Observational research in childhood respiratory diseases

Abstract

Acute respiratory infections (ARIs) are common during the first two years of life, when infants and young children experience six to eight ARIs annually. In this age group, ARIs are the most commonly managed problems in general practice. Between 3 and 6% of infants are hospitalised in their first year of life with a severe ARI illness. Information about the epidemiology of ARI in children is based on historic community-based studies, cohorts of hospitalised patients, emergency department or primary healthcare presentations, and more recently from birth cohort studies principally involving children at high-risk of asthma. However, these studies do not completely identify the burden of mild-to-moderate ARIs in the community.Recent community-based cohort studies have used sensitive polymerase chain reaction (PCR) assays. However, they have had one or more methodological limitations, including subject selection, length of study, non-representative sample populations, variable sampling frequency, and the lack of a control population. Furthermore, frequent detection of respiratory viruses in asymptomatic individuals questions their clinical and public health significance. Studies reporting the causal effect of individual respiratory viruses in ARI are needed to help address this question.The Observational Research in Childhood Infectious Diseases (ORChID) study was a four year prospective, community-based, longitudinal birth cohort study of ARIs in 158 healthy children from birth to two years of age. ORChID sought to minimise some of the methodological limitations of previous studies. Parents completed a daily symptom diary and collected weekly nasal swabs, which were tested against 17 respiratory viruses. Healthcare-seeking behaviour was recorded in a separate ‘burden’ diary.I found that young children experienced 0.56 (95% confidence interval (CI): 0.54, 0.59) ARIs per child-month. This equated to almost five cumulative months of respiratory symptoms during the first two years. Forty-eight percent of ARIs where a burden diary was completed initiated a visit to a family doctor. ARIs were associated with increasing age, the winter season, and childcare attendance.Studies examining respiratory viruses in neonates have largely been from neonatal units or neonates presenting to hospital with respiratory symptoms. I was able to show that respiratory virus infections were common (0.25 episodes per neonatal period, 95% CI: 0.18, 0.34), with diverse human rhinovirus (HRV) genotypes dominating (21/29; 72% of neonates with positive swabs). Almost 50% of respiratory virus infections in this period were asymptomatic. This subclinical shedding of all respiratory virus types complicates estimates of the true community burden of viral ARI in infants and young children.To explore this further, I investigated the relative contribution of individual respiratory viruses to ARIs by calculating the virus-specific attributable fractions in exposed (AFE) children of ARIs and lower respiratory tract infections (LRTIs). The overall incidence of virus infections was 978 (95% CI: 930, 1029) per 100 child-years in the first two years of life. Viruses were detected in 75% of ARI episodes, while 23% of weekly swabs were positive for viruses during asymptomatic periods. RNA viruses, including HRV, influenza, parainfluenza, respiratory syncytial virus (RSV), human metapneumovirus (HMPV), and human coronaviruses NL63 and OC43 were associated with a significantly increased risk of ARI symptoms. Support for causality was strongest for RSV (AFE 68%, 95% CI: 45%, 82%), and HMPV (AFE 69%, 95% CI: 43%, 83%) in children with LRTIs. In contrast, amongst the DNA viruses tested, only adenoviruses (AFE 29%, 95% CI; 12%, 42%) were significantly associated with an increased risk of ARI symptoms. Of HRV species, only HRV-C had a significant AFE result for LRTIs (AFE 22% (95% CI: 5%, 22%).I went on to examine the timing of detection for each of the 17 respiratory viruses tested for in the ORChID infant cohort. Determining timing of first virus detection episodes (fVDEs) for different respiratory viruses in infants and young children identifies risk periods and informs preventive interventions, including vaccination. The median age for first HRV infections was 2.9 months (25th–75th centiles: 1.6, 5.1), while for all other respiratory viruses combined the median age was >13.9 months. Overall 52% of first HRV detections were symptomatic, compared with 57-83% with the other first virus detections. Thus, infants and young children do not always experience respiratory symptoms with their first viral detection episode, and for some viruses, such as RSV, these commonly occur when maternal vaccines may no longer offer protection.Collectively, these findings highlight the important community-managed disease burden caused by respiratory viruses in early childhood. They also provide a wealth of information about the relationship between respiratory virus infection and symptoms of respiratory illness. The ORChID study uses modern molecular-based techniques over four respiratory seasons to address questions about respiratory virus acquisition and infection.