Kawasaki disease (KD) is the leading cause of acquired coronary disease in children and may lead to subsequent myocardial ischemia and infarction. Because coronary computed tomographic angiography (CTA) is the most sensitive noninvasive test in patients with atherosclerosis, the aim of this study was to retrospectively evaluate coronary CTA performed in patients with KD for aneurysm, stenosis, and calcified and noncalcified coronary artery disease (CAD). Clinical histories and prior stress and imaging test results were reviewed. Thirty-two patients underwent coronary CTA for KD, and 385 coronary segments were evaluated. Twenty-three of 32 patients had ≥1 diseased coronary segment. There were 20 aneurysms, 7 lesions, and 75 segments (20%) with nonobstructive CAD (16% noncalcified, 2% calcified, and 2% mixed). All nonobstructive and obstructive CAD was in patients with histories of acute-phase coronary artery dilatation or aneurysm (echocardiographic z score 4 to 44), and were almost always associated with normal stress imaging test results on follow-up. No lesion or CAD was found in coronary computed tomographic angiographic studies performed in a control group referred for other indications (n = 32, 422 segments evaluated). The median coronary computed tomographic angiographic dose-length product was 59 mGy cm (interquartile range 32 to 131), the median unadjusted radiation dose was 0.8 mSv (interquartile range 0.4 to 1.8), and the median age- and size-adjusted radiation dose was 1.3 mSv (interquartile range 0.7 to 2.3). In conclusion, high-risk patients with histories of KD had nonobstructive and obstructive CAD not visualized by other noninvasive imaging tests. In properly selected high-risk patients with KD, coronary CTA may identify a subset at increased risk for future coronary pathology who may benefit from medical therapy.
Kawasaki disease (KD) was first described in 1967 and is the leading cause of acquired coronary disease in children in the developed world. The small subset of patients who have persistent coronary aneurysms require lifelong follow-up, and the risk for coronary intervention continually increases with age. The current guidelines for the care and management of patients with KD recommend routine echocardiography and perfusion imaging for long-term evaluation of coronary pathology in high-risk patients who may require intervention. Coronary angiography is recommended within the first year for those with giant aneurysms at the highest risk for morbidity and mortality and in patients with symptoms suggestive of coronary ischemia. Coronary computed tomographic angiography (CTA) is the most sensitive noninvasive test for detecting a coronary stenosis and identifies obstructive and nonobstructive coronary artery disease (CAD) that may cause ischemia or predict a future cardiac event in patients with atherosclerosis. There are no previous reports of coronary CTA analysis of CAD with low Hounsfield unit (HU) attenuation in patients with previous KD. We investigated whether coronary CTA may demonstrate coronary disease in these patients not otherwise identified by currently recommended diagnostic techniques.
Methods
This was a retrospective descriptive cohort study of consecutive subjects who underwent coronary CTA for evaluation of KD associated CAD at a single institution from April 2007 to August 2013. The KD group was compared with a gender-, age-, and size-matched control group randomly selected from coronary computed tomographic angiographic studies performed for other indications. Institutional review board approval was obtained. All patients with KD were followed by a pediatric cardiologist and were treated according to national consensus guidelines from 1994 and revised in 2004 and clinical judgment. Patients were referred to coronary CTA on the basis of a clinician’s discretion. Charts were reviewed for KD history, including age at diagnosis, treatment, acute-phase reactants, and clinical course. All previous echocardiographic, angiographic, exercise tests or stress perfusion imaging results were reviewed and correlated with the anatomic coronary artery findings. Z scores were calculated for coronary artery measurements obtained by echocardiography during the acute phase of illness.
Patients <7 years of age were placed under general anesthesia with suspended respiration during image acquisition (n = 9). Patients >7 years of age were able to cooperate with a breath hold for image acquisition (n = 23). All patients were scanned on a first-generation dual-source (Siemens Definition, temporal resolution 83 ms; Siemens Healthcare, Forchheim, Germany) or second-generation dual-source (Siemens Definition Flash, temporal resolution 75 ms) scanner. Retrospective electrocardiographic gating (Mindose protocol) was used for patients on the first-generation dual-source scanner (from 2007 to 2009). Prospectively electrocardiographically triggered scan mode was used for patients who underwent coronary CTA on the second-generation dual-source scanner (from 2009 to 2013). A high-pitch scan mode (pitch 3.4) was used when the heart rate was <60 beats/min and heart rate variability was <10 beats/min. A prospective electrocardiographically triggered scan protocol was used for those with heart rates >60 beats/min or with heart rate variability >10 beats/min. Scanner output adjustment for body weight, β blockade, and contrast injection protocol were previously described. A pediatric and/or adult cardiologist experienced in coronary CTA was in attendance for all scans.
