Discussion

In this study, we evaluated left atrial and ventricular functional responses to exercise in patients with a remote history of Kawasaki disease and prior coronary artery aneurysm compared with age-matched healthy controls. While the Kawasaki disease cohort had a greater number of systolic and diastolic parameters that changed significantly from rest to post-exercise recovery, mixed-effects modelling revealed no significant difference in the magnitude of exercise-induced change between the two cohorts.

Exercise stress echocardiography is a safe and noninvasive evaluation of cardiac function. It has been shown to detect myocardial ischaemia with better sensitivity than electrocardiographic stress alone in paediatric patients with a history of Kawasaki disease. ^{Reference Pahl, Sehgal and Chrystof8} A prior study demonstrated exercise-induced diastolic abnormalities on regional tissue Doppler-based indices in Kawasaki disease patients with coronary aneurysms. ^{Reference Arnold, Goebel and Ulmer9} This study was published in an earlier era where the coronary artery aneurysm classification scheme was based on absolute coronary diameters rather than z-scores. ^{Reference Arnold, Goebel and Ulmer9,Reference Newburger, Takahashi and Gerber10} In that cohort, the majority of patients met criteria for giant coronary aneurysms based on the prior classification scheme (diameter >10 mm). In the present study, we evaluated diastolic function using approaches not employed by Arnold et al., including mitral valve E/e’ and left atrial strain. More recently, using contemporary z-score criteria, Tedla et al. demonstrated abnormal exercise stress echocardiograms in 3/29 patients with a history of Kawasaki disease, all 3 were classified as American Heart Association coronary artery risk level 4 or 5. ^{Reference Tedla, Burns and Tremoulet11} Notably, none of the 16 patients who were classified as risk level 2 to 3 had abnormal stress findings.

In contrast to these prior studies, our cohort was characterised by milder coronary involvement defined using contemporary z-score- based criteria, with an average maximal z-score of 3.8. The American Heart Association’s Kawasaki disease guideline recommends periodic assessment of inducible myocardial ischaemia via stress testing in patients with Kawasaki disease risk classification level ≥ 3 (coronary artery z-score ≥ 2.5). ^{Reference Jone, Tremoulet and Choueiter3} The guideline acknowledges that while testing for inducible ischaemia is suggested starting at risk classification 3, the patients at most risk for stenosis and obstruction are those with a coronary artery z-score ≥ 10. ^{Reference McCrindle, Manlhiot and Newburger12} There were no patients in the present study who had inducible ischaemia during the exercise stress echocardiography. Furthermore, the absence of a differential exercise response between groups in our study may reflect a lower burden of microvascular dysfunction and adverse remodelling in patients with less severe coronary involvement. ^{Reference Suzuki, Miyagawa-Tomita and Komatsu13}

Deformation imaging via left ventricular longitudinal strain and left atrial strain has become popular over the past decade due to its ability to provide a more sensitive assessment of subclinical myocardial dysfunction than conventional markers of myocardial function such as ejection fraction. ^{Reference Thomas, Edvardsen and Abraham14} In patients with acute Kawasaki disease, left ventricular longitudinal strain was decreased compared to controls despite no statistically significant differences in conventional systolic indices. ^{Reference McCandless, Minich and Wilkinson15} Abnormal stress imaging may have prognostic implications in higher-risk Kawasaki disease populations. In a study by Noto et al., Kawasaki disease patients with inducible regional wall motion abnormalities on dobutamine stress echocardiography had an approximately threefold higher risk of major adverse cardiovascular events compared to those without wall motion abnormalities over a 15-year follow-up period. ^{Reference Noto, Kamiyama and Karasawa4} The absence of significant differences between the Kawasaki disease and control cohorts, despite the use of deformation imaging following exercise stress, may provide additional evidence of preserved global systolic and diastolic function in Kawasaki disease patients with mild to moderate coronary artery involvement. These findings were corroborated by the lack of difference in MV E/e’ and further support the notion that routine exercise stress imaging may be of limited yield for surveillance in lower-risk Kawasaki disease patients. ^{Reference Jone, Tremoulet and Choueiter3,Reference Tedla, Burns and Tremoulet11}

