Background
Patients with Kawasaki disease (KD) are at risk for developing coronary artery lesions, but the association of noncoronary changes such as mitral regurgitation (MR) and/or pericardial effusion (PE) with cardiac mechanics in the acute phase of KD has not been previously described. The aim of this study was to test the hypothesis that these noncoronary markers for carditis are associated with abnormalities in strain (ε) and strain rate (SR) in patients with MR or PE not appreciated by conventional echocardiography.
Methods
Longitudinal and circumferential ε and SR analyses were retrospectively performed on patients with KD. Patients with and without MR or PE were compared. Strain values were also compared between patients with and without coronary artery lesions. Values for ejection fraction, shortening fraction, and clinical laboratory parameters were correlated with MR or PE. Follow-up echocardiographic outcomes were recorded at the first encounter after initial diagnosis. Follow-up ε and SR data were also obtained in the group with MR or PE and altered ventricular mechanics at diagnosis.
Results
Of the 110 patients reviewed, 92 had appropriate image quality for either longitudinal ε and SR or circumferential ε and SR analysis. Twenty-eight patients (30%) had either MR or PE. Longitudinal ε and SR were significantly decreased in patients with MR or PE compared with patients without MR or PE (ε: −16.4 ± 4.0% vs −19.0 ± 3.7%, P = .004; SR: −1.3 ± 0.7 vs −1.6 ± 0.4 sec −1 , P = .03). No significant difference in longitudinal ε or SR was noted between patients with and without coronary artery lesions (ε: −17.9 ± 4.1% vs −17.8 ± 3.8%, P = .50; SR: −1.5 ± 0.3 vs −1.6 ± 0.8 sec −1 , P = .50). In the group with abnormal coronary arteries, presence of MR or PE was correlated with decreased longitudinal ε (−16.1 ± 3.6% vs −18.9 ± 3.4%, P = .02), without a significant difference in longitudinal SR (−1.6 ± 0.4 vs −1.5 ± 0.4 sec −1 , P = .20). At approximately 3-week follow-up (21.3 ± 15.8 days), longitudinal ε and SR for the group with MR or PE had increased significantly compared with diagnosis (ε: −16.4 ± 4.3% vs −18.6 ± 0.5%, P = .03; SR: −1.3 ± 0.6 vs −1.8 ± 0.4 sec −1 , P = .008), coincident with resolution of MR or PE. In both groups, erythrocyte sedimentation rate and C-reactive protein were elevated (85.3 ± 36.2 mm/h vs 75.1 ± 33.1 mm/h [ P = .34] and 12.3 ± 6.7 vs 11.7 ± 8.2 mg/dL [ P = .83]), but only modest correlations were noted between longitudinal ε and elevated erythrocyte sedimentation rate ( r = 0.52, P = .01; confidence interval, 0.10–0.80) and C-reactive protein ( r = 0.50, P = .02; confidence interval, 0.10–0.80) in patients with MR or PE. Shortening fraction and ejection fraction were within the normal range in both groups.
Conclusions
Patients presenting with KD with MR or PE at diagnosis are likely to have altered ventricular mechanics compared with patients with KD without MR or PE despite normal conventional echocardiographic measures of function. There is no significant difference in ventricular mechanics when comparing patients with KD with coronary ectasia or aneurysms and those without coronary lesions. Presence of abnormal ε in patients with KD with altered ventricular mechanics correlates modestly with laboratory inflammatory markers. Peak systolic longitudinal ε and SR increased significantly at 3-week follow-up compared with initial diagnosis, coincident with resolution of MR or PE.
Kawasaki disease (KD) is a leading cause of acquired heart disease in the developed world, with the major morbidity and mortality associated with coronary artery lesions (CALs). The acute phase of the disease is characterized by generalized inflammation seen clinically, on laboratory markers, and by echocardiography. A majority of patients diagnosed with KD have evidence of endocarditis, pericarditis, or myocarditis as evidenced by histologic analysis. Echocardiographic evidence of carditis includes mitral regurgitation (MR), pericardial effusion (PE), and abnormal contractility. The long-term implications of these findings are not known, especially in the absence of CALs.
Although previous literature has shown that left ventricular (LV) contractility is diminished in acute KD prior to intravenous immunoglobulin (IVIG), ejection fractions (EFs) in patients with acute KD are typically normal. Crystal et al . showed that LV dilation (serving as a marker for carditis) is increased at diagnosis in patients with acute KD. Unlike the rapid effect of IVIG on function, they showed that LV end-diastolic dimension improves in the weeks after diagnosis. Multiple studies have demonstrated that ventricular strain (ε) is a useful tool in the detection of subtle abnormalities of cardiac function. Impairment of longitudinal ε in KD compared with normal controls has been reported ; as such, we have chosen to further investigate the relationship of noncoronary markers of carditis to LV mechanics in patients diagnosed with KD. MR and PE were used for our analysis because they have been previously established as useful echocardiographic parameters for noncoronary markers of carditis in KD and associated with CALs at diagnosis or in the acute phase.
