**Background** Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) is a rare inherited heart-muscle disorder, which is the most common cause of life-threatening arrhythmias and sudden cardiac death (SCD) in young adults and athletes. Early and accurate diagnosis can be crucial in effective ARVC management and prevention of SCD. **Methods** The genome Aggregation Database (gnomAD) population of 138,632 unrelated individuals was searched for previously identified ARVC variants, classified as pathogenic or unknown on the disease genetic variant database (http://www.arvcdatabase.info/), in five most-commonly mutated ARVC genes: *PKP2, DSP, DSG2, DSC2* and *JUP*. Minor allele frequency (MAF) of 0.001 was used to define variants as rare or common. **Results** The gnomAD data contained 32% (n=117) of the previously reported pathogenic and 57% (n=152) of the unknown ARVC variants. The cross-ethnic analysis of MAF revealed that 11 previously classified pathogenic and 57 unknown variants were common (MAF≥0.001) in at least one ethnic gnomAD population and therefore unlikely to be ARVC-causing. After applying our MAF analysis the overall frequency of pathogenic ARVC variants in gnomAD was one in 257 individuals, but a more stringent cut-off (MAF≥0.0001) gave a frequency of one in 845, closer to the estimated phenotypic frequency of the disease. **Conclusions** Analysis of large cross-ethnic population sequencing data can significantly improve disease variant interpretation. Higher than expected frequency of ARVC mutations suggests that a proportion of ARVC-causing variants may be inaccurately classified, implying reduced penetrance of some mutations, and/or a polygenic aetiology of ARVC.