Investigation of kawasaki syndrome risk factors in colorado

Risk factors for Kawasaki syndrome (KS) were evaluated through a case-control study during an investigation of a KS cluster in Denver, CO. KS was associated with a humidifier in the child’s room (odds ratio, 7.3; 95% confidence interval, 1.8 to 29.3) and possibly with an antecedent respiratory illness. The use of humidifiers should be further investigated as part of future studies of KS. Kawasaki syndrome (KS) is an illness of unknown etiology that generally affects children <5 years of age, 1 and it is the leading cause of acquired heart disease among children in the United States and Japan. 2 Outbreaks of KS have been reported in the United States, including Colorado. 3–6 Investigations of these outbreaks have shown that the occurrence of KS was associated with exposure to recently shampooed carpets, residing near bodies of water and/or an antecedent respiratory illness. 3, 4, 6 However, a case-control study investigating KS clusters in counties in New York and Michigan during the winter of 1982 to 1983 did not find an association between carpet shampooing or antecedent respiratory illness and KS onset. 5 To evaluate these and other potential risk factors, a matched case-control study was conducted during an investigation of a KS cluster that occurred from November 1997 through April 1998 in Denver, CO. Methods. The KS patients in the Colorado cluster were identified through review of cases reported to the Colorado Department of Public Health and Environment and through review of statewide hospital discharge records. The International Classification of Diseases, 9th revision, Clinical Modification code for KS (446.1) was used to identify KS patients in the state’s hospital discharge database. 7 KS patients were also identified by contacting the infection control nurse or medical records administrator of all hospitals in Colorado that admitted pediatric patients or had >100 beds. KS patients who met the CDC case definition for KS were included in our study. 8 The KS incidence for <5-year-old children during the 6-month period was calculated by using the Colorado census data as the population denominator. 9 The Greater Denver area included the contiguous counties of Adams, Arapahoe, Denver, Douglas and Jefferson. The controls were obtained through hospital databases of live births in the counties of each case residence. Database information was used to match two controls to each case patient by age (<18 month olds matched ±3 months; ≥18 month olds matched ±6 months), sex, race and area of residence. Cases and their matched controls were interviewed by telephone as a set by the same interviewer using a standardized questionnaire. Information about possible risk factors was collected from the primary caretaker(s) of each case and control. The following risk factors were assessed: (1) respiratory illness in the child (cough, sore throat or runny nose) within 30 days before KS onset; (2) humidifier in the child’s room within 30 days before KS onset; (3) carpets at home or elsewhere where the child spent significant time were shampooed within 30 days before the child’s KS onset; (4) body of water within 200 yards of the child’s residence; (5) animal (e.g. dog, cat, fish, bird) living in the child’s house or yard within 30 days before the child’s KS onset; (6) any other household members ill within ∼30 days before the child’s KS onset; and (7) vaccination received by the child within ∼30 days before KS onset. Odds ratios (ORs) with 95% confidence intervals (CIs) were estimated with the Mantel-Haenszel test for matched analysis. 10 Multiple conditional logistic regression was used for further analysis of risk factors that may be associated with the occurrence of KS. 11 Results. Thirty-seven patients who met the case definition for KS were identified in Colorado from November 1, 1997, through April 30, 1998. Twenty-two patients (59%) were boys, and 28 (76%) of the patients were <5 years of age at the time of illness onset. Age at onset ranged from 1 month to 13 years (median, 2 years). Thirty (81%) patients were white, 6 (16%) were black and 1 (3%) was an Asian/Pacific Islander. Twenty-two (59%) of the 37 cases occurred in the Greater Denver area, and 22 (59%) patients had illness onset during December through February. The 6-month incidence of KS among children <5 years of age in Colorado was 10.1 per 100 000 children, with incidences of 8.7, 38.6 and 11.3 for whites, blacks and Asians/Pacific Islanders, respectively. Thirty-five (95%) of the patients received intravenous immunoglobulin (IVIG) treatment, and 26 (74%) of these patients received it within 10 days of illness onset. Seven (27%) of the 26 patients whose echocardiogram results were evaluated had evidence of coronary artery aneurysm or dilatation. In these 7 patients the median number of days after illness onset of IVIG administration was 6 days (range, 4 to 10 days, with 5 of 7 patients receiving IVIG before the 10th day and