Table 5.

RRs and 95% CIs of UADT cancer according to quintiles of total meat, its subtypes, fish, and heme iron intake in the EPIC study stratified by smoking status

Lifelong nonsmokersFormer smokersCurrent smokers
Cases (N)RR (95% CI)aCases (N)RR (95% CI)aCases (N)RR (95% CI)a
Total meat (g/1,000 kcals)
 Q1 (Reference)211.00141.00311.00
 Q2210.97 (0.51–1.86)231.25 (0.63–2.49)571.09 (0.67–1.73)
 Q3251.13 (0.59–2.15)261.27 (0.64–2.54)871.40 (0.90–2.17)
 Q4140.64 (0.30–1.34)351.54 (0.78–3.02)1221.65 (1.07–2.54)
 Q5190.88 (0.43–1.81)501.97 (1.02–3.19)1311.60 (1.03–2.47)
Ptrend0.500.020.01
P for interaction with smoking = 0.01c
 Per 20 g/1,000 kcals, observedb0.98 (0.81–1.18)1.20 (1.04–1.38)1.12 (1.03–1.22)
 Per 20 g/1,000 kcals, predictedb0.86 (0.58–1.27)1.28 (0.97–1.69)1.30 (1.10–1.54)
Red meat (g/1,000 kcals)
 Q1 (Reference)171.00141.00471.00
 Q2211.59 (0.75–3.39)181.00 (0.47–2.15)530.74 (0.48–1.14)
 Q3191.28 (0.58–2.83)311.45 (0.70–2.97)750.77 (0.51–1.19)
 Q4221.37 (0.61–3.07)431.90 (0.93–3.87)1130.89 (0.59–1.36)
 Q5211.24 (0.53–2.92)421.70 (0.81–3.57)1400.89 (0.58–1.37)
Ptrend0.990.070.68
P for interaction with smoking = 0.10c
 Per 10 g/1,000 kcals, observedb1.03 (0.89–1.20)1.12 (1.00–1.25)1.03 (0.96–1.10)
 Per 10 g/1,000 kcals, predictedb0.90 (0.63–1.29)1.16 (0.92–1.45)1.04 (0.91–1.20)
Poultry (g/1,000 kcals)
 Q1 (Reference)261.00241.00781.00
 Q2160.56 (0.26–1.21)290.80 (0.40–1.59)1190.85 (0.61–1.20)
 Q3160.47 (0.22–1.04)270.70 (0.35–1.43)860.67 (0.47–0.97)
 Q4190.56 (0.25–1.21)370.99 (0.50–1.98)860.67 (0.46–0.96)
 Q5230.61 (0.28–1.32)310.87 (0.42–1.80)590.64 (0.42–0.96)
Ptrend0.810.760.04
P for interaction with smoking = 0.13c
 Per 5 g/1,000 kcals, observedb0.97 (0.86–1.09)1.03 (0.94–1.12)0.91 (0.84–0.98)
 Per 5 g/1,000 kcals, predictedb0.88 (0.64–1.20)1.11 (0.91–1.36)0.92 (0.78–1.10)
Processed meat (g/1,000 kcals)
 Q1 (Reference)201.00201.00321.00
 Q2321.64 (0.86–3.12)250.76 (0.41–1.43)590.98 (0.63–1.54)
 Q3180.96 (0.46–2.00)270.74 (0.40–1.39)1021.34 (0.87–2.05)
 Q4201.10 (0.53–2.29)371.01 (0.55–1.86)941.16 (0.75–1.80)
 Q5100.65 (0.26–1.60)391.00 (0.52–1.90)1411.89 (1.22–2.93)
Ptrend0.120.50<0.0001
P for interaction with smoking <0.0001c
 Per 10 g/1,000 kcals, observedb0.87 (0.69–1.08)1.11 (0.97–1.26)1.18 (1.10–1.27)
 Per 10 g/1,000 kcals, predictedb0.84 (0.54–1.30)1.16 (0.87–1.56)1.33 (1.15–1.55)
Fish (g/1,000 kcals)
 Q1 (Reference)151.00211.00621.00
 Q2171.74 (0.63–4.84)291.28 (0.63–2.60)911.07 (0.73–1.58)
 Q3181.57 (0.53–4.69)351.15 (0.54–2.44)900.91 (0.60–1.38)
 Q4302.31 (0.79–6.76)381.20 (0.55–2.58)1041.07 (0.70–1.64)
 Q5201.43 (0.46–4.40)250.81 (0.35–1.85)910.90 (0.57–1.42)
Ptrend0.840.200.52
P for interaction with smoking = 0.95c
 Per 10 g/1,000 kcals, observedb0.96 (0.79–1.16)0.85 (0.71–1.02)1.01 (0.91–1.11)
 Per 10 g/1,000 kcals, predictedb0.97 (0.59–1.64)0.77 (0.52–1.13)1.12 (0.91–1.37)
Haem iron (μg/1,000 kcals)
 Q1 (Reference)211.00141.00471.00
 Q2231.14 (0.61–2.14)261.50 (0.77–2.94)650.95 (0.64–1.40)
 Q3120.62 (0.29–1.33)362.03 (1.05–3.91)740.92 (0.62–1.37)
 Q4170.83 (0.40–1.73)281.58 (0.78–3.19)980.99 (0.67–1.46)
 Q5241.17 (0.57–2.39)422.14 (1.08–4.24)1271.19 (0.81–1.75)
Ptrend0.680.060.13
P for interaction with smoking = 0.32c
 Per 200 μg/1,000 kcals, observedb1.06 (0.86–1.32)1.16 (0.99–1.36)1.10 (1.01–1.21)
 Per 200 μg/1,000 kcals, predictedb1.03 (0.92–1.17)1.02 (0.91–1.14)1.05 (0.99–1.12)

aRRs are derived from multivariate Cox regression stratified by age at recruitment and center, and adjusted for nonconsumer status (0/1), sex, energy intake from fat and nonfat sources, education (none/primary, technical/professional, secondary school, university, not specified), alcohol consumption (g/d), drinking history (never, former, unknown), BMI (kg/m2), physical activity (inactive, moderately inactive, moderately active, and active), citrus and noncitrus fruits, and vegetables. Red meat, poultry, and processed meat were mutually adjusted.

bObserved, intake of meat/iron was estimated from the dietary questionnaire; predicted, intake of meat/iron was calibrated using data of the 24-hour diet recall of the calibration study participants.

cP values based on the likelihood ratio test are for comparisons of the model that included interaction terms between the respective dietary variable and smoking status with the model that did not include these interactions (with 2 degrees of freedom). P values (two-sided) < 0.05 are considered statistically significant.