Table 1B.

The cohort studies checklist consisted of 4 main items: selection bias, performance bias, attrition bias and detection bias. Based on the score of their respective subitems, the main items were scored as low, unclear and high risk of bias. For each main item, a low response indicated that the study was designed and conducted in a manner that minimized the risk of bias for that item. An unclear response was given when the information required to score an item was not reported or was not reported clearly. A priori, was decided that studies were excluded if more than 2 out of 4 main items would be scored as high risk of bias on the basis of low methodological quality.

Study	Selection bias	Performance bias	Attrition bias	Detection bias	Inclusion?
Park et al32	low	unclear	low	low	Yes
Franceschi, 201547	unclear	unclear	unclear	unclear	No
Oh et al31	low	unclear	low	unclear	Yes
Cho et al17	unclear	unclear	low	low	Yes
Warner et al35	unclear	unclear	low	low	Yes
Ryu, 201560	low	unclear	low	high	Yes
Taniguchi, 201561	unclear	unclear	low	unclear	No
Hug, 201553	unclear	unclear	unclear	low	No
Namdari, 201457	high	unclear	unclear	low	No
Moraiti, 201429	unclear	unclear	low	low	Yes
Warner and Parsons35	high	high	unclear	unclear	No
Boileau et al4	low	unclear	low	unclear	Yes
Lee, 201656	high	high	high	unclear	No
Gasbarro, 201648	unclear	unclear	unclear	unclear	No
Shin et al40	unclear	unclear	high	unclear	No
Paribelli, 201559	high	high	low	unclear	No
Franceschi, 201547	high	high	unclear	unclear	No