the remaining 2 receiving IVIG on Day 10); 2 were <6 months of age and 1 was 9 years of age. Six of the 7 patients received IVIG on or after the day of their first echocardiogram showing aneurysm or dilatation. We were able to find 2 matched controls for 19 of the 37 KS patients. The age and racial distribution, as well as the proportion of patients residing in the Greater Denver area, did not differ between these 19 KS patients and the remaining 18. Analysis of risk factors among the 19 KS patients and their matched controls indicated that KS was associated with the use of a humidifier in the child’s room (OR 7.3; 95% CI 1.8 to 29.3) within 30 days before KS onset (Table 1). A possible association was observed between antecedent respiratory illness and KS onset (OR 2.6; 95% CI 0.9 to 7.3). A high proportion of the case patients who had a humidifier in the room [10 (91%) of 11 patients] had a respiratory illness before KS onset. When taking respiratory illness into account with the use of multivariate analysis, humidifier use remained significantly associated with KS onset (OR 11.4; 95% CI 1.0 to 131.4). An association with KS was not seen for a history of recently shampooing carpets, residing near a body of water, recent vaccination and other possible risk factors examined. Unmatched analysis of the data showed similar results.TABLE 1: Selected possible risk factors for KS, Colorado, November 1997 to April 1998Discussion. Although much has been learned about KS since Dr. Tomisaku Kawasaki first described the illness in 1967, the causative agent remains elusive. 2 The reported increase in the number of KS cases observed in Colorado during the 6-month study period allowed us to further investigate possible environmental risk factors for KS. The demographic characteristics of the patients during the 6-month period in Colorado were consistent with those described in previous studies, except the incidence among Asians/Pacific Islanders was lower than the incidence among blacks. 12 Through analysis of data from CDC’s KS surveillance system, the rate of coronary artery abnormalities among KS patients treated with IVIG has previously been reported to range from 8.3 to 12.9%. 13 This Colorado cluster was associated with a comparatively high prevalence of coronary artery aneurysms and dilatations (27%), as reported in another study that investigated KS patients in Colorado during a time period encompassing part of the period of our investigation. 14 This high prevalence indicates the need for continued assessments of factors that may influence the occurrence of cardiac complications in patients who have echocardiogram evidence of coronary artery abnormalities at the time of or before IVIG treatment. The matched case-control study indicated that a humidifier in the child’s room before onset of KS was significantly associated with KS. Previous studies have found humidifier use to be associated with the development of asthma among children. 15, 16 This association may be because humidifiers provide favorable conditions for increases in house dust mite allergen or the growth of mold in the home; limited laboratory evidence indicates that mold may even grow in the humidifier itself. 15 A study has also suggested humidifier use as a predictor of the presence of airborne endotoxin in the home. 17 The findings of these studies underscore the plausibility that humidifier use may support the growth or facilitate the transportation of the agent causing KS. A possible association was also observed between antecedent respiratory illness and KS onset. In an earlier study KS patients were more likely to have an antecedent respiratory illness than matched controls. 3 It is possible that the agent causing KS could first settle in the upper respiratory tract before causing systemic illness. A recent study reporting increased IgA plasma cell infiltration of the upper respiratory tract and coronary arteries of KS patients supports the respiratory route as a potential portal of entry for the agent causing KS. 18 Previous studies identified recently shampooed carpets and residing near a body of water as risk factors for KS, but these findings were not confirmed in our study. 4, 6 Recently the occurrence of KS after vaccination has been reported to the national Vaccine Adverse Events Reporting System (T. Oussova, personal communication). However, vaccination ∼30 days before KS onset was not found to be associated with KS in our investigation. Our findings might have been affected by the small number of patients, and the retrospective nature of this investigation might have contributed to some recall bias. Nevertheless our study indicates that both the use of humidifiers and antecedent respiratory illness should be further investigated as possible risk factors for KS. Acknowledgments. We thank the personnel at the Colorado Department of Public Health and Environment, and at the participating Colorado hospitals, and Karen Mark and John O’Connor for technical assistance.