A remark on Apéry's numbers