Page 1 of 73 Infant RSV prophylaxis and nasopharyngeal microbiota until six years of life: a sub- analysis of a randomized controlled trial Wing Ho Man a,b,c , Nienke M. Scheltema a , Melanie Clerc d , Marlies A. van Houten b , Elisabeth E. Nibbelke a , Niek B. Achten a , Kayleigh Arp a , Elisabeth A.M. Sanders a , Louis J. Bont a , Debby Bogaert a,d Affiliations: a Department of Paediatric Immunology and Infectious Diseases, Wilhelmina Children’s Hospital/University Medical Center Utrecht, Utrecht, The Netherlands; b Spaarne Gasthuis Academy, Hoofddorp and Haarlem, The Netherlands; c Department of Pediatrics, Leiden University Medical Center, Leiden, The Netherlands; d Medical Research Council/University of Edinburgh Centre for Inflammation Research, Queen's Medical Research Institute, University of Edinburgh, Edinburgh, United Kingdom. Preferred degree (one only): W H Man MD, N M Scheltema MD, M Clerc PhD, M A van Houten MD, E E Nibbelke MSc, N B Achten MD, K Arp BASc, Prof E A M Sanders MD, Prof L J Bont MD, Prof D Bogaert MD Correspondence to: D. Bogaert, MD, PhD 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
73
Embed
Affiliations:€¦ · Web viewWord count main text: 3,983 / 3,500. Word count abstract: 375 / 250. Summary . Background Respiratory syncytial virus (RSV) infection during infancy
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1 of 58
Infant RSV prophylaxis and nasopharyngeal microbiota until six years of life: a sub-analysis of a randomized controlled trial
Wing Ho Mana,b,c, Nienke M. Scheltemaa, Melanie Clerc d, Marlies A. van Houtenb, Elisabeth E.
Nibbelkea, Niek B. Achtena, Kayleigh Arpa, Elisabeth A.M. Sandersa, Louis J. Bonta, Debby Bogaerta,d
Affiliations:a Department of Paediatric Immunology and Infectious Diseases, Wilhelmina Children’s
Hospital/University Medical Center Utrecht, Utrecht, The Netherlands;
b Spaarne Gasthuis Academy, Hoofddorp and Haarlem, The Netherlands;
c Department of Pediatrics, Leiden University Medical Center, Leiden, The Netherlands;
d Medical Research Council/University of Edinburgh Centre for Inflammation Research, Queen's
Medical Research Institute, University of Edinburgh, Edinburgh, United Kingdom.
Preferred degree (one only):
W H Man MD, N M Scheltema MD, M Clerc PhD, M A van Houten MD, E E Nibbelke MSc, N B
Achten MD, K Arp BASc, Prof E A M Sanders MD, Prof L J Bont MD, Prof D Bogaert MD
Correspondence to:
D. Bogaert, MD, PhD
Medical Research Council/University of Edinburgh Centre for Inflammation Research
Queen's Medical Research Institute, University of Edinburgh
Granulicatella, Pantoea agglomerans, and Roseomonas with reversible airway obstruction. In
contrast, several Haemophilus spp. were positively related with reversible airway obstruction, as well
as S. pneumoniae, Cupriavidus and a low-abundant Corynebacterium (figure 4). Correcting for the
use of palivizumab in the first year of life did not change the results.
15
311
312
313
314
315
316
Page 16 of 58
Discussion
In this single, assessor-blind, randomized, placebo-controlled trial, we demonstrate that palivizumab
during infancy in otherwise healthy preterms is linked with significant changes in respiratory
microbiota composition at one year and six years following the intervention.
Surprisingly, the nasopharyngeal microbial community at year one was still dominated by
Staphylococcus, a profile that has been associated extensively with a respiratory community of
younger infants.3,5,12,23 Only at 6 years of life we observed the ‘classical’ distribution of microbial
profiles as previously observed in younger children 1-2 years of life, including Moraxella,
Streptococcus, Haemophilus and Corynebacterium plus or minus Dolosigranulum-dominated profiles.
In contrast to the children in the previous studies however, we studied preterm born children. A recent
study also demonstrated that the maturation of the gut microbiota of preterm born infants lags behind
that of full term born infants and has not caught up yet at four years of age. 24 This might explain our
findings for the respiratory microbiota as well, and might also in part influence the observed
microbiota differences between the palivizumab and placebo arms of this study.
In children that had received palivizumab in their first year of life, Staphylococcus was less present
and abundant at age 12 months, and their microbial community was less stable over time, suggesting
that palivizumab might accelerate the maturation of the nasopharyngeal microbiota. Whether this is
caused by the prevention of RSV infection or may be related to infections by other viruses remains to
be elucidated. Also, it is unclear whether this faster maturation is beneficial to respiratory health or not
but a previous study suggested that an expedited maturation of the respiratory microbiota early in life
is related with increased susceptibility to respiratory disease later on.5
At six years, we find a higher abundance of Haemophilus spp. and a lower abundance of Moraxella
spp. in children who received palivizumab compared to those who received placebo. Similar effects
were found when comparing children without versus with PCR-confirmed RSV infection in the first
year of life, which is in line with the previous findings that accelerated microbiota maturation, and
consequently reduced stability is associated with increased abundance of non-typeable H.
influenzae.5,12
Although our study design does not allow us to fully unravel the underlying mechanisms, our data
suggest that RSV infection in otherwise healthy preterm infants may have long-term beneficial
ecological effects with reduction of Haemophilus spp. This effect could for example be mediated by
16
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
Page 17 of 58
the induction of local antiviral responses in the airway epithelium25 or by inducing an adaptive
immune response to the co-colonizing bacterial potential pathogens at time of RSV infection.
Reversible airway obstruction, a core measure for characterizing asthma,26 was positively associated
with microbiota community composition, with palivizumab-associated microbes like Haemophilus
spp. positively and Moraxella and Neisseria spp. and health-associated Corynebacterium and
Dolosigranulum spp. being negatively correlated with reversible airway obstruction. Additionally, we
demonstrated that reversible airway obstruction was positively associated with bacterial species that
were previously reported to be associated with asthma, i.e. Streptococcus spp., including S.
pneumoniae, and gram-negative oral bacteria,11,27 and negatively correlated with presumed
commensals of the nasopharyngeal niche, i.e. Corynebacterium, Dolosigranulum and
Staphylococcus.28,29 In all, we previously showed that palivizumab affects the spectrum of viral
infections in infancy, and prevents wheezing in early life. However, on the long term these effects
seem to diminish, which may in part be explained by long-term ecological effects, including
enrichment of more pro-inflammatory bacterial species like Haemophilus and a reduction in potential
beneficial species.
