Synergistic Effects of Early Psychosocial Factors and Polygenic Risk for Smoking: a Cross-Sectional Analysis of a Sample of Older Adults in the United States.

Genetic influence on smoking may be modified by negative early psychosocial factors (EPFs). However, few studies have examined this interaction. We hypothesized that negative EPFs will exacerbate the association between genetic risk for smoking and ever smoking. We used data on European Americans (EA) and African Americans (AA) from the Health and Retirement Study (HRS). Polygenic score for "ever smoking" (PGS) was assessed using PRSice from a 2010 GWAS conducted by the Tobacco and Genetics Consortium. We operationalized PGS as a binary variable: top 25% versus the rest. Ever smoking was assessed by asking participants if they had smoked 100 or more cigarettes in their lifetimes. EPFs included education levels of mother or father, perceived financial status, maternal warmth, and stressful events before age 18. We used logistic regression to assess the odds ratio (OR) for PGS, EPFs, and their interaction terms in relation to ever smoking, adjusting for age, gender, and the five principal components. Among the 6969 EA participants (mean age: 74.3 years, 57% women), 55% reported ever smoking. Among the 2141 AA participants (mean age: 59.6, 67.3% women), 61.1% reported ever smoking. Within the AA sample, high genetic risk (top 25%) was associated with a 18% higher likelihood of ever smoking for AA (95% confidence interval [CI] = 0.91-1.52). Within the EA sample, high genetic risk was associated with a 25% higher likelihood of ever smoking (95% CI = 1.11-1.41), and low maternal warmth and PGS showed significant additive interactions to increase odds of ever smoking. The expected OR for those exposed to both high genetic risk and low maternal warmth exceeded the sum of their individual effects (relative risk due to interaction [RERI] = 0.42 [0, 0.85], p = 0.03), resulting in 75% higher odds of smoking compared to individuals with none of these exposures. These synergistic effects observed within the EA sample were not observed within the AA sample. Several EPFs and PGS were associated with increased odds of ever smoking. Among EA participants, low maternal warmth was associated with exacerbated genetic predisposition to smoking. No synergy between EPFs and genetic risk was observed for AA participants.