Skip to main content
. 2018 Jan 11;18:71. doi: 10.1186/s12885-017-3970-5

Table 2.

Overall and stratified meta-analyses of relative risk estimates (associations) for physical activity and risk of colorectal cancer by BMI subgroups

Overall/stratified analysis Total number of estimates Number of casesa Pooled RR (95% CI) I2 (%) Pheterogeneity P across subgroups P across high and low BMI
Low BMI
 Overall 27 7045 0.74(0.66–0.83) 47.9 0.003
Sex
 Male 12 2590 0.76 (0.64–0.92) 53.1 0.015 0.551
 Female 8 1639 0.72 (0.57–0.91) 33.6 0.160
 Both 7 2816 0.70 (0.56–0.89) 50.2 0.061
Study Design
 Cohort 11 3073 0.76 (0.61–0.94) 59.3 0.006 0.088
 Case-control 16 3972 0.72 (0.62–0.84) 40.7 0.046
Cancer Site
 Colon 17 3436 0.69 (0.59–0.81) 41.5 0.038 0.410
 Colorectal 6 2854 0.74 (0.56–0.98) 55.6 0.047
 Rectal 4 755 0.94 (0.83–1.06) 0.0 0.942
Interactionb
 No 13 3128 0.78 (0.66–0.92) 37.3 0.004 0.462
 Yes 14 3917 0.70 (0.59–0.84) 57.6 0.085
Physical Activity Type
 Total 9 2681 0.74 (0.61–0.89) 32.7 0.156 0.605
 Recreational 12 3874 0.78 (0.64–0.94) 60.5 0.003
 Commuting 6 490 0.64 (0.51–0.79) 0.0 0.675
Physical Activity Measurement
 Adulthood/lifetimec 16 5151 0.72 (0.64–0.82) 16.0 0.271 0.497
 Past year or two 6 1319 0.75 (0.56–1.01) 67.8 0.008
 Regulard 5 575 0.77 (0.51–1.17) 63.8 0.026
Physical Activity Assessment
 MET-h/day or week 13 2188 0.73 (0.62–0.87) 52.4 0.014 0.969
 Kcal/week 2 1993 0.75 (0.54–1.04) 0.0 0.437
 Times/week or month 3 1939 0.75 (0.53–1.05) 78.1 0.010
 Othere 9 925 0.72 (0.54–0.98) 41.6 0.090
Physical Activity Reference Group
 No activity 9 4453 0.70 (0.57–0.85) 41.9 0.088 0.538
 Some activity 18 2592 0.76 (0.65–0.88) 47.8 0.013
Geographical Region
 North America 12 5026 0.82 (0.71–0.94) 47.7 0.033 0.122
 Europe 5 920 0.60 (0.38–0.93) 59.2 0.044
 Asia 8 824 0.66 (0.55–0.80) 7.4 0.373
 Australia 2 275 0.71 (0.38–1.30) 0.0 0.482
High BMI
 Overall 36 9407 0.65 (0.53–0.79) 89.0 <0.001 0.389
Sex
 Male 17 2928 0.56 (0.34–0.94) 74.9 <0.001 0.268 0.264
 Female 9 1568 0.61 (0.47–0.79) 92.1 <0.001 0.333
 Both 10 4911 0.83 (0.65–1.05) 82.8 <0.001 0.384
Study Design
 Cohort 15 4625 0.93 (0.78–1.13) 64.1 <0.001 0.002 0.178
 Case-control 21 4782 0.51 (0.39–0.66) 88.0 <0.001 0.053
Cancer Site
 Colon 20 5330 0.58 (0.44–0.79) 91.5 <0.001 0.282 0.435
 Colorectal 9 2841 0.74 (0.49–1.11) 85.9 <0.001 0.948
 Rectal 7 1236 0.74 (0.62–0.88) 4.1 0.395 0.052
Interactionb
 No 18 4881 0.87 (0.79–0.96) 0.0 0.761 0.001 0.220
 Yes 18 4526 0.48 (0.34–0.68) 93.4 <0.001 0.081
Physical Activity Type
 Total 13 4141 0.81 (0.68–0.96) 40.4 0.065 <0.001 0.424
 Recreational 19 4830 0.74 (0.61–0.91) 75.3 <0.001 0.810
 Commuting 4 436 0.21 (0.14–0.32) 81.9 0.001 0.001
Physical Activity Measurement
 Adulthood/lifetimec 20 7082 0.54 (0.40–0.73) 93.5 <0.001 0.022 0.179
 Past year or two 10 1985 0.75 (0.64–0.87) 3.5 0.408 0.780
 Regulard 6 340 1.00 (0.73–1.38) 21.7 0.270 0.366
Physical Activity Assessment
 MET-h/day or week 13 2411 0.55 (0.37–0.81) 90.8 <0.001 0.086 0.224
 Kcal/week 5 2058 0.57 (0.45–0.72) 20.3 0.285 0.251
 Times/week or month 5 3549 0.90 (0.67–1.19) 87.9 <0.001 0.478
 Othere 13 1389 0.77 (0.59–1.00) 49.8 0.030 0.778
Physical Activity Reference Group
 No activity 14 8521 0.77 (0.67–0.87) 41.2 <0.001 0.132 0.474
 Some activity 22 4745 0.57 (0.40–0.82) 92.4 0.054 0.199
Geographical Region
 North America 21 6990 0.75 (0.64–0.87) 69.6 <0.001 0.127 0.525
 Europe 7 1239 0.67 (0.45–0.99) 69.9 0.003 0.724
 Asia 6 583 0.36 (0.19–0.66) 92.0 <0.001 0.090
 Australia 2 595 0.98 (0.57–1.68) 45.1 0.177 0.520

Abbreviations: BMI body mass index, MET metabolic equivalent of task

aBMI subgroup-specific case numbers were not described for Gerhardsson (1990) [30] Slattery (1997) [36], thus total cases were used for case numbers from these two studies

bEstimates representing an interaction between physical activity and BMI on colorectal cancer risk

cStudies with over 10 years of physical activity measurement were included in this category

dTime period of measurement not specified

eOther measures include a study’s own physical activity index or unique classification of active vs. inactive individuals