory infections are transmitted by aerosol and droplets in close contact. Aim We investigated global M. pneumoniae incidence after implementation of non-pharmaceutical interventions (NPIs) against COVID-19 in March 2020. Methods We surveyed M. pneumoniae detections from laboratories and surveillance systems (national or regional) across the world from 1 April 2020 to 31 March 2021 and compared them with cases from corresponding months between 2017 and 2020. Macrolide-resistant M. pneumoniae (MRMp) data were collected from 1 April 2017 to 31 March 2021. Results Thirty-seven sites from 21 countries in Europe, Asia, America and Oceania submitted valid datasets (631,104 tests). Among the 30,617 M. pneumoniae detections, 62.39% were based on direct test methods (predominantly PCR), 34.24% on a combination of PCR and serology (no distinction between methods) and 3.37% on serology alone (only IgM considered). In all countries, M. pneumoniae incidence by direct test methods declined significantly after implementation of NPIs with a mean of 1.69% (SD ± 3.30) compared with 8.61% (SD ± 10.62) in previous years (p < 0.01). Detection rates decreased with direct but not with indirect test methods (serology) (–93.51% vs + 18.08%; p < 0.01). Direct detections remained low worldwide throughout April 2020 to March 2021 despite widely differing lockdown or school closure periods. Seven sites (Europe, Asia and America) reported MRMp detections in one of 22 investigated cases in April 2020 to March 2021 and 176 of 762 (23.10%) in previous years (p = 0.04). Conclusions This comprehensive collection of M. pneumoniae detections worldwide shows correlation between COVID-19 NPIs and significantly reduced detection numbers.


Introduction
Non-pharmaceutical interventions (NPIs) were suggested to reduce the spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) during the worldwide coronavirus disease (COVID-19) pandemic [1]. Many countries introduced NPIs in March 2020, which included physical distancing measures, personal protective measures (e.g. the use of masks, improved hand hygiene, respiratory etiquette), stay-at-home orders, school and day-care closures, closing borders and travel restrictions. The NPIs have been temporally associated with a global unprecedented suppression of influenza epidemics and other viral respiratory infections, such as respiratory syncytial virus (RSV) [2][3][4][5][6][7][8]. COVID-19 vaccinations were available as measures in addition to NPIs since December 2020 [9]. NAAT (PCR, real-time; in-house) [47] Yes [48] Switzerland in-house) [49] Yes [50] Zurich (B) d Hospital / clinical laboratory (tertiary centre) NAAT (PCR, real-time; in-house) g [49] Yes [50] St. a Stay-at-home orders for the general population (referred to as lockdown) according to an ECDC document [25] for Europe and to Wikipedia [26] for other UN regions, with adjustments made by the local participating author and considered until the end of the study period (31 March 2021). b Full and partial school closure duration in days according to [27] until 2 March 2021 (last update before end of study period).
c More detailed information including reporting characteristics, de-duplication and exclusion criteria are provided in Supplementary Table S2. ≥ 90% of data are from children and adolescents < 18 years of age.
e Data from several hospitals in the region of Ticino.
f Additional use of a specific in-house PCR [52].
g From 12 October 2020 to the end of the survey period additional testing with the FilmArray Respiratory Panel (bioMérieux/BioFire Diagnostics).
h In addition to PCR also serological data separately reported.
i Multiplex PCR testing before 2020 using the Respifinder (Pathofinder), and single PCR testing over the total survey period with a specific in-house PCR, as described previously [61].   Data from some countries during the first months in 2020 indicated that the introduction of NPIs also coincided with a reduction in Mycoplasma pneumoniae detections [2,6,10]. Mycoplasma pneumoniae is a major bacterial cause of respiratory tract infections in children and adults [11]. These infections occur both endemically in many different climates across the world and epidemically every few years. Previous epidemics in Europe were reported in 2010-2012, 2014-2015 and 2015-2017 [12][13][14][15]. Mycoplasma pneumoniae is transmitted by aerosol particles and respiratory droplets through close contacts within families, schools, military bases, institutions (residential care and nursing homes, homes for cognitively disabled people etc.) and among closed communities [15][16][17].
Diagnostic tests for M. pneumoniae include nucleic acid amplification tests (NAAT) such as PCR, antigen tests and culture from respiratory specimens (direct test methods) or serology (indirect test method) with varying sensitivities and specificities [11,18,19]. Realtime PCR applications are the most commonly used approach for detection of M. pneumoniae in clinical settings [20]. However, real-time PCR is not yet standardised across laboratories [20], and there are no internationally defined guidelines on the requirements for M. pneumoniae testing and surveillance [14]. Some countries collect laboratory reports on M. pneumoniae detections through national reference laboratories (e.g. England), but only few countries have a national surveillance (e.g. Denmark) [14]. To our knowledge, no analysis on the M. pneumoniae incidence from several United Nations (UN) regions has been published so far.
In this study, we used survey data on laboratory M. pneumoniae testing and detection before and during the COVID-19 pandemic across the world to assess the impact of NPIs on the global incidence of M. pneumoniae in the first year after the implementation of NPIs. Of particular interest was the impact of children returning to schools on M. pneumoniae incidence while maintaining other NPIs during the course of the pandemic, as children are believed to be the main drivers of M. pneumoniae transmission [16] and have greater difficulty adhering to physical distancing and personal protective measures. In this context, was also analysed the proportion of females in particular because of their assumed closer vicinity with children.  [23]. A pilot test was performed with 10 individuals (infectious diseases specialists and microbiologists) to ensure that the questions were understood and interpreted consistently and that collection of requested data was feasible within the survey time period. Details of the survey are shown in Supplementary Table S1.  a Three sites provided serological data in addition to PCR. b No distinction possible between detection methods, but predominantly serological data included.

