Cytokine Profile Changes in Infants With Respiratory Syncytial Virus Bronchiolitis With Nirsevimab

To the Editor,

Respiratory syncytial virus (RSV) is one of the leading causes of hospitalization (approximately 3% of cases) and pediatric intensive care unit (PICU) admission among infants and young children worldwide [1]. Each year, RSV is responsible for 33 million cases of RSV‐associated acute respiratory infection. Therefore, the development and administration of effective RSV prevention is critically needed. Two primary strategies are currently available to protect infants from RSV infection [2]. Passive protection can be achieved through maternal vaccination during late pregnancy and/or immunization with a monoclonal antibody after birth or at the onset of the RSV season in infants. In this context, nirsevimab was approved in 2023 to reduce the healthcare burden associated with RSV infection in infants, and it was the strategy adopted by the Spanish Health Ministry for all infants achieving coverage above 90%. Nirsevimab targets the RSV F protein, a key viral glycoprotein required for viral entry and membrane fusion, showing potent neutralizing activity and provides passive immunity in infants. Because it does not rely on priming adaptive T‐cell responses, nirsevimab avoids skewing the T1/T2 balance and prevents the antibody‐dependent enhancement phenomena observed with some vaccines. However, these mechanisms are not yet fully understood or characterized in the context of recurrent RSV reinfections. Therefore, the present exploratory study aimed to characterize the profile of pro‐inflammatory mediators (including cytokines, chemokines, and other pro‐inflammatory molecules) and to compare these profiles between children with RSV infection who did not receive nirsevimab and those who received prophylaxis but subsequently developed infection.

For this study, Hispanic infants were prospectively enrolled at Severo Ochoa University Hospital (Leganés, Madrid) and La Paz University Hospital (Madrid). The cohort included 38 hospitalized infants due to RSV infection, 18 of them without nirsevimab prophylaxis and 20 infants after receiving nirsevimab. This study was conducted in accordance with the Declaration of Helsinki and approved by the Clinical Research Ethics Committee of the Severo Ochoa University Hospital (Leganés, Madrid) and La Paz University Hospital (Madrid) (PI‐6300). Parents of all participating infants provided written informed consent prior to study enrollment. The following clinical parameters were collected: age, gender, personal and family history of atopy, viral coinfections, oxygen therapy, intensive care needs, and duration of stay. Nasopharyngeal aspirates (NPAs) for immunological and virological analyses were obtained by nasal wash in the first 24–48 h after admission. All biological samples were anonymized and identified using coded study IDs. Samples were processed at the Influenza and Respiratory Viruses Laboratory, National Centre for Microbiology (Madrid, Spain). Respiratory virus detection was performed by four independent real‐time multiplex PCR (RT‐PCR) assays using the SuperScript III Platinum One‐Step Quantitative RT‐PCR System (Invitrogen, Waltham, MA, USA) to detect 18 respiratory viruses. Cytokine concentrations in NPA supernatants were measured in duplicate using the MILLIPLEX MAP Human Cytokine/Chemokine Magnetic Bead Panel (Magpix, Merck Millipore, Massachusetts, USA) according to the manufacturer's instructions. The panel included the following analytes: sCD40L, EGF, FGF2, GM‐CSF, IFN‐α2, IFN‐γ, IL‐1β, IL‐2, IL‐4, IL‐5, IL‐8, IL‐9, IL‐10, IL‐12p70, IL‐13, IL‐17A, IL‐17E, IL‐33, IP‐10, MCP‐1, MIP‐1α, MIP‐1β, TNF‐α, TSLP, and VEGF. Calibration standards and quality control samples were also assayed in duplicate in each run. According to the manufacturer's validation data, the assay demonstrates high reproducibility, with intra‐assay coefficients of variation generally < 10% and inter‐assay coefficients of variation < 15% for most analytes. It is important to note that laboratory investigators which performed these cytokine measurements were blinded to the participants' nirsevimab prophylaxis status throughout the analyses. Additionally, the unsupervised heatmap and principal component analysis (PCA) were generated using the ClustVis web tool [3]. It is important to note that cytokine concentrations falling below the lower limit of detection (LLOD) were excluded from individual statistical comparisons, and individual valid data points were explicitly plotted in figures to ensure data transparency. For heatmap generation and hierarchical clustering using ClustVis tool, missing values below LLOD were imputed using the software's integrated SVD/NIPALS‐based imputation algorithm to enable complete matrix calculations. Data are presented as mean ± standard error of the mean (SEM) for variables with a normal distribution, or as median with interquartile range (Q1–Q3) for non‐normally distributed variables. The normality of each cytokine and clinical variable was assessed individually using the Shapiro–Wilk test. Comparisons between groups were subsequently performed using the unpaired Student's t‐test for normally distributed data or the Mann–Whitney U test for non‐normally distributed data. For clinical variables, p values < 0.05 were considered statistically significant. To account for multiple comparisons across the cytokine panel, p values were adjusted using the Benjamini–Hochberg false discovery rate (FDR) procedure. Considering the exploratory design of the study and the limited cohort size, an FDR threshold of 10% (Q < 0.10) was adopted to reduce the risk of false negatives (Type II errors) while retaining hypothesis‐generating discovery potential. Therefore, for cytokine analyses, adjusted p values (Q values) < 0.10 were considered suggestive of statistical significance. Otherwise, to assess the potential influence of clinical severity and other relevant clinical characteristics on cytokine concentrations, an additional multivariable general linear model was performed. Candidate covariates included clinically relevant variables and/or variables showing a p value < 0.10 in the univariate comparison between groups (Table 1), including indicators of respiratory disease severity such as hypoxia and mechanical ventilation. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA) and IBM SPSS Statistics version 29 (IBM, Armonk, NY, USA).

