Coronavirus and Kawasaki disease

March 2005

In the February issue of Infectious Disease News was a brief article about the identification of an apparently new strain of human coronaviruses that caused respiratory disease in children and an apparent relationship to the subsequent occurrence of Kawasaki disease (KD). The work was carried out at the Yale University School of Medicine; hence the coronavirus strain was identified as the New Haven strain, or HCoV-NH.

Since most of us as adult ID physicians are generally unfamiliar with a number of pediatric ID issues, including KD and its diagnosis and management, it seems appropriate to explore this new association more fully. The only exception to the previous generalization would likely be those among us who find themselves providing infectious disease consultation to pediatricians. (It would be of interest to know what proportion of adult ID physicians actually do that.)

The key articles reporting this work are by Frank Esper, MD, and a number of colleagues at Yale and are reported in the Journal of Infectious Diseases (2005; 191[4)]:492-502). In addition, interested readers should review the accompanying editorial by Kenneth McIntosh, MD, of Harvard Medical School, found on pages 489-491 of the same issue.

It started with SARS

In a real sense, the work of the Yale investigators grew out of the recent appearance of the severe acute respiratory syndrome (SARS) coronavirus and the remarkable feat of defining the genomic sequence of that virus. This quickly led to development of detection methods that eventually contained the SARS virus. As a result of all this research, however, it was possible to identify a gene that is highly conserved among the known human and animal coronaviruses and then to develop molecular probes that would identify target regions of that specific gene, a CoV replicase 1a gene. They then proceeded to screen respiratory specimens from 895 children younger than 5 years old with symptomatic respiratory disease that tested negative for RSV, influenza A and B, parainfluenza viruses 1 through 3 and adenovirus. The time frame for specimen collection was Jan. 1, 2002, to Feb. 14, 2003.

Of the 895 children, 79 (8.8%) were positive for the HCoV-NH virus; 67 of them tested positive only for that virus. Among those children, cough, rhinorrhea, tachypnea, fever and hypoxia were the most common clinical manifestations. Chest radiographs, when obtained, generally showed peribronchial cuffing, atelectasis and/or infiltrates. Two children died, although they had serious comorbidities. Most of the cases occurred during the usual respiratory disease season, ie, the first 10 weeks of the year.

Of interest is the fact that almost concurrently, a closely related and perhaps identical human coronavirus was identified by two groups of investigators in the Netherlands. Thus, this virus may already have a worldwide distribution. The investigators point out that there may be other, as yet unrecognized HCoVs that are as yet not described, a point also made by McIntosh in his editorial. These new viruses now join the human meta-pneumoviruses as recently described human respiratory pathogens. One hopes that subsequent studies will inform us further about the nature and extent of illness due to these agents in adolescent and adult populations.

Linked with KD?

More intriguing, however, was the second paper in this series, from much the same group of investigators, starting with one of the children in the first study who was positive for the New Haven coronavirus and who developed classic KD. They then analyzed respiratory secretions from 11 children with KD and 22 control children; eight (72.7%) of the KD children were positive for the HCoV-NH virus, but only one (4.5%) of the controls was. This association was highly significant, suggesting that HCoV-NH infection is somehow associated with KD.

KD, for those not fully familiar with this entity, is a systemic vasculitis of children that in its most serious manifestation results in aneurysms of the coronary arteries. It is said to be the most common cause of acquired heart disease in children, at least in the developed world, a “title” probably once held by rheumatic fever. “Classic” KD is defined as fever for five or more days, plus four of five of the following: (1) bilateral conjunctivitis; (2) erythema of the mouth or pharynx, strawberry tongue or stomatitis; (3) polymorphous rash; (4) erythema or edema of the hands or feet; and (5) nonsuppurative cervical lymphadenopathy. Alternatively, the diagnosis is also established if three of the criteria are met and there is evidence of coronary artery abnormalities.

The epidemiology of KD suggests a response to a respiratory infectious agent. For example, it tends to occur in the winter and spring months; it is rare in infants older than 3 months, suggesting protection by passively acquired maternal antibody; and it is often preceded by an apparent respiratory infection. A number of respiratory agents have been proposed in the past, including an unidentified retrovirus, parvovirus B19, Epstein-Barr virus, Chlamydia pneumoniae, toxin-producing Staphylococcus aureus or Streptococcus pyogenes and a few others as well. McIntosh points out in his editorial comments that the statistical strength of the initial association reported for several of these possible etiologies was almost as strong as that reported by Esper and his colleagues. None, however, have stood the tests of time and reproducibility by others in different locations.

Reasons to believe

McIntosh goes on to cite several reasons why the current coronavirus association may prove to be real. In addition to the epidemiologic evidence, immunopathologic studies have suggested that during the course of KD, some external agent causes a powerful immunoglobulin A (IgA) response in the respiratory tract, as well as in other organs, suggesting a response to a specific microbial infection. Finally, he pointed out the enormous size of the coronavirus genome, capable of producing highly varied pathogenic effects in both animals and man. He thus felt that there are reasons to be at least guardedly optimistic about the Yale findings.

Even if the coronavirus association proves to be correct, there is still much research yet to be done to understand the immunopathologic processes that lead to KD. Is it directly caused by the virus; is it an abnormal host response to the virus; is there an intermediate process involved? This promises to be a fascinating area as further research unfolds, so stay tuned!