Duration and severity of asthma‐like episodes in young children vary by viral and bacterial triggers

Abstract

Background

Asthma‐like episodes with cough, wheeze and/or breathlessness are frequent in young children and often triggered by respiratory tract infections. We aimed to examine whether specific microbial triggers of asthma‐like episodes were associated with episode duration and severity.

Methods

We analyzed 453 nasopharyngeal aspirates collected during acute asthma‐like episodes from 268 children aged 0–3 years from the COPSAC‐2010 cohort. Samples were tested for common pathogenic respiratory viruses by PCR and for bacteria by culture. Asthma‐like episodes were prospectively recorded in diaries. Associations between microbial triggers and episode duration and severity were analyzed by generalized estimating equations. Interactions between microbial triggers and selected asthma risk loci and polygenic risk scores (PRSs) were explored in relation to episode duration.

Results

Rhinovirus was associated with 27% longer episode duration compared to episodes without rhinovirus, predominantly driven by rhinovirus C (50% longer). Rhinovirus was also associated with a more than twofold increased risk of a severe asthma exacerbation (i.e., need for oral prednisolone, high‐dose inhaled corticosteroids, and/or hospitalization) (OR 2.26; 95% CI 1.10–4.64; p = .027), with rhinovirus C exhibiting an even higher risk (OR 4.38; 95% CI 1.71–11.25; p = .002). Episodes with Streptococcus pneumoniae, Haemophilus influenzae, and/or Moraxella catarrhalis were associated with 61% longer duration than episodes without these bacteria, which was particularly pronounced for M. catarrhalis (47% longer). In contrast, RSV; parainfluenza and influenza viruses were associated with shorter duration. Younger age correlated with longer episode duration and risk of a severe asthma exacerbation. Exploratory analyses suggested pathogen‐specific interactions with selected asthma risk loci and asthma exacerbartion PRS in relation to episode duration.

Conclusion

Specific microbial triggers and younger age were associated with the duration and severity of asthma‐like episodes in young children. These findings may improve prognostic assessment and support more individualized management.

This study investigates how specific viral and bacterial triggers and age are associated with the duration and severity of asthma‐like episodes in children aged 0–3 years. Rhinovirus, particularly rhinovirus C, and the bacterial pathogens Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis were associated with longer episode duration, with the bacterial association primarily driven by M. catarrhalis. In contrast, respiratory syncytial virus, influenza viruses, and parainfluenza viruses were associated with shorter duration. Rhinovirus, particularly rhinovirus C, and younger age were also associated with an increased risk of severe asthma exacerbations.

Abbreviations

ANOVA

analysis of deviance

CI

confidence interval

COPSAC

Copenhagen Prospective Studies on Asthma in Childhood

CS

continuous shrinkage

FDR

false discovery rate

GEE

generalized estimating equation

GWAS

genome‐wide association study

ICS

inhaled corticosteroid

IL

Interleukin

IQR

interquartile range

LASSO

least absolute shrinkage and selection operator

N

number

OR

odds ratio

PCR

polymerase chain reaction

PRS

polygenic risk score

RCT

randomized controlled trial

RSV

respiratory syncytial virus

RTIs

respiratory tract infections

SNP

single nucleotide polymorphism

Key message.

Specific microbial triggers are associated with the clinical course of asthma‐like episodes in young children. Rhinovirus, particularly rhinovirus C, was linked to both longer episodes and increased risk of severe exacerbation, whereas common pathogenic airway bacteria, driven by Moraxella catarrhalis, were associated with longer duration only. Gene–pathogen interactions further suggested that host genetics may modify pathogen‐specific effects. Together, these findings support more precise prognostic assessment and a more personalized approach to early‐life wheezy illness.

1. INTRODUCTION

Asthma‐like episodes, including symptoms with cough, wheeze, and/or breathlessness, are very common in young children but often diminish with age. ¹ , ² , ³ However, these represent a significant risk factor for later asthma development. ³ Asthma‐like episodes are a frequent cause of hospitalization among young children. ⁴ It is well‐established that such episodes in preschool children often are triggered by respiratory tract infections (RTIs). The role of viruses in triggering acute wheezy/asthma‐like episodes is well‐documented, which is why these often are referred to as viral wheeze. ⁵ The contribution of bacteria has been more controversial; however, it has been demonstrated that bacterial infections are significantly associated with the risk of acute asthma exacerbations in young children similar to and independent of viral infections. ⁶ Respiratory pathogens are detected in up to 95% of asthma‐like episodes. ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Moraxella catarrhalis, Haemophilus influenzae, and Streptococcus pneumoniae are known to cause RTIs, ¹¹ and have been associated with the development and exacerbation of childhood asthma. ¹² , ¹³ , ¹⁴

There is however a notable gap in knowledge regarding whether specific respiratory viruses and bacteria affect the duration of these episodes differently since previous studies have yielded inconsistent results. ¹⁰ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Research from the Copenhagen Prospective Studies on Asthma in Childhood (COPSAC)‐2000 cohort, a high‐risk birth cohort of children born to mothers with asthma, found that the duration of wheezy episodes in young children was independent of identified pathogenic viruses and bacteria. ¹⁰ In another study, nasopharyngeal bacterial colonization during such episodes was correlated with longer hospitalizations and increased risk of relapse. ⁸ One study reported that bacterial co‐infection in children with viral infections was associated with prolonged hospitalization and more severe episodes, ¹⁹ whereas others have shown the opposite. ⁹ Furthermore, the child's genetic makeup and the burden of asthma‐like episodes in early childhood have been linked to the risk of childhood asthma. ³ , ²⁰ , ²¹ However, it remains unclear whether risk genes, along with host and environmental factors, influence the duration of asthma‐like episodes and interact with pathogens.