Several limitations of our study should be recognized. First, our cohort of children that were treated
with palivizumab still contained several children (n=8) that had a symptomatic RSV infection in the
first year of life.14 In addition, we probably underestimated the true incidence of RSV infections in our
analyses comparing children with and without PCR-confirmed RSV infections, because these were
based on voluntary parental swab collection.14 Both phenomena may likely have led to an
underestimation of the true impact of proven RSV infection in infancy on respiratory microbiota later
in life, especially with regard to mild RSV infections. The fact that we find very similar results when
comparing children with and without palivizumab, with children without and with PCR-confirmed
RSV infection, however, supports the validity of our results; i.e. at age one, Staphylococcus is
overrepresented in both the placebo group and the children with proven RSV infection, and at age six,
Haemophilus is overrepresented and Moraxella is underrepresented in both the palivizumab group and
the children without proven RSV infection. Second, our study was primarily designed to study the
effects of early life RSV infection, but not of other viral infections on respiratory microbiota
composition at age six, whereas our results now indicate it might be important to study the potential
impact of all viral infections early in life on the respiratory ecosystem. Our data suggests that the
impact of RSV immunoprophylaxis on long-term respiratory microbiota composition is at least in part
17
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
Page 18 of 58
due to prevention of RSV infection since absence of PCR-confirmed RSV infection in the first year of
life was associated with individual bacterial taxa in a similar way to having received palivizumab. It is,
however, possible that these associations are mediated by other non-RSV viral infections, which as
reported previously, are increased observed in children receiving RSV prophylxis.30,31 Third, we lacked
power to analyze the effect of age and timing of viral infections, which might be extremely important
for long-term ecological and health effects. Fourth, our study only included three sampling moments
of respiratory microbiota, whereas we and others have demonstrated the value of frequent sampling in
studying long-term respiratory health and disease.3,4,28 Even though potential differences in baseline
microbiota between randomization groups cannot be ruled out, the fact that microbiota composition
during the first infections in early life were highly similar between groups, baseline differences
become highly unlikely. Fifth, our study sample was drawn from preterm born children between 32
and 35 weeks’ gestational age and is therefore not representative of the general population. A
comparator with term born babies might therefore be warranted for future studies. Sixth, it is highly
likely that all children, including those who received palivizumab, were infected with RSV at some
point beyond the first year of life. Our study design cannot ascertain this effect on the microbiota
composition at age six. It is, however, presumed that changes in respiratory microbiota composition in
early life -during the so-called “window of opportunity”- have more impact on respiratory health later
in life than microbiota changes later in life.28 Finally, 16S-rRNA sequencing can only examine the
bacterial microbiota, but not the viral or fungal microbiota, while an increasing body of evidence
suggests the importance of the respiratory virome and mycobiome in respiratory health and disease.28
This should be taking into consideration with new studies.
Nevertheless, our results suggest that albeit generating a major direct health benefit by prevention of
RSV infections, in early life, in this cohort of otherwise healthy preterm infants palivizumab seems to
affect the respiratory microbiota composition at age one and age six. At age six, palivizumab is
accompanied by the potentially unfavorable overrepresentation of Haemophilus spp. that, in turn, are
independent from the intervention, associated with reversible airway obstruction in our cohort. At the
least, our study findings suggest that viral infections in early life have an important role in shaping the
respiratory ecosystem long-term, possibly as a result of immune modulation during the essential phase
of early life immune maturation.32 These data may nuance the discussion regarding the effects of
universal prevention of RSV infection33, and provide a premise for further studies on early life
interactions between respiratory viruses, microbiota, and the host immune system, and their potential
18
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
Page 19 of 58
long-term consequences on the human ecosystem as well as asthma development. This is of critical
importance, especially in our population of otherwise healthy preterm children, as asthma is still one
of the leading and increasing causes of substantial disability in this group of children.
19
409
410
411
Page 20 of 58
Acknowledgements
The authors are indebted to all the participating children and their families for their commitment and
participation. We thank all members of the MAKI research team, including the staff of the pediatric
lung function laboratory, and the laboratory staff. We are grateful to Hicham el Madkouri for his
primary exploration of the data. This work was supported in part by the Netherlands Organisation for
Scientific Research (NWO-VIDI; grant 91715359) and CSO/NRS Scottish Senior Clinical Fellowship
award (SCAF/16/03).
Declaration of interests
E.A.M.S. declares to have received unrestricted research support from Pfizer, grant support for
vaccine studies from Pfizer and GSK. L.J.B. reports grants from AbbVie during the conduct of the
study and grants from MedImmune, Janssen, MeMed, and the Bill & Melinda Gates Foundation. D.B.
declares to have received unrestricted fees paid to the institution for advisory work for Friesland
Campina and well as research support from Nutricia. No other authors reported financial disclosures.
None of the other authors report competing interests.
20
412
413
414
415
416
417
418
419
420
421
422
423
424
425
Page 21 of 58
References
1 Shi T, McAllister DA, O’Brien KL, et al. Global, regional, and national disease burden estimates of acute lower respiratory infections due to respiratory syncytial virus in young children in 2015: a systematic review and modelling study. Lancet 2017; 390: 946–58.
2 Feldman AS, He Y, Moore ML, Hershenson MB, Hartert T V. Toward Primary Prevention of Asthma. Reviewing the Evidence for Early-Life Respiratory Viral Infections as Modifiable Risk Factors to Prevent Childhood Asthma. Am J Respir Crit Care Med 2015; 191: 34–44.
3 Teo SM, Mok D, Pham K, et al. The infant nasopharyngeal microbiome impacts severity of lower respiratory infection and risk of asthma development. Cell Host Microbe 2015; 17: 704–15.
4 Bisgaard H, Hermansen MN, Buchvald F, et al. Childhood Asthma after Bacterial Colonization of the Airway in Neonates. N Engl J Med 2007; 357: 1487–95.
5 Bosch AATM, Piters WAA de S, van Houten MA, et al. Maturation of the Infant Respiratory Microbiota, Environmental Drivers, and Health Consequences: A Prospective Cohort Study. Am J Respir Crit Care Med 2017; 196: 1582–90.
6 Hyde ER, Petrosino JF, Piedra PA, Camargo CA, Espinola JA, Mansbach JM. Nasopharyngeal Proteobacteria are associated with viral etiology and acute wheezing in children with severe bronchiolitis. J Allergy Clin Immunol 2014; 133: 1220–2.
