|Home | About | Journals | Submit | Contact Us | Français|
Risk factors for severe human rhinovirus (HRV) associated infant illness are unknown.
To examine the role of HRV in infant respiratory illness, and assess viral and host risk factors for HRV disease severity.
We utilized a prospective cohort of term, previously healthy infants enrolled during an inpatient or outpatient visit for acute upper or lower respiratory illness during fall-spring months 2004-2008. Illness severity was determined using an ordinal bronchiolitis severity score with higher scores indicating more severe disease. HRV was identified by real-time RT-PCR. The VP4/VP2 region from HRV positive specimens was sequenced to determine species.
Of 630 infants with bronchiolitis or URI, 162 (26%) had HRV; HRV was associated with 18% of bronchiolitis and 47% of URI. Among infants with HRV, 104 (64%) had HRV alone. Host factors associated with more severe HRV illness included maternal and family history of atopy (median score 3.5, IQR [1.0-7.8] vs. 2.0 [1.0-5.2], and 3.5 [1.0-7.5] vs.2.0 [0-4.0]). In adjusted analyses maternal history of atopy conferred an increase in risk for more severe HRV bronchiolitis (OR=2.39, 95% CI:1.14-4.99, p=0.02). In a similar model, maternal asthma was also associated with greater HRV bronchiolitis severity (OR=2.49, 95% CI: 1.10-5.67, p=0.03). Among HRV, 35% were HRVA, 6% HRVB, and 30% HRVC.
HRV was a frequent cause of bronchiolitis and URI among previously healthy term infants requiring hospitalization or unscheduled outpatient visits. Substantial genetic diversity was seen amongst the HRV, and predominant groups varied by season and year. Host factors including maternal atopy were associated with more severe infant HRV illness.
Human rhinoviruses (HRV) are the most common viral etiology of asthma exacerbations in adults and children2-11, and wheezing with HRV during infancy has been linked with the onset of asthma11-13. Although it is well-recognized that the majority of upper respiratory illness 1 episodes in adults and children are associated with HRV, HRV historically have not been thought to play a significant role in infant respiratory illness or morbidity. During recent years, in studies using sensitive reverse transcriptase-polymerase chain reaction (RT-PCR), HRV have been associated with a significant burden of disease in infants and young children14,15. Bronchiolitis, a lower respiratory infection in infants presenting with wheezing, rales, and respiratory distress, has typically been associated with respiratory syncytial virus (RSV) infection15-17, but a number of studies suggest that HRV are another important cause of bronchiolitis.15, 17-19
Host factors for HRV illness are poorly defined compared with RSV illness. Copenhaver and colleagues found that among high-risk infants of atopic parents, daycare attendance and siblings were risk factors for HRV-associated wheezing.20 However, the spectrum of HRV-associated illness among infants who are not at high risk for atopy, and host and viral risk factors for HRV-associated infant bronchiolitis are not well-defined. Some viral factors important in childhood respiratory morbidity have recently been elucidated in relation to wheezing illnesses in older children. These recent studies have found that a novel group of HRV, called HRV-C, is associated with a substantial burden of respiratory illness in sick children.1, 2, 21-31
We sought to assess the importance of HRV, including HRV-C, as a cause of URI or bronchiolitis among term, non-low birth weight, previously healthy infants without known risk factors for bronchiolitis and to examine host and viral factors that may contribute to HRV pathogenesis. A better understanding of host and viral risk factors for HRV-associated infant morbidity will have implications not only for understanding HRV pathogenesis, but for treatment and disease prevention strategies.
Clinical biospecimens and clinical data available from the Tennessee Children's Respiratory Initiative (TCRI) were used for this investigation, the methods for which have been previously described 32. The TCRI is a longitudinal prospective investigation of term, non low birth weight otherwise healthy infants and their biological mothers. The primary goals of the cohort are to investigate the acute and the long-term health consequences of varying severity and etiology of viral respiratory tract infections and other environmental exposures on the outcomes of allergic rhinitis and early childhood asthma, and to identify the profile of children at greatest risk of developing asthma following infant respiratory viral infection32. Infants aged ≤12 months were enrolled at the time of a clinical visit (hospitalization, emergency department (ED) visit or outpatient visit) for bronchiolitis or URI from Fall through Spring 2004-2008. Repeat infections were not included. Demographic and clinical data were collected, and nasopharyngeal swabs were obtained by trained nurses for viral testing, as described below. The children underwent an outpatient follow-up visit between ages 12 and 24 months in the Vanderbilt General Clinical Research Center (GCRC), where infant and maternal blood was obtained and an additional parental questionnaire completed.