In this study, we investigated whether specific pathogenic respiratory viruses and bacteria, identified during asthma‐like episodes in children aged 0–3 years from the unselected and population‐based COPSAC‐2010 cohort, were associated with (1) the duration of asthma‐like episodes and (2) the risk of progression to severe asthma exacerbation. Furthermore, we aimed to examine if genetic markers of childhood asthma, and host and environmental factors influenced episode duration and explored potential interactions.

2. METHODS

2.1. Study population

The COPSAC‐2010 study is a Danish, prospective, population‐based clinical mother–child cohort study that enrolled 700 children born between 2008 and 2011. ²² These children attended scheduled clinical visits at the COPSAC research unit at the ages of 1 week, one, three, six, 12, 18, 24, 30, and 36 months. Additionally, children were seen for acute care visits and treatment of any asthma‐like symptoms, which were defined as cough, wheeze, and/or breathlessness severely affecting the child's well‐being.

During pregnancy, the mothers participated in two double‐blinded, randomized controlled trials (RCT) of fish oil and vitamin D in a 2 × 2 factorial design. ²³ , ²⁴

The study received approval from the Local Ethics Committee (H‐B‐2008‐093) and the Danish Data Protection Agency (2015‐41–3696). Written and oral informed consent was provided by both parents before enrolment.

2.2. Asthma‐like episodes

Asthma‐like symptoms, including wheeze, cough and/or breathlessness, (yes/no) were monitored longitudinally from birth to age 3 years through daily diary cards completed by the parents. COPSAC physicians reviewed these diaries with the parents at every clinical visit (both scheduled and acute) to validate the reported symptoms. ²² An asthma‐like episode was defined by symptoms for at least 3 days and should be separated by at least three symptom‐free days.

2.3. Severe asthma exacerbation

A severe asthma exacerbation was based on acute care visits or medical record checks and diagnosed if the child (1) needed oral prednisolone or high‐dose inhaled corticosteroid (ICS) for acute asthma‐like symptoms, and/or (2) was hospitalized due to such symptoms. Only one of the two criteria had to be fulfilled.

2.4. Objective assessment and procedures

Parents were encouraged to bring their children to the clinic for an acute care respiratory visit if they experienced asthma‐like symptoms for three consecutive days. During such a visit, each child underwent a thorough examination by a trained COPSAC physician following a standardized protocol. Aspirates were collected using a soft suction catheter inserted through the nose into the hypopharynx. ¹² These samples were cultured within 24 h using standard methods on both non‐selective and selective media, as previously detailed. ²⁵ The bacteria were characterized at the species level. This study focuses on the primary airway bacterial pathogens M. catarrhalis, H. influenzae, and S. pneumoniae. Additional airway bacteria identified are detailed in Appendix S1 (Table S1).

Nasopharyngeal samples for viral identification were similarly collected using a soft suction catheter, while a few samples (2.9% ~ 13/450) were obtained using nasal swabs. These samples were analyzed for a range of viruses by an in‐house polymerase chain reaction (PCR), ²⁶ including rhinoviruses (A, B, C, and non‐typeable), enteroviruses, respiratory syncytial virus (RSV), parainfluenza viruses (1, 2, and 3), coronaviruses (229, OC43, and NL163), human metapneumovirus, adenovirus, influenza viruses (A – including subtypes H1N1 and H3N2, B, and C), bocavirus, WU virus, KI virus, and human parechovirus. ²⁶ , ²⁷

Aspirates were excluded from the analyses if the child was clinically diagnosed with croup or pneumonia, or if oral antibiotics had been prescribed ≤7 days before the sample collection date. Children aged 1–3 years with recurrent asthma‐like symptoms (See Appendix S1) were invited to participate in a RCT investigating the effect of azithromycin on episode duration (NCT01233297). ²⁸ Samples from children who received azithromycin in the trial were excluded from analyses.

2.5. Genetic, host and environmental factors

Genetics, baseline characteristics, and environmental factors were included in the analyses as potential predictors of episode duration and therefore also potential confounders of the possible association of the microbial trigger on episode duration.

Included genetic markers were known susceptibility loci (single nucleotide polymorphisms (SNPs) in/near genes) associated with severe childhood asthma exacerbations leading to hospitalization during the first 6 years of life. ²⁰ , ²⁹ Additionally, two polygenic risk scores (PRSs) were included. These were an asthma exacerbation PRS ³ , ²⁹ and a low lung function PRS ³⁰ (See Appendix S1 for details).

Host factors included were sex and age of the child during the asthma‐like episode. Additionally, allergic sensitization, examined at ages six and 18 months using both skin prick test and measurement of serum specific immunoglobulin E (sIgE), any diagnosis of asthma/persistent wheeze from birth to age 3 years, maternal and paternal asthma were evaluated as possible predictors of episode duration (See Appendix S1 for details on these factors). Environmental factors included exposure to second‐hand smoking from age 0–3 years and social circumstances (See Appendix S1 for details on these factors).

We also investigated whether interventions given during pregnancy with fish oil and high‐dose vitamin D influenced episode duration.

2.6. Statistics

The primary analysis of this study was the association between viral and bacterial triggers and the primary outcome; duration of asthma‐like episodes in the first 3 years of life. To account for repeated measurements in the same child, we used generalized estimating equation (GEE) models using a Poisson distribution to analyze episode duration. Effect estimates from these models were expressed as percentage increase or decrease in episode durations, calculated as (e^β─1) × 100. Each viral and bacterial trigger was treated as a dichotomous variable (present/not present), independent of the co‐presence of other agents. The comparison groups were episodes without presence of the specific trigger investigated. Further, the age‐related impact of the specific pathogens was explored through interaction models by adding cross‐products of age (in years as a numerical variable) with the specific pathogens. GEE models were also applied to analyze the relation between (1) host and environmental factors and (2) genetic markers and episode duration. We adjusted the main analyses for host or environmental factors that were significantly associated with episode duration. Only the age of the child at the time of sampling was found to be significantly associated with episode duration. We also examined interactions between child age and maternal asthma in relation to episode duration.

