Table 2. Subgroup analyses of relative risk of colon and rectal cancer.

	No. of reports	Relative risk	(95%CI)	I2	P-value for heterogeneity
Colon cancer
Primary meta-analysis	35	1.30	1.22–1.39	41.70%	0.006
Sex
Men	19	1.30	1.18–1.42	55.80%	0.002
Women	11	1.28	1.19–1.41	2.30%	0.426
Study design
Case–control studies	14	1.46	1.27–1.68	7.70%	0.368
Cohort studies	21	1.27	1.18–1.36	50.40%	0.005
Ethnicity
Asia population	6	1.58	1.11–2.26	8.90%	0.36
Western population	29	1.29	1.21–1.38	45.50%	0.005
Sedentary behaviour domain
Occupational	25	1.30	1.20–1.40	49.60%	0.003
Recreational	10	1.32	1.17–1.49	13.00%	0.323
Study quality
⩾5	18	1.25	1.17–1.35	9.80%	0.338
<5	17	1.38	1.23–1.54	59.10%	0.001
Controlling BMI in models
Yes	17	1.24	1.14–1.35	12.60%	0.306
No	18	1.36	1.24–1.49	56.50%	0.002
Controlling physical activity in models
Yes	9	1.30	1.16–1.46	0.00%	0.505
No	26	1.31	1.21–1.42	50.90%	0.002
Rectal cancer
Primary meta-analysis	28	1.05	0.98–1.13	19.90%	0.175
Sex
Men	16	1.09	1.03–1.16	0.00%	0.454
Women	8	0.98	0.87–1.10	0.00%	0.579
Study design
Case–control studies	13	1.06	0.85–1.33	39.90%	0.068
Cohort studies	15	1.06	1.01–1.12	0.00%	0.472
Ethnicity
Asia population	5	0.87	0.62–1.22	0.00%	0.411
Western population	23	1.06	0.99–1.14	22.40%	0.164
Sedentary behaviour domain
Occupational	22	1.05	0.98–1.14	24.30%	0.148
Recreational	6	1.03	0.80–1.31	14.30%	0.323
Study quality
⩾5	12	1.07	1.00–1.15	0.00%	0.474
<5	16	1.04	0.93–1.17	34.70%	0.084
Controlling BMI in models
Yes	11	1.02	0.91–1.16	0.00%	0.459
No	17	1.06	0.97–1.16	31.80%	0.102
Controlling physical activity in models
Yes	3	1.34	1.03–1.73	0.00%	0.857
No	25	1.03	0.96–1.11	21.00%	0.173