Association of alcohol with lung cancer risk in men with different growth hormone receptor genotypes

To test whether genetic variants of the growth hormone receptor gene (GHR) modulate the effect of lifestyle variables on lung cancer (LC) risk.

This population-based cohort study involved 6,439 men from the Japan-Hawaii Cancer study drawn from the Kuakini Honolulu Heart Program who were cancer-free at baseline examination (1965-1968; age 45-68 years) and followed-up until December 1999. We determined the association of GHR SNP rs4130113 genotypes GHR-AA (common allele A homozygotes) and GHR-G (minor allele G carriage) with alcohol drinking, BMI, physical activity and cigarette smoking in relation to LC and non-small cell LC (NSCLC). Results were expressed as hazard ratios and 95 % CIs estimated from Cox proportional hazard models.

Over mean 26.7 ± 7.4 SD years follow-up, 190 LC cases, including 133 NSCLC cases, were diagnosed. After adjusting for age, education, alcohol intake, BMI, physical activity, cigarette smoking, green tea consumption and dietary saturated fat, main-effect Cox models showed that compared with GHR-AA, GHR-G was associated with protection against LC: HR = 0.75 (95 % CI, 0.56-1.00). Full Cox models showed GHR-G interacted with alcohol intake only (β = 1.171; p = 0.0003; drinks per week: β = 0.279; P = 0.0024). Stratified analyses showed that for GHR-AA, drinkers had reduced LC risk (HR = 0.54; 95 % CI, 0.35-0.85), and that <2 drinks/week had the strongest protection (HR = 0.38; 95 % CI, 0.18-0.83). In contrast, for GHR-G, alcohol drinkers had increased LC risk (HR = 1.70; 95 % CI, 1.07-2.69) which was dose-dependent (P for trend = 0.005). Results for NSCLC were similar.

In men with the GHR-AA genotype, alcohol drinking at a low dose poses significantly less risk of LC compared with non-drinkers and higher alcohol consumption., the overall relationship being U-shaped. In contrast, in GHR minor allele carriers, alcohol posed a progressively greater risk of LC as amount consumed increased.