A note on directed adjusted profile likelihood