Evidence for Gene-smoking interactions for hearing loss in Japanese American families

Background/Purpose:

Hearing loss is the most common sensory impairment, affecting about 13% of the US population.  Hearing is a complex process and is strongly influenced by genetic variation, with over 400 known susceptibility genes. The importance of these genes varies by population. An association between smoking and hearing loss is also well-established, but gene-by-smoking interactions have not been reported.

Methods:

In a Japanese American sample of 133 subjects in 17 families, the association between smoking and self-reported hearing loss or deafness (HLD) was assessed in a logistic regression analysis adjusted for family structure.  Non-parametric linkage analysis (NPL) was performed using MERLIN software and 374 genome-wide polymorphic markers in these families.  Ordered subset analysis (OSA) was then used to assess evidence for interactions between smoking and genetic variation on HLD.  OSA reduces genetic heterogeneity by finding the subset of families that maximize evidence for linkage. Families were ranked by the proportion of ever smoking family members, and then sequentially added to the linkage analysis until the maximum LOD was achieved. Statistical significance was assessed by permutation testing. 

Results/Outcomes:

The association between HLD and smoking was confirmed in our sample, with ever smokers 4 times more likely than nonsmokers to report HLD (OR=4.40; 95% CI, 1.31 to 14.78). Evidence for linkage with HLD genes was observed in multiple genomic regions and was stronger in subsets of families with a larger percent of ever smokers. Five regions are interpreted as suggestive of gene-by-smoking interactions: (1) 5 families (n=43 subjects) at D15S816 marker (chr15:95019784bp, build 37) with maxLOD=2.88, p-value=0.019; (2) 3 families (n=27) at D12S1045 (chr12:130397738bp) with maxLOD=2.60, p-value=0.020;  (3) at D20S171 (chr20:57808030bp) with maxLOD=3.03, p=value=0.043; (4) 2 families (n=13) at D11S1986 (chr11:111223605bp) with maxLOD=2.09, p-value=0.051; and (5) at GATA137H02 marker (chr7:18284528bp) with maxLOD=3.32, p-value=0.086. 

Conclusion:

To our knowledge this is the first report of possible gene-by-smoking interactions in HLD, and will require confirmation and additional work to understand the basis of these findings. Hearing loss and deafness are important public health issues and understanding the contributions of both genetic and environmental factors may inform policy regarding screening.