Clinical evidence of maternal atopy was determined by clinical symptoms of an atopic disease as assessed by the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire (allergic rhinitis, asthma, and/or atopic dermatitis)32. Laboratory evidence of maternal atopy was ascertained by skin testing or specific IgE32. Maternal asthma was ascertained by using the ISAAC questionnaire32. Family history of atopy was determined by report of first degree relatives with an atopic disease.
Acute respiratory illness severity was determined using an ordinal bronchiolitis score that incorporates admission information on respiratory rate, flaring or retractions, room air oxygen saturation, and wheezing into a score ranging from 0 to 12 (12 being most severe).33, 34
Nasal and throat swabs were obtained from ill infants by trained nurses and placed together into both lysis buffer and viral transport medium. Specimens were taken immediately to the lab on ice, divided into aliquots and stored at -80°C until processed. RNA was extracted from 200 μl of pooled nasal/throat swab media on a Roche MagNApure LC automated nucleic acid extraction instrument and real-time RT-PCR for the detection of HRV was performed as described.35-38 Specimens were also tested for RSV, influenza viruses A and B, parainfluenza viruses types 1 - 3, human metapneumovirus, and coronaviruses (including OC43, 229E, and NL63) 35, 39-45Conventional RT-PCR was performed on HRV-positive samples using primers that amplify a fragment of ~ 548 nt encompassing the VP4/VP2 region.46 Amplified fragments were sequenced in the Vanderbilt DNA Sequencing Core, edited and aligned with MacVector version 11.1 (MacVector). Phylogenetic analysis was performed using Mega version 4 using 500 bootstrapped replicates and the neighbor joining algorithm with HRV87 as outgroup.47 Nucleotide identity between strains within each HRV species (A, B, or C) was calculated and mean diversity between HRV species was compared using a 2-tailed t test assuming unequal variance.
Host factors were presented using median and interquartile ranges [IQR] or frequencies and proportions as appropriate and were compared between HRV+ and HRV- (other or no study virus) using chi-squared test for categorical variables and Wilcoxon-rank sum test for continuous variables. To assess for adjusted effects of host factors on the bronchiolitis severity score we used a multivariable ordinal logistic regression (proportional odds model). Factors included were infant illness age, gender, gestational age, race, smoking, and co-infection status and maternal history of atopy. Separate adjusted models were used to assess the role of maternal asthma, family history of atopy, and laboratory and clinically assessed maternal atopy. To assess broadly the scope of HRV-associated illness, main findings are presented based on the 162 infants with HRV. Because coinfections with other viruses may affect clinical presentation, a sensitivity analysis restricted to the HRV group without coinfection for other tested viruses was performed to assess for host factors mentioned above and severity of bronchiolitis.
We examined the variation of HRV species by year graphically. Multinomial logistic regression was used to assess independent association between host factors and HRV species (A, B or C). Proportional odds model was used to examine whether HRV species were associated with outcome of respiratory illness severity independent of infant age, race and sex. The independent association of HRV species with length of stay outcome was assessed using the same covariates and the proportional odds model. R (version 2.10.1, www.r-project.org) was used for all statistical analyses. A p value <0.05 was considered significant.
630 infants with either URI or bronchiolitis were enrolled over four respiratory viral seasons from 2004-2008. 162 (26%) of these tested positive by real-time RT-PCR for HRV infection and are the focus of this investigation. Of all enrolled infants with bronchiolitis, 18% had HRV; of all infants with URI, 47% had HRV. Among the 162 infants studied with HRV, 81 (50%) had bronchiolitis and 81 (50%) had URI. One hundred four subjects had HRV only with no other virus detected, 44 (42%) of whom were hospitalized. Of the children infected with HRV alone, 37 (36%) were female, 32 (31%) were black, 18 (17%) Hispanic, 38 (37%) white, and 16 (15%) other race. Median infant age at enrollment for patients with HRV only was 20 weeks [IQR 8, 39] and maternal age 24 [21.0, 30.0]. Eighty-one infants (78%) with HRV only had public insurance, 15 (14%) private, 1(1%) other, and 7 (7%) were uninsured.
Table 1 depicts demographic and clinical differences in infants who had HRV-associated bronchiolitis versus HRV-associated URI. Infants with HRV-associated bronchiolitis vs. infants with HRV-associated URI were more likely to be white (57% vs. 30%, p overall for race distributions =0.003) and less likely to be on Medicaid (59% with HRV-associated bronchiolitis vs. 84% with HRV-associated URI, p=0.005), more likely to have a family history of atopy (77% vs. 59%, p=0.018), and more likely to have coinfection with another virus (42% vs. 11%, p<0.001). Fifty-four percent of infants with HRV-associated bronchiolitis had atopic mothers, vs. 40% of infants with HRV-associated URI (p=0.059).
