In an editorial that was also published in thebmj they point out alternative explanations for their results: http://www.bmj.com/content/351/bmj.h4141 http://www

In an editorial that was also published in thebmj they point out alternative explanations for their results: http://www.bmj.com/content/351/bmj.h4141 http://www.bmj.com/content/351/bmj.h4141 TL;DR: "As the authors acknowledge, a cause and effect relation cannot be inferred from their work. In this prospective study, Lv and colleagues have shown temporality of association, but we need to evaluate additional criteria to judge the strength of evidence. Their findings should be considered hypothesis generating, not definitive, and will undoubtedly encourage further work."