Kawasaki disease (KD) is a self-limited systemic vasculitis of unknown etiology that affects children. The diagnosis of KD is based on established clinical criteria. The prevalence of coronary artery (CA) involvement in untreated cases of KD is approximately 15–25 % [24], which decreases to less than 10 % when intravenous gamma globulin (IVIG) is administered within 10 days after onset of fever [18, 24]. Children with CA aneurysms are at risk for stenosis, obstruction, myocardial ischemia/infarct, and sudden death. The rate of CA aneurysm regression is shown to be associated primarily with the initial aneurysm size.

Giant aneurysms (GAN), measuring 8 mm in diameter or larger, rarely regress and exhibit the highest risk for subsequent stenosis [1, 10]. Most of the available data in the literature originates from Japan and hence pertains primarily to a population of Asian ethnicity. The current study aimed to determine the prevalence, management, long-term follow-up evaluation, and outcomes related to GAN in the province of Quebec because little data are available from a Caucasian-based population.

Methods

Type of Study and Population of Interest

The current study adopted a retrospective case series approach to identify and extract relevant data from all charts pertaining to patients with a diagnosis of KD and GAN in Quebec from the first diagnosis in 1976 until 2008. In the current study, GANs were defined as CA aneurysms measuring 8 mm or larger.

Identification of Relevant Patient Charts

In 2009, the Quebec Kawasaki Disease Registry was initiated to document and study the epidemiology of KD in Quebec, Canada. Based on an estimate of 2,200 hospitalizations with a diagnosis of KD, obtained from the Quebec Ministry of Health, since the first known case in 1976 until the end of 2008, this retrospective case study reviewed 1,975 charts (89.8 %). The remaining estimates of KD hospitalizations include rehospitalizations for previously known patients with a history of KD.

Because all subjects with suspected KD require echocardiography and a pediatric cardiology consultation, this retrospective study covered all pediatric cardiology centers and their ancillary clinics operating in the province of Quebec. From our total KD database, patients with CA larger than 5 mm in diameter (based on transthoracic echocardiography reports) were reviewed to identify better the cases of GAN, i.e. the target population of the study, and to ensure that no significant cases were missed among the non-GAN population with CA measuring 6–8 mm.

A standardized data collection record including a description of the acute phase, diagnostic timing, initial therapy, clinical and noninvasive imaging follow-up evaluation, and invasive diagnostic and therapeutic management was used to review the relevant charts. The length of the follow-up period was calculated using the date of KD diagnosis and the last documented follow-up assessment. Ethics board approval was obtained from each participating center.

Results

Cohort Description

From a total of 1,975 KD cases identified by complete hospital chart reviews under the Quebec Kawasaki Disease Registry study, 145 patients (7.3 %) had CA aneurysms measuring 5 mm or larger. Of these 145 patients, 21 (14.5 %) had aneurysms 6–8 mm in size, only 1 of whom had a history of late CA thrombus formation and myocardial ischemia. No deaths and no percutaneous or surgical interventions of the CA were reported.

The population described in this report comprises the 38 (26.2 %) of 145 or 38 (1.9 %) of 1,975 cases with GAN (≥8 mm), who were at highest risk for CA complications, as described in Japanese literature [35].

Demographic data on the GAN cases are summarized in Table 1, and individual data are presented in Table 2. Overall, the series had a male predominance with a predominantly Caucasian ethnicity (95 %). The age at diagnosis was 5.52 ± 4.04 years (Fig. 1). The KD diagnosis was retrospectively made at autopsy in 1 case (Patient 37), a posteriori by echocardiography in 4 of 38 live diagnoses (10.8 %), beyond 21 days of fever in 1 (2.7 %) of 37 cases, during the inflammatory phase in 32 (86.5 %) of 37 cases, and in GAN unrecognized on echocardiography but confirmed at autopsy in 1 (2.7 %) of 37 cases (patient 38, Table 1). Coronary artery aneurysms (CAAs) were documented by echocardiography at diagnosis in 25 (67.5 %) of 37 cases, whereas 6 (16.2 %) of 37 patients had no initial dilation, and 4 (10.8 %) of 37 patients had no echocardiogram available at the presumed onset of disease. The follow-up period was 9.26 ± 6.9 years (median, 9.25 years; range, 0.45–24.7 years), during which GAN regressed to a non-GAN size (<8 mm) in 19 (55.9 %) of 34 early survivors (i.e., excluding cases of early death).

