A one-pot, visual, multiplex microfluidic digital PCR chip for simultaneous point-of-care detection, genotyping, and macrolide resistance assessment of Mycoplasma pneumoniae

Abstract

Mycoplasma pneumoniae (M. pneumoniae) is a leading cause of community-acquired pneumonia, especially concerning for young children. Effective management and infection control necessitate rapid, convenient, and user-friendly methods for early diagnosis, including genotype and macrolide resistance monitoring during outbreaks. Current clinical methods are often time-consuming, complex, and lack the capability for simultaneous detection. To address these challenges, we developed a one-pot visual multiplex microfluidic digital polymerase chain reaction (MMDPCR) assay for the simultaneous detection of M. pneumoniae infection, P1 typing, and macrolide resistance, utilizing four special fluorescence probes. The developed multiplex microfluidic digital polymerase chain reaction (MMDPCR) assay provides absolute quantification of clinical samples with a detection time of approximately 1.5 h. The MMDPCR assay demonstrated high specificity and superior sensitivity for M. pneumoniae, with a low detection limit of 10–100 copies and no cross-reactivity with other respiratory pathogens. We validated the assay’s performance using 192 clinical samples collected during the 2023–2024 Beijing epidemic. Results showed a positivity rate of 56.25% and a high prevalence (99%) of the A2063 macrolide resistance mutation. Among positive samples, Type I accounted for 75.9% and Type II for 24.1%, showing 100% concordance with qPCR findings. This novel multiplex detection platform holds immense potential for point-of-care testing of M. pneumoniae, offering a rapid and comprehensive tool for analysing its molecular epidemiological characteristics. Its ability to simultaneously provide information on infection, genotype, and macrolide resistance is crucial for guiding clinicians in the rational use of antibiotics and strengthen public health surveillance efforts.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00604-025-07666-0.

Introduction

Mycoplasma pneumoniae is among the smallest prokaryotic organisms implicated in community-acquired pneumonia. It is transmitted via respiratory droplets and has an incubation period of 1–3 weeks, which facilitates clustered outbreaks—particularly among children in confined environments [1]. M. pneumoniae is responsible for 10%−40% of community-acquired pneumonia cases, with this broad range largely attributable to variations in epidemic years and geographic heterogeneity [1]. According to global surveillance data from 2023, infection rates in northern China increased by 40% during autumn and winter seasons compared with the same period in 2022 [2]. Several European countries have reported an increase in M. pneumoniae cases in recent years [3, 4]; data from the U.S. Centers for Disease Control and Prevention indicate that M. pneumoniae-related cases represent 15.6% of paediatric outpatient visits [5].

The absence of a cell wall renders M. pneumoniae resistant to β-lactam antibiotics [6]. Macrolide antibiotics remain the first-line treatment for paediatric patients; however, macrolide resistance rate has reached 99% during epidemic periods [7]. Additionally, genotypes have been reported associated with clinical manifestations. P1-restriction fragment length polymorphism (RFLP) genotype II strains were found more virulent and induced severe cases [8]. Since 2015, a gradual increase in the prevalence of P1 genotype II strains has been observed in several countries, whereas genotype I strains were shown more commonly associated with higher resistance rates [9]. For this reason, an integrated method capable of simultaneously identifying genotypes and resistance profiles would provide clinicians with crucial information to guide appropriate antibiotic use [10].

Conventional polymerase chain reaction (PCR), while a cornerstone of molecular diagnostics, often faces limitations in absolute quantification and the detection of rare targets. Its reliance on standard curves and susceptibility to inhibition from complex sample matrices can lead to variability in results [11]. Furthermore, real-time PCR, though offering quantitative data, can struggle with low-concentration samples and has limited multiplexing capabilities [12].

To overcome these challenges, microfluidic digital PCR (MDPCR) has emerged as a powerful alternative. MDPCR partitions a sample into thousands of tiny droplets, with each droplet effectively acting as an individual reaction chamber. This compartmentalization allows for the absolute quantification of nucleic acids by counting the number of positive droplets, eliminating the need for standard curves and making it highly robust against inhibitors. Its superior sensitivity enables the detection of minute amounts of target DNA or RNA, making it ideal for applications like rare mutation detection, pathogen quantification especially the mixture infection in early infection [13, 14].

