extract-reads will only output trimmed sequences from reads that contain the primers. So there is no need to confirm that the primers were in those sequences — they were there, but then were trimmed out and the internal region (simulated amplicon) output.
Thanks for your quick help. Do I also need to remove taxonomy for gene Id which are not present in the ref.seq.qza and can I know how many sequences are present in ref.seq.qza?