Table 1 Baseline characteristics of cases with giant coronary artery aneurysm
Full size table
Table 2 Detailed clinical data of cases with giant aneurysms
Full size table
Fig. 1
figure 1

Histogram of patients’ age distribution

Full size image

Immunoglobulin

The therapeutic management at diagnosis included IVIG within 10 days of fever for 15 (39.5 %) of 38 patients, beyond 10 days for 11 (29 %) of 38 patients, and not at all for 11 (29 %) of 38 patients. The use of IVIG was undetermined in 1 (2.6 %) of 38 live diagnoses because the diagnosis was determined in a remote country with no documentation regarding the use of IVIG (patient 1; Fig. 2). Of the 11 patients who did not receive IVIG, 5 had their diagnosis determined before 1988, previous to the endorsement of IVIG 2 g/kg by the American Academy of Pediatrics [4], 4 had their diagnosis determined a posteriori, and 2 received their diagnosis beyond 10 days after onset of fever (13 and 22 days, respectively). Of the 26 patients who received IVIG therapy, 10 (38.4 %) received more than one dose, 7 received two doses, 1 received three doses, and 2 received six doses. Documented adjuvant therapy with steroids was used to treat four patients: for resistance to IVIG (2 cases), for resistance versus relapse (1 case), and for prolonged fever before the diagnosis of KD [1 case, patient 9 (Table 1)]. Three of these four patients had received multiple doses of IVIG.

Fig. 2
figure 2

Intravenous immunoglobulin during acute therapy and subsequent cardiac testing. *Not applicable due to early death of patients 35 and 38 or due to a posteriori diagnosis at autopsy of patients 36 and 37. †Exercise stress electrocardiography (ECG), myocardial nuclear imaging, and coronary angiography

Full size image

Anticoagulation

In addition to amino salicylic acid or other antiplatelet agents, anticoagulation therapy was used for 20 (58.8 %) of 34 early survivors. Of 11 patients identified as having primary coronary thrombi (Fig. 3), 3 (27.3 %) had an early occurrence (<1 month after KD diagnosis) and 8 (72.7 %) had a late occurrence (>1 month after KD diagnosis). Seven of these patients were not receiving anticoagulation therapy when the thrombi were diagnosed. One case exhibited mural thrombus formation in the right CA aneurysm after therapeutic withdrawal of warfarin subsequent to eye trauma [3], and then 2 years later in the left CA aneurysms after head trauma, which required another therapeutic withdrawal of warfarin.

Fig. 3
figure 3

Distribution of primary coronary artery thrombi with respect to early occurrence [within 1 month of Kawasaki disease (KD) onset] versus late occurrence (beyond 1 month of KD onset) and whether the patients were receiving anticoagulation therapy or not

Full size image

The time from onset of fever to the identification of coronary thrombi varied from 2 weeks to 13 years (median, 2 months). There also were four cases of secondary coronary thrombi (i.e., after scar occlusion of the CA, diagnosed during elective coronary angiography). Two of these patients (50 %) were receiving anticoagulation therapy at the time of diagnosis.

Noninvasive Testing

All the patients underwent serial resting transthoracic echo- and electrocardiography testing as part of their regular follow-up assessment. In addition, 33 (97 %) of 34 patients had one or more of the following investigations (Fig. 4): exercise stress test for 26 (78.8 %) of the 33 patients, 26.9 % of whom had abnormal results, and nuclear myocardial perfusion imaging for 23 (69.7 %) of the 33 patients, 56.5 % of whom had abnormal perfusion. Of these patients, 22.7 % (5/22) had abnormal readings on both exercise stress testing and nuclear myocardial perfusion imaging.

Fig. 4
figure 4

Distribution and categories of longitudinal cardiac testing. Additional noninvasive testing included exercise stress electrocardiography (ECG), nuclear myocardial perfusion imaging, multihead computed tomographic (CT) angiography, or magnetic resonance imaging. Additional invasive testing included selective coronary angiography

Full size image

The relationship between the adequacy of IVIG therapy, coronary complications, and longitudinal myocardial testing is illustrated in Fig. 2. Of the 11 patients whose IVIG therapy was delayed, 8 had abnormal myocardial test results. Coronary artery obstruction was demonstrated by selective CA angiography in five patients.