We hypothesized that (1) MR and/or PE is associated with altered myocardial deformation, and (2) altered myocardial deformation correlates with laboratory markers of inflammation in the acute phase.
Methods
We retrospectively reviewed echocardiograms and clinical parameters of patients admitted for KD between 2002 and 2012. Inclusion criteria were (1) age < 18 years at diagnosis, (2) clinical and laboratory diagnosis supportive of typical or atypical KD as established by the American Heart Association criteria for KD, and (3) completion of initial echocardiographic examination within 24 to 48 hours of diagnosis. We defined atypical KD as ≥5 days of fever as well as two or three clinical criteria as previously established by the American Heart Association for KD. Patients with structural heart disease or echocardiograms of inadequate image quality for deformation analysis were excluded. Approval for this study was obtained from the Institutional Review Board at Children’s National Medical Center.
Conventional Two-Dimensional Echocardiography
All two-dimensional transthoracic echocardiographic examinations were performed with a standardized protocol using a commercial echocardiographic system (Philips Medical Systems, Best, The Netherlands). Conventional echocardiographic parameters included EF, shortening fraction (SF), presence or absence of MR (greater than trivial or physiologic on the basis of qualitative assessment of color Doppler imaging, as defined by published guidelines), PE (>1 mm in diastole in all dimensions), and the presence of CALs (either aneurysms or ectasia). Coronary ectasia was defined as a Z score ≥ 2. EF was calculated by the single-plane method of disks in the apical four-chamber view, and SF was calculated by standard LV M-mode measurements of at the level of the papillary muscles. All conventional echocardiographic measurements were obtained from reports generated by experienced pediatric cardiologists using a standardized laboratory protocol for measurement of coronary artery diameters, classification of coronary lesions, and measurement of ventricular function (EF and SF).
Myocardial Deformation Analysis
Echocardiograms in Digital Imaging and Communications in Medicine format were analyzed offline using Velocity Vector Imaging (syngo version 3.1; Siemens Healthcare, Malvern, Pennsylvania) by a single physician (L.G.) blinded to other patient data. Image analysis was performed on images with an average viewed frame rate of 75 ± 17 Hz. Endocardial tracings of the left ventricle were performed in the parasternal short-axis and apical four-chamber views for each subject if available. Parasternal short-axis tracings performed at the level of the papillary muscles were used to obtain peak circumferential ε and strain rate (SR). Apical four-chamber views with good visualization of the left ventricle from base to apex were used to obtain peak systolic longitudinal ε and SR. Three separate measurements for global longitudinal and circumferential ε and SR were recorded on the same two-beat clip and averaged for each patient.
Clinical Data
Clinical parameters recorded for each patient included initial white blood cell count, absolute neutrophil count, hemoglobin, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), platelet count, and albumin level. Duration of fever, classification of typical versus atypical KD, and number of doses of IVIG were obtained from the medical record.
Follow-Up Analysis
Echocardiographic outcomes for all patients at first available follow-up were gathered. Regression or worsening of CALs as well MR or PE on first follow-up echocardiogram was documented.
Additionally, peak systolic longitudinal ε and SR analyses were performed in the group with initially diagnosed MR or PE at first follow-up echocardiography. All speckle-tracking analyses were performed as described earlier.
Statistical Analysis
Comparison was made between patients with MR or PE on initial echocardiography and patients without MR or PE regardless of coronary status. Additional analysis was performed for patients with or without CALs. Statistical analyses were performed using MedCalc for Windows version 12.2.1 (MedCalc Software, Ostend, Belgium). Statistical significance was reflected by a P value < .05. Unpaired two-tailed t tests were used to compare longitudinal ε and SR as well as circumferential ε and SR between groups. All data analyzed using the two-sample unpaired t test satisfied the D’Agostino-Pearson test for normal distribution (with P values > .05). Unpaired two-tailed t tests were used to compare echocardiographic and laboratory data between patients with or without MR or PE. Unpaired two-tailed t tests were also used to compare echocardiographic and laboratory data in patients with or without coronary abnormalities. The χ 2 test was used to compare categorical variables. Longitudinal ε and SR in the group with MR or PE was correlated with laboratory parameters using the Pearson correlation coefficient. Intraobserver agreement was assessed using the Bland-Altman comparison method on 15 randomly selected patients in each group.