Quality control
Entries were included if they met the following quality control criteria for valid datasets: (i) verification of the participant, laboratory and institution via provided link and/or references in PubMed, (ii) validation of the information and/or references about the test method, and (iii) data check for multiple entries from the same institutions (double reporting), invalid or incomplete data, and inconsistent entries. In case of inconsistency or multiple entries from the same institutions, participants were contacted by email to request clarification and/or adapt entries to exclude double reporting. Criteria for de-duplication and exclusion criteria are listed in Supplementary Table S2.  (Table 1). Participants were asked whether a positive serology was confirmed by a fourfold increase in IgG levels measured in convalescent samples (as serological gold standard for M. pneumoniae infection [11]).

Stay-at-home order and school closure periods
Periods of stay-at-home orders for the general population (referred to as lockdowns) in Europe were obtained from the Response Measures Database (RMD) of the European Centre for Disease Prevention and Control (ECDC) [25] and those in other UN regions from a collection of pandemic lockdown dates in Wikipedia [26], with adjustments made by the participants. The total duration in days until the end of the study period was calculated for each site. School closure duration in days (full and partial closure in total) was determined according to the United Nations Children's Fund (UNICEF) global school closures database until 2 March 2021 (last update before the end of the study period) [27].

Statistical analysis
Incidence was defined as the number of new cases over a specified period of time within a community [28]. Given the missing population denominators we were not able to report incidence rates. We compared M. pneumoniae detections between April 2020 and March 2021 with total numbers observed from April 2017 to March 2020. Fisher›s exact test was used to compare proportions with corrections for multiple testing. Spearman rank correlation coefficient (R, rho) was used for analyses of correlation. All reported p values are two-tailed with statistical significance defined as p < 0.05. Data were analysed using R software (version 4.0.5) [29].

Survey entries and detection methods
We received entries from 48 sites, of which 29 were entered via the online survey and 19 via email to authors. Of the 12 experts collating laboratory detections of M. pneumoniae in Europe and Israel for the ESGMAC in a previous study (January 2011-April 2016) [14], eight provided information for this survey. An overall response rate could not be calculated because the survey was widely disseminated through societies, social media and further dissemination among participants themselves. We excluded 11 entries because of invalid or incomplete data (n = 7), inconsistent data (n = 2; positive test numbers by month did not match with total numbers per year) or double reporting (n = 2; congruent data from same institutions). Thus, 37 valid datasets from separate sites in 21 countries from four UN regions were eligible for inclusion (Europe: n = 12; Asia: n = 5; America: n = 2; Oceania: n = 2), 29 from hospital laboratories, two from national reference laboratories and six from national and/or regional surveillance systems (Figure 1).
Demographic characteristics and laboratory information of participating sites are shown in Table 1. The detection method varied between sites: 29 (78.38%) sites reported exclusively PCR (n = 17 multiplex); three sites used exclusively serology (enzyme-linked immunosorbent assay (ELISA)), three sites reported combined PCR and serology (no distinction possible between detection methods, but predominantly serology), one site used a combination of direct test methods (i.e. PCR, antigen test or culture) and one site used exclusively rapid antigen testing. Three sites reported only the number of positive tests over the entire study period (Saxony (Germany) and national surveillance systems of Belgium and Finland), and another three sites provided serological data in addition to PCR.  Table S2 lists the reporting characteristics per site).