Controlling for age, race, smoking, and gender, maternal history of atopy conferred an increase in risk for more severe HRV bronchiolitis (OR=2.39, 95%CI: 1.14-4.99, p=0.02). In a similar model, maternal asthma was also associated with HRV bronchiolitis severity (OR=2.49, 95%CI: 1.10, 5.67, p=0.03). Figure 1 depicts the odds ratios of risk factors for HRV-associated disease severity.
Of 455 infants enrolled with bronchiolitis, 18% had HRV (9% HRV only without coinfections). Of 175 infants enrolled with URI, 47% had HRV (36% HRV only). Of 104 infants studied with HRV only (no coinfections), 39% (41) had bronchiolitis and 61% (63) had URI. Clinical features of infants with HRV and RSV coinfection were similar to those of infants with RSV alone, thus coinfections were excluded from HRV analysis (data not shown). Demographic and clinical differences between infants with HRV alone vs. infants who did not have HRV (other or no study virus detected) are compared in Table 2. Infants with HRV alone were more likely to be older (20 [IQR 8, 39] vs. 11 [IQR 6, 25] weeks, p<0.001). There were racial distribution differences (p=0.005); a higher proportion of infants with HRV infection were black (31% vs. 21% white). Infants with HRV alone were also more likely to be outpatients (58% vs. 27%, p<0.001) and attend daycare (33% vs. 21%, p=0.012), compared to infants with respiratory illness associated with other or no study virus. Infants with HRV alone were more likely to have had a lower median bronchiolitis severity score (2.0 [IQR1.0, 5.5] vs. 5.0 [IQR 2.0, 8.0], p<0.001), require less supplemental oxygen (18% vs. 51%, p<0.001), and be diagnosed with an URI (61% vs. 20%, p<0.001). Duration of hospitalization for infants with HRV alone was 2.0 days [IQR 2.0, 3.0]) compared with infants with other or no viral etiologies of respiratory illness whose duration of hospitalization was 3.0 days [2.0, 4.0] (p=0.13). Infants with HRV alone were more likely to have a maternal history of asthma (29% vs. 18%, p=0.009), compared with infants with respiratory illness associated with other or no study virus.
Of the HRV positive specimens, 57 (35%) were species HRVA, 9 (6%) were HRVB, 49 (30%) HRVC, and 47 (29%) could not be sequenced (U=untyped). Though the number of infants hospitalized with HRVB was small (n=7), 56% required supplemental oxygen (vs. 25% HRVA vs. 18% HRVC, p=0.059). Infants with HRVB had a longer duration of stay (median 4 days [IQR 3, 5.5] for HRVB vs. 2 days [2, 4] for HRVA vs. 2 days [1.25, 3] for HRVC, p=0.038). Among patients with HRVC, there were racial distribution differences, with a higher proportion of infants who had HRVC infection being of black race (41% HRVC vs. 0% HRVB vs. 26% HRVA, p=0.019). Maternal atopy was identified in 67% of infants with HRVB, compared with 46% with HRVA and 45% with HRVC (p=0.47). When comparing HRVA to HRVC, infants with HRVC were less likely to attend daycare than those with HRVA (20% with HRVC vs. 38% with HRVA, p=0.048).
In our multinomial logistic regression analysis, there was no difference in infant age at illness, or sex, between HRV species. In this adjusted analysis, the odds ratio for the association of race with HRV species was for white vs. black for HRVC vs HRVA, OR=0.39 (95% CI: 0.15, 1.00, p=0.049). In an ordinal logistic regression model adjusting for age, gender, and race, HRV species factor were not associated with increased bronchiolitis severity (p=0.078). In a similar adjusted regression model, species was significantly associated with length of stay (p=0.024). HRV species B were associated with increased length of stay compared with species A (OR=6.26, 95CI: 1.11, 35.48). Our results are limited by sample size, particularly for our multinomial logistic regression analyses that use multiple equations with only 9 patients in the HRV B species group.
There was marked variability between HRV species by season and by year, as illustrated in Figure 2a and Figure 2b. In addition to peaks of HRVA or HRVC during certain seasons and years, there was also a peak in untyped HRVs seen in January 2008.
