Kawasaki Disease Lacks Association With Human Coronavirus NL63 and Human Bocavirus

Abstract

To the Editors: Numerous reports have proposed an infectious etiology for Kawasaki disease (KD), but none of the discussed agents has been confirmed so far. Recently, an association of KD with the emerging respiratory pathogens, human coronavirus NL63 (HCoV-NL63), and human bocavirus (HBoV) was suggested.1,2 However, ensuing studies could not find any correlation.3 All trials employed PCR methods, though viral nucleic acids can usually be detected only during viremia or shortly thereafter. Hence, we used a serologic approach to examine a possible correlation between HCoV-NL63 or HBoV infections and KD. Twelve children (10 male, 2 female) suffering from complete KD and 9 children (5 male, 4 female) with incomplete KD were identified according to diagnostic guidelines presented elsewhere.4 Patients were aged 8 months to 14.9 years (mean: 4.5 years). Samples were obtained 3 to 10 days (mean: 5.5 days) after onset of fever, before medical treatment was commenced. As a control group, 33 children (18 male, 15 female; 6 months-14.5 years, mean: 5.2 years) were enrolled. All children originated from South-Eastern Germany and sera were collected between January 2006 and May 2008. The collection of specimens was approved by the University of Regensburg Ethics Committee and written parental consent was obtained. HCoV serology was conducted by use of a novel line-immunoassay, which allows the detection of antibodies to HCoV-strains 229E, NL63, OC43, HKU1, and SARS-CoV.5 Screening of HBoV-specific humoral responses was performed by ELISA as described previously.6 Additionally, all control sera and 19 of 21 samples obtained from KD children were screened for HBoV-DNA by quantitative real-time PCR as published elsewhere.6 Statistical analysis was performed by standard two-tailed χ2 test. A P ≤ 0.05 was considered statistically significant. Seroprevalences of HCoVs and HBoV in KD children and in healthy controls are summarized in Table 1. IgG specific for HCoV-NL63 was found in 48% of children with KD but also in 67% of healthy children (P = 0.27). Comparison of antibody responses to any other HCoV in KD children and in healthy controls did not reveal any statistical relevance. Our data are in line with previously published findings, revealing that HCoV-NL63 and HCoV-229E seroconversion occurs on average before children reach the age of 3.5 years.7 Recent coronaviral infections characterized by virus-specific IgM were neither overtly elevated in healthy controls (0%–3%) nor in KD children (5%–10%). However, due to the short time between onset of fever and sampling, some acute infections could have been missed. HCoV IgA-seroprevalence was not correlated with any of the 2 study groups. None of the tested individuals was detected positive for SARS-CoV.TABLE 1: HCoV and HBoV Serology and PCR Detection of HBoV in KD Children and Healthy ControlsHBoV-specific IgG and IgA were observed in 90% and 48% of samples obtained from KD patients, respectively. A comparable seroprevalence of bocavirus-specific IgG (82%, P = 0.63) and IgA (58%, P = 0.66) was found in the control group, confirming a previously reported high IgG seroprevalence in children between 4 and 5 years of age.6 For the first time, IgA-antibodies against HBoV were detected in both healthy children and KD patients. Yet, IgA's role during viral infection and pathogen clearance remains to be elucidated. HBoV-specific IgM was not found in any of the samples, whereas in another trial IgM was detected in 67% of individuals suffering from productive HBoV infection.6 Potential acute infections in the study groups were additionally excluded by PCR analysis. We detected productive infection in 1 healthy child (4.9 × 103 genome equivalents/mL serum), whereas no viral DNA was found in sera collected from children with KD (Table 1, HBoV PCR). In summary, our serologic data strongly suggest that there is no association between infections with HCoVs and/or HBoV and KD in children. It yet remains to be determined whether KD is caused by a single pathogen or whether it is the result of interaction of more than 1 etiologic agent. Furthermore, genetic predisposition and environmental factors might promote development of full blown KD. Christian Lehmann, PhD Richard Klar, MSc Juha Lindner, PhD Petra Lindner, PhD Hans Wolf, PhD Institute of Medical Microbiology and Hygiene University of Regensburg Regensburg, Germany Stephan Gerling, MD, PhD Pediatric Clinic St. Hedwig Hospital Barmherzige Brüder Regensburg, Germany