Detections before and after the introduction of non-pharmaceutical interventions
There was a significant reduction of M. pneumoniae detections after the introduction of NPIs ( Figure  2). Among total detections, 1,714 (5.60%) derived from April 2020 to March 2021 compared with 28,903 (94.40%) from April 2017 to March 2020 (Table  2). Mycoplasma pneumoniae testing and detection in children/adolescents and females per year is shown in Table 3. The annual proportion of children/adolescents and females with detections before and during the COVID-19 pandemic was 55.16% vs 49.77% (p < 0.01) and 53.01% vs 50.86% (p = 0.15), respectively. Detailed graphs for each site and country are shown in Supplementary Figures S1-S6. The difference in detections before and during the COVID-19 pandemic was more obvious for direct test methods (Figure 2A) than indirect test methods ( Figure 2B). This is supported by a direct comparison of detections with PCR and single-sample serology (IgM, IgG and IgA) from the three sites that reported data separately for each method, which did not show any correlation between those two test methods (Figure 3). Grey backgrounds indicate local stay-at-home order (lockdown) periods. Another site from Germany (Homburg) did also provide PCR and serological data separately but numbers by month were not available.   There was no correlation of the duration of lockdown or school closure periods with direct M. pneumoniae detection rates from April 2020 to March 2021. Several sites reported a longer duration of lockdown than school closure periods, which suggested that children returned to schools while lockdown continued for some time ( Table 1). The re-opening of schools had no observable impact on the incidence of M. pneumoniae as direct detections remained remarkably low throughout the period April 2020 to March 2021. Detections were very low or absent even in countries where no school closures or official lockdowns were enforced (e.g. Japan, Taiwan; see Supplementary Figure  S3 for M. pneumoniae detections in Asia).

Macrolide resistance
As a consequence of the significant decrease in M. pneumoniae detections after the introduction of NPIs, only few cases were investigated for macrolide resistance. In total, seven sites from Europe, Asia and America reported MRMp rates from April 2017 to March 2021 (Table 4). Macrolide resistance determination was reported as part of national surveillance of positive samples (Japan, Cuba) or only on positive samples identified at the reference laboratory and/or upon physician request. The MRMp detections among investigated cases are shown as absolute numbers in Figure  4A and as percentages in Figure 4B. The highest MRMp rate was found in Taiwan