In this study, we developed a multiplex microfluidic digital polymerase chain reaction (MMDPCR) assay -based detection method for identifying M. pneumoniae infection, genotypes and resistance profiles directly from clinical samples in a single tube at one time. Our MMDPCR assay offers absolute quantification of clinical samples with a swift detection time of approximately 1.5 h. It demonstrates high specificity and superior sensitivity for M. pneumoniae, achieving a low detection limit of 10–100 copies without any cross-reactivity with other common respiratory pathogens. We rigorously validated the assay’s performance using 192 clinical samples collected during the 2023–2024 Beijing epidemic. This novel multiplex detection platform holds immense potential for point-of-care testing of M. pneumoniae which is vital for guiding clinicians in the rational use of antibiotics and bolstering public health surveillance efforts.

Methods

Ethical approval

This study was approved by the Ethics Committee Board of Capital Institute of Pediatrics with number SHERLL2023092, and the written informed consent was obtained from all participants included in this study. All specimens were obtained as part of routine diagnostic and treatment procedures; no additional sampling was required. All patient data were reported anonymously.

Bacterial strains

Two standard strains of M. pneumoniae, M129 (accession number: U00089.2) and FH (accession number: CP010546), along with two clinical isolates, R1984 and 23M574, were preserved in our laboratory. The culture medium comprised pleural pneumonia-like organism (PPLO) powder, foetal bovine serum, yeast extract, and glucose. Cultures were incubated at 37 °C; bacterial growth was monitored using the phenol red colorimetric method.

Twelve common respiratory pathogens were used for specificity testing: Haemophilus influenzae, Mycobacterium tuberculosis, Staphylococcus aureus, Pseudomonas aeruginosa, Klebsiella pneumoniae, Escherichia coli, Mycoplasma hominis, Streptococcus pneumoniae, Legionella pneumophila, respiratory syncytial virus, adenovirus, and Candida auris (Table 1).

Table 1.

Bacterial/Virus strain or DNA used in this study

Species	Strain/DNA
Haemophilus influenzae	ATCC10211
Mycobacterium tuberculosis DNA	ATCC25618/H37Rv
Staphylococcus aureus	ATCC29213
Pseudomonas aeruginosa	PA14
Klebsiella pneumoniae	CMCC46117
Escherichia coli	ATCC25922
Mycoplasma hominis	ATCC 23114
Streptococcus pneumoniae	ATCC49619
Legionella pneumophila	ATCC33152
Respiratory syncytial virus	Clinical isolate DNA
Adenovirus DNA	Clinical isolate DNA
Candida auris DNA	B8441 DNA
M.pneumoniae-M129	ATCC29342
M.pneumoniae -FH	ATCC15531
M.pneumoniae -23M574	Clinical isolate
M.pneumoniae -R1984	Clinical isolate

Bioinformatic analysis and design of primers and probes

To establish a specific test method for M. pneumoniae, whole-genome sequences of 17 type I and 12 type II strains were downloaded from the National Center for Biotechnology Information (NCBI) database. Multiple sequence alignment was performed using MAFFT to identify conserved and type-specific genetic regions.

Based on the alignment results, primers and probes targeting the identified specific genes were designed for strain typing. For the detection of macrolide resistance, a minor groove binder (MGB) probe was specifically designed to target the A2063G mutation site within the 23S rRNA gene. The thermodynamic stability and potential for non-specific structures (e.g., dimers, hairpins) of all primers and probes were evaluated in silico using the Oligo software (Molecular Biology Insights, Inc., USA). The specificity of the primers was further verified by BLAST analysis against a database of other common respiratory pathogens using the Primer-BLAST tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). All primers and probes were synthesized and purified by high-performance liquid chromatography (HPLC) by Sangon Biotech (Shanghai, China).

Microfluidic system for high-throughput droplet imaging

The device of MMDPCR was assembled by TargetingOne Corporation (Xinyi, Beijing, China), with minor modifications on microfluidic chip for high-throughput droplet analysis, especially for the multiplexing detection (Supplementary material). It integrates three units:

• Microdroplet Generation Unit: in this step, microdroplets was generated by using a microfluidic chip with a cross-junction design to generate micron-sized "water-in-oil" or "oil-in-water" microdroplets.