7 Mansbach JM, Hasegawa K, Henke DM, et al. Respiratory syncytial virus and rhinovirus severe bronchiolitis are associated with distinct nasopharyngeal microbiota. J Allergy Clin Immunol 2016; 137: 1909–1913.e4.
8 Rosas-Salazar C, Shilts MH, Tovchigrechko A, et al. Nasopharyngeal microbiome in respiratory syncytial virus resembles profile associated with increased childhood asthma risk. Am. J. Respir. Crit. Care Med. 2016; 193: 1180–3.
9 De Steenhuijsen Piters WAA, Heinonen S, Hasrat R, et al. Nasopharyngeal microbiota, host transcriptome, and disease severity in children with respiratory syncytial virus infection. Am J Respir Crit Care Med 2016; 194: 1104–15.
10 Huang YJ, Boushey HA. The microbiome in asthma. J Allergy Clin Immunol 2015; 135: 25–30.
11 Lynch JP, Sikder MAA, Curren BF, et al. The influence of the microbiome on early-life severe viral lower respiratory infections and asthma-Food for thought? Front. Immunol. 2017; 8: 156.
12 Biesbroek G, Tsivtsivadze E, Sanders EAM, et al. Early Respiratory Microbiota Composition Determines Bacterial Succession Patterns and Respiratory Health in Children. Am J Respir Crit Care Med 2014; 190: 1283–92.
13 Sze M a., Dimitriu P a., Hayashi S, et al. The lung tissue microbiome in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012; 185: 1073–80.
14 Blanken MO, Rovers MM, Molenaar JM, et al. Respiratory Syncytial Virus and Recurrent Wheeze in Healthy Preterm Infants. N Engl J Med 2013; 368: 1791–9.
15 Scheltema NM, Nibbelke EE, Pouw J, et al. Respiratory syncytial virus prevention and asthma in healthy preterm infants: a randomised controlled trial. Lancet Respir Med 2018; 6: 257–64.
16 O’Brien KL, Nohynek H, World Health Organization Pneumococcal Vaccine Trials Carriage Working Group. Report from a WHO Working Group: standard method for detecting upper respiratory carriage of Streptococcus pneumoniae. Pediatr Infect Dis J 2003; 22: e1-11.
17 Prevaes SMPJ, de Winter-de Groot KM, Janssens HM, et al. Development of the Nasopharyngeal Microbiota in Infants with Cystic Fibrosis. Am J Respir Crit Care Med 2016; 193: 504–15.
18 Carvalho M da GS, Tondella ML, McCaustland K, et al. Evaluation and improvement of real-time PCR assays targeting lytA, ply, and psaA genes for detection of pneumococcal DNA. J Clin Microbiol 2007; 45: 2460–6.
19 Team RC. R: A Language and Environment for Statistical Computing. 2015. https://www.r-project.org/ (accessed April 8, 2017).
20 Man WH, Clerc M, de Steenhuijsen Piters WAA, et al. Loss of Microbial Topography between Oral and Nasopharyngeal Microbiota and Development of Respiratory Infections Early in Life. Am J Respir Crit Care Med 2019; 200: 760–70.
21 Man WH, van Houten MA, Mérelle ME, et al. Bacterial and viral respiratory tract microbiota and host characteristics in children with lower respiratory tract infections: a matched case-control study. Lancet Respir Med 2019; 7: 417–26.
22 Paulson JN, Stine OC, Bravo HC, Pop M. Differential abundance analysis for microbial marker-gene surveys. Nat Methods 2013; 10: 1200–2.
23 Salter SJ, Turner C, Watthanaworawit W, et al. A longitudinal study of the infant nasopharyngeal microbiota: The effects of age, illness and antibiotic use in a cohort of South East Asian children. PLoS Negl Trop Dis 2017; 11: e0005975.
24 Fouhy F, Watkins C, Hill CJ, et al. Perinatal factors affect the gut microbiota up to four years after birth. Nat Commun 2019; 10: 1517.
25 Chen Y, Hamati E, Lee P-K, et al. Rhinovirus induces airway epithelial gene expression through double-stranded RNA and IFN-dependent pathways. Am J Respir Cell Mol Biol 2006; 34: 192–203.
27 Huang YJ, Nelson CE, Brodie EL, et al. Airway microbiota and bronchial hyperresponsiveness in patients with suboptimally controlled asthma. J Allergy Clin Immunol 2011; 127: 372-381.e1-3.
28 Man WH, de Steenhuijsen Piters WAA, Bogaert D. The microbiota of the respiratory tract: gatekeeper to respiratory health. Nat Rev Microbiol 2017; 15: 259–70.
29 Liu G, Tang CM, Exley RM. Non-pathogenic Neisseria: members of an abundant, multi-habitat, diverse genus. Microbiology 2015; 161: 1297–312.
30 Achten NB, Wu P, Bont L, et al. Interference Between Respiratory Syncytial Virus and Human Rhinovirus Infection in Infancy. J Infect Dis 2017; 215: 1102–6.
31 Blanken MO, Korsten K, Achten NB, et al. Population-Attributable Risk of Risk Factors for Recurrent Wheezing in Moderate Preterm Infants During the First Year of Life. Paediatr Perinat Epidemiol 2016; 30: 376–85.
32 Lambrecht BN, Hammad H. The immunology of the allergy epidemic and the hygiene hypothesis. Nat Immunol 2017; 18: 1076–83.
33 Hayden FG, Herrington DT, Coats TL, et al. Efficacy and Safety of Oral Pleconaril for Treatment of Colds Due to Picornaviruses in Adults: Results of 2 Double-Blind, Randomized, Placebo-Controlled Trials. Clin Infect Dis 2003; 36: 1523–32.
34 Clarke KR. Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol 1993; 18: 117–43.
35 Joshi N, Fass J. Sickle: A sliding-window, adaptive, quality-based trimming tool for FastQ files (Version 1.33) [Software]. 2011. https://github.com/najoshi/sickle (accessed Dec 31, 2018).
38 Caporaso JG, Kuczynski J, Stombaugh J, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010; 7: 335–6.
39 Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 2011; 27: 2194–200.
40 Rognes T, Flouri T, Nichols B, Quince C, Mahé F. VSEARCH: a versatile open source tool for metagenomics. PeerJ 2016; 4: e2584.
41 Westcott SL, Schloss PD. De novo clustering methods outperform reference-based methods for assigning 16S rRNA gene sequences to operational taxonomic units. PeerJ 2015; 3: e1487.