Table 1.

Comparison of clinical characteristics of patients immunized with nirsevimab prior to sample collection.

Characteristics	Nirsevimab = Yes (n = 20)	Nirsevimab = No (n = 18)	p
Male sex	10/20 (50%)	8/18 (44.4%)	0.73
Prematurity	0/20 (0%)	3/18 (16.7%)	0.057
Breastfeeding	13/20 (65%)	16/18 (88.9%)	0.084
Breastfeeding (months)	4.8 ± 3.3	3.04 ± 1.7	0.10
Cesarean section	2/20 (10%)	5/18 (27.8%)	0.15
Daycare attendance	5/20 (26.2%)	5/18 (27.8%)	0.92
Age at daycare start (months)	7.7 ± 1.5	5.4 ± 3.0	0.28
Food allergy	1/20 (5.3%)	2/18 (11.1%)	0.51
Atopic dermatitis	1/20 (5.3%)	1/18 (5.6%)	0.96
Siblings under 5 years	13/20 (65%)	13/18 (72.2%)	0.63
Passive smoking	2/20 (10%)	4/18 (22.2%)	0.30
Family history
Maternal asthma	4/20 (20%)	2/18 (11.1%)	0.45
Maternal atopy	6/20 (30%)	2/18 (11.1%)	0.15
Paternal asthma	3/19 (15.8%)	2/18 (11.1%)	0.67
Paternal atopy	5/19 (26.3%)	5/18 (27.8%)	0.92
Sibling asthma	1/16 (6.3%)	1/17 (5.9%)	0.96
Sibling atopy	2/15 (13.3%)	3/17 (17.6%)	0.73
Hospitalization details
Length of stay (days)	6 ± 3.3	6.7 ± 3.2	0.52
Age at admission (months)	5.8 ± 3.9	4.5 ± 3.2	0.28
Fever > 38°C	10/20 (50%)	8/18 (44.4%)	0.73
Fever duration (days)	2.9 ± 1.4	1.8 ± 0.9	0.057
Hypoxia (Sat < 95%)	16/20 (80%)	18/18 (100%)	0.045
Hypoxia duration (days)	6.1 ± 2.9	5.8 ± 2.8	0.76
High‐flow oxygen therapy	10/20 (50%)	11/18 (61.1%)	0.49
High‐flow duration (days)	4.6 ± 1.6	3.8 ± 1.5	0.27
PICU admission	2/20 (10%)	5/18 (27.8%)	0.15
PICU stay (days)	4.5 ± 2.1	3.8 ± 2.0	0.70
Mechanical ventilation	1/20 (5%)	5/18 (27.8%)	0.055
Mechanical ventilation duration (days)	3	2.8 ± 1.9	0.92
Positive viral detection	20/20 (100%)	18/18 (100%)	—
Viral coinfection	10/20 (50%)	8/18 (44.4%)	0.73
RSV infection	20/20 (100%)	18/18 (100%)	—
HRV infection	6/20 (30%)	6/18 (33.3%)	0.82
HAdV infection	1/20 (5%)	2/18 (11.1%)	0.48
hMPV infection	0/20 (0%)	0/18 (0%)	—
HBoV infection	1/20 (5%)	0/18 (0%)	0.33
Influenza (FLU) infection	0/20 (0%)	0/18 (0%)	—
PIV infection	1/20 (5%)	1/18 (5.6%)	0.93
CoV infection	3/20 (15%)	0/18 (0%)	0.087
EV infection	0/20 (0%)	0/18 (0%)	—
Leukocytes (cells/mm3)a	12.856 ± 13.08	12.175 ± 5.08	0.89
C‐reactive protein (mg/L)a	36.7 ± 41.1	8.9 ± 6.6	0.079
Procalcitonin (ng/mL)a	0.11 ± 0.05	0.99 ± 1.15	0.15

Note: Quantitative variables are shown as mean ± SD.