• PCR & Temperature Control Unit: This step is to perform the PCR amplification reaction on the generated microdroplets. A precision thermal unit with a semiconductor heater provides stable and uniform temperature for the PCR reaction.

• Microdroplet Detection Unit: Reads the amplified droplets. A detection chip singles out each droplet, and a high-throughput optical system measures its fluorescence for final digital quantification.

One-pot multiplex microfluidic digital PCR (MMDPCR)

The MMDPCR assay was performed using a commercial kit (Xinyi, Beijing, China) in a 30-μL total reaction volume based on the above assembled microfluidic system. Each reaction mixture contained 6 μL of PCR Supermix A, 1.5 μL of PCR Supermix B, 1.5 μL of each forward and reverse primer (final concentration: 500 nM), 0.75 μL of each probe (final concentration: 250 nM), and 2 μL of template DNA.

The prepared reaction mixtures were loaded onto the assembled microfluidic system for droplet generation and then subjected to thermal cycling, which included an initial pre-denaturation step at 95 °C for 10 min, followed by 40 cycles of denaturation at 94 °C for 30 s and annealing/extension at 55–66 °C for 1 min. A final cooling step was performed at 12 °C for 5 min.

Following thermal cycling, the entire droplet emulsion was loaded into the sample storage reservoir of the custom-built microfluidic imaging system. The system then automatically performed the sequential batch-wise imaging process described in the above content. Fluorescence signals from positive and negative droplets in each batch were analyzed using the Chip Reader device (Xinyi, Beijing, China) to determine the final concentration of the target molecules. The entire post-PCR analysis was conducted within the closed fluidic system, preventing potential aerosol contamination.

Sensitivity and specificity evaluation

The sensitivity of MMDPCR assay was assessed using serially diluted DNA from the M129 and FH reference strains and clinical isolates. Specificity was verified by testing 12 common respiratory pathogens (Table 1). Analytical sensitivity was determined using tenfold serial dilutions of strain DNA, ranging from 1 × 10⁰ to 1 × 10^–6 ng/µL.

Multiplex qPCR assay

The qPCR assay was performed using the same primer and probe mix as the MMDPCR assay, within a reaction volume of 20 μL. Each reaction contained 10 μL Super Real PreMix (TIANGEN, China), 0.4 μL each of the Mpn137 and F539 forward primers, 0.4 μL each of the Mpn137 and F539 reverse primers, 0.4 μL each of the Mpn137 and F539 probes, 0.28 μL of the 23S rRNA 2063 forward and reverse primers, 0.2 μL of probe 2063-P1, 0.36 μL of probe 2063-P2, 4.48 μL ddH₂O, and 2 μL DNA template. The thermal cycling protocol included initial denaturation at 95 °C for 20 s, followed by 40 cycles of denaturation at 95 °C for 1 s and annealing/extension at 62 °C for 20 s.

Clinical sample collection and detection

In total, 192 clinical specimens (bronchoalveolar lavage fluid: n = 156, throat swabs: n = 6, and sputum samples: n = 30) with suspected M. pneumoniae infection were prospectively collected during the 2023–2024 epidemic period at the affiliated children’s hospital of the Capital Institute of Pediatrics in Beijing.

Total nucleic acids were extracted using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions, with minor modifications: (1) proteinase K digestion was extended to 20 min at 70 °C; and (2) final elution was performed using 50 μL ddH2O. DNA concentrations were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), with the purity criterion set at an A260/A280 ratio of 1.8–2.0. Extracts were stored at −80 °C until analysis.

DNA of all 192 collected clinical samples were simultaneously tested using the newly established MMDPCR assay and the verified real-time PCR method.

Statistical analysis

Statistical significance was calculated by ChiPlot (https://www.chiplot.online). The heat map was initially drafted by ChiPlot and fine-tuned by Adobe Illustrator. The two-tailed Mann–Whitney U test was employed to compare two groups, with a P-value < 0.05 serving as a threshold for significance. Three technical replicates were carried out to enhance the statistics.

