Table 3.
Number of risks | n | HR (95% CI) | p | PAR (%)† |
---|---|---|---|---|
Risk genotypes£ | ||||
0 | 1,495 | reference | 66.8 | |
1 | 491 | 9.57 (3.08 – 29.67) | 9.20E-05 | |
Risk lifestyles ¶ | ||||
0 | 1,850 | Reference | 26.2 | |
1 | 136 | 6.63 (2.30 – 19.11) | 0.0005 | |
Risk genotypes plus lifestyle factors § | ||||
0 | 1,394 | Reference | ||
Risk genotypes only | 456 | 8.55 (2.27 – 32.24) | 1.53E-03* | 73.3 |
Risk lifestyles only | 101 | 4.97 (0.52 – 47.76) | 0.1653 | |
Both risks of genotypes and lifestyles | 35 | 58.76 (13.15 – 262.68) | 9.73E-08* | |
p trend | 2.00E-06 |
CI, confidence interval; HR, hazard ratio; PAR, population attributable risk. Numbers in bold face are statistically significant.
†PAR(%) reflects, in total Hispanic African women, a risk of colorectal cancer attributable to the risk genotypes and the risk lifestyles, both singly and in combination.
£The number of risk genotypes (IFT172 rs780104 GG; GCKR rs6753534 CC; and NRBP1 rs704791 TT) was defined as follows: 0 (none/1 risk allele) vs. 1 (2/3 risk alleles).
¶The maximum combined number of lifestyles (> 15.9% of calories from monounsaturated fatty acid [FA]/day, ≥ 25 cigarettes/day, ≤ 38 years old at menopause, > 12.4% of calories from saturated FA/day, ≤ 4.7% of calories from polyunsaturated FA/day, and ≤ 55.6 mg of dietary vitamin K) was 3. The number of lifestyles was determined on the basis of analysis for the combined lifestyle factors ( Figure S2B ) and defined as follows: 0 (null/1/2 risk lifestyles) vs. 1 (3 risk lifestyles).
§The combined number of risk genotypes and risk lifestyles was based on risk genotypes defined as 0 (none/1 risk allele) and 1 (2/3 risk alleles), and risk lifestyles defined as 0 (null/1/2 risk lifestyles) and 1 (3 risk lifestyles). The ultimate number of risk genotypes combined with risk lifestyles was defined as 0 (no risk genotypes and risk lifestyles); and risk genotypes (only risk genotypes) and risk lifestyles (only risk lifestyles), separately and together.
*p values with false discovery rate < 0.05 were shown after multiple comparison corrections via the Benjamini-Hochberg method.