All coronary computed tomographic image data sets (KD and control groups) were evaluated using curved multiplanar reformatted images with the use of a commercially available image postprocessing workstation (Vitrea 4; Vital Images, Minnetonka, Minnesota). Coronary arteries were divided into 16 segments per patient, as recommended by the Society of Cardiovascular Computed Tomography. An adult cardiologist (MN) experienced in coronary CTA evaluated all coronary artery segments in the first 4 cm of each vessel for the presence of stenosis on the basis of the percentage of luminal narrowing (none, mild [<50%], moderate [50% to 69%], or severe [>70%]), as well as the presence of noncalcified (low attenuation, 30 to 149 HU), calcified (high attenuation, >149 HU), and mixed CAD. The segments contained in the proximal 4 cm of each coronary artery considered abnormal were additionally evaluated with a plaque software evaluation tool on the basis of HU attenuation (surePlaque; Vital Images) designed for patients with CAD. The outer vessel boundary and inner luminal boundary of the abnormal segments were automatically traced and subsequently manually adjusted in a cross-sectional view. Single-pixel HU attenuation was measured in multiple points within each segment with possible coronary disease. The median and range of HU attenuation for all abnormal segments were calculated. Aneurysm was defined as a coronary vessel >3 mm in patients <5 years of age and >4 mm in patients >5 years of age. A coronary artery z score was calculated for all echocardiographic measurements meeting this criterion and was compared with measurements obtained from the coronary computed tomographic angiographic scans.
Descriptive statistics are displayed as medians and interquartile ranges for continuous variables. CAD was analyzed per segment; the number of affected segments compared with total segments analyzed is described as a percentage. Noncalcified and calcified CAD, lesion, and aneurysm size were described per patient and compared with controls. Categorical variables are displayed as number and percentages. Groups were compared using Pearson’s chi-square or Fisher’s exact tests. Continuous variables were analyzed using Student’s t test or analysis of variance for normally distributed variables and Kruskal-Wallis tests for continuous variables not normally distributed. Aneurysm measurement by echocardiography versus coronary CTA was compared using a correlation coefficient. A p value <0.05 was considered significant, and p values were 2 sided where possible. All statistical calculations and plots were done with Stata version 11.2 (StataCorp LP, College Station, Texas).
Results
Thirty-two coronary computed tomographic angiographic studies were performed in 32 patients a median of 7.75 years (interquartile range 1 to 12.9) after diagnosis of KD. During this time period, there were 658 outpatient encounters for the diagnosis of KD. For those eventually referred for coronary CTA, the median age at the time of KD diagnosis was 2.35 years (interquartile range 0.55 to 4.5; range 2 months to 12.5 years). The referral indications for coronary CTA were symptoms (n = 1), as a substitute for invasive coronary angiography 1 year after diagnosis in patients with multiple or giant aneurysms (n = 4), and for evaluation before discharge or transfer to adult cardiology for adolescent and young adult patients considered at high risk by the referring cardiologist on the basis of coronary artery involvement or clinical course (n = 27). Clinical features considered high risk independent of coronary findings were prolonged or profound inflammation, including diagnosis after 10 days of illness, severe elevation of acute-phase reactants (C-reactive protein >6 mg/L, sedimentation rate >75 μg/L), requiring >1 dose of immunoglobulin, steroids, or infliximab for resolution of symptoms. Initial laboratory data were not available for 6 patients diagnosed at outside institutions and referred for follow-up after the acute phase of illness. The control group consisted of 32 patients randomly selected from age-, size-, and gender-matched patients who underwent coronary CTA for other indications. Control group and KD group comparisons are listed in Table 1 .