42 Quast C, Pruesse E, Yilmaz P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 2012; 41: D590–6.
43 Subramanian S, Huq S, Yatsunenko T, et al. Persistent gut microbiota immaturity in
Demographic characteristics of the children analyzed on year 1 (A) and year 6 (B). Baseline characteristics at randomization have been published and are available at NEJM.org.14
A
Placebo Palivizumab
n 66 79
Male (%) 34 (51·5) 51 (64·6)
Height in cm (median [IQR]) 116 [111, 119] 116 [113, 120]
Weight in kg (median [IQR]) 20·1 [18·6, 21·7] 20·4 [18·8, 21·8]
Any wheezing year 1 (%) 30 (46·9) 25 (31·6)
Recurrent wheezing year 1(%) 11 (16·7) 9 (11·4)
Pets (%) 31 (49·2) 35 (44·3)
Breastfeeding (%) 45 (75·0) 48 (67·6)
Maternal smoking during pregnancy (%) 10 (16·1) 16 (20·8)
Parental atopy (%) 35 (54·7) 53 (69·7)
Atopy Mother (%) 24 (38·1) 34 (43·0)
Asthma Mother (%) 12 (18·8) 9 (11·4)
Hay fever Mother (%) 15 (23·8) 19 (24·4)
Eczema Mother (%) 12 (18·8) 18 (22·8)
23
531532
533
534
535
536537
538
Page 24 of 58
Atopy Father (%) 18 (29·0) 31 (41·3)
Asthma Father (%) 6 ( 9·7) 10 (13·0)
Hay fever Father (%) 9 (14·5) 20 (26·0)
Eczema Father (%) 6 ( 9·5) 11 (14·7)
Parental smoking (%) 26 (39·4) 31 (39·2)
Smoking Mother (%) 11 (18·0) 13 (16·5)
Smoking Father (%) 18 (28·6) 26 (34·7)
24
539
540
Page 25 of 58
B
Placebo Palivizumab
n 166 176
Male (%) 81 (48·8) 109 (61·9)
Height in cm (median [IQR]) 116 [113, 119] 116 [113, 120]
Weight in kg (median [IQR]) 20·4 [18·9, 22·3] 20·9 [19·1, 22·6]
Any wheezing year 1 (%) 85 (51·5) 61 (35·1)
Recurrent wheezing year 1(%) 40 (24·2) 23 (13·1)
Pets (%) 79 (48·8) 80 (45·5)
Breastfeeding (%) 121 (74·7) 117 (72·2)
Maternal smoking during pregnancy (%) 23 (14·6) 24 (14·0)
Parental atopy (%) 98 (60·5) 110 (63·6)
Atopy Mother (%) 55 (34·2) 73 (41·5)
Asthma Mother (%) 17 (10·4) 18 (10·2)
Hay fever Mother (%) 33 (20·5) 46 (26·1)
Eczema Mother (%) 25 (15·3) 40 (22·7)
Atopy Father (%) 62 (38·5) 65 (37·8)
Asthma Father (%) 17 (10·7) 25 (14·4)
Hay fever Father (%) 41 (25·3) 37 (21·3)
Eczema Father (%) 19 (11·7) 26 (15·1)
Parental smoking (%) 57 (34·3) 61 (34·7)
Smoking Mother (%) 23 (14·6) 24 (13·8)
25
541
Page 26 of 58
Smoking Father (%) 43 (26·9) 47 (27·8)
26
542
12 months NP swab
Availablen = 170
Participantsprimary study
n = 429
Participantsfollow-up study
n = 395
6 yearsNP swab
Availablen = 349
High quality DNAn = 145
High quality DNAn = 342
Palivizumabn = 79
Placebon = 66
Palivizumabn = 176
Placebon = 166
Primary analysis
Secondary analysis
16S sequencingn = 170
16S sequencingn = 349
Palivizumabn = 74#
Placebon = 66
No RSVn = 132
Proven RSV
n = 14
Palivizumabn = 168#
Placebon = 166
No RSVn = 342
Proven RSV
n = 34
Combinedn = 274
6 yearsSpirometry
Availablen = 290
Palivizumabn = 137
Placebon = 137
Page 27 of 58
Figure 1. Flow diagram for samples analyzed.
NP = nasopharynx. # We excluded 5/79 and 8/176 children from the palivizumab group in our posthoc
analyses of the children at age 12 months and 6 years, respectively, as they developed an RSV
infection during their first year of life. Including these samples yielded similar results (appendix).
27
543
544
545
546
547
548
Page 28 of 58
Figure 2. Microbiota profiles at age 12 months and 6 years.
(A) Three-dimensional NMDS plots depicting the individual nasopharyngeal microbiota composition
at year one (data points, n=145) colored by treatment group: placebo (brown, n=66) and palivizumab
(dark green, n=79). The difference in total microbiota composition in both groups is significant
(adonis, R2=1·3%, p=0·0185).
(B) Three-dimensional NMDS plots depicting the individual nasopharyngeal microbiota composition
at year 6 (data points, n=342) colored by treatment group: placebo (brown, n=166) and palivizumab
(dark green, n=176). The difference in total microbiota composition in both groups is smaller
compared to 12 months (R2 0·6% versus 1·3%), though still not significant (adonis, R2=0·6%,
p=0·0575). Ellipses represent the standard deviation of all points within a cohort. The stress-value
using the first two dimensions was 0·25, whereas this dropped to 0·18 when using three dimensions.
Because a stress of <0·2 indicates a reasonable interpretability,34 we decided to depicts the samples
across these three dimensions (NMDS1-NMDS3). The figures also depict the biomarkers species
(determined by random forests analysis on hierarchical clustering results) colored by phylum (Green
circles = Bacteroides). To avoid OTUs with identical annotations, we refer to OTUs using their
taxonomical annotations combined with a rank number based on the abundance of each given OTU.
28
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
Page 29 of 58
A
B
29
567
Page 30 of 58
Figure 3. Differential abundance of bacterial taxa between treatment groups.
Effect sizes are depicted for the 20 most differentially abundant taxa in either the palivizumab group (right) or placebo group (left). Log2 fold changes (including 95% confidence intervals) were obtained by metagenomeSeq analysis and corrected for multiple comparisons (Benjamini-Hochberg). (A) depicts the results for samples (n = 145) obtained at 12 months of life, whereas (B) depicts the results for samples obtained 6 years of life (n = 342). To avoid OTUs with identical annotations, we refer to OTUs using their taxonomical annotations combined with a rank number based on the abundance of each given OTU.
A
B
30
568
569570571572573574575
Page 31 of 58
31
576577
Page 32 of 58
Figure 4. Sparse random forests models associating microbial species with reversible airway obstruction.