^a Blood samples were available for 10 infants in the immunized group and 8 infants in the nonimmunized group.

Table 1 shows the clinical data of immunized and nonimmunized patients. No significant differences were observed between the immunized and nonimmunized patients, except for hypoxia, which was more frequent in nonimmunized children (100% vs. 80%, p < 0.05, Table 1), and a slight, nonsignificant trend toward an increased need for mechanical ventilation (5% vs. 27.8%, p = 0.055, Table 1). Interestingly, nirsevimab appeared to exert a protective role against hypoxia (odds ratio = 0.47, 95% CI: (0.33–0.67), p < 0.05, data not shown).

Regarding the comparison of cytokine profiles between the two groups (Figure 1), four cytokines were found to be differentially expressed (Figure 1A). Related to T1 responses, IL‐2 (1.63 ± 0.19 vs. 1.02 ± 0.13 pg/mL, Q < 0.1, Figure 1A) and IL‐12p70 (8.30 ± 1.00 vs. 5.14 ± 0.69 pg/mL, Q < 0.1, Figure 1A) were significantly decreased in infants who experienced infection after nirsevimab administration, whereas IP‐10 (CXCL10) was significantly increased in this group (458.80 ± 171.70 vs. 95.97 ± 48.91 pg/mL, Q < 0.1, Figure 1A). Additionally, IL‐1β was significantly reduced in immunized infants (445.44 ± 101.13 vs. 1570.69 ± 343.89 pg/mL, Q < 0.1, Figure 1A). To further assess the potential influence of clinical severity on the observed cytokine differences, multivariable analyses were performed. After adjustment for relevant clinical variables, the differences between the immunized and nonimmunized groups remained significant for all cytokines except IL‐2 (p = 0.056, data not shown). Notably, the difference in IL‐2 levels was no longer statistically significant when mechanical ventilation was included in the model, suggesting that the observed IL‐2 difference may be influenced by mechanical ventilation.

Figure 1.

Different cytokine profile in infants with RSV infection with/without nirsevimab prophylaxis. Nasopharyngeal aspirate samples from infants with RSV were analyzed by Luminex. (A) IL‐2 (nonimmunized, n = 18; immunized, n = 20), IL‐12p70 (nonimmunized, n = 18; immunized, n = 19), IP‐10 (CXCL10) (nonimmunized, n = 18; immunized, n = 19), and IL‐1β (nonimmunized, n = 18; immunized, n = 18) levels were compared between infants with nirsevimab (immunized) and those without prophylaxis (nonimmunized). Infants with prior nirsevimab exposure showed significantly reduced levels of IL‐2, IL‐12p70, and IL‐1β, accompanied by an increase in IP‐10. (B) For IL‐5 (nonimmunized, n = 15; immunized, n = 11) and IL‐17A (nonimmunized, n = 10; immunized, n = 7), a reduction was observed of these cytokines in immunized children, although without statistical differences. (C) Principal component analysis (PCA) showing the distribution of individual patients according to their overall cytokine profile. (D) Unsupervised hierarchical clustering heatmap showing cytokine expression profiles across individual patients. Missing values in the PCA and heatmap were imputed using the algorithm implemented in ClustVis. Data are shown as mean ± SEM. *Q < 0.1. [Color figure can be viewed at wileyonlinelibrary.com]

Other important cytokines associated with the immune response, including IL‐5 (2.26 ± 0.86 vs. 1.01 ± 0.58 pg/mL, Q = 0.1994, Figure 1B) and IL‐17A (20.85 ± 2.35 vs. 12.85 ± 2.83 pg/mL, Q = 0.1994, Figure 1B), also showed decreased levels in immunized infants, although these changes did not reach statistical significance after multiple‐comparison adjustment. Taking all these cytokines into account, the unsupervised heatmap and PCA revealed a distinct cytokine profile between the two groups (Figure 1C,D). In contrast, no significant differences were observed in the remaining pro‐inflammatory mediators evaluated (with the exception of sCD40L and FGF2, which were not detected), including key molecules involved in antiviral defenses and T1 responses (IFN‐γ and TNF‐α), as well as epithelial alarmins associated with epithelial damage signaling (IL‐25, IL‐33, and TSLP) (data not shown).

