Table 3. Results of sensitivity analysis.

Studies omitted	RR	95% CI	I2	P for effect
van Vlerken 2012	1.37	1.18–1.60	74%	<0.001
Bajaj 2012	1.42	1.21–1.66	73%	<0.001
de Vos 2013	1.41	1.20–1.66	74%	<0.001
Min 2014	1.38	1.18–1.62	74%	<0.001
Kwon 2014	1.37	1.17–1.60	73%	<0.001
Mandorfer 2014	1.42	1.21–1.67	73%	<0.001
Dultz 2015	1.36	1.17–1.59	73%	<0.001
Merli 2015	1.43	1.23–1.67	71%	<0.001
Cole 2016	1.39	1.19–1.63	74%	<0.001
Kim 2017	1.38	1.18–1.62	74%	<0.001
Miozzo 2017	1.43	1.22–1.67	70%	<0.001
Hung 2018a	1.40	1.18–1.67	74%	<0.001
Hung 2018b	1.41	1.19–1.68	74%	<0.001
Smith 2018	1.38	1.18–1.62	74%	<0.001
Tergast 2018	1.37	1.17–1.60	74%	<0.001
Janka 2019	1.36	1.16–1.59	72%	<0.001
Nardelli 2019	1.37	1.17–1.60	73%	<0.001
Lewis 2019	1.31	1.14–1.50	62%	<0.001
Dam 2019	1.42	1.22–1.67	73%	<0.001
Hung 2019	1.40	1.19–1.67	74%	<0.001
De Roza 2019	1.36	1.17–1.59	73%	<0.001

Abbreviations: CI, confidence interval; RR, risk ratio.