Twelve taxa were associated with reversible airway obstruction as selected by a cross-validated
random forests analysis using all samples (n=274). Taxa are ranked in descending order based on their
importance to the accuracy of the model. Variable importance was estimated by calculating the mean
increase in node purity after randomly permuting the values of each given variable (mean ± standard
deviation, 100 replicates). A higher value increase in node purity represents a higher variable
importance. The direction of the associations was estimated post-hoc using Pearson’s correlation (red
= positive association reversibility; blue = negative association with reversibility). To avoid OTUs
with identical annotations, we refer to OTUs using their taxonomical annotations combined with a
rank number based on the abundance of each given OTU. Whether children had received RSV
prophylaxis versus placebo in the first year is not significantly contributing to the model (gray).
32
578579
580
581
582
583
584
585
586
587
588
589
590
Page 33 of 58
Supplementary appendix
Sampling at 12 months
Only during the last year of recruitment, we sampled children at 12 months. We saw no differences in
baseline characteristics between these last 40% children and the first 60% children of which we did
not have a 12-month sample (p>0·20 for all characteristics).
Additional methods for bioinformatics analysis
Raw sequences were trimmed using an adaptive, window-based trimming algorithm (Sickle, Q>20,
length threshold of 150 nucleotides).35 We aimed to further reduce the number of sequence errors in
the reads by applying an error correction algorithm (BayesHammer, SPAdes genome assembler
toolkit).36 Forward and reverse reads were then assembled into contigs using PANDAseq.37 Merged
reads were demultiplexed using QIIME (v1·9).38 After removal of singleton sequences, we removed
chimeras using both de novo and reference (against Gold database) chimera identification (UCHIME
algorithm in VSEARCH).39,40 VSEARCH abundance-based greedy clustering was used to pick OTUs
at a 97% identity threshold.41 Taxonomic annotation was executed using the RDP-II naïve Bayesian
classifier on SILVA v119 training set.42 Taxonomic assignment was validated by blasting against the
NCBI database, using a 100% identity cut-off. We generated an abundance-filtered dataset by
including only those OTUs that were present at or above a confident level of detection (0·1% relative
abundance) in at least two samples.43 In addition, to ensure our data was of the highest quality, we
identified and removed 58 potential contaminants using the Decontam R-package, as previously
described (supplementary figure 4).21 Also, we only kept samples that contained at least 9,000 reads.
To avoid OTUs with identical annotations, we refer to OTUs using their taxonomical annotations
combined with a rank number based on the abundance of each given OTU. In case when an OTU
could not be confidently annotated as either of two species, both species are indicated and separated
by a solidus. The raw OTU-counts table was used for calculations of alpha-diversity (local diversity)
and analyses using the metagenomeSeq package.22 The OTU-proportions table was used for all other
downstream analyses, including hierarchical clustering and random forests modelling. Beta-diversity
was assessed using the Bray-Curtis dissimilarity metric.
Sequence reads were submitted to the National Center for Biotechnology Information Sequence Read
Archive (accession number SRP141698).
33
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
Page 34 of 58
Quality control of 16S-rRNA gene amplicon sequencing
In total, 24 DNA isolation and PCR blanks were sequenced along with the study samples. All blanks
yielded <4,500 reads and the number of reads was more than one order of magnitude lower compared
to that of the samples (median 551 vs 37,529 reads; supplementary figure 5a). Hierarchical
clustering also clearly separated the blanks from samples (supplementary figure 5b). These results
robustly indicate that our strict sequencing protocol resulted in no apparent contamination.
Characterization of nasopharyngeal microbiota
Across all analyzed samples, 19,516,175 reads in total (mean 36,412 ± 19,513 reads/sample) were
retained for downstream analyses and these were binned into 309 operational taxonomic units (OTUs),
further referred to as bacterial taxa. For all samples the Good’s estimator of coverage was above
99·9%. The taxon annotated as “Streptococcus (6)” had a strong correlation with lytA qPCR Ct-values,
confirming its origin Streptococcus pneumoniae (Spearman’s rho -0·81, p<0·001). The dominant
phyla were Proteobacteria (47·5%), Firmicutes (34·8%), and Actinobacteria (15·9%). Hierarchical
clustering showed the presence of 15 distinct microbiota profiles (supplementary figure 1b) driven
by Moraxella catarrhalis/nonliquefaciens (28·3%), Staphylococcus (16·2%), Corynebacterium
propinquum & Dolosigranulum pigrum (14·4%), Haemophilus (11·2% of samples), Streptococcus
Viruses and microbiota composition in the first year of life
In the subset of RTI samples, we detected a virus in 58/66 (87·9%) samples. RSV was detected in 1/66
of the subset of children (1·5%). The most common other virus was HRV, detected in 48/66 (72·7%)
of the children, followed by adenovirus (11/66 [16·7%]), and coronavirus (9/66 [13·6%];
supplementary table 1). The nasopharyngeal microbiota composition during those respiratory
infections was typified by microbiota dominated of Staphylococcus, followed by a.o. Klebsiella, S.
pneumoniae, Rothia and Moraxella spp. (supplementary figure 1a): microbiota did not differ
between the RSV prophylaxis group and the placebo group (adonis, R2 0·6%, p=0·981).
34
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
Page 35 of 58
Viruses in the first year of life
From the analyzed samples at 12 months, a total of 80/145 (55·1%) otherwise healthy preterm born
children experienced a PCR-confirmed viral infection in the first year of life. RSV was detected in
14/145 children 9·6%). The most common other virus was HRV, detected in 66/145 (45·5%) of the
children, followed by adenovirus (22/145 [15·2%]), and bocavirus (18/145 [12·4%]; supplementary
table 2a). There was a similar distribution of early life viral infection within our subset of samples at
age 12 months.
From the analyzed samples at 6 years, a total of 182/342 (53·2%) otherwise healthy preterm born
children experienced a PCR-confirmed viral infection in the first year of life. RSV was detected in
34/342 children 9·9%) and was detected significantly more in the placebo group compared to the RSV
prophylaxis group (8/166 [4·5%] and 26/176 [15·7%], respectively, p=0·001; supplementary table
2b). The most common other virus was HRV, detected in 158/342 (46·2%) of the children. HRV was
detected significantly more in the RSV prophylaxis group compared to the placebo group (87/176
[49·4%] and 71/166 [42·8%], respectively, p=0·027). The other viruses were detected equally across
treatment groups in the first year of life in the subset of the samples analyzed at year six.
