Kawasaki disease (KD) is an acute febrile illness predominately affecting children <5 years of age (75%–80%). KD patients develop a systemic autoimmune-like vasculitis leading to widespread inflammation of blood vessels and the small- to medium-sized arteries [1
], which also demonstrates a predilection for the coronary arteries in particular. When left untreated, 20%–25% of KD patients develop coronary artery lesions (CAL) such as dilation, aneurysm, or fistula that can result in potentially fatal thrombosis or sudden cardiac failure [2
]. Prompt detection of KD is required so that treatment is initiated with intravenous immunoglobulin (IVIG) before the 10th day after febrile onset to significantly reduce the risk of CAL formation in KD patients [3
The cause of KD remains unknown and there is currently no test available to confirm its diagnosis. Therefore, KD continues to be diagnosed clinically through the exclusion of other similar illnesses. To satisfy a diagnosis of complete KD according to American Heart Association (AHA) guidelines [4
], a patient must present with a high-grade fever that lasts 5 days or longer and is resistant to both antibiotics or antipyretics, in addition to four out of the following five principal diagnostic features: (1) a polymorphous rash; (2) changes in the extremities; (3) oral changes; (4) conjunctivitis; and (5) cervical lymphadenopathy. The principal symptoms of KD often develop over a week and do not always appear together. Incomplete cases are mostly diagnosed from the detection of CAL by echocardiogram [5
Over the last few decades, KD has replaced rheumatic fever as the leading cause of acquired heart disease in the developed world. The highest rates of KD are observed in East Asia, particularly in Japan, Korea, and Taiwan where between 67.3 and 243.1 in 100,000 children <5 years develop the disease each year [6
]. Even when residing in the West, East Asian children are 10 to 20 times more likely to develop KD than children of other ethnic backgrounds. KD rates are markedly lower in both Europe and North America (4.9–26.2 in 100,000) [9
], while the highest rates of KD in the US are reported from the State of Hawaii (50.4 in 100,000) that has the largest predominate Asian population [10
]. The incidence of KD has been increasing in East Asia and also appears to be increasing across the world.
This has led to speculation that the genetic variations between differing ethnic groups may account for the significant ethnic disparity in KD rates. Several susceptibility loci in KD have been identified through the meta-analysis of previous genome-wide association study (GWAS) investigations, such as BLK
, and ITPKC
]. However, these genetic susceptibility markers confer risk to KD development in a passive manner [19
], and therefore sensitize individuals to the currently unknown environmental agents that induce KD in susceptible children [20
]. As noted by Onouchi [19
], the ethnic disparities in KD rates also “… give the impression that KD is a type of endemic disease or a disease related to a lifestyle commonly found among the East Asian populations
The clinical and epidemiological findings of KD that closely resemble an infectious disease have also given rise to suspicion that microorganisms could be causative agents of KD. Various bacterial and viral ailments such as measles or scarlet fever are often initially misdiagnosed in KD patients, which can lead to the unnecessary use of antibiotics or—most importantly—potential delay in required time-sensitive IVIG treatment. Subsequent efforts to isolate an infection since Kawasaki first established the illness internationally during the 1970s have been consistently met with conflicting results and no significant associations upon further replication. However, significantly higher rates of concurrent viral infections are observed in acute KD patients compared to healthy controls [21
]. Researchers also note positive confirmation of pathogenic organisms in suspected KD should be interpreted with caution and not lead to immediate elimination of KD in the differential diagnosis [22
], particularly detection via polymerase chain reaction (PCR).
Earlier authors initially suggested mercury (Hg) from pollution and seafood may be a cause of KD and account for the significantly higher rates of KD in Asian children [25
]. These authors noted the striking clinical similarities between KD and infantile acrodynia, a childhood mercurial reaction that can mimic the complete clinical picture of KD and is therefore included in its differential diagnosis [5
]. One reported case of acrodynia in a 13-year old female playing with elemental Hg fulfilled all six principal diagnostic findings for KD and was diagnosed after very high urine Hg levels were determined (580 µg/L; normal <10 µg in 24 h) [27
]. Significantly elevated urine Hg levels from n
= 6 acute KD patients compared to age-matched hospitalized controls (13.6 vs.
1.2 µg in 24 h; p
= 0.018) had been previously reported by Orlowski and Mercer [28
]. Two other investigations reported normal urine or hair Hg levels [29
], although proper control groups were not used.
