The proposed excuse is "limited space" in the publishing journals and of course the incentive to hide adverse events for marketing reasons. I doubt there are many high impact journals left that have not gone online where "limited space" is lie. Who cares if the PDF of the study is 150KB or 200KB...