In Reply: We agree with Drs Koenen and Galea and Drs Lotrich and Lenze regarding the importance of consistent environmental measures in studies of gene-environment interaction. However, it is also important to acknowledge the myriad ways an environmental exposure is defined and potentially adjusted to identify an interaction effect in gene-environment interaction studies when assessing statistical significance. For example, Lotrich and Lenze are critical of our meta-analysis for excluding studies with single heterogeneous life stressors such as job loss, pregnancy, hip fracture, or heart disease. However, exposure to a single life event was not significantly associated with depression in the original study of Caspi et al,1 in which (as shown in Figure 1 of that article) an interaction with the 5-HTTLPR genotype occurred only among persons with 3 or more stressful life events.
Merikangas KR, Lehner T, Risch NJ. Gene-Environment Interactions and Depression—Reply. JAMA. 2009;302(17):1859-1862. doi:10.1001/jama.2009.1579