“…After removing duplicates (n=98), 188 records were screened using titles and abstracts of which 38 studies retained for the full-text review. Finally, a total of 15 studies were included in the systematic review ( Watanabe et al, 2001 ; Toledo et al, 2003 ; Takahashi et al, 2010 ; Wang et al, 2010 ; Gupta et al, 2013a ; Gupta et al, 2013b ; Gupta et al, 2013c ; Ip et al, 2013 ; Ip et al, 2014 ; Sahin et al, 2014 ; Bhatia et al, 2015 ; Stice et al, 2015 ; Aizawa et al, 2016 ; Gupta et al, 2016 ; Bhatia et al, 2018 ), of which 10 studies ( Watanabe et al, 2001 ; Toledo et al, 2003 ; Takahashi et al, 2010 ; Wang et al, 2010 ; Gupta et al, 2013c ; Ip et al, 2013 ; Ip et al, 2014 ; Sahin et al, 2014 ; Stice et al, 2015 ; Aizawa et al, 2016 ) included in the meta-analysis. The major reasons for exclusion were not containing the outcome of interest (n=9), not related to exposure of interest, and being in vitro study (n=5) ( Figure 1 ).…”