I would worry about publication bias here, since it's a metastudy and few people would publish such a study showing that, would you know, there's nothing going on...