Wheezing is a frequent manifestation of lower respiratory tract infection (LRTI) in infants and young children. Viral infections are the commonest cause of acute wheezing. Several respiratory viruses, including respiratory syncytial virus (RSV), influenza viruses, parainfluenza viruses, enteroviruses, human coronaviruses, human metapneumovirus and human bocavirus have been associated with wheezy illness [1–5].

With the improvement of molecular techniques the frequency of HRV detection in clinical samples has increased dramatically [6–8] providing increasing evidence that HRV infection may be associated with LRTI including bronchiolitis, pneumonia, asthma exacerbations or influenza-like illnesses [9]. A recent population-based study showed that HRV was detected in 26% of children under 5 years of age hospitalized with respiratory symptoms or fever [10]. Subsequently, studies in high income countries have confirmed the importance of HRV as a cause of severe LRTI in young children requiring hospitalization [11, 12]. It is unclear if the diverse spectrum of clinical illnesses associated with HRV infection is related to host factors, the infecting HRV type or both. Recent evidence suggests that infection with HRV-C may result in more severe disease [12–16].

Since the first isolation of HRV in 1953 [17] approximately 100 serotypes have been described and new types are being discovered indicating that this genus is considerably more varied than previously recognized. Based on sequence analysis, antiviral susceptibilities and receptor usage HRV was until recently divided into 2 groups; HRV-A and HRV-B [18]. However, a third and possible fourth grouping, HRV-C and HRV-D, have been identified after sequence analysis of HRV types identified some which did not cluster with HRV-A, HRV-B or other species within the genus Enterovirus [15, 16, 19–23]. HRV-C has a global distribution, with a prevalence intermediate with HRV-A and HRV-B [24].

The importance of HRV as a cause of acute wheezing illness in infants and young children has not been studied in African children. The aim of this study was to investigate the prevalence of HRV in African children with acute wheezing.

HRV was detected throughout the year although there was a peak in spring (September-November of 2004 and 2005) and another in February-April 2005 (autumn) (Figure 1).

Co-infection with hMPV, HBoV and HCoV-NL63 was found in 8 (6.3%), 6 (4.7%) and 1 (0.8%) of the 128 HRV-positive samples respectively. Of the 92 HRV-negative samples 5 (5.4%) were hMPV positive, 10 (10.9%) HBoV positive and 3 (3.3%) HCoV positive. Of the 58 samples screened for the common respiratory viruses by general shell vial culture 14 (24.1%) were positive; 4 (6.9%) HRV-positive and 10 (17.2%) HRV-negative.

Sequencing of the 5' UTR was performed on 71/128 (55.5%) positive samples using convenience sampling ensuring each month of the study period was included, but without prior knowledge of patient data. Figure 2 shows the phylogenetic distribution. HRV-C can be divided into 2 variants, Ca or Cc, with either HRV-A-like or HRV-C-like 5'UTR's respectively. HRV-Ca sequences (n = 24) were interspersed within the HRV-A region of the phylogenetic tree. The remaining HRV-C samples (n = 13) formed a separate branch with the HRV-C-like 5'UTR. HRV-C was the most prevalent (37/71; 52.1%) followed by HRV-A (26/71; 36.6%) and HRV-B (8/71; 11.3%). HRV-C was found throughout the year while HRV-A had peaks in the spring and autumn seasons and HRV-B was sporadically detected (Figure 3).

Presenting symptoms or signs in HRV-infected and uninfected children were similar, except for fever which was less common in HRV-infected group of children and vomiting which was more common (Table 1). Most children with HRV infection had mild illness, although 45 (35.2%) required hospitalisation and 3 (2.3%) were admitted to the intensive care unit. The median (IQR) duration of hospitalisation was 1 (1-4) days; there was no difference in the duration of hospitalisation between HRV-infected and uninfected children. There were no in-hospital deaths. Of those admitted to hospital and where a HRV sequencing result was available (n = 15) 8 (25.8%) were typed as species C, 5 species A (16.1%) and 2 species B (6.4%).