As a sensitivity analysis to assess the robustness of microbial trigger selection, we applied least absolute shrinkage and selection operator (LASSO) regression restricted to microbial triggers only, followed by age‐adjusted GEE models accounting for repeated samples within children (See Appendix S1 for details).

The main microbial trigger analyses were adjusted for age at sampling, which was the only host or environmental factor significantly associated with episode duration. For the analyses of progression to severe asthma‐like episodes, the associations between microbial triggers of episode duration and the risk of a severe asthma exacerbation were analyzed using GEE models to compute the odds ratios (ORs) with a 95% CI, which account for repeated measurements in the same child. These analyses were likewise adjusted for age at sampling. Asthma/persistent wheeze was not included as an adjustment variable because the diagnosis and the occurrence of severe asthma‐like episodes partly reflect the same early‐life disease course, which could lead to overadjustment. Binomial distribution with a logit link function (family = binomial) was utilized. Significant host factors were assessed in similar analyses.

Effect modification by co‐infection was assessed by interaction tests between selected microbial triggers associated with episode duration. These analyses were performed only when at least ten episodes were present in each exposure group: positive for both pathogens, positive for either pathogen alone, and negative for both pathogens. Gene–pathogen interactions were assessed between genetic markers and microbial triggers associated with episode duration. p‐values from both co‐infection and gene–pathogen interaction analyses were false discovery rate (FDR) adjusted.

We also examined interactions between significant specific microbial triggers of episode duration and asthma diagnosis at any timepoint from age 0–3 years and performed stratified analyses based on asthma diagnosis and maternal asthma.

A two‐sided significance level of 0.05 was used in all analyses. Statistical analyses were performed using R, version 4.4.2 (R Core Team, Vienna, Austria).

3. RESULTS

3.1. Study population and sample inclusion

Nasopharyngeal samples were obtained during 774 asthma‐like episodes from 350 children in the first 3 years of life. After excluding samples based on prespecified criteria, 453 episodes were included in the main analyses from 268 children (Figure 1). Of these children, 159 contributed one sample, 109 contributed 2–5 samples (Figure S1). Children excluded due to missing diary data were largely comparable to included children, except for a lower prevalence of asthma/persistent wheeze during age 0–3 years (Table S2).

FIGURE 1.

Study sample selection of nasopharyngeal aspirates from children with asthma‐like episodes analyzed for viruses and bacteria. *Some of the samples have overlapping exclusion criteria, e.g., diagnosis of pneumonia and antibiotics prescribed at sample date.

3.2. Airway pathogen prevalence

A virus was detected in 74.2% of the samples (Table S3). The most prevalent viruses detected were rhinoviruses (36.7%), enteroviruses (19.8%), and RSV (15.6%). A bacterium (M. catarrhalis, H. influenzae, or/and S. pneumoniae) was cultured positive in 79.0% of the samples (Table S2). The most prevalent bacteria were M. catarrhalis (63.8%), H. influenzae (28.2%), and S. pneumoniae (28.0%). Co‐occurrence of virus and bacterium (of the above) was observed in 60.1% of the samples. However, the presence of viruses and bacteria was not found to be significantly associated, as indicated by Fisher's Exact Test (p = .20). Exclusive viral or bacterial (of the above) detection occurred in 14.7% and 19.0% of the samples, respectively. No pathogens were detected in 6.4% of the samples (Table S3). In a heatmap, we visualized correlations of specific pathogens (Figure S2).

3.3. Age and seasonal variations

Significant age‐related differences in pathogen prevalence were observed. Influenza viruses were more prevalent in the third year of life, whereas adenovirus and bocavirus were more prevalent in the second year. In contrast, pathogenic bacteria (S. pneumoniae, H. influenzae, and/or M. catarrhalis), particularly H. influenzae, were less prevalent in the third year (p‐interaction <.05; Figure 2A).

FIGURE 2.

Microbial prevalence in nasopharyngeal samples during asthma‐like episodes, stratified by age (A), and season (B). The y‐axis shows the proportion of the samples positive for the pathogen(s). p values in right corner represent the statistical significance of the associations between the age groups (A) or seasons (B), determined through logistic regression analyses followed by an analysis of deviance (ANOVA). A p value <.05 suggests that the prevalence is likely influenced by the age of the child in Figure 2A an by season of sample day in Figure 2B. *S. pneumoniae, H. influenzae and/or M. catarrhalis. **See Table S1 for list of all bacteria.

Clear seasonal patterns were also identified. Overall viral prevalence was lower in summer. RSV, coronaviruses, and influenza viruses peak in winter; parainfluenza viruses and H. influenzae peak in spring; enteroviruses and bocavirus peak in autumn (p‐interaction <.05; Figure 2B). Season was not associated with episode duration (p = .078).

3.4. Host and environmental factors, asthma risk genes vs. episode duration and severity

The median duration of the asthma‐like episodes was 12 days with active symptoms (interquartile range (IQR) 7–21). Additionally, 6.8% of the asthma‐like episodes (N = 31) progressed to a severe episode, requiring treatments such as oral prednisolone, high‐dose inhaled corticosteroids ICS, or hospitalization. Episodes occurring earlier in life were associated with longer episode duration (18% increase per 1‐year decrease in age). Asthma/persistent wheeze showed a trend toward longer episode duration, whereas other environmental and host factors, as well as the prenatal fish oil and high‐dose vitamin D interventions, were not significantly associated with episode duration (Table 1). There was no significant interaction between maternal asthma and age in relation to episode duration (p = 0. 59). Further, asthma‐like episodes experienced earlier in life were at greater risk of progressing into a severe episode (OR 1.71, 95% CI 1.09–2.69, p = .020). Further, children diagnosed asthma/persistent wheeze were at greater risk of progressing into a severe asthma‐like episode (OR 6.72, 95% CI 2.30–19.67, p < .001). Asthma risk SNPs and PRSs were not associated with longer episode duration, although CDHR3 showed a trend toward longer episodes (18% per risk allele; p = .056; Table S4). Further, interleukin (IL)‐13 were link to shorter episodes (−15% per risk allele; p = .043). Neither SNPs nor PRSs were associated with progression to severe episodes (data not shown).