Relationship between microbiota composition at age one and six
We confirmed that microbiota maturation continues from age 12 months to age 6 years with
significant differences in microbiota community compositions between both age groups (R 2=9·5%,
p<0·0001). Especially the biomarker taxa E. faecium, M. osloensis, Chryseobacterium, Rothia,
Brevundimonas, and S. salivarius diminished over time, whereas Haemophilus spp., M.
catarrhalis/nonliquefaciens, C. propinquum, D. pigrum, S. pyogenes, and M. lacunata increased
significantly with age (supplementary figure 7). Interestingly, although most children had a different
microbiota profile at age 6 years compared to age 12 months (n=118 paired samples, supplementary
figure 8), we still saw that there was a higher but not significant correlation between paired microbiota
of the same child at 12 months and 6 year of life when compared to unpaired samples (median Bray-
Curtis similarity 0·101 and 0·044, respectively; p=0.056; supplementary figure 9a) suggesting the
existence of a modest intra-individual core microbiome. This was strengthened by the fact that more
stable microbiota development between year 1 and year 6 samples was associated with the presence of
Moraxella catarrhalis/nonliquefaciens and Corynebacterium propinquum & Dolosigranulum pigrum-
dominated clusters (median Bray-Curtis similarity 0·455 and 0·348, respectively, versus 0·001-0·191
for all other clusters; supplementary figure 9b), corroborating previous studies suggesting the
35
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
Page 36 of 58
presence of these microbiota profiles in early life are associated with stable and resilient microbiota
development over time.
Interestingly, when stratifying for treatment groups, the existence of a core microbiome was only
evident for the placebo group (median Bray-Curtis similarity 0·126 and 0·039, for intra-individual
versus inter-individual concordance, respectively; p=0·03583; supplementary figure 9a) but not for
the palivizumab group (median Bray-Curtis similarity 0·054 and 0·045, respectively; p=0·475;
supplementary figure 9a), suggesting that children receiving placebo had more stable microbiota
development compared to the palivizumab group.
Findings excluding children within the palivizumab group who developed an RSV infection during their first year of life
We excluded 5/79 and 8/176 children from the palivizumab group in a posthoc analyses of the
children at age 12 months and 6 years, respectively, as they developed an RSV infection during their
first year of life. At 12 months of life, we observed a significant difference in overall microbial
community structure between the palivizumab and the placebo group (R2=1·3%, p=0·0285). On
cluster level, the Staphylococcus-dominated profile was positively associated with the placebo group
(chi-square, p=0·00381; OR 0·27, 95% CI 0·10-0·67). On individual taxon level, Staphylococcus
abundance was not different in the placebo group compared to palivizumab (metagenomeSeq, log2 fold
change 1·4, q=0·122, figure 10a). We further observed a significantly higher abundance of
Helcococcus, Dolosigranulum pigrum, Lactobacillus spp., Streptococcus spp. and a range of gram-
negative spp. including Klebsiella and several oral bacteria in the palivizumab group.
When evaluated at six years of age, we still observed a small though not significant difference in
overall microbial community structure between the otherwise healthy preterm infants who were
treated with palivizumab and those who were treated with placebo (R2=0·7%, p=0·0425). On cluster
level, the Haemophilus-dominated profile at age six years was positively associated with the
palivizumab group (chi-square, p=0·02571; OR 1·91, 95% CI 1·08-3·41). On individual bacterial
taxon level, Haemophilus spp. as well as S. pyogenes were positively associated with the palivizumab
group (figure 10b), whereas Moraxella, Corynebacterium and Neisseriaceae spp. were negatively
associated with palivizumab in the first year of life.
36
679
680
681
682
683
684
685
686
687688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
Page 37 of 58
Supplementary table 1. Prevalence of respiratory virus detection in the 66 RTI samples analyzed in this study.
None of the differences between treatment groups were statistically significant (Chi-square all p>0·4).
All RTI samples(n=66)
Placebo(n=33)
Palivizumab(n=33)
Virus n % n % n %
Any virus 58 87·9 28 84·8 30 90·9
Multiple viruses 20 30·3 8 24·2 12 36·4
Respiratory syncytial virus (RSV) 1 1·5 1 3·0 0 0
RSV with any other virus 1 1·5 1 3·0 0 0
Human rhinovirus 48 72·7 23 69·7 25 75·8
Adenovirus 11 16·7 4 12·1 7 21·2
Bocavirus 7 10·6 2 6·1 5 15·2
Coronavirus 9 13·6 3 9·1 6 18·2
Parainfluenza 6 9·1 3 9·1 3 9·1
Human metapneumovirus 2 3·0 1 3·0 1 3·0
Influenza 0 0 0 0 0 0
37
708709
710
711
Page 38 of 58
Supplementary table 2. Prevalence of respiratory virus detection in the first year of life.
The total number of PCR detected viral infections in the first year of life of the subset of samples that we analyzed at age 12 months (A) and 6 years of life (B).
RSV with any other virus 22 6·4 15 9·0 7 4·0 0·081
38
712
713714
715
716
Page 39 of 58
Human rhinovirus 158 46·2 71 42·8 87 49·4 0·027
Adenovirus 60 17·5 22 13·3 38 21·6 0·063
Bocavirus 52 15·2 24 14·5 28 15·9 0·842
Coronavirus 44 12·9 16 9·6 28 15·9 0·122
Parainfluenza 30 8·8 10 6·0 20 11·4 0·124
Human metapneumovirus 13 3·8 7 4·2 6 3·4 0·906
Influenza 7 2·0 6 3·6 1 0·6 0·066
39
717
Page 40 of 58
Supplementary figure 1. Microbiota profiles during early respiratory infections, at and age 12 months and 6 years.
(A) Bar chart showing average profiles of respiratory samples obtained during early respiratory infections in the first year of life (n=66), and during routine sampling at age 12 months (n=145) and 6 years (n=342). The legend shows the biomarker taxa over time.(B) Hierarchical clustering of all study samples identified 15 profiles. Biomarker species defined by random forests analysis for these 15 profiles identified by hierarchical clustering were from left to right: Chryseobacterium, Klebsiella, Enterococcus faecium, Moraxella osloensis, Streptococcus salivarius, Staphylococcus, Moraxella lincolnii, Streptococcus pneumoniae, Rothia & Streptococcus pneumoniae, Streptococcus pyogenes, Haemophilus, Moraxella lacunata, Corynebacterium propinquum & Dolosigranulum pigrum, M. catarrhalis/nonliquefaciens and Brevundimonas. The figure visualizes from top to bottom the clustering dendrogram, including information on the treatment allocation, sample type, and a bar chart of the relative abundance for each of the biomarker species in the individual samples. A
(A) NMDS plots depicting the individual nasopharyngeal microbiota composition at year one (data points, n=145) colored by early life RSV infection: with RSV (purple, n=14) and without RSV (orange, n=132).(B) NMDS biplots depicting the individual nasopharyngeal microbiota composition at year six (data points, n=342) colored by early life RSV infection: with RSV (purple, n=34) and without RSV (orange, n=308). Ellipses represent the standard deviation of all points within a cohort. The stress-value using the first two dimensions was 0·25, whereas this dropped to 0·18 when using three dimensions. Because a stress of <0·2 indicates a reasonable interpretability, 34 we decided to depicts the samples across these three dimensions (NMDS1-NMDS3). The figures also depict the biomarkers species (determined by random forests analysis on hierarchical clustering results) colored by phylum (Green diamonds = Proteobacteria, orange triangles = Firmicutes, purple squares = Actinobacteria, pink circles = Bacteroides).