Left atrial reservoir strain increased significantly from rest to post-exercise recovery in the Kawasaki disease cohort but not in the control group. Left atrial functional remodelling is recognised as an important compensatory mechanism in the early stages of left ventricular diastolic dysfunction to preserve cardiac output. ^{Reference Thomas, Marwick and Popescu16} Left atrial strain is divided into three phases. Reservoir strain measures left atrial filling during ventricular systole. Conduit strain measures passive left atrial emptying during ventricular diastole. Contractile strain measures active left atrial emptying during late ventricular diastole when the atrium contracts. In the early stages of left ventricular diastolic dysfunction, both left atrial reservoir and conduit strains decrease due to increased atrial afterload and decreased passive ventricular filling, respectively. ^{Reference Thomas, Marwick and Popescu16,Reference Sabatino, Di Salvo and Prota17} The left atrial contractile strain increases in early stages of diastolic dysfunction to compensate for poor passive left ventricular filling. ^{Reference Singh, Addetia and Maffessanti18} Rest to post-exercise left atrial conduit and contractile strain did not differ significantly between the two cohorts. Given the well-described inverse relationship between left atrial reservoir strain and diastolic dysfunction severity, ^{Reference Singh, Addetia and Maffessanti18} the observed increase in left atrial reservoir strain in the Kawasaki disease cohort is unexpected and likely a chance finding, as mixed-effects modelling demonstrated a nonsignificant interaction.

There are a few limitations of our retrospective study. The relatively small sample size limits statistical power and raises the possibility of type II error, particularly for detecting subtle differences in the magnitude of exercise-induced strain changes. Our sample size, however, is consistent with previously published paediatric Kawasaki disease stress echocardiography literature, where single-centre cohorts have ranged from roughly 20 to 60 patients. ^{Reference Pahl, Sehgal and Chrystof8,Reference Arnold, Goebel and Ulmer9,Reference Tedla, Burns and Tremoulet11} Our cohort had mild to moderate coronary artery involvement, which may limit generalisability in patients with more severe coronary artery aneurysms. Due to our cross-sectional design, we cannot address longitudinal changes in myocardial mechanics over time. Importantly, stress echocardiography in our study was performed during early to mid-recovery rather than peak exercise. Immediate post-exercise imaging is the most widely used protocol in published Kawasaki disease stress echocardiography studies, ^{Reference Pahl, Sehgal and Chrystof8,Reference Tedla, Burns and Tremoulet11} and the American Society of Echocardiography recommends post-exercise imaging be completed within 1–2 minutes of peak exercise ^{Reference Pellikka, Arruda-Olson and Chaudhry7} as transient strain impairments and diastolic filling abnormalities are most apparent at peak stress. Our delay in obtaining stress imaging, as evidenced by a post-stress heart rate around 90 beats per minute, may have precluded our ability to find meaningful differences. Notably, left ventricular longitudinal strain has been shown to increase with exercise, ^{Reference Cifra, Mertens and Mirkhani19} yet our cohorts demonstrated lower systolic strain indices post-exercise. We suspect that this finding may not reflect true systolic dysfunction in either cohort, but rather, the strain values at mid-recovery may be attenuated by increased parasympathetic tone and reduced preload. Future studies employing supine bicycle protocols with peak stress imaging may provide a more sensitive assessment of myocardial function.

In paediatric patients with remote history of Kawasaki disease and prior coronary artery aneurysm, exercise stress echocardiography demonstrated no significant differences in the magnitude of exercise-induced change in patients with mild to moderate coronary artery involvement. Further studies are needed to determine whether patients with more severe coronary artery disease develop long-term abnormalities in myocardial function during exercise.