Results

Design strategy of the multiplex MDPCR assay

To achieve integrated detection of positive or negative, genotyping, and macrolide resistance simultaneously, we abandoned the traditional method that relied on the P1 gene polymorphism, and instead sought other genome differences between type I and type II strains [15]. Whole-genome alignment of 17 type I and 12 type II clinical isolates revealed two type-specific regions: A 791-bp segment in Mpn137 (nt 177,739–178,529; M129 reference), exclusive to type I (Fig. 1A); A 621-bp segment in F539-03310 (nt 703,619–704239; FH reference), specific to type II (Fig. 1B). These type-specific sequences were subsequently utilised as molecular targets for strain differentiation. Furthermore, given the well-documented association between the A2063G mutation and macrolide resistance—with reported prevalence exceeding 90%—we designed a pair of MGB probes targeting this resistance-conferring mutation to enable simultaneous detection (Fig. 1C,Table 2).

Fig. 1.

Design strategy of the multiplex MDPCR assay. (A) The alignment of type 1 specific primers and probe sequence among different types of M.pneumoniae strains. (B) The alignment of type 2 specific primers and probe sequence among different types of M.pneumoniae strains. (C) The alignment of macrolides susceptible and resistant primers and probes sequence among different types of M.pneumoniae strains

Table 2.

Primers and probes sequence used in this study

Gene	Primer/probe	Sequence (5’−3’)
Mpn137	Mpn137-F	CGATATATCTTCTTCTCATGGCTTCT
	Mpn137-R	CCTGATTAAATGGGAACCCTTACT
	Mpn137-P	FAM-AGCAATACCTCCTTTCTTGTTGTGGT-BHQI
F539	F539-F	TGTCAGGTTCCTTTCAGTCAA
	F539-R	CGTATTGGCACTAACCTTTTTAAAT
	F539-P	VIC-TGCACTTCGCTACTACCAGGAATGTT-BHQ1
23 s RNA	2063-F	GGTGAAGACACCCGTTAG
	2063-R	AAGCTACAGTAAAGCTTCACGG
	2063-P1	CY5-ACGGGACGGAAAGAC-MGB
	2063-P2	ROX-ACGGGACGGGAAGAC-MGB

Establishment and optimization of the multiplex MDPCR assay

The assay (Fig. 2) integrates clinical sample processing with molecular detection in a unified workflow. For sample preparation, DNA is extracted directly from clinical specimens (BALF, sputum, or pharyngeal swabs) or following enrichment in Mycoplasma-specific culture medium. In the molecular detection phase, the MMDPCR system partitions the PCR mixture into approximately 40,000 nanoliter-sized droplets. Multiplex target detection is achieved through four distinct fluorescence channels: FAM signals identify type I strains, VIC signals indicate type II strains, while Cy5 and ROX channels discriminate wild-type 23S rRNA from A2063G mutants, respectively. A microdroplet reader then automatically analyses post-amplification fluorescence signals. This streamlined approach enables direct, simultaneous strain typing and macrolide resistance profiling from clinical specimens within a single assay.

Fig. 2.

Experimental flow chart of MMDPCR assay. (A) Sample processing and DNA extraction: clinical specimens (BALF, sputum, or throat swabs) were enriched in Mycoplasma culture medium, followed by genomic DNA extraction via centrifugal lysis and column-based purification. (B) MMDPCR detection and analysis: target DNA-containing PCR master mix was partitioned into microdroplets. Post-amplification, four fluorescence channels were utilized to determine M. pneumoniae type and 23S rRNA gene mutation status. Positive droplets were automatically quantified using threshold segmentation algorithms

To determine the optimal reaction conditions in a single-tube system, we evaluated various annealing temperatures that might enable efficient amplification by each primer and probe pair. Primers and probes designed for P1 typing exhibited lower melting temperatures (Tm) than those targeting the 2063 locus; thus, they required a reduced annealing temperature range (P1 probes: 58–60 °C; 2063 probes: 61–65 °C). To achieve thermal compatibility in a single reaction, we tested a range of annealing temperatures targeting the 2063 locus (55, 56.5, 58, 59.5, 61, 62.5, 64, and 66 °C) for the typing probes. Testing annealing temperatures (55–66 °C) showed optimal signal separation at 58 °C for both typing probes (Fig. 3A). Validation confirmed successful amplification for reference/clinical strains (M129, FH, 23M574, R1984) under these conditions (Fig. 3B).

Fig. 3.