HRV was the commonest viral infection in young African children presenting to a tertiary children's hospital in South Africa with acute wheezing. HRV has previously been reported in older children with wheezing or asthma exacerbations [5, 29]. However, there is now accumulating evidence that HRV-associated illness occurs in young children with LRTI with many studies from other areas of the world reporting HRV detection rates from 24% to 48% [11, 30, 31], and as high as 78% in predisposed children [32]. Consistent with these, 72% of the HRV-positive children were under 2 years. This data is also consistent with prior studies that report mild to moderate illness in most children [11], although in this study a third required hospitalization. The study suggests that HRV is responsible for a large burden of illness in young African children with associated implications for health care utilisation.

HRV has also been associated with recurrent wheezing and with the development of asthma in older children [32–35]. HRV-induced wheezing during the first 3 years of life was associated with a 10 fold risk of developing asthma by six years in a USA cohort of infants at high risk for developing asthma [32]. Asthma is the most common chronic childhood illness in South African children affecting 10-20% of adolescents; the prevalence has increased over the last decade [36]. The role of HRV in the inception and development of asthma and in triggering acute asthma episodes in African children needs further study.

Consistent with other studies, HRV was identified throughout the year with peaks of activity in the spring of 2004 and 2005 and autumn of 2005 [3, 10, 35]. Interestingly this pattern was more prominent with HRV-A than HRV-C and HRV-B which were detected throughout the year or sporadically. Seasonal variation is a common feature of many respiratory viruses, including RSV, influenza, parainfluenza virus, and although the causes of this variation are largely unknown, one possible explanation may be as a result of viral interference where one respiratory virus predominates and prevents other viruses from establishing infection at the same time. As the study was of 18 months duration any inferences about seasonality are difficult to make. However, the study period included 2 winter seasons. Further long term studies of the epidemiology of viral infections and HRV specifically in African children are needed.

Phylogenetic analysis of the 5'UTR of HRV-positive samples showed the circulation of all 3 species in Cape Town, South Africa. In this study, and as reported by Huang et al. [19], HRV-C was the most prevalent followed by HRV-A and HRV-B. This distribution pattern is different from other studies where HRV-A was more frequently detected; this may vary depending on the population [21, 24, 37]. Consistent with previous studies HRV-C was the common type in children with wheezing and asthma [13, 21, 37]. Typing using the 5'UTR may underestimate the true number of HRV-C samples as there is evidence that recombination with HRV-A species at this site is a common occurrence, resulting in these types grouping with HRV-A species instead of HRV-C [19, 38]. However, limited sequencing of the VP4/VP2 region (data not shown) and inclusion of sequences of HRV-C with HRV-A-like 5'UTR's in the phylogenetic analysis confirmed that HRV-C was the most common species in this population, with HRV-Ca more prevalent (65%) than HRV-Cc. Further confirmation is that all HRV-Ca sequences from this study grouped phylogenetically with GenBank-derived HRV isolates that have been confirmed to be true HRV-C types based on the classification proposed by Simmonds et al [27]. As most children presented with mild or moderate illness, and as only a subgroup of HRV isolates were sub-typed, we were unable to investigate the association between HRV species and clinical severity; this requires further investigation.

The study has some limitations. A control group of asymptomatic or clinically well children from the same time period were not included. However, previous studies have reported that HRV is more frequently found in sick children than in asymptomatic children [4, 11]. HRV can be shed from the nasophaynx for 10 to 14 days in immunocompetent individuals with extended shedding in immunocompromised patients [9]. The immune status of children enrolled in this study was not recorded, although it can be inferred that some children may have been HIV-positive as South Africa is a high HIV burden country. Thus detection of HRV in some children with wheezy illness may therefore be an incidental finding due to extended shedding. Due to resource limitations, only a subset of specimens could be tested for the common respiratory viruses (RSV, influenza virus A and B, parainfluenza viruses 1, 2, 3 and adenovirus) using the relatively insensitive general shell vial culture assay which does not provide sufficient information to determine the role of co-infection and disease severity. Finally, the study did not test for all known respiratory viruses (e.g. human coronavirus 229E and OC43, and human parainfluenza type 4) and recently discovered viruses (e.g. human coronavirus HKU1 and polyomaviruses K1 and Wu) which may also play a role in wheezy illness.

HRV may be the commonest virus in young African children with acute wheezing. Infection is associated with mild or moderate clinical disease. Longitudinal studies with frequent samplings and HRV type identification are needed to investigate the role of HRV in wheezing illness in African children and their long term outcome.