Table 3.
Multivariable associations of lifestyle and anthropometric factors with risk of colorectal cancer according to molecular subtypes in the NHS and HPFSa
| Risk factors | All CRC (n = 3063) |
Subtype 1b (Conventional pathway) (n = 498) |
Subtype 2b (Serrated pathway) (n = 138) |
Subtype 3b (Alternate pathway) (n = 367) |
Subtype 4b (Other pathway) (n = 172) |
P
overall
heterogeneity c |
|---|---|---|---|---|---|---|
| Smoking | ||||||
| Pack-years of smoking, per 20 pack-year | ||||||
| HR (95% CI) | 1.09 (1.05 to 1.12) | 0.94 (0.84 to 1.04) | 1.29 (1.14 to 1.45) | 1.04 (0.95 to 1.14) | 1.23 (1.05 to 1.43) | — |
| Ptrend | <.001 | .24 | <.001 | .42 | .01 | <.001 |
| Pheterogeneity (compared to subtype 1)c | — | — | <.001 | .15 | .005 | — |
| Pack-years of smoking before age 30 years, per 20 pack-year | ||||||
| HR (95% CI) | 1.18 (1.06 to 1.33) | 0.76 (0.52 to 1.10) | 1.52 (1.02 to 2.26) | 1.04 (0.76 to 1.42) | 1.60 (1.02 to 2.51) | — |
| Ptrend | .004 | .14 | .04 | .81 | .04 | .03 |
| Pheterogeneity (compared to subtype 1)c | — | — | .01 | .19 | .01 | — |
| Alcohol intake, per 14 g/d | ||||||
| HR (95% CI) | 1.11 (1.07 to 1.15) | 1.09 (0.98 to 1.20) | 1.14 (0.94 to 1.38) | 1.22 (1.07 to 1.38) | 1.24 (1.03 to 1.49) | — |
| Ptrend | <.001 | .10 | .19 | .002 | .02 | .43 |
| Pheterogeneity (compared to subtype 1)c | — | — | .66 | .16 | .22 | — |
| BMI, per 5 kg/m2 | ||||||
| Men | 1282 | 224 | 30 | 157 | 62 | — |
| HR (95% CI) | 1.30 (1.20 to 1.40) | 1.42 (1.21 to 1.67) | 0.82 (0.42 to 1.60) | 1.00 (0.79 to 1.27) | 1.27 (0.91 to 1.77) | — |
| Ptrend | <.001 | <.001 | .55 | .99 | .15 | .06 |
| Pheterogeneity (compared to subtype 1)c | — | — | .12 | .02 | .56 | — |
| Women | 1781 | 274 | 108 | 210 | 110 | — |
| HR (95% CI) | 1.16 (1.10 to 1.22) | 1.10 (0.88 to 1.38) | 1.06 (0.89 to 1.25) | 1.23 (1.08 to 1.41) | 1.06 (0.88 to 1.27) | — |
| Ptrend | <.001 | .39 | .54 | .002 | .55 | .40 |
| Pheterogeneity (compared to subtype 1)c | — | — | .76 | .40 | .77 | — |
| BMI at age 18/21, per 5 kg/m2 e | ||||||
| Men | 1282 | 224 | 30 | 157 | 62 | — |
| HR (95% CI) | 1.07 (1.00 to 1.14) | 1.20 (1.03 to 1.41) | 0.89 (0.73 to 1.09) | 1.17 (0.99 to 1.39) | 1.18 (0.89 to 1.56) | — |
| Ptrend | .06 | .02 | .27 | .07 | .24 | .11 |
| Pheterogeneity (compared to subtype 1 c | — | — | .02 | .82 | .90 | — |
| Women | 1781 | 274 | 108 | 210 | 110 | — |
| HR (95% CI) | 1.21 (1.12 to 1.30) | 1.05 (0.77 to 1.44) | 1.12 (0.89 to 1.41) | 1.30 (1.18 to 1.45) | 1.34 (1.15 to 1.60) | — |
| Ptrend | <.001 | .75 | .34 | .007 | .01 | .50 |
| Pheterogeneity (compared to subtype 1)c | — | — | .76 | .26 | .23 | — |
| Waist circumference, per 10 cm | ||||||
| Men | 1282 | 224 | 30 | 157 | 62 | — |
| HR (95% CI) | 1.30 (1.22 to 1.38) | 1.34 (1.19 to 1.52) | 1.00 (0.69 to 1.46) | 1.05 (0.88 to 1.26) | 1.37 (1.07 to 1.74) | — |
| Ptrend | <.001 | <.001 | .97 | .57 | .01 | .08 |
| Pheterogeneity (compared to subtype 1)c | — | — | .15 | .02 | .90 | — |
| Women | 1781 | 274 | 108 | 210 | 110 | — |
| HR (95% CI) | 1.07 (1.02 to 1.12) | 0.97 (0.80 to 1.18) | 0.99 (0.86 to 1.14) | 1.08 (0.97 to 1.20) | 0.96 (0.79 to 1.18) | — |
| Ptrend | .007 | .77 | .89 | .16 | .71 | .61 |
| Pheterogeneity (compared to subtype 1)c | — | — | .88 | .35 | .95 | — |
| Height, per 10 cm | ||||||
| HR (95% CI) | 1.17 (1.10 to 1.23) | 1.03 (0.89 to 1.18) | 1.52 (1.21 to 1.90) | 1.22 (1.01 to 1.47) | 1.21 (0.91 to 1.61) | — |
| Ptrend | <.001 | .71 | <.001 | .03 | .19 | .04 |
| Pheterogeneity (compared to subtype 1)c | — | — | .004 | .14 | .31 | — |
| Physical activity, per 7.5 METS-hr/wkd | ||||||
| HR (95% CI) | 0.98 (0.96 to 1.00) | 0.98 (0.95 to 1.01) | 0.95 (0.89 to 1.02) | 0.95 (0.91 to 0.98) | 0.98 (0.93 to 1.04) | — |
| Ptrend | .02 | .20 | .16 | .01 | .53 | .73 |
| Pheterogeneity (compared to subtype 1)c | — | — | .72 | .39 | .88 | — |
Inverse probability weighting was applied to reduce a bias due to the availability of tumor tissue after cancer diagnosis (see Statistical Analysis subsection for details). Age- and cohort-stratified Cox proportional hazards model was used with further adjustment for race (White or nonWhite), height (continuous), family history of colorectal cancer (yes or no), history of lower gastrointestinal endoscopic exams (yes or no), body mass index (continuous), pack-years of smoking (continuous), physical activity (continuous), alcohol intake (continuous), and regular aspirin use (yes or no). BMI = body mass index; CI = confidence interval; CIMP = CpG island methylator phenotype; CRC = colorectal cancer; HPFS = Health Professionals Follow-up Study; HR = hazard ratio; NHS = Nurses’ Health Study; METS = metabolic equivalent task score; MIS = microsatellite instability.
Subtype 1 (conventional pathway; non-MSI-high, CIMP-lowor negative, BRAF-wild-type, and KRAS-wild-type); Subtype 2 (serrated pathway; any MSI status, CIMP-high, BRAF-mutated, and KRAS-wild-type); Subtype 3 (alternate pathway; non-MSI-high, CIMP-lowor negative, BRAF-wild-type, and KRAS-mutated); Subtype 4 (other marker combinations).
Heterogeneity across the 4 subtypes was tested by a weighted likelihood ratio test.
For physical activity, the follow-up started in 1986 in NHS.
BMI at age 18/21, BMI at age 18 for female and age 21 for male.