The genetic variance of HRV strains in this cohort is illustrated in Figure 3. While some clinical differences were seen between patients infected with different HRV groups, there were also broad genetic differences among viruses. The VP4/VP2 sequence regions correlate with the serotype classification of HRV 46, 48, 49. HRVC strains were defined using a nucleotide identity of <90%, based on the calculated genetic relatedness of established serotypes as previously described50. The HRVC genotypes had a mean nucleotide identity of 66.2% (minimum 23.3% and maximum 99.6%), compared to HRVA (mean 72.0%, minimum 31.0% and maximum 99.8%) and HRVB (mean 73.5%, minimum 54.7% and maximum 99.8%). Thus HRVC viruses exhibited a substantial amount of diversity among themselves, significantly greater than that observed for HRVA (mean diversity HRVA vs. HRVC p<0.001) or HRVB (mean diversity HRVB vs. HRVC p<0.001).
HRV is recognized as a virus of older children associated primarily with URI,6, 9, 51 but the role of HRV in term, non low birth weight, previously healthy infants who are not selected to be at high risk for atopy has not been as well studied. We detected HRV in 18% of bronchiolitis cases and 47% of URI among infants enrolled over four respiratory viral seasons. Thus, HRV appears to be an important cause of infant bronchiolitis. Kusel et al. reported that among 263 infants with ARI during the first year of life, half were associated with HRV with HRV being the most common virus detected in children with upper and lower respiratory illness.52 However, this finding has not been consistent. Another group identified HRV in only 9% of infants with bronchiolitis,14 and others have detected HRV in 7-17% of bronchiolitis in children <2 years of age53, 54. This variability in apparent HRV burden may depend on both the study population and the years of study, as suggested by the differential circulation of HRV in each of the seasons studied in this investigation.
Infants with HRV alone were more likely to be older, black, attend daycare, and have a maternal asthma history, compared with infants with respiratory illness associated with other viruses or with no virus detected. Maternal atopy conferred over twice the risk of more severe HRV-associated respiratory illness as determined by the bronchiolitis severity score, independent of age, race, smoking, and gender. Korppi et al. compared infants <24 months of age with HRV vs. RSV-associated wheezing and reported that infants with HRV were older and more often had atopic dermatitis and eosinophilia.55 These data are consistent with our findings that infants with HRV-associated respiratory illness are older and that an atopic profile in the mother may be an important risk factor for more severe HRV-associated infant disease. Other cohort studies, such as the COAST study, have elegantly described risk factors for infant virus-associated wheezing and subsequent early childhood wheezing/asthma; however, all infants in those cohorts had one or both parents with atopy or asthma and thus constituted a high-risk population.11, 12, 20 In contrast, our study included infants with no history of maternal atopy as defined by the ISAAC questionnaire; 56% (354/630) of the infants enrolled did not have a history of maternal atopy. While there are not immediate clinical implications from our study for asthma development, children in this cohort are being followed until the sixth year of life and the outcomes of asthma and allergic rhinitis. This is the first study, to our knowledge, to find that maternal atopy is an independent risk factor for HRV disease severity among term non low birth weight infants who are not selected to be at high risk for atopy, suggesting that there is a genetic predisposition to HRV severity, and that risk factors for atopy may be linked with risk factors for innate immunity to HRV. Airway epithelial cells from asthmatic subjects have been shown to exhibit aberrant responses to HRV infection in vitro56, 57; our findings suggest that these mechanisms may affect clinical disease.
We found a large proportion of HRV to be the newly described group HRVC. The diversity seen amongst HRVC strains was greater than that seen amongst HRVA strains, consistent with other studies.1, 2 All of the recently described HRV for which the VP4/VP2 sequences are available fell into the same group, including the QPM strains from Queensland that were originally classified as a subgroup of HRVA 21, 23, 31, 50, 58-63. Infants with HRVB were more likely to require supplemental oxygen and have a longer duration of hospitalization compared to infants with HRVA or HRVC. Consistent with these findings, infants with HRVB tended to have higher bronchiolitis severity scores, although the number of HRVB strains was small as in other studies,1, 2, 64. HRVC was more commonly seen in black infants, compared with HRVA or HRVB. Our prior study also found HRVC to be significantly more common in black children <5 years of age who were hospitalized with ARI or fever during one year and in one study site, but this effect was not significant during both study years and in both sites in that investigation.2 One cannot assume ethnic differences in predisposition to HRVC infection based on these limited data, as the effects could have merely been seen based on viral circulation patterns amongst different geographic or socioeconomic regions, but these findings suggest that both host and viral genetic factors may be important in disease pathogenesis.
