If accelerated data show a significant change or failure of any attribute in one or more batches, an applicant should submit intermediate data for all three batches. In addition, the submission should contain a failure analysis (i.e., discussion concerning the observed failure(s)).