Optimizing of the MMDPCR assay. (A) Optimization of annealing temperature for genotyping probes. The red horizontal line indicates the fluorescence threshold for positive droplet calling. The boxed area represent the optimal amplification temperature (B) The final established amplification results for each primer and probe channel. The black dots represent negative droplets, while the colored dots represent positive droplets with different fluorescence. (FAM fluorescence for type 1, VIC fluorescence for type 2, ROX fluorescence for macrolide-resistant and Cy5 fluorescence for macrolide-sensitive)

Analysis of MMDPCR assay sensitivity and specificity

To evaluate the specificity of the established MMDPCR assay, we tested a panel of common clinical respiratory pathogens: H. influenzae, M. tuberculosis, S. aureus, P. aeruginosa, K. pneumoniae, E. coli, M. hominis, S. pneumoniae, L. pneumophila, M. genitalium, adenovirus, and C. auris. Only the type I strain M129 and the type II strain FH produced high-intensity positive fluorescence signals. The FAM signal copy number for M129 was 77,283.3, and the VIC signal copy number for FH was 172,949.9. All other strains yielded negative results (Fig. 4A). More than 30,000 droplets and over 3,000 positive threshold events were generated for M129 and FH, effectively excluding interference from non-target bacterial species and environmental factors.

Fig. 4.

The specificity and sensitivity of the MMDPCR assay for M. pneumoniae. (A) The specificity of the MMDPCR assay detection. The black dots represent negative droplets, the blue dots represent positive droplets for M.pneumoniae type 1 strain M129, and the green dots represent positive droplets for M.pneumoniae type 2 strain FH. (B) The sensitivity of the MMDPCR assay detection. FAM fluorescence for type 1, VIC fluorescence for type 2, ROX fluorescence for macrolide-resistant and Cy5 fluorescence for macrolide-sensitive. (C)The corresponding linear experimental results of detection copies and serial dilution concentration. Data points represent the copies of three replicates (n = 3), the plot features the 95% confidence band of the best-fit line

To assess the sensitivity of the MMdPCR assay, we tested tenfold serial dilutions of DNA from type I, type II, macrolide-susceptible, and macrolide-resistant M. pneumoniae strains under the previously optimised experimental conditions. The original DNA concentration was approximately 1 ng/μL. The lowest detectable concentrations were 23.9 copies for type I, 29.7 copies for type II, 20.8 copies for the susceptible strain, and 34.5 copies for the macrolide-resistant strain (Fig. 4B). The assay demonstrated strong linearity across the dilution series, with DNA concentrations ranging from 10^–6 to 10⁰ ng/μL. Regression analysis yielded high correlation coefficients (R² values): 0.9941 for type I, 0.9985 for type II, 0.9949 for the susceptible strain, and 0.9970 for the macrolide-resistant strain (Fig. 4C).

Application of the established MMDPCR assay to clinical samples

To demonstrate the substantial clinical utility and reliability and providing critical information for diagnosis and surveillance, the optimized MMdPCR assay was used to analyze 192 suspected clinical specimens during the 2023–2024 M. pneumoniae epidemic yea. The assay achieved a 56.25% positivity rate (108/192 cases), with results demonstrated 100% concordance with real-time PCR (108/192 cases) (Fig. 5A), confirming its diagnostic reliability. Quantitative analysis revealed high pathogen burdens (> 10⁴ copies) in most infections, while its detection of low-load samples (< 10² copies) demonstrated exceptional sensitivity for early-stage or subclinical infections (Fig. 5B). Critically, the assay provided actionable epidemiological insights: Type I strains dominated infections (75.9%, 82/108), suggesting potential heightened virulence or transmissibility, while Type II strains remained in circulation (24.1%, 26/108). Analysis of macrolide resistance revealed a critical overlap: the A2063G mutation was present in 99.1% (107/108) of all samples. As clearly depicted in Fig. 5B, this resistance was overwhelmingly associated with Type I strains, with 82 of the 107 macrolide-resistant strains belonging to this type. The single macrolide-sensitive strain was identified as Type II. This finding necessitates urgent reconsideration of empirical macrolide therapy in favor of alternatives like tetracyclines or fluoroquinolones.

Fig. 5.