TABLE 1.

Environmental and host factors in the included population and their associations to duration of asthma‐like episode during the first 3 years of life.

Measure	Distribution among included episodes, mean (SD) or n (%)	Episode duration, median days (IQR)	Episodes (N)	Children (N)	Percent change in episode duration (95% CI)	p Value
Age at episode (per decrease in year)	1.17 (0.68)	12 (7–21)	453	268	18% (2 to 37)	.031
Sex		12 (7–21)	453	268	‐	‐
Male	275 (60.5%)	12 (7–21.75)	274	155	20% (−16 to 24)	.831
Female	179 (39.5%)	11 (6.5–21)	179	113	Reference	‐
Social circumstances (per SD increase)	−0.12 (0.96)	12 (7–21)	453	268	2% (−7 to 13)	.644
Paternal asthma		11 (7–21)	445	268	‐	‐
Yes	80 (18.0%)	11.5 (7–19.25)	80	48	−1% (−21 to 24)	.903
No	365 (82.0%)	11 (7–21)	365	214	Reference	‐
Maternal asthma		12 (7–21)	453	268	‐	‐
Yes	98 (21.6%)	12.5 (7–24.75)	98	52	14% (−10 to 44)	.287
No	355 (78.4%)	11 (7–20.5)	355	216	Reference	‐
Exposure to second‐hand smoking 0–3 years (per doubling in months)	7.26 (11.6)	12 (7–21)	453	268	9% (−7 to 27)	.275
Allergic sensitization		11 (7–20.25)	404	268	‐	‐
Yes	44 (10.9%)	12 (8–20)	44	29	−2% (−24 to 27)	.904
No	360 (89.1%)	11 (7–21)	360	209	Reference	‐
Asthma/persistent wheeze during age 0–3 years		12 (7–21)	447	268	‐	‐
Yes	231 (51.7%)	12 (7–24)	231	104	16% (−3 to 40)	.110
No	216 (48.3%)	11 (7–18)	216	159	Reference	‐
Fish oil intervention		12 (7–21)	452	268	‐	‐
Yes	209 (46.2%)	11 (7–20)	209	126	−4% (−20 to 15)	.674
No	243 (53.8%)	12 (7–23)	243	141	Reference	‐
Vitamin D intervention		11 (7–21)	364	268	‐	‐
Yes	154 (42.3%)	11 (6.25–20)	154	97	−4% (−22 to 17)	.681
No	210 (57.5%)	12 (7–21)	210	124	Reference	‐

Note: Values with p < 0.05 are shown in bold.

Abbreviations: CI, confidence interval; IQR, interquartile range; N, number.

3.5. Microbial triggers vs. episode duration and severity

Age‐adjusted analyses showed that several microbial triggers were significantly associated with episode duration in children aged 0–3 years. Rhinoviruses were associated with 27% longer episode duration, primarily driven by rhinovirus C (50% longer duration). The presence of any pathogenic bacteria (S. pneumoniae, H. influenzae, and/or M. catarrhalis) was associated with 61% longer duration than episodes without these bacteria, and M. catarrhalis alone with 47% longer duration (Table 2). In age‐adjusted analyses, rhinoviruses, particularly type C, were associated with increased risk of progressing to a severe asthma‐like episode (rhinoviruses: OR 2.26, 95% CI 1.10–4.64, p = .027 and rhinovirus C: OR 4.38, 95% CI 1.71–11.25, p = .002 – Table 3). In contrast, RSV; parainfluenza and influenza viruses; and viral infection without bacterial co‐detection were associated with shorter episode duration (−25% to −36%) (Table 2, Figure 3). No significant interactions were found between age of the child and microbial triggers in the duration analyses, except for M. catarrhalis, which showed a significant interaction with age (p‐interaction = .041; Table S5). Further, similar results were found when analyzing samples with presence of only viruses and only bacteria (Table S6). However, results were limited by numbers. When adjusting for co‐occurrence of other microbial triggers similar results were found as well (Table S7). When adjusting for sample day, defined as the timing of sample collection relative to episode start, the main associations were largely unchanged. However, the association between RSV and shorter episode duration was slightly attenuated and no longer statistically significant, suggesting that this signal may partly reflect differences in sampling time during the episode (Table S8). In a LASSO sensitivity analysis restricted to microbial triggers, rhinovirus and M. catarrhalis were selected as the strongest predictors of episode duration. In the subsequent age‐adjusted GEE model, rhinovirus was associated with 24% longer episode duration (95% CI, 3% to 50%; p = .026), while M. catarrhalis was associated with 45% longer duration (95% CI, 21% to 74%; p < .001).

TABLE 2.

Association between viral and bacterial triggers and the duration of asthma‐like episodes analyzed by age‐adjusted generalized estimating equation models.