Invasive Testing and Interventions

One or more selective CA angiographies were performed in 31 (91.2 %) of 34 cases (excluding the 3 postmortem diagnoses). Coronary artery stenosis was diagnosed for 11 (35.5 %) of 31 patients and occlusion for 9 (29 %) of 31 patients. Of the 34 patients with a live diagnoses (11.8 %), 4 underwent percutaneous transluminal coronary revascularization, with a total of 8 stents placed. Three patients had drug-eluting stents, and one patient had an emergency bare metal stent placement while receiving Coumadin (international normalized ratio 2.5–3.0). The latter patient (patient 2) returned 2 months later with in-stent stenosis and intra-aneurysm clotting. An attempt at repermeabilization was complicated by clot embolization into the circumflex. A massive infarction required mechanical ventricular assistance followed by cardiac transplantation.

Finally, 2 (8.8 %) of 34 patients underwent primary CA bypass surgery. One patient required a double bypass graft of the anterior and posterior interventricular branch 23 months after KD (patient 4), and another patient underwent triple bypass grafting of the marginal left anterior descending and diagonal branches 13 years after KD (patient 11). At this writing, both patients continue to be asymptomatic with no graft stenosis. Up to the last follow-up visit, no deaths or acute coronary events had occurred.

Mortality

Five deaths (13 %) occurred in this series. One early death (patient 35) was recorded in 1985, 21 days after onset of fever, secondary to an extensive myocardial infarct. Another early death was recorded in 1980 (patient 37), 21 days after onset of fever, for a patient with an a posteriori diagnosis of KD at autopsy, described as “panarteritis nodosa” involving the coronary, adrenal, pancreatic, and renal arteries. Still another early death (patient 38) occurred 1.8 months after onset of fever, secondary to gastrointestinal vascular complications in 1988. A late cardiac death in 1997 claimed patient 36, who had a retrospective diagnosis of KD at presentation, with CA aneurysms and CA thrombus identified by echocardiography and a presumed acute phase that had occurred 1 year earlier. The autopsy report for this patient also confirmed extensive myocardial infarction and thrombosed aneurysms of the proximal right and left coronary arteries. Another late cardiac death occurred 5.7 years after KD (patient 23), 5 years from the last cardiac catheterization, 1.6 years from the last exercise stress test, and 7.5 months from the last resting transthoracic echocardiography. The relationship between patient survival and different outcomes (primary coronary thrombus, any coronary intervention, nonaccidental death, and any cardiac incident) is portrayed by Kaplan–Meier curves in Fig. 5.

Fig. 5
figure 5

Kaplan–Meier actuarial freedom from cardiac events (regression lines are mean, and 5th-95th confidence intervals)

Full size image

Discussion

Since the first described case of KD in 1961 [4], the management strategies and long-term follow-up evaluation of patients with KD have improved significantly. Also, identification of GAN has increased over the years with advancements in noninvasive coronary artery imaging, which has enabled the medical community to stratify their patients according to risk, as proposed in the 2004 American Heart Association (AHA) guidelines [24].

Our retrospective study enabled us to observe the evolution of KD management and investigations in the province of Quebec over the years. We found that the prevalence of GAN in our KD population was approximately 1.9 % (38/1975), which is in concert with another Canadian series (2.2 %) [18], and slightly higher than what has been reported historically (1 %) [24]. With the advent of IVIG therapy, the current prevalence of GAN is even lower (0.29 %) according to Japanese surveys [36].

Whereas KD is diagnosed before the age of 5 years in the majority of KD cases [24], the diagnosis in half of our cases was after the age of 5 years. In fact, the risk of CA involvement appears to be higher in infants and older children [12].

The well-known male predominance in cases with GAN was replicated in our survey [16, 18, 32]. From a genetic ethnic perspective, children of Asian descent are known to have the highest risk for the development of KD [13, 14]. However, it is unclear whether they are at higher risk for CAA abnormalities than Caucasians. The unique aspect of our study was the predominantly Caucasian ethnicity of the population.

Findings have demonstrated that CAA regression is related to the initial size of the coronary aneurysm [9, 23], with the greatest risk for stenosis and obstruction in GAN [35]. We observed echocardiographic reduction in the size of GAN to a non-GAN definition in 55.9 % of cases (19/34), similar to what was reported previously in another Canadian series [18]. Regression of GAN may be falsely reassuring because the mechanism of echocardiographic regression appears to be associated with intimal narrowing caused by myointimal proliferation, which may subsequently lead to stenosis and obstruction [8, 28, 33]. In addition, despite apparent regression of CAAs, endothelial and smooth muscle dysfunction persists [11, 31, 38]. All these factors influence our practice and follow-up strategies for these patients, with the aim to prevent severe cardiac events.

