Finding Kawasaki disease

What lessons can we learn from the study of Kawasaki disease (KD) around the globe? Epidemiologic tools have traditionally been used to determine the etiology of infectious diseases, but, 50 years later, the application of these approaches to the study of KD has only yielded limited insights.[1,2] Although KD is the most common cause of acquired heart disease in children in countries where rheumatic heart disease has waned, the etiology remains a mystery.[3] The current paradigm is that KD is an immunologic reaction to an environmental trigger in genetically susceptible children.[4] Genes in pathways that modulate calcium signaling, activation of FCγ receptors, prostaglandin secretion, and myofibroblast transformation have been implicated in host susceptibility and coronary artery aneurysm formation.[5–9] Recent genome-wide association studies from Asia have uncovered genetic influences specific to Asian populations that may contribute to the increased incidence of KD in children of Asian descent.[6,10] However, while progress has been made in understanding genetic susceptibility, progress on etiology has proved a more daunting task. The lack of well-preserved autopsy tissues and the inaccessibility of the coronary arteries, the main target of inflammation, have certainly impeded progress. Recently, new clues from the study of KD epidemiology may help us focus the search for the KD trigger.

EXPLAINING THE LARGE EPIDEMICS IN JAPAN

Epidemiologists often focus on outbreaks or epidemics in their search for clues to etiology. Investigation of KD outbreaks in the USA variously suggested an association with carpet cleaning, residence near a body of water, humidifier use in the child's room, and antecedent respiratory infection.[11–13] However, none of these observations has been consistently replicated. The observations about seasonality and clustering of cases were first made by Yanagawa and colleagues in Japan in the 1980s.[14] A later refinement of those observations clearly established the seasonality of KD in Japan and in the Western US.[15] Temporospatial clustering of KD cases was also observed, which suggested an environmental exposure acting on a regional scale.[16] In Japan, three large nationwide epidemics were observed in 1979, 1982, and 1986. A new analysis of atmospheric variables in relation to these epidemics has uncovered a potential new clue: winds blowing from Central Asia in a southwesterly direction across Japan were highly associated with the dramatic peaks in KD activity, while a reversal of wind direction coincided with the cessation of the epidemics.[17,18] Further analyses demonstrated that the annual peaks in KD disease activity in Japan, Hawaii, and southern California also coincided with this tropospheric wind pattern. These observations raise several questions. First, will study of these wind patterns allow predictions of peak and trough KD activity in a given region? Second, from analysis of these wind patterns, can we make estimates of the incubation period between environmental exposure and onset of fever? Clearly, knowing when to look would help the search for the etiologic agent. Third, is it possible that these large-scale tropospheric wind currents associated with KD are actually carrying aerosols harboring the agent? If so, then determining the source regions of these aerosols would provide clues as to where to look for the agent. Experiments are in progress to attempt to answer these questions.

Old disease or new disease

To better focus our search for the etiology of KD, it would be very helpful to know if this is an old or newly emergent agent. Time series of KD cases from different countries suggest that, at least in some parts of Asia, KD is newly emergent. In Japan, where meticulous hospital records allowed this type of retrospective inquiry, it seems quite clear that the clinical entity of KD did not exist prior to World War II.[19–21] In India, this question framed a debate that was filmed in different parts of the country (YouTube, “Kawasaki Disease in India”).[22,23] In the West, infantile periarteritis nodosa (IPN), a vasculitis of young infants, was clearly observed in Western Europe, South America, and the USA as early as the turn of the last century.[24] Although IPN and KD are pathologically indistinguishable, it is difficult to determine whether they share a common etiology.[25] To answer this question, it would be helpful to have additional time series from countries around the globe and for researchers to attempt to find “the first case” in their region. Interviews with senior pediatricians, review of hospital and clinic records, and review of autopsy cases with coronary artery aneurysms can be productive avenues to pursue. As morbidity and mortality from vaccine-preventable diseases falls in different parts of the world, KD seems to emerge. But, whether this is due to increased awareness with increased case ascertainment or is due to rising numbers of cases is a matter of debate. One aspect on which we can all agree is that tracking and reporting KD from different countries, as in the article from Pakistan in this issue of the Annals,[26] will contribute to our understanding of how KD operates in different parts of the globe and may improve our search for the etiologic agent.