Results
We originally evaluated 110 patients and excluded 18 patients (16%) from the analysis to include high-quality imaging clips for ε analysis. Ninety-two patients (84%) had adequate echocardiographic images for either longitudinal or circumferential deformation analysis. IVIG treatment was received by approximately 85% of patients within the first 24 to 48 hours of diagnosis. Forty-seven percent of children with KD received IVIG before echocardiography (equally divided in the groups with or without MR or PE). All patients had clinical diagnoses of KD on the basis of clinical and laboratory criteria. There were no significant differences in demographic or anthropometric data between the two groups ( Table 1 ). Duration of illness at diagnosis was similar for both groups (8.3 days [CI, 6.8–9.7 days] vs 9.5 days [CI, 7.0–12.0 days], P = .40).
| Variable | MR or PE present (n) | Neither MR nor PE present (n) | P |
|---|---|---|---|
| Age at diagnosis (y) | 3.8 (28) | 3.2 (64) | .43 |
| BSA (m 2 ) | 0.71 (26) | 0.69 (58) | .84 |
| Male | 64% (28) | 56% (64) | .47 |
| Atypical | 44% (25) | 48% (52) | .74 |
| Day of illness | 8.3 (24) | 9.5 (46) | .42 |
| >1 dose of IVIG | 23% (26) | 32% (56) | .41 |
Conventional Echocardiographic Parameters
Conventional echocardiographic data are shown in Table 2 . Patients were divided into groups on the basis of the presence or absence of MR or PE. MR or PE was present in 28 (30%) of 92 patients analyzed. Of the 28 patients, eight had MR, and 20 had PE. MR was classified as mild in seven of eight patients analyzed, with only one patient classified with moderate to severe MR. “MR or PE” was used to denote patients with carditis, and our results include patients with either MR or PE or both. Separate analysis on the entire patient group was then performed on the basis of the presence or absence of CALs. There were no significant differences in SF. EF was higher in the group without MR or PE ( Table 2 ) but was within the normal range for age in both groups.
| Variable | MR or PE present (n) | Neither MR nor PE present (n) | P | CALs present (n) | CALs absent (n) | P |
|---|---|---|---|---|---|---|
| EF (%) | 61.5 ± 3.6 (23) | 65.2 ± 5.7 (55) | .001 | 64.7 ± 6.3 (37) | 63.6 ± 4.3 (41) | .36 |
| SF (%) | 35.2 ± 5.5(25) | 36.7 ± 4.1 (58) | .17 | 37.5 ± 5.4 (39) | 35.1 ± 3.6 (43) | .02 |
Strain and SR Analysis
Of the 92 patients, 88 underwent longitudinal ε and SR analysis, and 68 patients had adequate image quality for circumferential ε and SR analysis. There were no significant demographic differences between the patients who underwent longitudinal ε and SR analysis versus circumferential ε and SR analysis ( Table 3 ).
| Variable | Longitudinal ε and SR group (n) | Circumferential ε and SR group (n) | P |
|---|---|---|---|
| Age (y) | 3.3 (84) | 3.4 (64) | .74 |
| BSA (m 2 ) | 0.68 (78) | 0.73 (54) | .36 |
| Male (%) | 39% (84) | 42% (64) | .89 |
| Atypical (%) | 51% (71) | 47% (57) | .73 |
| Day of illness | 7.8 (64) | 8.5 (49) | .56 |
| >1 dose of IVIG | 73% (54) | 72% (41) | .90 |
Patients with MR or PE (including patients with and without coronary anomalies) had decreased longitudinal ε and SR compared with patients with KD without MR/PE. Circumferential ε and SR were not significantly different between groups. There was no significant difference in longitudinal or circumferential ε or SR when patients with or without CALs were compared ( Table 4 ).
| Variable | MR or PE present (n) | Neither MR nor PE present (n) | P | CALs present (n) | CALs absent (n) | P |
|---|---|---|---|---|---|---|
| Longitudinal ε (%) | −16.4 ± 4.0 (27) | −19.0 ± 3.7 (61) | .004 | −17.9 ± 4.1 (42) | −17.8 ± 3.8 (46) | .54 |
| Longitudinal SR (sec −1 ) | −1.3 ± 0.7 (27) | −1.6 ± 0.4 (61) | .03 | −1.5 ± 0.3 (42) | −1.6 ± 0.8 (46) | .46 |
| Circumferential ε (%) | −21.6 ± 5.9 (21) | −22.8 ± 5.5 (47) | .52 | −21.6 ± 5.8 (34) | −23.2 ± 5.2 (34) | .27 |
| Circumferential SR (sec −1 ) | −2.0 ± 0.6 (21) | −2.2 ± 0.7 (47) | .26 | −2.1 ± 0.6 (34) | −2.3 ± 0.8 (34) | .50 |
Patients with abnormal coronaries had decreased longitudinal ε if MR or PE was present versus absent. These groups did not show a significant difference in longitudinal SR. Presence or absence of MR or PE for patients in the group with normal coronaries did not differ significantly for longitudinal ε ( Figure 1 ).
Bland-Altman analysis showed high intraobserver reproducibility with no significant bias and narrow limits of agreement for longitudinal and circumferential ε and SR on 15 randomly selected patients in each group ( Figure 2 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