Previous examinations of Hg levels in the hair or urine of KD patients only involve a fairly small total of samples (n
= 16). Recently, we reviewed how ITPKC susceptibility loci for KD would render individuals vulnerable to even low-level exposure to Hg [20
], while Hg exposure can be difficult to detect in children through blood or urinalysis alone even when chronically exposed [31
]. Therefore, we decided to investigate whether variations of Hg exposure among ethnic groups were associated with KD risk in US children <5 years of age by using blood Hg levels from a large sample of infants and young children aged 1–5 years in the general U.S. population.
KD is an autoimmune-like vasculitis driven by markedly elevated inflammatory and immune cells. An infectious agent has not been consistently isolated in KD and the cause of KD remains unknown despite nearly half a century of extensive international investigation. The prevailing theory for causation involves a pathogen or other environmental agents that trigger an allergic or hypersensitivity type of response, which only results in clinical KD among susceptible children through mediating factors such as genetics and possibly other idiosyncrasies. Serum immunoglobulin E (IgE) and interleukin-4 (IL-4) are significantly elevated in acute KD patients [37
], which supports a hypersensitivity theory. IgE and IL-4 levels are used as biomarkers for the severity of asthma and allergic diseases. In addition, children with KD are significantly more likely to develop asthma and common allergies both before and after the onset of KD, while children with an onset of allergic diseases during childhood are also significantly more likely to develop KD [38
]. Pollution and passive exposure to environmental tobacco smoke are well-known for their associations with the development and exacerbation of asthma and allergic diseases [40
]. Seasonal pollen exposure was recently observed to be associated with KD in Japan and attributed as a delayed hypersensitivity reaction [44
Rodó et al.
] previously reported seasonal shifts in large-scale wind currents originating from China and Central Asia significantly associated with downstream KD cases in Japan, Hawaii, and Southern California. In their most recent study, yeast pollution from the soil of cereal crops in the heavy agricultural and industrial region of Manchuria in China appeared to account for the significant association of KD cases in neighboring Japan [46
], which suggests KD is caused by pre-formed toxins. Metal contamination with Cd and Hg in particular is one of the most pressing issues currently facing soil quality in China as a result of rapid industrialization [47
]. Previously, the contamination of soil for rice crops with Cd from mining pollution resulted in an appearance of Itai-Itai disease in Japan from the 1940s to the 1960s [48
]. Local seafood highly contaminated with methylmercury led to the initial appearance of Minamata disease in Japan during the 1950s and the 1960s [49
], which also corresponds to the initial appearance of KD in Japan after the end of the 1940s during the 1950s and 1960s [50
]. Lastly, toxic metals impact soil microbiota and yeast growth [51
There is no physiological role of Cd, Hg, or Pb in the body, which are considered to have no safe doses for exposure and are of notable public health concern, particularly in young children who are most susceptible [52
]. Hypersensitivity or idiosyncratic reactions to Hg such as infantile acrodynia often occur at levels considered to be normal, low, or subtoxic doses [31
]. We observed that increased blood Hg was significantly associated with increasing ethnic KD risk (Figure 1
), although our average blood Hg levels in all ethnic groups are considered to be normal or subtoxic (<0.1 µg Hg intake per kg bodyweight per day) [53
]. Researchers now stress the use of Hg and Se molar ratios in blood or examined tissues to more precisely determine methylmercury toxicity from seafood consumption, as there appears to be a direct relation between the toxicity of methylmercury and Se content [54
]. However, our average blood Se levels did not significantly differ between ethnicity in our current study, while decreasing Hg exposure appeared to be far more important than increasing Se intake as has been previously reported by other authors [55
We used total Hg in the blood as it reflects recent Hg exposure, particularly from fish consumption. The half-life of methylmercury is much longer in the blood than other forms of Hg such as the inorganic forms [56
], while methylmercury in the blood is converted into other forms of Hg after its absorption in the body. Asian children in the U.S. between 1 and 2 years of age have the highest average dietary intake of methylmercury by body weight per day compared to the rest of the U.S. population as a result of significantly higher rates of seafood consumption [57
], which also parallels the peak age of incidence for KD in East Asia [6
], in addition to the US [10
]. Dietary seafood intake significantly increases Hg levels in the blood of young children [54
], particularly large predatory fish such as tuna, swordfish, and shark that contain the highest methylmercury levels as a result of aquatic bioaccumulation. Our results suggest that the supplementation of large predatory fish with seafood that instead contains lower levels of Hg may be preferable, particularly in vulnerable populations such as young children and the developing fetus.