Measure	Negative			Positive			Age‐adjusted GEE model
	Median days (IQR)	Episodes (N)	Children (N)	Median days (IQR)	Episodes (N)	Children (N)	Percent increase in episode duration (95% CI)	p Value
Virus (Number of episodes)							N = 450
Rhinoviruses (any)	10 (7–18)	285	204	13 (7–24)	165	125	27% (1.05–1.53)	.012
Rhinovirus A a	11 (7–20)	336	227	13 (7–24)	77	68	14% (0.9–1.44)	.293
Rhinovirus B a	11 (7–21)	402	250	15 (4.5–38)	11	11	38% (0.77–2.46)	.284
Rhinovirus C b	10.5 (7–19)	296	209	13 (8.75–28.25)	40	37	50% (1.03–2.18)	.034
Rhinovirus non‐typeable	12 (7–21)	413	255	13 (7–21)	37	35	−12% (0.68–1.13)	.306
RSV	12 (7–23)	380	230	11 (7.25–16)	70	69	−25% (0.62–0.91)	.003
Coronaviruses (any)	12 (7–21)	428	260	11.5 (8–20.25)	22	20	7% (0.69–1.65)	.765
Parainfluenza viruses (any)	12 (8–22)	406	248	7 (6–12)	44	43	−33% (0.49–0.91)	.011
Influenza viruses (any)	12 (7–21)	437	265	11 (8–15)	13	13	−28% (0.54–0.96)	.026
Human Metapneumovirus	12 (7–21)	427	253	12 (8.5–13.5)	23	21	−13% (0.63–1.21)	.412
Adenovirus	11 (7–21)	430	259	17.5 (11.25–23.25)	20	19	5% (0.84–1.3)	.695
Enteroviruses (any)	11 (7–21)	361	233	13 (7–23)	89	81	6% (0.84–1.33)	.629
Bocavirus	12 (7–21)	434	259	10 (8–14)	16	16	−25% (0.55–1.02)	.063
WU virus	11 (7–21)	439	262	17 (12.5–29)	11	11	34% (0.88–2.03)	.172
Any virus	12 (7–24)	116	96	11.5 (7–21)	334	221	−6% (0.77–1.14)	.522
Multiple viruses	11 (7–20)	327	215	13 (7–23)	123	104	1% (0.84–1.21)	.907
Only virus (Without S. pneumoniae, H. influenzae and/or M. catarrhalis)	12 (7–23)	373	242	9 (7–14)	63	52	−36% (−47 to −22)	<.001
Only virus c	12 (7–21)	418	256	12.5 (7.25–21)	18	17	−11% (−33 to 18)	.401
Bacteria (Number of episodes)							N = 439
S. pneumoniae	12 (7–21)	316	216	12 (7–23)	123	107	9% (−10 to 31)	.397
H. influenzae	12 (7–21)	315	215	11 (6–21.25)	124	103	−12% (−27 to 6)	.175
M. catarrhalis	10 (6–15)	159	117	13 (7.75–24)	280	200	47% (23 to 77)	<.001
Any bacteria (including S. pneumoniae, H. influenzae and/or M. catarrhalis)	9 (6–14.25)	92	74	13 (8–23.5)	347	230	61% (34 to 94)	<.001
Multiple bacteria (including S. pneumoniae, H. influenzae and/or M. catarrhalis)	12 (7–20.5)	283	199	12 (6–24)	156	125	6% (−12 to 28)	.544
Any bacteria c	12 (7–20)	23	22	12 (7–21)	416	256	16% (−13 to 55)	.310
Multiple bacteria c	12 (7–18.5)	103	83	12 (7–22)	336	224	6% (−16 to 34)	.639
Only bacteria (including S. pneumoniae, H. influenzae, and/or M. catarrhalis)	11 (7–20)	353	232	15 (8–26)	83	72	22% (−2 to 51)	.070
Only bacteria c	11.5 (7–21)	330	220	12.5 (8–24)	106	91	8% (−12 to 33)	.442
Co‐occurrence and no pathogens (Number of episodes)							N = 436
Co‐occurrence (Including virus and S. pneumoniae, H. influenzae, and/or M. catarrhalis)	11 (7–20)	174	133	12 (7–21.75)	262	193	16% (0.96–1.4)	.135
Co‐occurrence c	12 (7–23)	129	109	11 (7–21)	307	212	−4% (0.79–1.17)	.680
No pathogens (Without virus and S. pneumoniae, H. influenzae, and/or M. catarrhalis)	12 (7–22)	408	248	9 (6–15)	28	26	−32% (−52 to −6)	.020
No pathogens c	12 (7–21)	431	258	6 (6–12)	5	5	−23% (−67 to 80)	.550

Note: Values with p < 0.05 are shown in bold, while values with p < 0.10 are shown in italics.

Abbreviations: CI, confidence interval; GEE, generalized estimating equation; IQR, interquartile range; N, number.

^a Number of episodes included was 413.

^b Number of episodes was 336.

^c See Table S1 for list of all bacteria.

TABLE 3.

Association between viral and bacterial triggers (p < .05, Table S4) and risk of progressing to severe asthma‐like episodes (i.e., received oral prednisolone or high‐dose inhaled corticosteroid) analyzed using generalized estimating equation models, including age‐adjusted analyses.

Measure	Severe asthma‐like episode	Risk of progressing to a severe episode		Risk of progressing to a severe episode; age‐adjusted
	N (%) a	OR (95% CI)	p Value	OR (95% CI)	p Value
Virus (Number of episodes)		N = 450		N = 450
Rhinoviruses (any)	17 (3.8%)	2.40 (1.18–4.91)	.016	2.26 (1.10–4.64)	.027
Rhinovirus C b	7 (2.1%)	4.62 (1.82–11.75)	<.001	4.38 (1.71–11.25)	.002
RSV	5 (1.1%)	1.09 (0.41–2.91)	.860	1.10 (0.42–2.93)	.843
Parainfluenza viruses (any)	2 (0.4%)	0.64 (0.15–2.72)	.548	0.66 (0.16–2.82)	.577
Influenza viruses (any)	0 (0%)	0 (0–0)	<.001	0 (0–0)	<.001
Bocavirus	1 (02%)	0.93 (0.12–7.32)	.946	0.43 (0.12–7.49)	.965
Bacteria (Number of episodes)		N = 439		N = 439
M. catarrhalis	22 (5.0%)	1.42 (0.67–3.03)	.362	1.39 (0.66–2.96)	.390
Any bacteria (S. pneumoniae, H. influenzae, and/or M. catarrhalis)	27 (6.2%)	1.86 (0.64–5.41)	.257	1.74 (0.57–5.26)	.329
Co‐occurrence and no pathogens (Number of episodes)		N = 436		N = 436
No pathogens (Without virus and S. pneumoniae, H. influenzae, and/or M. catarrhalis)	0 (0%)	0 (0–0)	<.001	0 (0–0)	<.001

Note: Values with p < 0.05 are shown in bold.

