Blood Tests Fail to Flag Drug-Resistant Mycoplasma in Children

Routine blood tests did not distinguish children with Mycoplasma pneumoniae carrying the 2063/2064 mutation, associated with reduced macrolide susceptibility, from those whose infection lacked the detected mutation. In a retrospective single-centre study of 161 children with nucleic-acid-confirmed M. pneumoniae infection, neither white blood cell counts, inflammatory proteins, nor composite indices such as the neutrophil-to-lymphocyte ratio differed between the mutation-detected and mutation-undetected groups at diagnosis. The researchers did observe a modest fall in the monocyte-to-lymphocyte ratio over time among 64 mutation-detected children with complete paired samples, but follow-up analyses showed that the change was poorly able to classify individual cases and appeared concentrated among hospitalised patients.

The findings indicate that common haematological measurements can reflect the presence of infection without revealing the pathogen’s resistance-associated mutation. The study enrolled children treated at the Fourth Affiliated Hospital of Soochow University in Suzhou, China, during 2024. All were aged 14 years or younger, had positive respiratory pathogen nucleic-acid tests for M. pneumoniae, and had complete clinical and laboratory information. Investigators separated the infection cohort into 134 children in whom the 2063/2064 mutation was detected and 27 in whom it was not detected. A separate group of 27 healthy children provided a reference for distinguishing infection-associated blood changes from normal values. Children with coinfections or missing data were excluded.

At diagnosis, the mutation-detected group had a median white blood cell count of 7.46 × 10^9 cells per litre, compared with 7.03 × 10^9 in the mutation-undetected group. Median lymphocyte counts were 1.87 and 1.78 × 10^9 cells per litre, respectively, while monocyte counts were 0.56 and 0.61 × 10^9 cells per litre. None of these differences was statistically significant. The same was true of the neutrophil-to-lymphocyte ratio, monocyte-to-lymphocyte ratio, systemic immune-inflammatory index, and high-sensitivity C-reactive protein. These results apply specifically to the mutation status and population examined; the retrospective design and low statistical power for several secondary indices mean that modest differences cannot be completely excluded.

Blood markers were more clearly associated with infection itself. Compared with healthy controls, children with M. pneumoniae infection had higher white blood cell, neutrophil, and monocyte counts and markedly higher high-sensitivity C-reactive protein. The median C-reactive protein concentration was 9.38 mg/L in the infection cohort, compared with 0.28 mg/L in controls. The infected children were also younger, with a median age of seven years versus nine in the control group, so the groups were not demographically equivalent in every respect. The authors noted that inflammation can be triggered by pathogen adherence, tissue injury, and immune responses, but these processes do not necessarily correspond directly to the molecular changes responsible for reduced macrolide sensitivity.

To examine whether change over time was more informative than a single measurement, researchers compared blood samples taken at the first visit after respiratory symptoms began with samples obtained when the M. pneumoniae and 23S ribosomal RNA mutation results were reported. The median interval to diagnosis was seven days in the mutation-detected group and six days in the mutation-undetected group, without a significant difference. Complete paired measurements were available for 64 mutation-detected children. Among them, the monocyte-to-lymphocyte ratio fell from a median of 0.32, with an interquartile range of 0.24 to 0.48, at the first visit to 0.27, with an interquartile range of 0.19 to 0.37, at diagnosis. The paired comparison gave a p-value of 0.004.

Other laboratory measures did not change significantly between the two sampling points in the mutation-detected subgroup. These included white blood cell, neutrophil, lymphocyte and monocyte counts, the neutrophil-to-lymphocyte ratio, the systemic immune-inflammatory index, and high-sensitivity C-reactive protein. In the smaller mutation-undetected group, none of the measured haematological or inflammatory indices showed a significant paired change, including the monocyte-to-lymphocyte ratio. The longitudinal analysis therefore identified a temporal pattern within one subgroup rather than a paired contrast directly demonstrating that the change was specific to resistant infection.

Statistical checks tempered the apparent significance of the monocyte-to-lymphocyte ratio decline. A bootstrap analysis using 20,000 resamples produced a 95% confidence interval for the mean paired difference of −0.0988 to 0.0155, which included zero. The authors explained that the bootstrap and paired rank test addressed different features of the data, but together the results indicated uncertainty in the size and stability of the change. Post-hoc power was low for the inflammatory indices: 0.09 for the monocyte-to-lymphocyte ratio, 0.11 for the neutrophil-to-lymphocyte ratio, 0.14 for the systemic immune-inflammatory index, and approximately 0.17 for high-sensitivity C-reactive protein. Thus, the absence of statistically significant group differences may partly reflect insufficient precision, though it does not establish a useful diagnostic indicator.

Exploratory receiver operating characteristic analyses also showed that the change in monocyte-to-lymphocyte ratio had little ability to distinguish mutation status when used alone. The area under the curve was 0.508, with a 95% confidence interval of 0.344 to 0.672, for the signed change from first visit to diagnosis. For the absolute magnitude of that change, the area under the curve was 0.523, with a 95% confidence interval of 0.349 to 0.699. An exploratory threshold based on absolute change yielded 85.9% sensitivity but only 29.4% specificity, implying a high false-positive rate. Regression analysis likewise found no independent association between mutation-detected status and either age or the monocyte-to-lymphocyte ratio at diagnosis.

Hospitalisation-stratified analysis suggested that clinical circumstances influenced the longitudinal result. Of the 64 mutation-detected children with paired measurements, 49 were hospitalised and 15 were not. The monocyte-to-lymphocyte ratio declined significantly among hospitalised children, from a median of 0.35 to 0.30, but not among those treated without hospital admission, when it changed from 0.39 to 0.37. Hospital admission may reflect greater disease severity, longer symptoms, different treatment exposure, or closer monitoring, none of which was directly measured in a way that could separate these influences. An exploratory multivariable model found no statistically significant association with mutation status for sex, age, hospitalisation, monocyte-to-lymphocyte ratio change, C-reactive protein change, or fever duration.

Overall, the study provides no support for using a static blood count or a single inflammatory index as an early warning test for the 2063/2064 mutation, and it does not show that serial monocyte-to-lymphocyte ratio measurements can replace molecular susceptibility testing or assessment of disease severity. The retrospective, single-centre design, unequal mutation-status groups, incomplete treatment records, lack of standardised severity scores, and insufficient paired data in the mutation-undetected group limit generalisation. Prospective studies with larger and more balanced cohorts, paired sampling in both groups, and complete information on treatment and disease course would be needed to determine whether changes in blood indices add useful information alongside established clinical measures.

Subject of Research: Evaluation of routine blood indices for identifying the 2063/2064 mutation in paediatric Mycoplasma pneumoniae infection.

Article Title: Analysis of early haematological warning indicators for drug-resistant Mycoplasma pneumoniae infection

Article References: Yang, Y., Sun, T., Fu, C., & Yan, X. (2026). Analysis of early haematological warning indicators for drug-resistant Mycoplasma pneumoniae infection. Frontiers in Pediatrics, 14, Article 1854263. https://doi.org/10.3389/fped.2026.1854263

Image Credits: AI Generated

DOI: 10.3389/fped.2026.1854263

Keywords: Mycoplasma pneumoniae, macrolide resistance, paediatric pneumonia, blood counts, monocyte-to-lymphocyte ratio, inflammatory markers, 23S rRNA, diagnostic accuracy, retrospective study, Analysis, early, haematological