NIOSHTIC-2 Publications Search

Design and analysis issues in studies of gene-environment.

Paolo V; Schulte P; Carreon T; Bailer AJ; Medvedovic M

EPICOH 16th International Conference on Epidemiology in Occupational Health, Barcelona, Spain, September 14, 2002. Rome, Italy: International Commission on Occupational Health, 2002 Sep; :1-9

20027176

Making sense from population studies incorporating genes and environment has a rich history, but one fraught with problems. The nature-nurture debate has given way to the fundamental realization that both genes and environment play a role in disease, especially cancer. Interaction has been defined in several ways, for example as "the coparticipation of two or more agents in the same casual mechanism leading to disease development" (Yang and Khoury, 1997). According to a more generally accepted definition, interaction occurs when the effect of multiple factors differs from the effect predicted on the basis of the separate effects of each of the factors. Interactions need to be explicitly considered in the design and analysis of epidemiologic studies. If only the association between a genotype and disease is examined, the effect of biologic interaction cannot be appreciated (Khoury et al., 1993). Epidemiologists need to measure relevant risk factors, in addition to genotype(s). Also, it is likely that each exposure interacts with genes along a pathway (typically: DNA repair genes), so that multiple interactions are likely to be the norm. Evidence for a three-way interaction as showed by Taylor et a1., (1998) for NAT2 and NAT1 genotypes and smoking as risk factors for bladder cancer. Depending on the type of model for gene-environment interaction, risk estimates can vary dramatically. At least five different types of gene-environment interaction have been described (Khoury and Wagener, 1995). These have been generally considered with the dichotomous conditions, a single susceptibility genotype (present or absent), and a single enviromenta1 factor (present or absent). For example, in one analysis (Ottman, 1996) the risk to an exposed person with a rare (I % prevalence) susceptibility genotype ranges from 8% (under a multiplicative model of interaction) to 95.6% (when we assume that the exposure has no effect in susceptibles, but the genotype raises risk in unexposed as well as exposed persons).

Case-studies; Mortality-rates; Mortality-data; Mortality-surveys; Morbidity-rates; Epidemiology; Statistical-analysis; Cancer-rates; Cancer

20020914

Abstract

2002

EID

EPICOH 16th International Conference on Epidemiology in Occupational Health, Barcelona, Spain, September 14, 2002

OH