Reliability and usefulness of random fecal alpha-1-antitrypsin concentration: further simplification of the method