Reproducibility of ventricular arrhythmias at exercise testing for prediction of non-ischaemic left ventricular scar in athletes

Aims The non-ischaemic left ventricular scar (NILVS) is an emerging substrate of ventricular arrhythmias (VA) in the athlete. We tested the diagnostic value of VA reproducibility at repeated exercise testing (ET). Methods We included consecutive athletes who underwent cardiac magnetic resonance (CMR) for evaluation of VA and two con- and results secutive ET off-therapy within 12 months. Those with a positive family history of premature sudden death/cardiomyopathy, syncope/pre-syncope, ECG or echocardiography abnormalities, or low-risk VA (exercise-suppressed infundibular or fascicular VA) were excluded. The NILVS was defined as subepicardial/midmyocardial stria of late gadolinium enhancement involving >5% of the LV mass. Reproducibility was defined as the occurrence of VA with the same pattern and behaviour during repeated ET. Of 325 athletes who underwent CMR for evaluation of VA, 75 were included, and 30 showed NILVS involving three [2–4] LV segments. At first ET, athletes with NILVS showed a higher prevalence of exercise-induced VA (93% vs. 53%, P < 0.001), while other VA characteristics did not differ between groups. At repeated ET, reproducibility was observed in 97% of athletes with vs. 13% without NILVS (P < 0.001). The remaining 87% of athletes with normal CMR either did not show any VA at repeated ET (59%) or showed arrhythmias with different patterns, mostly infundibular. Reproducibility yielded a positive predictive value for NILVS of 83% and a negative predictive value of 98%. Conclusion VA reproducibility at repeated ET predicted an underlying NILVS in athletes with VA and otherwise normal clinical workup. This finding may improve risk stratification and appropriate CMR referral of athletes with apparently idiopathic VA.