Japanese series have shown that GAN progresses to stenosis or obstruction in 46–61 % of patients [16, 29]. The prevalence of CA stenosis and obstruction in our series, however, was relatively lower (35.5 and 29 %, respectively). Our approach to CA intervention in the recent years has been driven partly by the heavy macro-calcification we observed [17], resulting in insidious progression to “asymptomatic” stenosis, typically compensated by formation of collateral CA circulation [6, 34].

The management strategies for KD also have evolved over the last few decades. More recently, the number of percutaneous interventions in the pediatric population has increased due to more experience with adults and the availability of low-profile devices for younger patients [2, 29]. We describe eight percutaneous interventions in four patients using different techniques of percutaneous transluminal coronary revascularization and coronary stent placement.

Because children with KD may display symptoms usually atypical for myocardial ischemia [15], our centers tend to perform at least one diagnostic cardiac catheterization with selective CA angiography when CA aneurysms or significant dilations are present in order to document the extent of CA aneurysms, the timing, and the degree of stenosis. Repeat CA angiography typically is alternated with myocardial perfusion nuclear imaging in the most severe cases [5, 37] and when macro-calcifications are apparent on chest X-ray or fluoroscopy [17]. Despite the concern for inherent complications and exposure to radiation [26], other groups seem to rely on repeated selective CA angiography (at 3 and 9 months after diagnosis, then every 2–3 years thereafter) to follow patients with GAN [29].

The latest AHA guidelines recommend the routine use of z-scores in the evaluation and treatment of KD and define CAA dilations as those with a diameter z-score of 2.5 or higher [24]. These recommendations are an improvement over the initial definitions proposed in the past because they take into account growth variation based on body surface area [7, 21, 25] instead of relying on absolute cutoff values [27] for CA size categorization. For the definition of GAN, z-score calculation of CA size also may become necessary, as suggested in a recent report [20]. Accordingly, the severity of the CAA aneurysm is less likely to be underestimated, especially in younger patients. With the increased use of z-scores, our group established z-score equations for all CA segments, including the circumflex and the distal right CA segments, which were unavailable in the literature previously [21, 25].

The overall mortality rate for the current KD population was 0.3 % (5/1975), similar to that reported in the literature from Japan (0.2 %) [12] and usually observed in the acute phase or in cases with CA sequelae [22]. In our series, except for one case, fatality occurred within the first year. Most deaths were attributable to extensive myocardial infarction secondary to coronary obstruction.

Another important factor in the management of GAN is the use of anticoagulation in addition to antiplatelet therapy. This has been demonstrated to decrease the incidence of myocardial infarction in patients with GAN [30]. This practice was encouraged in the last AHA guidelines. The use of low-molecular-weight heparin also seems to be an appropriate alternative to warfarin [19].

The current 2004 AHA guidelines regarding GAN follow-up evaluation leave much to the discretion of the treating physician, and management is based largely on institutional preferences. We therefore propose a multicenter, prospective study investigating the current management and follow-up evaluation of patients with GAN, which would add to future clinical guidelines and help to develop a more regimented/structured approach to follow-up evaluations and investigations in this population. Lack of structured data regarding adult follow-up evaluation and long-term complications also is an issue that needs to be included in future research.

Study Limitations

This study spanned a few decades, during which we observed advances in the diagnosis and the acute and long-term management of patients with GAN. We also witnessed an increase in awareness of this disease. Significant advancements in CA imaging also have occurred, especially with the introduction of two-dimensional echocardiography in the 1980s, which has continued to improve over time. Consequently, this may mean that cases of GAN were missed before this era in our series.

Conclusion

Giant aneurysms occurred for nearly 2 % of KD patients in the province of Quebec in this retrospective cohort, which included the pre-IVIG era. It affected Caucasians in this population but had no specific predilection for Asians. The management of GAN was based on institutional preferences due to the absence of evidence-based recommendations. These initial Quebec-based observations warrant a larger collaborative prospective study with regard to management, follow-up evaluation, and long-term outcome of all patients with KD in a predominantly Caucasian population but with a rich ethnic mixture.