Application of the MMDPCR assay in clinical samples. (A) Comparison of MMDPCR and real-time PCR detection results. (B) The number of different genotypes, macrolide-sensitive and macrolide-resistant among the tested clinical samples. And the number of samples with extremely low, low, medium and high copy numbers among the tested samples

Discussion

M. pneumoniae is an important bacterial pathogen responsible for respiratory tract infections; it has exhibited a cyclical epidemic pattern of every 3–7 years [16, 17]. Between 2020 and 2021, the incidence of M. pneumoniae substantially declined after implementation of non-pharmaceutical interventions to control coronavirus disease 2019 (COVID-19) [18]. However, beginning in 2023, a large-scale resurgence with high morbidity due to M. pneumoniae was observed in Europe and Asia; other respiratory pathogens continued to show consistently lower incidences [17]. Several hypotheses have been proposed to explain this shift in epidemiological trends, including the concept of herd immunity [19]. This situation underscores the urgent need for robust surveillance of M. pneumoniae epidemics; it also supports the development of rapid and straightforward detection methods.

Molecular typing and macrolide resistance profiling are critical indicators for effective surveillance of M. pneumoniae. To improve the speed and practicality of monitoring epidemic patterns, we developed a MMDPCR assay integrating detection, genotyping, and macrolide resistance assessment into a single reaction system. The application of this assembled Multiplex Microfluidic Digital PCR not only enables precise quantification of copy numbers; it also facilitates differentiation of co-infections involving multiple M. pneumoniae genotypes in complex clinical cases which will have a good benefit for epidemic surveillance and rational use of antibiotics of clinicians. The platform’s clinical impact stems from its ability to deliver comprehensive pathogen profiling – simultaneously identifying strain type, resistance status, and bacterial load within 2 h. Although some newly developed methods, like the LAMP [20], MCDA-CRISPER [21], and RAA [22] et al., can be completed within 30 to 60 min, they cannot provide the results of genotypes and drug resistance simultaneously. Compared with the previous method, which first detection exist of the pathogen and then molecular characteristic (about 4–5 h), the integrated multiplexing new method significantly reduced the turnaround time. This enables fast detection and supports real-time surveillance of strain dynamics and resistance evolution.

As one of the smallest self-replicating organisms, M. pneumoniae possesses a compact and highly stable genome [23]. The P1 adhesin gene functions as the primary virulence factor of M. pneumoniae, with gene polymorphisms—particularly recombination events involving RepMP2/3/4 repeat units—resulting in the formation of two genotypes: I and II [24]. With advances in sequencing technology and the successful sequencing of additional clinical isolates, it has been observed that, beyond the P1 gene, several other genes also exhibit the characteristic features of type I and type II strains [25]. These observations provide alternative targets for P1 typing, supplementing the conventional approach of HaeIII enzyme digestion after PCR amplification of the P1 gene. Although the multiple-locus variable number tandem-repeat analysis (MLVA) typing system allows further subdivision of these two genotypes, a strong correlation persists between MLVA profiles and P1 types. For instance, MLVA type M4572 is primarily associated with P1-I, whereas MLVA types M3562 and M3662 are predominantly linked to P1-II [10, 24]. Based on these data, we developed a novel P1 typing method to specifically target the Mpn137 gene in type I strains and the F539 gene in type II strains. Gene alignment analyses confirmed the specificity and feasibility of this approach.

Macrolide antibiotics remain the first-line treatment for M. pneumoniae infection in paediatric populations [26]. However, macrolide-resistant M. pneumoniae (MRMP) has become a serious clinical concern. In China, resistance rates have exceeded 99% during epidemic years [7]; the A2063G/C mutation in domain V of the 23S rRNA gene is the most frequently reported site and constitutes a key contributor to high-level resistance (MIC ≥ 64 μg/mL) [27]. To improve the understanding of resistance patterns, we designed an MGB TaqMan probe to target this mutation site; our probe was integrated with the P1 typing system into a single assay. Through systematic optimization of reaction parameters, the newly developed assay exhibiting high specificity and a relatively low limit of detection (LOD) ranging from 10 to 100 copies per reaction. When compared to existing methods, the current assay demonstrates superior or comparable analytical sensitivity. While the commercial MycoSEQ™ kit has a reported LOD of 10 CFU/mL, other advanced techniques, like the MP-MCDA-CRISPR assay reaches 50 fg per reaction [21], the RAA assay was 2.23 copies per reaction [22], and the duplex real-time PCR assay reports LDLs of approximately 8.1 fg and 9.3 fg DNA for M. pneumoniae type 1 and type 2, respectively [28].