A recent multi-year population-based study of HRVC in two U.S. locations found that HRVC was associated with medically diagnosed wheezing and asthma more frequently than other HRV species. 50 Other reports also suggest that viruses in the HRVC group are more likely to be associated with wheezing 21, 23, 31, 58, 61, 62. However, we did not find a significant association between HRVC and bronchiolitis severity in the current study. It is possible that clinical features amongst children who are ill with certain strains or species of HRV differ based on age or other demographic or epidemiologic factors, and we only enrolled children <12 months of age. In addition, specific genotypes within species A, B, or C may be more likely to be associated with increased disease severity. Further studies are required to address these possibilities.
When examining frequency and proportions of HRV species by year in our study, there are clearly annual variations and alternation of the two major species during a single year, and sometimes recurrence at certain times of the year. Of the studies that have tested for HRVC to date, data on seasonality of HRVC is not conclusive. While many studies have detected HRVC as the predominant HRV species in the fall, suggesting it may play a role in the so-called September asthma epidemic21-23, other studies in different regions and years have found otherwise. Han et al reported HRVC to be more common in the spring and HRVA in the fall in South Korea in 2006, but co-circulation occurred during both seasons28. Consistent with our studies, Lau et al found HRVA and HRVC to alternate as the most common HRV species at different times during the peak seasons65. Alternate disease activity by species and seasons suggests possible viral interference or serological cross-protection between HRVA and HRVC65. While one study reported HRVC year-round without peaks in the spring and fall 25, most studies do suggest that HRV species vary by season, year, and geographic location. This variation underscores the importance of studies over multiple years and seasons to understand fully the geographical and seasonal HRV epidemiology.
Despite the strengths of our prospective cohort study, there are limitations that should be noted. First, we did not test concurrent healthy controls to determine the prevalence of asymptomatic HRV infection, which is well described. 52,53,24 Since HRV can cause asymptomatic infection, our data suggest but do not prove that HRV infection was the etiologic agent for the respiratory illnesses, bronchiolitis and URI. However, we performed highly sensitive molecular testing for the spectrum of viruses known to cause infant respiratory viral illness, including influenza, RSV, HMPV, PIV1-3 and HCoV including OC43, 229E, and NL63 (data not shown) with no other virus detected in 73.5% of the HRV-positive children, strongly suggesting that HRV was the causative pathogen. Other studies support these findings.54-55 In addition, this study cannot delineate population-based rates of HRV-associated LRI and URI, as the cohort did not equally enroll hospitalized and non-hospitalized infants, with nearly two-thirds of the cohort being hospitalized infants, thus over representing infants with bronchiolitis. A final limitation was that we evaluated only one geographic site. However, we report a comprehensive study over 4 years in a site that captures over 90% of Nashville Davidson county infant hospitalizations.
In this study, HRV were frequently associated with bronchiolitis and clinically significant URI in previously healthy term infants. Predominant HRV groups varied by season and year, with substantial genetic diversity. HRVB was associated with more severe disease in univariate analyses. Maternal atopy and asthma were associated with more severe infant HRV illness. This finding suggests that there are underlying susceptibility factors for HRV illness severity that are strongly linked with asthma susceptibility. As HRV is the virus most commonly associated with acute asthma exacerbations in children, future reports from this cohort, as longer term outcomes such as recurrent wheezing and asthma are identified, will help to clarify the complex relationship between host susceptibility and HRV infection.
The authors thank the families enrolled in the Tennessee Children's Respiratory Initiative.
Funding Declaration: This work was supported by KL2 RR24977-03 (to EKM), Thrasher New Investigator Award (to EKM), Thrasher Research Fund Clinical Research Grant (to TVH), NIH mid-career investigator award K24 AI 077930 (to TVH), UL1 RR024975 (Vanderbilt CTSA), and NIH K01 AI070808 mentored clinical science award (to KNC). TVH is also supported by NIH U01 HL 072471, R01 AI 05884, R01 HS018454, R01 HS019669 and NIH K12 ES 015855. JVW has served as a consultant for MedImmune, Novartis, and serves on the scientific advisory board for Quidel.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Clinical Implications: Maternal atopy is a risk factor for severe rhinovirus illness in infants, suggesting that there are underlying susceptibility factors for HRV illness severity that are strongly linked with asthma susceptibility.
Capsule Summary: Human rhinoviruses were frequently detected in previously healthy term infants with bronchiolitis and upper respiratory illness. In infants with rhinovirus, maternal atopy was a risk factor for greater disease severity.