2022 Emil Christensen – MD

Otitis media is one the most common bacterial infections in early childhood, as well as one of the most common causes of antibiotic use and surgical treatment in children: In Denmark, two out of three children experience at least one episode during childhood 5, while more than one in four children undergoes tympanostomy tube insertion in full anesthesia 3. This high burden of disease has previously been shown to be unevenly distributed, both globally and within different populations 6. Numerous risk factors play a part in this variability, including both modifiable and non-modifiable factors 7. Understanding the causal mechanisms behind these risk factors are of great interest, as it may open new possibilities of disease prevention independently of the risk factors themselves. As such, the overarching aim of this thesis is to elucidate the roles and interplay of some of the largest risk factors for otitis media, notably the environment, the human microbiota and host factors.

Three separate studies form the backbone of this thesis, all of them reliant on data collected from the COPSAC2010 cohort comprising 700 children. This data includes: Bacterial culturing and 16s rRNA gene amplicon sequencing of more than 1,800 hypopharyngeal and 2,600 fecal samples obtained from 1 week to 6 years of age; immune mediator assays of airway mucosal lining fluid sampled at 1 month of age; otitis media symptom diaries registered daily until 3 years of age; information on tympanostomy tube insertions in the first 3 years of life obtained from national registers; and last but not least, information on numerous risk factors obtained through interviews with the primary caregivers.

In Study I, we set out to examine the associations between pathogenic bacterial airway colonization in early life and later risk of acute otitis media and tympanostomy tube insertion, and whether such associations were modulated by an insufficient local immune mediator response to bacterial colonization. This was spurred on by previous findings, that airway colonization is associated with later otitis media 8, as well as a distinct local immune response 9. For this study, we relied on data from bacterial cultures of the airway samples, as they allowed us taxonomic identification at species level, where the association between colonization and disease had previously been well-established. While we did find that airway colonization in early life was associated with an increased risk of otitis media, the associations were far from consistent. Furthermore, the associations were independent of the local immune response to colonization.

7

In Study II, we therefore decided to shift our attention away from the local immune response. Instead, we further examined the associations between the early airway microbiota and otitis media, as well as the possible mediation of known risk factors through the early airway microbiota. For this study, we utilized data from 16S rRNA gene amplicon sequencing, to examine as much as possible of the airway microbiota, rather than just the usual suspects. Furthermore, we constructed a composite outcome of otitis media severity to reduce the risk of type II errors. We found airway beta diversity and specific taxa abundances to be significantly associated with later otitis media severity. Likewise, multiple risk factors were significantly associated with otitis media severity. Of those, only the presence of older children in the household and lower maternal education level were additionally associated with changes in the early airway microbiota. The effect of older children was partly found to be mediated through the airway microbiota. We found no evidence of mediation for other risk factors.

In Study III, we focused on the microbiotal impact of siblings, as Study II had shown this association to be particularly strong. The presence or absence of siblings has previously been linked to numerous health outcomes, yet research into the role of siblings on the developing microbiota has for the most part been incidental. As we were no longer using an airway infection as endpoint, we included fecal samples from the cohort as well. We found siblings to be one of the most important discriminatory determinants of the developing microbiota, both in the airways and the gut, with significant differences in alpha diversity, beta diversity and relative abundances of the most abundant taxa. Most of these associations were particularly apparent during the first year of life. Furthermore, we found that the age gap to closest older sibling was by far the most important determinant of effect size, compared to number of older siblings, despite the latter metric being used much more frequently in microbiological and epidemiological research.

In conclusion, we found strong evidence for a link between environment – i.e., the presence or absence of siblings – and otitis media, mediated through the microbiota. As for host factors, we found no evidence of modulation by an insufficient local immune mediator response to bacterial colonization. However, care should be taken in extrapolating our results beyond the local immune response that was sampled for our study. Further research with higher taxonomic resolution may show even stronger associations between risk factors, microbiota, and otitis media, and could ultimately pave the way for intervention studies. In such future studies, it would be justified and even prudent to consider the age gap to closest older sibling rather than number of older siblings.