Based on the newly established Multiplex Microfluidic Digital PCR, we investigated the 2023–2024 M. pneumoniae epidemic in Beijing. We observed a positivity rate of 56.25% and macrolide resistance rate of 99%, both substantially higher than average rates reported over the previous 5 years. For example, in 2019, data from Beijing indicated an M. pneumoniae positivity rate of 17.59% in children. This rate considerably declined in 2020 and 2021, to 8.9% and 4.95%, respectively. However, in 2023, the positivity rate sharply increased, reaching 25.4% among outpatients, 48.4% among hospitalised patients, and 61.1% among patients with respiratory diseases.

Furthermore, the rate of MRMP has exhibited a consistent annual rise since 2008 [29, 30]. These epidemiological findings indicate that the spread of MRMP strains has been the primary driver of the recent outbreak observed in Beijing. This conclusion is supported by a recent local study reporting a MRMP rate of 99% during the same period [7], alongside corresponding increases in the minimum inhibitory concentration (MIC) of erythromycin [31]. In alignment with this trend, a study by Yanming Sun et al. documented a slightly lower—though still predominant—MRMP prevalence of 95.8% throughout a comparable epidemic wave, further corroborating the pervasiveness of macrolide resistance [32]. Moreover, data from an international surveillance study spanning more than 20 countries identified MRMP as the predominant circulating strain since 2023 [19], Collectively, these findings underscore the critical importance of integrating routine macrolide resistance screening into clinical practice to guide appropriate therapeutic interventions. Furthermore, the predominance of type I strains (75.9%) observed in this study is consistent with earlier reports, indicating that type I remains the dominant genotype in the current epidemic, despite its slightly lower prevalence [5, 33]. The persistent dominance of type I may reflect genotype displacement driven by antimicrobial selection pressure—a phenomenon previously documented during the post-pandemic resurgence of M. pneumoniae, where reduced immune priming and antibiotic overuse favoured the emergence of macrolide-resistant genotypes [34].

However, this study had some limitations. Although this assay is highly effective for detection of low-copy targets, its operational complexity and associated costs may hinder implementation in resource-limited settings [35]. Additionally, due to the limited number of fluorescence channels and the potential for cross-reactivity among primers and probes, the current assay does not detect all known macrolide resistance mutation sites (e.g., A2064G and A2617G), resulting in incomplete coverage of these less common mutations. Future iterations of the assay will aim not only to broaden mutation coverage but also to simplify operational procedures and reduce costs, facilitating its adoption in routine clinical practice. Specifically, the use of novel probes based on silicon-doped carbon quantum dots (Si-QDs) and single-layer graphene oxide (GO) has demonstrated excellent selectivity, sensitivity, stability, low photobleaching, and cost-effectiveness, offering a viable alternative to expensive TaqMan probes [36].

In conclusion, the developed Multiplex Microfluidic Digital PCR assay provides a robust tool for the simultaneous characterisation of M. pneumoniae infection dynamics, strain typing, and antibiotic resistance profiles. The assay’s precise quantification capabilities can support the rational use of macrolide antibiotics in clinical practice; facilitate assessments of treatment efficacy; and provide valuable molecular epidemiological data to guide tracing, prevention, and control of nosocomial infection outbreaks.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

JY and GHX designed the study and revised the manuscript. YYX, YLF, HQZ and JXF performed the experiments. YYX, FZ, LG, TTF and YY collected and processed the clinical samples., YYX,YHP,SS,CY, JHC and ZYX, ZHY and YHK analysed the results. GHX,YYX and JY wrote and reviewed the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by grants from the National Natural Science Foundation of China (32170201), Beijing Natural Science Foundation (7242015, 7232007, L232071), National Natural Science Foundation of China for Key Programs of China Grants (82130065), National Key Research and Development Projects of the Ministry of Science and Technology of the People’s Republic of China (2024YFC2309204), Beijing High-Level Public Health Technical Talent Project (2023–02-08), Beijing Municipal Public Welfare Development and Reform Pilot Project for Medical Research Institutes (JYY2023-10) and Beijing Hospitals Authority’s Ascent Plan (DFL20241301).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.

Consent for publication

All the authors have read and approved the final manuscript.

Clinical trial number

Not applicable.

Statement on animal welfare

Not applicable.