Abbreviations: CI, confidence interval; N, number; OR, odds ratio.

^a Numbers of episodes related to the microbial trigger that was defined as a severe asthma‐like episode.

^b Number of episodes in these models was 336.

FIGURE 3.

The duration of an asthma‐like episode in days by microbial trigger during asthma‐like episodes. Percentages in parentheses indicate the proportion of nasopharyngeal samples positive for the specific pathogen assessed during the asthma‐like episode. (A) Viruses. (B) Bacteria. (C) Only viruses, only bacteria, no pathogens, and co‐occurrence of virus and pathogenic bacteria. *With S. pneumoniae, H. influenzae, and/or M. catarrhalis. **Virus(es) and S. pneumoniae, H. influenzae, and/or M. catarrhalis. ***Without S. pneumoniae, H. influenzae, and/or M. catarrhalis. ****Without virus(es), S. pneumoniae, H. influenzae, and/or M. catarrhalis.

When evaluating other cultivated bacteria (Table S1) with a presence >2% (n = 12), Moraxella nonliquefaciens was significantly associated with 44% (95% CI −49% to −14%, p = .002) shorter episode duration.

Episodes triggered by only viruses were linked with quicker resolution of symptoms. Though, co‐detection of viruses and bacteria was not significantly associated with longer episode duration, nor was detection of multiple viruses or bacteria. More, episodes without a detected microbial trigger were associated with a 33% shorter duration (Table 2, Table S5).

3.6. Pathogen‐pathogen and pathogen‐gene interactions

Initial analysis revealed nominally significant interactions between pathogens in relation to episode duration, specifically between S. pneumoniae and M. catarrhalis, H. influenzae and M. catarrhalis, and H. influenzae and RSV (Table S9, Figure S3). These interactions were no longer significant after FDR adjustment (Table S9).

Significant interactions between microbial triggers and genetic markers were found between parainfluenza viruses and CDHR3, parainfluenza viruses and asthma exacerbation PRS, bocavirus and TSLP (FDR‐adjusted p‐interaction <.05) (Table S10, Figure S4). Exploratively, we removed CDHR3 from the asthma exacerbation PRS to assess whether the observed interaction was driven by this risk SNP, which resulted in a non‐significant interaction (p = .068).

3.7. Asthma‐pathogen interactions

No significant interactions between pathogens and asthma were found. However, stratified analyses by asthma diagnosis indicated that the associations of viruses with episode duration were more pronounced in children with asthma, whereas the associations of the bacteria were similar in children with and without asthma (Table S11).

In line, maternal asthma did not interact with pathogens present in the airway (Table S12).

4. DISCUSSION

4.1. Primary findings

Viruses and/or bacteria were detected in nearly all samples collected during asthma‐like episodes age 0–3 years, and episodes triggered by rhinoviruses, driven by rhinovirus C were associated with longer duration and increased severity. More, pathogenic bacteria (S. pneumoniae, H. influenzae, and/or M. catarrhalis), particularly M. catarrhalis, were associated with longer episode duration. Conversely, RSV, parainfluenza and influenza viruses were associated with shorter duration. Interestingly, episodes triggered by only virus and without a detectable pathogen were linked to shorter symptom duration.

Additionally, we identified significant gene–pathogen interactions between parainfluenza viruses with both CDHR3 and the asthma exacerbation PRS, as well as bocavirus with TSLP, indicating that these genetic markers modified the associations between specific pathogens and episode duration. Further, IL13 was associated with shorter episode duration. More, younger age was associated with longer episode duration and increased risk of progressing into an asthma exacerbation. Our results suggest that duration of asthma‐like episodes and severity depends on the microbial pathogen and the child's age.

4.2. Strengths and limitations

The major strength of this study is the unique, longitudinal daily diary recordings of asthma‐like episodes from birth up to the age of 3 years in combination with nasopharyngeal sampling for airway pathogens, which enabled us to quantify the effect of microbial triggers on the duration of these episodes. This approach minimizes recall bias and allows for the investigation of episode duration. Additionally, having information on both viruses and bacteria present in the upper airways and detailed information on genetic markers of childhood asthma risk enabled us to explore pathogen‐pathogen and pathogen‐gene interactions. It is also advantageous that the children were examined thoroughly by COPSAC physicians at the COPSAC clinic using standardized operating procedures during these acute asthma‐like episodes, distinguishing episodes from conditions like pneumonia and croup. Moreover, we have detailed information on both host and environmental factors.

Our analyses were performed within cases, and as such we did not have a healthy control group, nor would our primary outcome of episode duration be defined in such a control group. In addition, each microbial trigger was compared with episodes without that specific trigger, rather than with pathogen‐free episodes or episodes with a specific alternative pathogen. Therefore, pathogen‐specific estimates may partly reflect differences in the composition of the reference groups. Because pathogen‐free episodes were rare, such comparisons would be underpowered and potentially unstable. Further, co‐infection was frequent, which potentially was amplified by upper airway colonization, was ameliorated by our adjusted and interaction analyses but should be kept in mind when interpreting the results.