To our knowledge, this is the first exploratory study reporting changes in the cytokine profile of infants infected with RSV with or without nirsevimab prophylaxis. Although our small cohort showed only mild differences in disease severity between the groups (with a lower percentage of hypoxia in immunized children), some studies have reported that hospitalized children who received nirsevimab presented with a less severe clinical course [4]. Interestingly, analysis of NPA samples revealed a consistent pattern in infants who had received nirsevimab but subsequently developed infection: reduced levels of IL‐2, IL‐12p70, and IL‐1β, accompanied by a tenfold increase in IP‐10 compared with nonimmunized infants. These changes may result from selective modulation of both innate and adaptive mucosal immune responses in the presence of nirsevimab. IL‐2 is known to support T‐cell proliferation and survival, whereas IL‐12p70 plays a central role in polarizing T cells toward IFN‐γ–producing responses [5, 6]. Therefore, the reduction in IL‐2 and IL‐12p70 suggests a potential attenuation of T1 response in the upper airway during infection. Previous studies have associated elevated mucosal or systemic IL‐12 and IL‐2 levels with more pronounced cellular inflammatory responses in severe RSV disease, suggesting that reduction of these signals may mitigate immunopathology, albeit at the potential cost of dampening certain adaptive immune features. Although the difference in NPA IL‐2 concentration was modest, levels of this cytokine are comparable with those in the lower respiratory tract of infants with RSV‐positive bronchiolitis [7], indicating that NPA cytokine levels provide a simple and useful means of assessing overall respiratory tract inflammation. Given that NPA represent diluted secretions, subtle changes in these concentrations can reflect meaningful immunological and pathophysiological variations during acute infection. However, the multivariable analysis provides additional insight into the potential influence of disease severity on the observed cytokine profile. While the differences between the immunized and nonimmunized groups remained significant for the other cytokines after adjustment for relevant clinical variables, the difference in IL‐2 levels was no longer significant when mechanical ventilation was included as a confounding factor. This finding suggests that the lower IL‐2 levels observed in the immunized group may be partly influenced by mechanical ventilation rather than being solely attributable to nirsevimab administration. This interpretation is biologically plausible, as mechanical ventilation has been shown to modify the pulmonary inflammatory response in infants with severe RSV bronchiolitis. Hennus et al. reported a variation in the concentrations of several inflammatory mediators in NPA after initiation of mechanical ventilation in infants with RSV bronchiolitis [8]. However, direct evidence specifically linking mechanical ventilation to changes in IL‐2 concentrations is limited. Indeed, a previous study assessing IL‐2 in nasopharyngeal secretions of infants with RSV bronchiolitis did not find a significant association between IL‐2 levels and disease severity [9]. Therefore, our finding regarding IL‐2 should be considered exploratory and requires confirmation in larger cohorts. Otherwise, the observed decrease in IL‐1β may indicate attenuation of inflammasome‐associated innate inflammation [10]. This cytokine has been involved in epithelial injury, neutrophil recruitment, and poorer clinical outcomes in infants with severe bronchiolitis. Reduced IL‐1β levels in nirsevimab‐exposed infants may therefore reflect decreased innate cell activation or lower amplification of danger signals in the mucosa. In contrast, IP‐10 was observed to be increased in NPA samples from the immunized group. IP‐10 is an IFN‐inducible chemokine involved in the recruitment of activated T cells and natural killer cells and is consistently induced during RSV infection. Importantly, experimental evidence suggests that IP‐10 may contribute to effective antiviral immunity during RSV infection. In a murine model, blockade of IP‐10 resulted in impaired viral clearance, reduced dendritic cell accumulation and maturation, and fewer RSV‐specific CD8⁺ T cells in the lungs, supporting a role for the IP‐10/CXCR3 axis in promoting antiviral responses [11]. Interestingly, antibody‐mediated modulation of this pathway may also occur during RSV infection. Vissers and colleagues reported that the presence of antibodies enhanced IP‐10 production by immune cells from infants exposed to RSV, despite reducing the rate of RSV infection, suggesting that antibodies may modulate the innate immune response during RSV infection [12]. These findings may provide a potential biological explanation for the higher IP‐10 concentrations observed in nirsevimab‐treated infants, as the presence of anti‐RSV antibodies may contribute to a differential IP‐10 response during infection. However, this interpretation remains speculative. Moreover, IP‐10 levels may be influenced by viral burden, disease severity, and the timing of sampling. Quantitative viral load measurements were not available in our cohort, and the exact time from symptom onset to NPA collection was not systematically recorded. Therefore, the increased of IP‐10 concentrations observed in nirsevimab‐treated infants should be interpreted as an association rather than evidence of a direct immunomodulatory effect of nirsevimab.