| Parameter | Case (N=32) | Control (N=32) | p-value |
|---|---|---|---|
| Age at scan (years) | 12.9 (6.3, 14.9) | 12.9 (6.3, 14.8) | 0.8202 |
| Age at diagnosis (years) | 2.35 (0.55, 4.5) | — | |
| Male | 23 (71.9) | 23 (71.9) | 1.0000 |
| Interval between diagnosis and scan (years) | 7.75 (1.0, 12.9) | — | |
| Height (cm) | 149.9 (117.5, 164.7) | 157.5 (121.0, 168.9) | 0.6257 |
| Weight (kg) | 37.0 (23.8, 62.5) | 46.5 (23.1, 63.9) | 0.8936 |
| BMI (kg/m 2 ) | 19.6 (16.4, 24.1) | 18.9 (16.9, 20.9) | 0.6070 |
| Oral metroprolol (dose in mg) | 50 (11, 100) | 50 (0, 75) | 0.1026 |
| Oral metroprolol (dose mg/kg) | 1.28 (0.63, 1.61) | 0.76 (0, 1.47) | 0.1769 |
| Intravenous metroprolol (dose in mg) | 10 (0, 15) | 6 (0, 10) | 0.1855 |
| Intravenous metroprolol (dose in mg/kg) | 0.27 (0, 0.61) | 0.15 (0, 0.44) | 0.3345 |
| Contrast (ml) | 64 (35, 83.5) | 65 (37.5, 90) | 0.9146 |
| Contrast (ml/kg) | 1.47 (1.25, 1.92) | 1.41 (1.33, 1.77) | 0.9148 |
| Heart Rate (beats per minute) | 56 (51, 74) | 59 (52, 73) | 0.4849 |
| Scan dose length product (DLP) | 59 (31.5, 131) | 50 (25, 130.5) | 0.6544 |
| Scan CT dose index/volume (CTDIvol) | 4.08 (3.07, 10.96) | 4.69 (1.62, 9.84) | 0.3753 |
| Effective dose, unadjusted (mSv) ∗ | 0.82 (0.43, 1.8) | 0.7 (0.35, 1.82) | 0.6544 |
| Effective dose, age/size-adjusted (mSv) | 1.325 (0.729, 2.296) | 0.995 (0.583, 2.236) | 0.2868 |
| Scan length (cm) | 12.0 (10.2, 13.5) | 12.7 (10.3, 14.5) | 0.3821 |
| Coronary CTA Scan protocol | 0.1512 | ||
| Prospective ECG triggered (helical) | 9 (28%) | 11 (34%) | |
| Prospective ECG triggered (axial) | 13 (41%) | 6 (19%) | |
| Retrospective ECG gated | 10 (31%) | 15 (47%) | |
| CT tube potential (kilovolts) | 0.7879 | ||
| 70 kVp | 1 (3%) | 1 (3%) | |
| 80 kVp | 20 (63%) | 23 (72%) | |
| 100 kVp | 10 (31%) | 8 (25%) | |
| 120 kVp | 1 (3%) | 0 |
The proximal 4 cm of all coronary segments in the KD and control group coronary computed tomographic angiographic studies were evaluated for the presence of stenosis and calcified and noncalcified CAD. In the 32 patients with KD, 22 (69%) had ≥1 coronary segment with CAD. Three hundred eighty-five segments were included in the proximal 4 cm of the 32 data sets. Seventy-five segments (20%) were determined to have calcified and noncalcified CAD by visual estimate. Of those, 61 segments had noncalcified CAD, 7 segments had calcified CAD, and 7 segments had mixed CAD. All segments qualitatively determined to have CAD were also documented by the automated plaque analysis tool. Twenty coronary aneurysms and 7 coronary lesions (2 occlusion, 3 severe, 1 moderate, and 1 mild) were present. Representative examples are shown in Figures 1-4 . Of the 23 patients with abnormal findings, 8 patients had CAD and aneurysms, 7 patients had isolated CAD, 2 patients had lesions with additional CAD but no aneurysm, 5 patients had CAD, aneurysms, and coronary lesions, and 1 patient had an isolated aneurysm. Patients with KD with normal results on coronary CTA had acute-phase echocardiograms showing normal coronary arteries or diffuse dilatation of the coronary arteries without aneurysms (z = 2.5 to 9) that resolved during follow-up. All patients with abnormal results on coronary CTA had acute-phase echocardiograms showing profound coronary artery dilatation or aneurysm, with echocardiography-based coronary artery z scores ranging from 4.7 to 44.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
