To our knowledge, this is the first population-based cohort study to examine the relationship between KD and the subsequent risk of allergic diseases (asthma and allergic rhinitis). Previous studies used either cross-sectional [16
] or case–control designs [19
] to investigate the prevalence of asthma/allergic diseases in KD patients and non-KD controls.
Previous studies reported that the majority of the increased asthma and allergy admissions occurred prior to the KD illness, suggesting that these allergy/asthma are unlikely to reflect immune dysfunction resulting from KD itself. It may also suggest that a distinct immune phenotype may be associated with an increased risk of both KD and asthma/allergy [19
]. In the acute stage of KD, there is sustained neutrophil activation [22
], with increased release of human neutrophil elastase and matrix metalloproteinases [23
], and similar patterns may be important in childhood asthma [24
]. Polymorphisms in the mannose-binding lectin (MBL) gene have been reported to be associated with susceptibility to KD [25
]. A role for MBL polymorphisms in susceptibility to asthma and allergic disease has also been reported [26
]. This suggests that immunogenetic variations in the innate immune response may contribute to the shared risk of KD and asthma/allergy [19
In this population-based cohort study, we found that KD patients were 1.52 times more likely than the comparison cohort to develope allergic diseases. These data may reflect the immunological consequences of KD, and/or an underlying susceptibility to both KD and asthma/allergy. The investigation of children following the KD illness does not differentiate between the two possibilities.
The mechanisms by which KD may increase the risk of future development of allergic diseases remain unclear. The immune system is highly activated during KD with a myriad of immunoregulatory changes including Th1 immune related response (such as IFN-gamma, TNF-alpha, IL-1, and IL-10) [27
] and Th2 immune related response (such as eosinophil, IL-4, IL-5 [30
]. There is evidence that the scar from prior BCG vaccination and the tuberculin skin test becomes inflated in patients during acute KD and is a marker of Th1-mediated delayed hypersensitivity reaction [32
]. Thus, if KD is indeed a strong Th1 trigger and Th1 is associated with a reduced risk of allergic diseases, we would expect that children with KD will have a lower risk of developing allergic diseases. Our data did not provide support for this hygiene hypothesis, where an inverse relationship between increased Th1 cell-mediated inflammation (KD) and decreased Th2-cell mediated allergy is observed [33
]. By contrast, we observed a positive association between KD and the risk of allergic diseases. The reasons for the differences in the findings are unknown. KD results in an acute inflammation. It has been shown that infections (inflammations) are able to alter the function of regulatory T cells [35
]. Our finding of an increased risk of developing allergic diseases following KD may therefore plausibly be related to the effect of regulatory T cells. Clearly, more work will be needed before the influence of immune activation of Th1/Th2 on the risk of developing allergic diseases is understood.
The major strength of our study is the use of a computerized database, which is population based and is highly representative and allows a clear observation of the temporal relationship between KD and allergic diseases. Because we included a national sample of KD patients between 1997 and 2005, and because the control subjects in this study were selected from a simple random sampling of insured general population, we can rule out the possibility of selection bias.
Several limitations of the present study should be noted. First, KD cases were identified using medical diagnoses recorded in the claim database. Data on the accuracy of KD discharge diagnoses is not available in Taiwan. Potential inaccurate data in claims record could also lead to possible misclassification of KD. Given the comprehensiveness of the NHI database and its previous use for incidence study for KD [12
], we are confident that the effect of misclassification of case status on our results is most likely of minor importance. Second, diagnoses of allergic diseases rely on administrative claims data may be less accurate than those obtained according to standardized criteria and misclassification is possible. However, this misclassification is likely to be nondifferential and therefore would tend to underestimate rather than overestimate the true relative risk of alleric diseases among KD patients. Third, any study investigating an increased incidence of one disease (allergic disease) conditional on the diagnosis of another disease (KD) in clinical settings may suffer from detection bias. In this study, the fact that KD patients had more physician visit on average than patients in the comparison cohort could make KD patients more likely to have allergies diagnosed. In addition, physicians might be more inclined to find allergic diseases in patients who had KD due to the fact that previous studies have shown the correlation between KD and allergic disease [16
]. Therefore, the possibility that our findings have arisen from detection bias could not be excluded. Fourth, although we adjusted for several potential confounders in the statistical analysis, a number of possible confounding variables, including body mass index and family history of allergic diseases, which are associated with allergic diseases were not included in our database. Lastly, as with any observational study, residual confounding by unmeasured factors which are different between case cohort and comparison cohort is also possible.