Respitatory virus infections during the first year of life a longitudinal community-based dynamic birth cohort study

Abstract

Background: Viral acute respiratory infections (ARIs) are the commonest illnesses experienced by all age groups, especially in infants where infection rates are highest. Nevertheless, during the molecular era, outside of hospital-based studies, little is known about the current aetiology and community burden of viral ARIs in infants and young children. The observational research in childhood infectious disease (ORChID) project is a prospective community-based birth-cohort study of healthy Australian infants and children. It began in 2010 to investigate respiratory virus infections until two-years of age. My PhD established laboratory quality control techniques for studies of this nature and describes the respiratory viruses and molecular epidemiology of human rhinoviruses (RV) during the first year of life in a nested subgroup of this cohort. The hypotheses were: I. During the first year of life and in otherwise healthy infants, RV is the most commonly detected respiratory virus in respiratory secretions. II. Various factors impact upon successful viral detection, including the ability of parents to collect appropriate samples and other laboratory-based technical issues. III. Repeated detection of RV-RNA in respiratory secretions over periods of more than 4-weeks results from genotype replacement and new infection events rather than from prolonged shedding of the same genotype. IV. Many RV genotypes circulate in a single location in 1 year period. Methods The ongoing ORChID study completed sample collection at the end of 2014. Parents were approached antenatally and asked to collect weekly anterior nasal swabs from the time of their child’s birth until their second birthday. Swabs were mailed to the laboratory where they were stored at -80oC until analysis. Parents also completed a daily symptom diary, which was submitted monthly. My PhD focused on samples collected from an infant subgroup within this cohort and the first two-years of the ORChID study. Samples were extracted using an automated robotic system after spiking each sample with equid herpes virus (EHV-1). The extraction quality and presence of human DNA in extracts were assessed by real-time PCR for EHV-1 and endogenous retrovirus 3 (ERV-3) respectively. Respiratory virus PCR testing included: RV, influenza viruses (IFVs: A/B), parainfluenza viruses (PIV: 1-3), respiratory syncytial virus (RSV; A/B), human metapneumovirus (hMPV); human coronaviruses (hCoV; NL63, 229E, OC43 and HKU); human polyomaviruses (PyV: WU and KI), adenovirus (AdV), and human bocavirus (hBoV). In a subset of 3366 nasal swab samples, the impact of ERV-3 load upon respiratory virus detection was determined. Mould was observed incidentally in some samples reaching the laboratory. The impact of different mould levels upon ERV-3 and respiratory virus detection was therefore investigated. The influence of sequence variation upon target sequences was assessed for HAdV detection. Two new HAdV real-time PCR assays utilising combinations of degenerate oligonucleotides were tested in parallel with a previously designed and well established real-time PCR assay. ORChID (n=8800) and routine clinical (n=779) samples were then tested and the results compared. The nature and shedding patterns of respiratory viruses, including the molecular epidemiology of RV was investigated in the nested infant cohort. Viral protein regions 4 and 2 were targeted to investigate RV-genotypes. Simple descriptive statistics and regression models analysed associations and outcomes of interest. Results My nested subgroup of 72 infants provided 3446 swabs. Of these, RV (19.1%) had the highest detection rates followed by PyV-KIV (1.7%), hBoV (1.6%), AdV (1.1%), PyV-WUV (0.9%), RSV-A (0.6%), hCoV-OC43 (0.3%), PIV-3 (0.3%), hCoV-NL63 (0.3%), RSV-B (0.2%), hMPV (0.2%), PIV-1 (0.1%), IFV-A (0.09%), IFV-B (0.06%), hCoV-229E (0.06%), PIV-2 (0.03%) and hCoV-HKU1 (0.03 %). Failure to detect ERV-3 was associated with 60% reduction in virus detection rates in nasal swabs. Mould was observed in 23% of samples and associated with delays in transportation, season and reduced ERV-3 and respiratory virus detection. Degenerate oligonucleotides may overcome season-to-season variation in viral gene targets. Compared with the established assay, the new HAdV assays provided similar qualitative, but superior quantitative results. Serial RV-detection for more than 3-4 weeks was from genotype replacement rather than prolonged shedding. Although detected in asymptomatic infants, an association was found between RV and respiratory symptoms, especially for the RV-C species. Conclusions Longitudinal studies help further understand respiratory virus detection, viral shedding and disease burden in the community. The quality of nasal swab collection and transportation can be monitored in real-time using the human DNA marker ERV-3. Gene target variation is a potential problem for longitudinal studies and was addressed successfully in HAdV real-time PCR assays by using combinations of degenerate oligonucleotides. This developmental work allowed me to show that in otherwise healthy Australian infants RVs were the dominant respiratory pathogens, followed by the DNA respiratory viruses. The apparent prolonged shedding of RVs over more than 3-4 weeks was from genotype replacement rather than persistent infection. RV-C appeared more pathogenic than the other RV-species and if confirmed this will help identify a viral target for future novel therapeutic and public health