Smoke Exposure, Cytokine Levels, and Asthma Visits in Children Hospitalized for Bronchiolitis

We enrolled children aged 31 days to 2 years who were hospitalized at a free-standing children’s hospital with a diagnosis of bronchiolitis between January 2013 and April 2014. After giving consent, parents completed a survey about their children’s health, family demographics, and children’s exposure to tobacco smoke; marijuana exposure was added after it was legalized in the state. Nasal wash samples were collected from each child. Six months after discharge, parents completed follow-up surveys, and chart reviews were completed by a trained abstractor; a report of a follow-up visit for asthma was recorded. Nasal wash samples were tested for IL-6, IL-13, TNF-α, and IFN-γ by using enzyme-linked immunosorbent assay. This study was approved by the Colorado Multiple Institutional Review Board.

Proportions and frequencies were used to describe categorical data; medians and interquartile ranges were used to describe continuous variables. χ² or Fisher’s exact tests were used for categorical variables and Wilcoxon rank tests were used for continuous variables to compare cytokine with cigarette smoke and marijuana exposures on baseline demographics and respiratory outcomes. Analyses were done by using SAS 9.4 (SAS Institute, Inc, Cary, NC).¹⁶ 

We approached 180 families for enrollment; 99 (55%) agreed to participate, and 74% of these completed the follow-up survey at 6 months. Nasal wash samples were obtained from 98 children (99%). Fifty-six families were asked about exposure to marijuana. Table 1 includes the baseline demographics of the participants. IL-6 and TNF-α, which had 24% (n = 24) and 34% (n = 34) of data values below the limit of detection (LOD), respectively, were categorized as undetectable (< LOD), low (25th percentile–75th percentile for IL-6; 35th percentile–75th percentile for TNF-α), and high (>75th percentile). IL-13 and IFN-γ, which had 93% (n = 91) and 88% (n = 87) below the LOD, respectively, were categorized as low (< LOD) and high (≥ LOD).

Fifty percent of those with elevated IL-13 (n = 4) had a follow-up appointment for asthma compared with 4.2% (n = 3) of those with low levels (P = .02; Table 2). There were no differences in cytokine levels or asthma visits by tobacco smoke exposure (Table 3).

Of the 56 subjects who were asked about marijuana exposure, 12.5% (n = 7) reported exposure to marijuana (Table 3). Patients who were exposed to marijuana were more likely to have low detectable levels of TNF-α than those who were not exposed to marijuana (86% [n = 6] vs 35% [n = 17]; P = .03). There was no statistically significant relationship between marijuana smoke exposure and visits for asthma.

Children who were hospitalized for bronchiolitis with elevated IL-13 at admission were more likely to have a visit for asthma, suggesting that IL-13 levels at admission might be used to predict which children were most likely to develop asthma. The patients with higher IL-13 levels were also older, which may have affected the diagnosis of asthma. If researchers in future studies confirm the association with higher IL-13 and later asthma development, hospitalization for bronchiolitis could represent an opportunity to identify children who are at high risk. Future researchers should confirm the relationship between cytokines, smoke exposure, and asthma. Then, interventions that are used to prevent asthma from developing (such as medications)¹⁷  or aggressive interventions for environmental contributors (such as tobacco smoke) could be investigated.

There was not a significant association between reported tobacco smoke exposure and cytokine levels or asthma visits; this may be because smoke exposure was reported by families, and self-reported exposure underestimates actual smoke exposure as measured by cotinine in biologic samples.¹⁸ 

We found that in children who had detectable levels of TNF-α, reported marijuana smoke exposure was associated with lower rather than higher TNF-α levels. Although the relationship is complicated, this may support studies revealing that tetrahydrocannabinol suppresses TNF-α maturation and secretion in a murine model.¹⁹  Further research is needed to understand the impact of marijuana smoke exposure on children.

We had a small sample size with limited statistical power; we were unable to control for confounding variables. We only included children who were hospitalized with bronchiolitis; thus, our cohort was likely more severely ill and not representative of the general population. Cytokine levels may vary depending on the severity of illness, virus, fluid sampled, and laboratory technique; these were not controlled for. We did not have biologic measures of either tobacco or marijuana smoke exposure and thus likely misclassified some children as unexposed. Using the 6-month follow-up, we may not have identified all children who will get asthma, and an asthma diagnosis at this age is subjective. It is unclear if using cytokines in clinical care is feasible or practical.

Our study revealed an association between IL-13 and later follow-up visits for asthma in children who are hospitalized with bronchiolitis, which could be used to identify children who are at high risk. Further research is needed to understand this relationship and determine if identifying children who are at risk could lead to interventions to prevent asthma.