Discussion
This global survey showed that all countries experienced a decrease in M. pneumoniae incidence by direct test methods in April 2020-March 2021, relative to the previous three years. This decline corresponded with the timing of the implementation of NPIs against COVID-19 in March 2020 in each country. We also observed a decrease in MRMp rates in April 2020 to March 2021. The MRMp rates before the COVID-19 pandemic were lower in Europe than in America or Asia, consistent with previous reports [11].
A reduction in M. pneumoniae detections after the introduction of NPIs was observed with direct test methods such as PCR but not with serology. This effect could be explained by the long-lasting nature of antibodies against M. pneumoniae. Mycoplasma pneumoniae-specific antibodies (IgM and IgG) persist for months to years after infection, and significantly longer than M. pneumoniae DNA in the upper respiratory tract [30,31]. Based on these kinetics, we would expect a decline in positive IgM serology in the second year of the COVID-19 pandemic, but not necessarily in IgG serology as M. pneumoniae-specific IgG antibodies can persist lifelong [30]. There is also the possibility of false-positive results caused by limited assay performance [32] as serological detections are reported from single-sample serology, which was in most cases not confirmed by the detection of a significant antibody level change in convalescent sera. In addition, PCR and serology (IgM and IgG) can be positive in asymptomatic carriers [11]. The detection of specific antibodysecreting cells by enzyme-linked immunospot (ELISpot) assay may allow for differentiation between infection and carriage [24], and a combination of clinical features and biomarkers can help identify patients at high risk for M. pneumoniae community-acquired pneumonia [15]. However, no clinical features were reported in this study and cases were defined by local practice.
Our findings are in line with several reports about a worldwide reduction in infections with respiratory and gastrointestinal pathogens after the introduction of NPIs [2,3,[5][6][7][33][34][35][36][37]. The incidence of invasive bacterial diseases caused by Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria meningitidis that are transmitted via the respiratory route were also considerably reduced during the early months of the COVID-19 pandemic [38]. The interruption of direct person-to-person transmission was suspected to be the most plausible explanation for the reduction in respiratory infections. These remained low even after the re-opening of schools, except for rhinovirus [6,[39][40][41].
Direct detections of M. pneumoniae between April 2020 and March 2021 were significantly below levels of non-epidemic periods of M. pneumoniae across countries despite widely differing lockdown or school closure periods, and even in countries where no official lockdowns or school closures were enforced.    and change in healthcare utilisation (e.g. telemedicine visits). After the reopening of schools, direct M. pneumoniae detections remained low. This was also observed at sites where lockdown and restrictions for the adult population continued while children returned to schools. Children have greater difficulty adhering to physical distancing and personal protective measures so that M. pneumoniae transmission may be less effectively prevented in schools than in the adult population. Unfortunately, we did not have information on the age distribution in children to look at the pre-school and school age groups separately. The low incidence despite the re-opening of schools might suggest that adults play a more important role in transmission of M. pneumoniae than previously thought. This is supported by the observed decrease in the proportion of children and adolescents with M. pneumoniae detection during the COVID-19 pandemic. Notably, there was no change in the proportion of females with M. pneumoniae infection before and during the COVID-19 pandemic. Reduced transmission by shielding of adults (regardless of school closures) was also discussed as possible reason for the decrease in invasive pneumococcal disease [38]. Interestingly, nasopharyngeal pneumococcal carriage in children was only slightly reduced during the first year of the COVID-19 pandemic and the reduction in invasive pneumococcal disease was therefore attributed to the suppression of specific respiratory viruses such as RSV and influenza, which are often implicated as co-pathogens with S. pneumoniae [42]. Mycoplasma pneumoniae is also frequently detected with other viruses in the upper respiratory tract [15,[43][44][45], but the role of co-detections in M. pneumoniae respiratory disease remains unclear [44]. A direct biological effect of SARS-CoV-2 on M. pneumoniae by interference or interaction could be another explanation. To our knowledge, data supporting this hypothesis do not exist so far. Further, transient herd immunity from the recent epidemic period in April 2019-March 2020 in several countries in Europe and Asia could have led to a decreased M. pneumoniae incidence during the COVID-19 pandemic [12]. However, the incidence was also reduced in countries that had not experienced a recent epidemic (e.g. Norway).
The study has a number of limitations. Firstly, because of the variable reporting methods and testing criteria at each site, conclusions based on the analysis across countries must be considered with caution. Data obtained from a single hospital laboratory from a specific region may not be fully representative of the country as a whole. No information about catchment area and numbers of laboratories within regions were available. The study also lacks representation from Africa and South America (no survey response and/or no testing for M. pneumoniae reported). Secondly, defining study-wide case definitions and de-duplication criteria was not feasible given the heterogeneous nature of data collection between sites. De-duplication methodologies were therefore set at site level. Thirdly, as mentioned previously, serological detections were not confirmed by antibody changes in paired sera in most cases. Fourthly, analysis of the local clinical testing pathway for M. pneumoniae was not possible within this study. Decision-making to test or not to test with specific methodologies during the COVID-19 pandemic may have impacted which individuals and sites offered testing at which time. The number of tests increased in one fifth of the sites during the period April 2020 to March 2021 and also the incidence was significantly lower compared with the pre-pandemic period; hence, we do not believe that the overall reduction in M. pneumoniae detections can solely be accounted for by reduced testing. Nor was there an indication that M. pneumoniae testing was reduced because of shifting laboratory resources towards SARS-CoV-2 testing during the whole first year after the introduction of NPIs covered by this study. Finally, an overall survey response rate could not be calculated because of the widespread dissemination of the survey. Incomplete response to a survey can introduce a bias related to differences in incidence between the responders and the non-responders [21,46]. However, this risk seems minimal as our survey dealt with microbiological laboratory data and generated a large and varied sample [46]. This study is another example of how pandemicfocused public health measures may have prevented infections caused by other respiratory pathogens. The COVID-19 pandemic resulted in restrictive NPIs such as lockdowns and school closures, which are unsustainable in the longer term. The results of this study suggest that even less restrictive NPIs such as personal protective and physical distancing measures might have prevented transmission of M. pneumoniae in the community.
The study also highlights the importance of establishing international working groups to investigate pathogen epidemiology where surveillance systems are lacking. It underlines the need for an international case definition for infection with M. pneumoniae (detection method and clinical criteria). The influence of the detection method for epidemiological surveillance of M. pneumoniae is shown in the discrepancy between PCR and single-sample serology in this study. Serological surveillance of M. pneumoniae may be only accurate by using paired sera in order to detect a fourfold increase in IgG levels [11]. However, such procedures are timeconsuming and are not useful for acute patient care. A more rapid response to public health measures may be obtained by surveillance of M. pneumoniae using PCR. Finally, epidemiological surveillance should also include antimicrobial resistance testing of M.
pneumoniae. This study represents the most comprehensive estimate of global resistance documented to date and is important for clinicians and infectious disease surveillance considering that macrolides remain the main global treatment option for children with M. pneumoniae infection.

Conclusion
The results of this study from diverse geographical locations and healthcare settings suggest that the implementation of NPIs against COVID-19 probably restricted transmission of M. pneumoniae, leading to a significant reduction in M. pneumoniae infections in many countries across the world from April 2020 to March 2021. The retention of some NPIs after the COVID-19 pandemic e.g. improved hand hygiene, respiratory etiquette or physical distancing in the community, or the use of masks in health care institutions may help reduce the burden of M. pneumoniae infections. The large collaborative network established for this study allows to assess the resurgence of M. pneumoniae infections at a later time.