We did not obtain samples from every asthma‐like episode in the cohort. Milder episodes may be less likely to prompt clinic visits, potentially biasing our data toward more severe cases. While we have excluded children clinically diagnosed with pneumonia, we cannot conclude whether the bacterial colonization is causing the asthma‐like symptoms. Exploratory analyses of interactions between pathogens were FDR‐adjusted to limit false positives, but some comparisons were limited by small sample sizes, increasing the risk of false negatives. Additionally, our characterization of bacteria relies on traditional culturing methods, limiting detection to only select bacteria without quantifying their abundances. Moreover, we assume that nasopharyngeal bacterial profiles reflect those of the lower airways, an assumption that may not hold true. Finally, this study included multiple analyses across microbial triggers, host and environmental factors, and interaction terms, increasing the risk of chance findings. Although FDR correction was applied to the interaction analyses, the overall approach should be considered exploratory, and secondary findings should be interpreted cautiously. However, the main associations with rhinovirus and M. catarrhalis were supported by sensitivity analyses, including LASSO‐based microbial trigger selection.

4.3. Interpretation

The duration of asthma‐like episodes in young children was shown to be dependent on the presence of certain microbial pathogens. We found that rhinovirus was associated with longer episode durations and a higher risk of progressing to a severe episode compared to episodes without rhinovirus, particularly driven by type C. Rhinovirus‐induced wheezy episodes and bronchiolitis, especially from type C, have been linked with more severe symptoms compared to other viruses and asthma risk. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ We recently reported from the same cohort that rhinovirus and enterovirus illnesses in early life increased the risk of pneumonia until age 13 years. ³¹

Genetic markers of childhood asthma were generally not major determinants of episode duration; however, IL13 was associated with shorter episodes, while CDHR3 showed a tendency toward longer episodes. The observed association between IL13 and shorter episode duration was unexpected, given the established role of IL13 in type 2 airway inflammation and asthma susceptibility, and should therefore be interpreted cautiously. ³² CDHR3 encodes for the transmembrane protein cadherin‐related protein‐3, which is highly expressed in the airway epithelium and functions as a receptor for rhinovirus C, thereby increasing susceptibility to rhinovirus C infections. ³³ Children with CDHR3 risk allele(s) are prone to more frequent rhinovirus C‐induced wheezy episodes ²⁰ and therefore also prolonged episodes. Moreover, the presence of the CDHR3 risk allele(s) interacted with parainfluenza viruses in the duration of episodes, suggesting that the presence of CDHR3 alleles not only affects rhinovirus C susceptibility. A similar pattern was observed with the asthma exacerbation PRS, but this was mainly driven by a high influence of CDHR3 in the PRS.

RSV detection, which peaks during winter, was associated with shorter asthma‐like episode duration compared to those episodes without RSV. This contrasts with previous studies in which RSV‐induced bronchiolitis was associated with longer hospital stays compared to rhinovirus. ¹⁵ , ³⁴ However, studies focusing exclusively on bronchiolitis in children <2 years are not directly comparable to our findings. In the high‐risk COPSAC‐2000 cohort study, RSV was not associated with duration of asthma‐like episodes. ¹⁰ Notably, the RSV association was slightly attenuated and no longer significant after adjustment for sample day, which could be explained by more severe symptoms at episode onset prompting parents to seek medical attention earlier. Bocavirus‐positive episodes showed a tendency toward shorter duration, but this did not persist after age adjustment. Previous literature on bocavirus is limited, although it is known to trigger both bronchiolitis and wheezing in children. ³⁵ In addition, asthma‐like episodes triggered by influenza and parainfluenza viruses were associated with faster resolution. While these viruses are established triggers of wheeze and asthma exacerbations, ³⁶ less is known about their relationship with episode severity. None of the other viruses, nor presence of multiple viruses, were found to influence episode duration.

Importantly, shorter episode duration should not be interpreted as lower clinical severity, as duration and severity reflect different aspects of illness. Recent data from hospitalized children have shown that RSV and influenza infections were associated with more severe acute disease, including increased risk of intensive care unit admission. ³⁷ Thus, our findings should be interpreted in the context of prospectively recorded asthma‐like symptom duration in a cohort setting, and the influenza association should be interpreted cautiously given the limited number of influenza‐positive episodes.

Interestingly, our heatmap visualizing the correlations among specific pathogens highlights a general trend where viruses, except for enteroviruses, exhibit negative associations with other viral pathogens. This could suggest interference or competitive exclusion among viruses. However, the exception of enteroviruses might be explained by cross‐reactivity in diagnostic tests with rhinoviruses, leading to positive results for both. In contrast, positive associations were observed among bacteria, which could indicate a symbiotic or commensal relationship. Alternatively, this might reflect an underlying host susceptibility to bacterial infection or colonization or even shared environmental risk factors that favor multiple bacterial pathogens.

