. 2020 Sep 29;146(12):3123–3135. doi: 10.1007/s00432-020-03398-1

Table 2.

Subgroup analysis and meta-regression

Subgroup analysis	No. of studies	No. of patients	Pooled HR (95% CI)		Overall effect P value	Meta regression (P value)	Heterogeneity
Subgroup analysis	No. of studies	No. of patients	Fixed	Random	Overall effect P value	Meta regression (P value)	I² (%)	P value
Region						0.015
Western	14	1565	0.986 (0.975, 0.997)	0.844 (0.725, 0.982)	0.028		60.90%	0.002
Eastern	9	704	0.385 (0.289, 0.513)	0.354 (0.244, 0.515)	< 0.001		35.30%	0.136
Sample size						0.988
< 100	16	1185	0.986 (0.975, 0.997)	0.668 (0.544, 0.821)	< 0.001		76.80%	< 0.001
> 100	7	1084	0.589 (0.462, 0.751)	0.589 (0.462, 0.751)	< 0.001		0.00%	0.561
NOS score						0.503
< 7	13	1017	0.522 (0.404, 0.675)	0.496 (0.332, 0.740)	0.001		53.00%	0.012
> 7	10	1252	0.986 (0.975, 0.997)	0.771 (0.645, 0.921)	0.004		76.10%	< 0.001
Detection method						0.813
Whole-tissue sections	14	1288	0.985 (0.974, 0.996)	0.505 (0.353, 0.722)	< 0.001		78.10%	< 0.001
Tissue microarrays	9	981	0.963 (0.883, 1.051)	0.664 (0.429, 1.028)	0.066		70.70%	< 0.001
Treatment						0.716
Yes	3	207	1.012 (0.923, 1.108)	0.547 (0.411, 0.729)	< 0.001		81.60%	0.004
No	20	2062	0.985 (0.974, 0.996)	0.573 (0.135, 2.119)	0.404		74.90%	< 0.001

HR hazards ratio, NOS Newcastle–Ottawa Quality assessment scale