A
42
735
736737738739740741742743744745746747
Page 43 of 58
B
43
748
Page 44 of 58
Supplementary figure 3. Differential abundance of bacterial taxa between children without and with proven RSV infection.
An analysis similar to that of figure 2b was performed stratifying the cohort at age 6 years alternatively, i.e. the associations between the 20 most differentially abundant taxa and either children with (n=34, left) and without proven RSV (n=308, right) infection early in life. Log 2 fold changes (including 95% confidence intervals) were obtained by metagenomeSeq analysis and corrected for multiple comparisons (Benjamini-Hochberg).
44
749
750751
752753754755756
757
Page 45 of 58
Supplementary figure 4. Fifty-eight OTUs were identified as contaminants.
The frequency of each OTU is depicted as a function of the bacterial biomass. The dashed black line shows the model of a noncontaminant sequence feature for which frequency is expected to be independent of the input DNA concentration. The red line shows the model of a contaminant sequence feature, for which frequency is expected to be inversely proportional to input DNA concentration, as contaminating DNA will make up a larger fraction of the total DNA in samples with very little total DNA.
45
758
759760761762763764
Page 46 of 58
46765
Page 47 of 58
Supplementary figure 5. Samples are distinct from blanks.
(A) The number of sequences in the DNA isolation blanks and PCR blanks (grey) were an order of magnitude lower compared to the samples of children who received either placebo (brown) or palivizumab (green). (B) visualizes the hierarchical clustering dendrogram, which clearly separates the blanks (red) from the samples (grey).
A
47
766
767768769770771
Page 48 of 58
B
Supplementary figure 7. Association of the biomarker taxa with age.
Associations between the biomarker taxa and either year one (left) or year six (right). Log 2 fold changes (including 95% confidence intervals) were obtained by metagenomeSeq analysis and corrected for multiple comparisons (Benjamini-Hochberg).
48
772
773774775
Page 49 of 58
49
776
Page 50 of 58
Supplementary figure 8. Individual microbial developmental trajectories between age one and six.
(A) Visualization of the relation between the overall microbiota profile of the same participant over time as a parallel alluvial diagram. The alluvial diagram depicts the direct links between the microbiota profile at age one (left) and at age six (right). Green lines represent participants that have the same profile at both ages (n=9/118 [7·6%]) and brown lines represent participants that have different profiles at both ages (n=109/118 [92·4%]). Participants with a similar profile at both ages were distributed evenly across treatment groups. (B) In the placebo group, 5/53 (9.4%) participants had the same profile at both ages 48/53 (90·6%) participants had different profiles at both ages. (C) In the palivizumab group, 4/65 (6·2%) participants had the same profile at both ages 61/65 (93·8%) participants had different profiles at both ages.
50
777778
779780781782783784785786787
Page 51 of 58
A
51
Page 52 of 58
B
C
52
788
Page 53 of 58
Supplementary figure 9. Similarity in total microbiota composition between year one and six.
(A) Boxplots depicting the Bray-Curtis similarities between the microbiota composition at the age of one and six in paired samples of the same child (Within child) or in unpaired samples (Between children). (B) Boxplots depicting the Bray-Curtis similarities between the microbiota composition at the age of one and six in paired samples of the same child stratified for the microbial clusters determined in supplementary figure 1.
A
B
53
789
790791792793794795796
Page 54 of 58
54
797
798
Page 55 of 58
Supplementary 10. Differential abundance of bacterial taxa between treatment groups.
We repeated an identical analysis as in Figure 2, but including the samples of children within the palivizumab group who still developed an RSV infection during their first year of life. Associations between the 20 most differentially abundant taxa and either the palivizumab group (right) or placebo group (left). Log2 fold changes (including 95% confidence intervals) were obtained by metagenomeSeq analysis and corrected for multiple comparisons (Benjamini-Hochberg). (A) depicts the results for samples obtained at 12 months of life, whereas (B) depicts the results for 6 years of life. To avoid OTUs with identical annotations, we refer to OTUs using their taxonomical annotations combined with a rank number based on the abundance of each given OTU.
A
B
55
799
800801802803804805806807
Page 56 of 58
56
808
Page 57 of 58
Supplementary references
1 Shi T, McAllister DA, O’Brien KL, et al. Global, regional, and national disease burden estimates of acute lower respiratory infections due to respiratory syncytial virus in young children in 2015: a systematic review and modelling study. Lancet 2017; 390: 946–58.
2 Feldman AS, He Y, Moore ML, Hershenson MB, Hartert T V. Toward Primary Prevention of Asthma. Reviewing the Evidence for Early-Life Respiratory Viral Infections as Modifiable Risk Factors to Prevent Childhood Asthma. Am J Respir Crit Care Med 2015; 191: 34–44.
3 Teo SM, Mok D, Pham K, et al. The infant nasopharyngeal microbiome impacts severity of lower respiratory infection and risk of asthma development. Cell Host Microbe 2015; 17: 704–15.
4 Bisgaard H, Hermansen MN, Buchvald F, et al. Childhood Asthma after Bacterial Colonization of the Airway in Neonates. N Engl J Med 2007; 357: 1487–95.
5 Bosch AATM, Piters WAA de S, van Houten MA, et al. Maturation of the Infant Respiratory Microbiota, Environmental Drivers, and Health Consequences: A Prospective Cohort Study. Am J Respir Crit Care Med 2017; 196: 1582–90.
6 Hyde ER, Petrosino JF, Piedra PA, Camargo CA, Espinola JA, Mansbach JM. Nasopharyngeal Proteobacteria are associated with viral etiology and acute wheezing in children with severe bronchiolitis. J Allergy Clin Immunol 2014; 133: 1220–2.