Our study has some limitations, such as the small number of patients which may not have sufficient power for analyzing so many variables and the lack of viral load data, given that respiratory samples are difficult to homogenize. Moreover, age‐related immune maturation may also influence cytokine production during infancy [13]. Although age at admission did not differ significantly between groups, the cohort of included infants ranging from 1 to 12 months of age, and giving the modest sample size, stratification or multivariable adjustment by age would substantially reduce statistical power and may result in unstable estimates. Therefore, a potential effect of age on cytokine responses cannot be completely excluded in this study. Additionally, it should also be considered that the nirsevimab‐immunized group comprised only infants who developed breakthrough RSV infection despite prophylaxis. These infants may represent a biologically distinct subset with potentially different susceptibility to infection or host immune responses. Therefore, although the cytokine profile observed in this cohort is different, it should not be generalized to all infants receiving nirsevimab prophylaxis. Additionally, as no a priori sample size or power calculation was performed, and given the exploratory and hypothesis‐generating nature of this study, an FDR < 0.10 was applied to balance Type I and Type II errors. Consequently, these findings should be interpreted as suggestive of distinct NPA cytokine profiles rather than definitive proof of causal immunomodulation. Larger, independent, and adequately powered prospective cohorts are required to confirm these preliminary observations. Nevertheless, taken together, the observed cytokine profile could be consistent with a model in which passive immunization with a potent anti‐F monoclonal antibody (nirsevimab) reduces viral replication and tissue damage, thereby lowering broad proinflammatory and inflammasome‐derived mediators (IL‐1β, IL‐12p70, IL‐2), while IP‐10 may modulate antiviral cellular responses.

Author Contributions

José Antonio Cañas: investigation, writing – original draft, methodology, writing – review and editing, formal analysis, data curation, supervision. Antonio Serrano‐Santiago: investigation, writing – original draft, methodology, writing – review and editing, formal analysis, data curation. Patricia Alonso‐López: methodology, writing – review and editing, visualization, data curation. Daniel Rodríguez‐González: methodology, visualization, writing – review and editing, data curation. Gema Guillén‐Sánchez: methodology, visualization, writing – review and editing, data curation. Sonia Alcolea: methodology, writing – review and editing, formal analysis, data curation, visualization. Inmaculada Casas: methodology, visualization, writing – review and editing, data curation. María Luz García‑García: conceptualization, funding acquisition, writing – original draft, writing – review and editing, formal analysis, data curation, supervision. Cristina Calvo: conceptualization, funding acquisition, writing – original draft, methodology, writing – review and editing, formal analysis, data curation, supervision. Victoria del Pozo: conceptualization, methodology, writing – review and editing, funding acquisition, writing – original draft, formal analysis, data curation, supervision.

Acknowledgments

This study has been funded by Instituto de Salud Carlos III (ISCIII) through the project PI24/00212 and co‐funded by the European Union. Also, the work has been supported by ISCIII‐FIS (Fondo de Investigación Sanitaria ‐ Spanish Health Research Fund) [Grant Numbers PI18CIII/00009, PI21CIII/00019, PI21/00377, PI21/00896 and FI19/00067]; Miguel Servet Program [Grant Number CP23/00017]; CIBER de Enfermedades Respiratorias (CIBERES); research support grants provided by the Spanish Society of Allergology and Clinical Immunology (SEAIC) Foundation [Grant Number 24B_02]; Conchita Rábago Foundation (FCR); FEDER funds (Fondo Europeo de Desarrollo Regional), and grant from Investigator Studies Program (MISP) Merck Sharp & Dohme Corp.

Ethics Statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Clinical Research Ethics Committee of the Severo Ochoa University Hospital (Leganés, Madrid) and La Paz University Hospital (Madrid) (PI‐6300).

Consent

Parents of all participating infants provided written informed consent prior to study enrollment.

Conflicts of Interest

The authors declare no conflicts of interest.

Cañas J. A., Serrano‐Santiago A., Alonso‐López P., et al., “Cytokine Profile Changes in Infants With Respiratory Syncytial Virus Bronchiolitis With Nirsevimab,” Pediatric Pulmonology 61 (2026): e71900, 10.1002/ppul.71900.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.