Table 6

Three-way interaction for NAT2 genotype, CYP1A2 phenotype, and red meat preference by smoking status

CYP1A2 Phenotypea
≤Median>Median
NAT2DonenessbNcORd (95% CI)NAT2DonenessNOR (95% CI)
Never smokers
 Slow/intermediateRare/medium32.571.0Slow/intermediateRare/medium32.530.9 (0.5–1.8)
 Slow/intermediateWell-done18.121.9 (0.8–4.8)Slow/intermediateWell-done17.191.3 (0.6–2.9)
 RapidRare/medium16.260.9 (0.4–2.1)RapidRare/medium16.280.8 (0.3–1.9)
 RapidWell-done9.150.8 (0.3–2.2)RapidWell-done9/61.5 (0.4–5.0)
P for interactione = 0.81
Ever smokers
 Slow/intermediateRare/medium40.551.0Slow/intermediateRare/medium61.701.3 (0.7–2.4)
 Slow/intermediateWell-done13.220.9 (0.4–2.2)Slow/intermediateWell-done25.330.6 (0.2–1.5)
 RapidRare/medium27.351.2 (0.6–2.6)RapidRare/medium11.210.9 (0.4–1.9)
 RapidWell-done10.121.3 (0.5–3.5)RapidWell-done12.28.8 (1.7–44.9)
P for interaction = 0.01
  • a As determined by the urinary ratio (17U + 17X):137X; median = 5.1.

  • b Rare/medium includes did not eat, rare, medium-rare, and medium. Well-done includes well-done and very well-done.

  • c Number of cases/number of controls.

  • d ORs (95% CIs) adjusted by unconditional logistic regression for age, sex, ethnicity, lifetime recreational physical activity, lifetime aspirin use, body mass index 5 years ago, years of schooling, intakes of nonstarch polysaccharides from vegetables and calcium from foods and supplements, and pack-years of cigarette smoking, and number of cigarettes, cigars, and pipes smoked during the 2 weeks preceding the caffeine test (smokers model only).

  • e Based on the likelihood ratio test comparing the model with interactions with one containing only main effects for the three variables.