Background A 12-nucleotide RIPOR2 in-frame deletion was recently identified as a relatively common and highly penetrant cause of autosomal dominant non-syndromic sensorineural hearing loss, type DFNA21, in the Netherlands. The associated hearing phenotype is variable. The allele frequency (AF) of 0.039% of this variant was determined in a local cohort, and the reported phenotype may be biased because studied families were identified based on index patients with hearing loss (HL). In this study, we determine the AF in a cohort from a different geographical region of the Netherlands. Additionally, we examine the hearing phenotype in individuals with the variant but not selected for HL. Methods The AF was determined in participants of the Rotterdam Study (RS), a large cohort study. The phenotype was characterised using individual clinical hearing data, including audiograms. Results The observed AF in the RS cohort was 0.072% and not statistically significantly different from the previously observed 0.039%. The AF in the two cohorts combined was 0.052%. Consistent with previous findings, we found a highly variable audiometric phenotype with non-penetrance of HL in 40% of subjects aged 55–81, which is higher than the 10% at age 50 previously observed. Conclusion We found an overall higher AF and lower penetrance than previously reported, confirming that DFNA21 is relatively common in the Netherlands. This supports its potential suitability as a target for therapeutic development. Studying possible modifying factors is essential to explain the phenotypical variability and to identify patients eligible for such a therapy.