Skip to main content
. 2020 Jan 22;2020:1487609. doi: 10.1155/2020/1487609

Table 2.

Results of subgroup meta-analysis and meta-regression analysis.

Subgroup No. of studies No. of patients Random-effects model Heterogeneity Meta-regression
P value
HR (95% CI) P value Ph I 2 (%)
Cancer type 0.860
NSCLC 3 320 2.50 (1.07–6.07) 0.035 <0.001 90
CRC 3 500 1.95 (1.34–2.84) <0.001 0.735 0
GC 2 256 2.32 (1.48–3.64) <0.001 0.636 0
Others 4 552 1.90 (1.00–3.61) 0.05 <0.001 83.3

Ethnicity 0.111
Asian 11 1283 2.35 (1.17–3.11) <0.001 0.009 57.6
Caucasian 1 345 1.03 (0.90–1.17) 0.659

Sample size 0.066
≥100 9 1484 1.83 (1.33–2.50) <0.001 <0.001 74.8
<100 3 144 3.60 (1.52–8.52) 0.004 0.012 77.6

Analysis method 0.772
Univariable 6 936 2.00 (1.23–3.25) 0.005 <0.001 81.8
Multivariable 6 692 2.32 (1.44–3.76) 0.001 0.001 76.7

Cut-off value 0.208
Median 9 1416 2.19 (1.39–3.45) 0.001 <0.001 88.9
Mean and others 3 212 1.91 (1.31–2.79) 0.001 0.464 0

Study design 0.924
Retrospective 10 1372 2.14 (1.41–3.26) <0.001 <0.001 87.3
Prospective 2 256 2.32 (1.48–3.64) <0.001 0.636 0

NSCLC, non-small-cell lung cancer; CRC, colorectal cancer; GC, gastric cancer; Ph, the P value of Q test for heterogeneity test.