7 Mansbach JM, Hasegawa K, Henke DM, et al. Respiratory syncytial virus and rhinovirus severe bronchiolitis are associated with distinct nasopharyngeal microbiota. J Allergy Clin Immunol 2016; 137: 1909–1913.e4.
8 Rosas-Salazar C, Shilts MH, Tovchigrechko A, et al. Nasopharyngeal microbiome in respiratory syncytial virus resembles profile associated with increased childhood asthma risk. Am. J. Respir. Crit. Care Med. 2016; 193: 1180–3.
9 De Steenhuijsen Piters WAA, Heinonen S, Hasrat R, et al. Nasopharyngeal microbiota, host transcriptome, and disease severity in children with respiratory syncytial virus infection. Am J Respir Crit Care Med 2016; 194: 1104–15.
10 Huang YJ, Boushey HA. The microbiome in asthma. J Allergy Clin Immunol 2015; 135: 25–30.
11 Lynch JP, Sikder MAA, Curren BF, et al. The influence of the microbiome on early-life severe viral lower respiratory infections and asthma-Food for thought? Front. Immunol. 2017; 8: 156.
12 Biesbroek G, Tsivtsivadze E, Sanders EAM, et al. Early Respiratory Microbiota Composition Determines Bacterial Succession Patterns and Respiratory Health in Children. Am J Respir Crit Care Med 2014; 190: 1283–92.
13 Sze M a., Dimitriu P a., Hayashi S, et al. The lung tissue microbiome in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012; 185: 1073–80.
14 Blanken MO, Rovers MM, Molenaar JM, et al. Respiratory Syncytial Virus and Recurrent Wheeze in Healthy Preterm Infants. N Engl J Med 2013; 368: 1791–9.
15 Scheltema NM, Nibbelke EE, Pouw J, et al. Respiratory syncytial virus prevention and asthma in healthy preterm infants: a randomised controlled trial. Lancet Respir Med 2018; 6: 257–64.
16 O’Brien KL, Nohynek H, World Health Organization Pneumococcal Vaccine Trials Carriage Working Group. Report from a WHO Working Group: standard method for detecting upper respiratory carriage of Streptococcus pneumoniae. Pediatr Infect Dis J 2003; 22: e1-11.
17 Prevaes SMPJ, de Winter-de Groot KM, Janssens HM, et al. Development of the Nasopharyngeal Microbiota in Infants with Cystic Fibrosis. Am J Respir Crit Care Med 2016; 193: 504–15.
18 Carvalho M da GS, Tondella ML, McCaustland K, et al. Evaluation and improvement of real-time PCR assays targeting lytA, ply, and psaA genes for detection of pneumococcal DNA. J Clin Microbiol 2007; 45: 2460–6.
19 Team RC. R: A Language and Environment for Statistical Computing. 2015. https://www.r-project.org/ (accessed April 8, 2017).
20 Man WH, Clerc M, de Steenhuijsen Piters WAA, et al. Loss of Microbial Topography between Oral and Nasopharyngeal Microbiota and Development of Respiratory Infections Early in Life. Am J Respir Crit Care Med 2019; 200: 760–70.
21 Man WH, van Houten MA, Mérelle ME, et al. Bacterial and viral respiratory tract microbiota and host characteristics in children with lower respiratory tract infections: a matched case-control study. Lancet Respir Med 2019; 7: 417–26.
22 Paulson JN, Stine OC, Bravo HC, Pop M. Differential abundance analysis for microbial marker-gene surveys. Nat Methods 2013; 10: 1200–2.
23 Salter SJ, Turner C, Watthanaworawit W, et al. A longitudinal study of the infant nasopharyngeal microbiota: The effects of age, illness and antibiotic use in a cohort of South East Asian children. PLoS Negl Trop Dis 2017; 11: e0005975.
24 Fouhy F, Watkins C, Hill CJ, et al. Perinatal factors affect the gut microbiota up to four years after birth. Nat Commun 2019; 10: 1517.
25 Chen Y, Hamati E, Lee P-K, et al. Rhinovirus induces airway epithelial gene expression through double-stranded RNA and IFN-dependent pathways. Am J Respir Cell Mol Biol 2006; 34: 192–203.
27 Huang YJ, Nelson CE, Brodie EL, et al. Airway microbiota and bronchial hyperresponsiveness in patients with suboptimally controlled asthma. J Allergy Clin Immunol 2011; 127: 372-381.e1-3.
28 Man WH, de Steenhuijsen Piters WAA, Bogaert D. The microbiota of the respiratory tract: gatekeeper to respiratory health. Nat Rev Microbiol 2017; 15: 259–70.
29 Liu G, Tang CM, Exley RM. Non-pathogenic Neisseria: members of an abundant, multi-habitat, diverse genus. Microbiology 2015; 161: 1297–312.
30 Achten NB, Wu P, Bont L, et al. Interference Between Respiratory Syncytial Virus and Human Rhinovirus Infection in Infancy. J Infect Dis 2017; 215: 1102–6.
31 Blanken MO, Korsten K, Achten NB, et al. Population-Attributable Risk of Risk Factors for Recurrent Wheezing in Moderate Preterm Infants During the First Year of Life. Paediatr Perinat Epidemiol 2016; 30: 376–85.
32 Lambrecht BN, Hammad H. The immunology of the allergy epidemic and the hygiene hypothesis. Nat Immunol 2017; 18: 1076–83.
33 Hayden FG, Herrington DT, Coats TL, et al. Efficacy and Safety of Oral Pleconaril for Treatment of Colds Due to Picornaviruses in Adults: Results of 2 Double-Blind, Randomized, Placebo-Controlled Trials. Clin Infect Dis 2003; 36: 1523–32.
34 Clarke KR. Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol 1993; 18: 117–43.
35 Joshi N, Fass J. Sickle: A sliding-window, adaptive, quality-based trimming tool for FastQ files (Version 1.33) [Software]. 2011. https://github.com/najoshi/sickle (accessed Dec 31, 2018).
38 Caporaso JG, Kuczynski J, Stombaugh J, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010; 7: 335–6.
39 Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 2011; 27: 2194–200.
40 Rognes T, Flouri T, Nichols B, Quince C, Mahé F. VSEARCH: a versatile open source tool for metagenomics. PeerJ 2016; 4: e2584.
41 Westcott SL, Schloss PD. De novo clustering methods outperform reference-based methods for assigning 16S rRNA gene sequences to operational taxonomic units. PeerJ 2015; 3: e1487.
42 Quast C, Pruesse E, Yilmaz P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 2012; 41: D590–6.
43 Subramanian S, Huq S, Yatsunenko T, et al. Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature 2014; 510: 417.