Colonization with S. pneumoniae, H. influenzae, and/or M. catarrhalis was found to prolong episodes of asthma‐like symptoms, with M. catarrhalis identified as the primary contributor of this association. More, we found that episodes triggered by only virus (without detection of these bacteria) were significantly shorter. This observation is intriguing given that azithromycin has been shown to shorten the duration of asthma‐like episodes, though presence of pathogenic bacteria did not modify the treatment effect. ²⁸ In line, a study has shown that nasopharyngeal bacterial colonization during wheezy episodes in young children was correlated with longer hospital stays and higher risk of relapse of wheezy episodes. ⁸ However, another study demonstrated that preschool children with asthma had shorter duration of wheezing in acute exacerbations when colonized with S. pneumoniae. ⁹ It has previously been shown that asymptomatic neonates colonized with these bacteria are at increased risk of developing childhood asthma up to age seven. ¹² , ¹³ Additionally, colonization with these pathogens has been linked to both airway inflammation, ¹³ airway inflammatory immune response, ³⁸ and systemic low‐grade inflammation, ³⁹ which could contribute to more severe asthma‐like episodes. More, the association could also be partly driven by concomitant protracted bacterial bronchitis. ⁴⁰ In contrast to the present findings, the previous COPSAC‐2000 study found that duration of wheezy episodes was independent of the specific microbial trigger. ¹⁰ This discrepancy may partly reflect differences in cohort composition and microbiological resolution. COPSAC‐2000 is a high‐risk cohort of children born to mothers with asthma, whereas COPSAC‐2010 is population‐based. Host susceptibility and asthma‐related factors may therefore have had a stronger influence on episode duration in COPSAC‐2000, potentially attenuating pathogen‐specific differences. However, stratified analyses by maternal asthma did in the present study not show significant interactions, suggesting that maternal asthma alone is unlikely to explain the observed pathogen‐specific associations. Furthermore, the pathogen classification differed between the studies. COPSAC‐2000 analyzed viruses in broader groups, including picornaviruses, RSV, coronaviruses, and “other viruses”, whereas the present study assessed individual viral triggers in greater detail, including rhinovirus species. This higher pathogen resolution may have allowed detection of pathogen‐specific associations that were not identifiable in the previous study.

It has been suggested that there may be synergy between respiratory viral and bacterial pathogens when these co‐occur. A study reported that children <3 years with recurrent wheezing, who were hospitalized due to virus‐induced wheezing, experienced prolonged hospitalizations when bacterial colonization was also present. ¹⁹ Conversely, our study did not show any statistically significant interactions between the different pathogens after FDR‐adjustment, but our episodes were not restrained to children with recurrent wheeze and hospitalization, which reflects more severe cases.

Our study revealed that younger age led to longer and more severe asthma‐like episodes. In contrast, episode duration was not linked to younger age in the high‐risk COPSAC‐2000 cohort, ¹⁰ indicating that the influence of age is more pronounced in the general population. Though, no significant interaction was observed between age and maternal asthma or asthma.

Asthma has previously been associated with longer episodes in the COPSAC‐2000 cohort, ¹⁰ but in the present study this association was only borderline, though we saw an increased risk of progressing into a severe exacerbation. A potential mechanism involved could be that children with asthma may mount a different and perhaps less efficient immune response when exposed to pathogens. ²⁷ Except for a nominal association between IL13 and shorter episode duration, no other host or environmental factors or genetic markers of childhood asthma were clearly associated with episode duration, in line with earlier results. ³ , ¹⁰ This suggests that classic asthma‐associated factors, such as parental asthma, allergic sensitization, and genetic markers for childhood asthma, which are associated with the overall burden of asthma‐like symptoms, ³ may not be the main determinants of episode duration in preschool children. Thus, early‐life asthma diagnosis may have a unique impact on symptom persistence. Lastly, no significant interaction was observed between asthma diagnosis and specific respiratory pathogens in relation to longer episodes, which suggests that children with asthma may generally have a different response to respiratory pathogens, rather than exhibiting a uniquely severe reaction to specific ones.

In conclusion, our findings highlight that specific bacteria and viruses detected during asthma‐like episodes are associated with duration and severity of asthma‐like episodes in preschool children, which may aid prognostic assessment and support the need for rapid diagnostic testing.

These findings suggest that specifically targeting rhinovirus C and M. catarrhalis could lead to better management and possibly prevention of longer and/or more severe wheezing episodes in young children.

GOVERNANCE

We are aware of and comply with recognized codes of good research practice, including the Danish Code of Conduct for Research Integrity. We comply with national and international rules on the safety and rights of patients and healthy subjects, including Good Clinical Practice (GCP) as defined in the EU's Directive on Good Clinical Practice, the International Conference on Harmonization's (ICH) good clinical practice guidelines, and the Helsinki Declaration. Privacy is important to us, which is why we follow national and international legislation on General Data Protection Regulation (GDPR), the Danish Act on Processing of Personal Data, and the practice of the Danish Data Inspectorate.

AUTHOR CONTRIBUTIONS

Signe Kjeldgaard Jensen: Writing – review and editing. Anton Kjellberg: Writing – review and editing. Julie Nyholm Kyvsgaard: Writing – original draft; conceptualization; methodology; visualization; formal analysis; investigation. Jonathan Thorsen: Writing – review and editing; supervision; data curation. Thea K. Fischer: Writing – review and editing. Nicklas Brustad: Writing – review and editing. Tamo Sultan: Writing – review and editing. Bo Lund Chawes: Writing – review and editing; supervision. Ann‐Marie Malby Schoos: Writing – review and editing. Casper‐Emil Tingskov Pedersen: Writing – review and editing. Nilo Vahman: Writing – review and editing. James E. Gern: Writing – review and editing. Karen A. Krogfelt: Writing – review and editing. Jakob Stokholm: Writing – review and editing; conceptualization; supervision. Klaus Bønnelykke: Writing – review and editing; supervision.

FUNDING INFORMATION

The funder of this study is the VP‐legat | Fabrikant Vilhelm Pedersen og hustrus legat. The funder had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. All authors had access to the raw data and had final responsibility for the decision to submit for publication.

CONFLICT OF INTEREST STATEMENT

JT has received a speaking fee from AstraZeneca. All authors declare no potential, perceived, or real conflict of interest regarding the content of this manuscript. The funding agencies did not have any role in the design and conduct of the study; collection, management, and interpretation of the data; or preparation, review, or approval of the manuscript. No pharmaceutical company was involved in the study.

Supporting information

Appendix S1.

Kyvsgaard JN, Thorsen J, Jensen SK, et al. Duration and severity of asthma‐like episodes in young children vary by viral and bacterial triggers. Pediatr Allergy Immunol. 2026;37:e70499. doi: 10.1111/pai.70499

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.