This current study is accompanied by certain limitations. Hospitalization rates for KD were underestimated as a result of underreporting in all ethnicities except for Caucasians. In addition, demographic data regarding self-reported ethnicity from the KID and NHANES cohorts is generalized and does not elaborate on specific Asian ethnicity, even though East Asians have the highest KD risk even compared to other Asian ethnicities. Our investigation involved separate cohorts for KD risk or blood values from two different time periods, and therefore did not examine Hg levels or other blood samples in KD patients. We also did not include a complete accounting of KD rates in the U.S. since 1997 and instead included the available reporting years for 1997, 2000, and 2006 that were then subsequently averaged together. However, our collected KD rates should accurately reflect the current incidence once combined, as KD rates have remained relatively stable in the U.S. from 1997 to 2007 [36
]. In addition, average blood Hg levels in children aged 1 to 5 years have not changed significantly since 1999 [60
], which is when the NHANES first began testing for this metal.
Our current study was modeled after our previous examination of ethnic KD risk with the consumption of soy and isoflavones [61
], which could possibly be mediated by FCGR2A susceptibility in KD [62
]. However, we did not examine the impact of dietary soy or seafood consumption upon ethnic KD risk among children in either of our cohorts. Prenatal and postnatal sources of Hg exposure were also not examined in our cohorts. Blood Hg was not examined in women who are pregnant or of childbearing age, although it is recognized that exposure to methylmercury from seafood consumption and other forms of Hg cross into the placenta and result in prenatal exposure [63
]. Blood values from the NHANES are not determined in children during their first year of life, which is likely a result for blood draw limitations in this age group. In addition, we were not able to examine the role of genetic susceptibility markers in either of our cohorts. Average blood Cd was considered too close to the lower detection limit in our cohort (0.12–0.17 µg/L vs.
0.12 µg/L), which then led us to conduct a more reliable examination by using detection rates of Cd in the blood. Lastly, our findings require additional replication with larger samples of KD patients and controls, including in other ethnic and regional populations.
As this study involves a population-based comparison of two separate datasets for KD risk or blood levels respectively in the general U.S. population, we can only currently afford speculation with regards to potential mechanisms or dietary factors that may link exposure to Cd and Hg with ethnic KD risk. Several animal studies have demonstrated that even low-level exposure to mercurials induces autoimmune vasculitis in susceptible rodent strains [64
], which is characterized by T cell-dependent polyclonal activation of B lymphocytes, marked increases of serum immunoglobulins such as IgE and proinflammatory cytokines such as TNFα, detectable antinucleolar autoantibodies (ANoA), and widespread vascular immune complex deposition that may also be specific to various target organs such as in Hg-induced glomerulonephritis [67
]. Many of these processes are also observed in KD, although certain investigations are limited in KD patients and still require further study. In addition, certain reported mechanisms of Hg-induced autoimmunity in animal models still require verification in humans. This has led several researchers to conduct examinations of Hg-exposed populations from seafood consumption or mining in affected regions, including among workers and local communities in Amazonian Brazil [71
]. Some of these authors have further reported interactions between Hg exposure and infectious agents, as we also previously reviewed regarding animal models and synergistic toxicity between metals [20
We suggest a more in-depth study of Cd and Hg toxicokinetics conducted in a large pediatric population with a high burden of KD risk that is multiethnic and primarily Asian. Such a prospective study should include parameters of cumulative exposure in addition to paired blood, urine, and hair levels, as blood Hg levels primarily represent recent exposure while urine Hg levels largely reflect current excretion and hair Hg levels mostly represent cumulative long-term exposure. The highly erratic nature of Hg levels in infantile acrodynia often results in dramatic day-to-day variation in blood and urine samples, which has led many authors to emphasize the need of 24-h collections for urine determinations of Hg and an observation of their trends over the span of several weeks or months [75
]. Lastly, sources of Hg exposure in acrodynia patients are often occult or remain unidentified, which results from the myriad of exposure sources to this toxic metal in pediatric populations. Although pollution and the contamination of seafood and cereal crops may account for the primary exposure source in the general or patient population—lack of controlling for other Hg exposure sources or paired biomarkers in the blood, urine, and hair may result in the inability to significantly identity true associations that then become masked in the data.