Apo E genotype, diabetes, and peripheral arterial disease in older men: The Honolulu Asia-Aging study

The epsilon 4 allele of the gene coding for apolipoprotein (apo) E is assoc iated with an atherogenic lipid profile that has been linked to increased r isk of coronary artery disease (CAD). Apo E genotype may also be associated with peripheral arterial disease (PAD). If present, this association may b e modified by diabetes, which is also associated with dyslipidemia that pre disposes to macrovascular disease. Observable associations between both Apo E genotype and diabetes with PAD may be confounded by smoking, a potent PAD risk factor that is unrelated to lipids. From 1991 to 1993, apo E genotype s (2/3, 3/3, 3/4), PAD (defined as ankle-brachial index [ABI] <0.9), diabet es (prevalent and newly-diagnosed), and smoking history (ever/never) were d etermined for 3,161 Japanese-American men aged 71-93. Data on hypertension and other potential confounders were also collected. Logistic regression wa s used to determine odds ratios (OR) between groups cross-categorized by ap o E genotype and diabetes with prevalence of PAD, within strata of smoking. In each smoking stratum, non-diabetic apo epsilon 3/3 carriers were consid ered the reference. Among ever-smokers, there was no association between ap o E acid PAD, regardless of diabetes status. Among never-smokers there appe ared to be both apo E- and diabetes effects on PAD prevalence. Compared to the non-diabetic epsilon 3/3 group, the ORs of PAD were 2.3 (1.2-4.4) and 2 .0 (1.1-3.4) for epsilon 3/3 newly-diagnosed and epsilon 3/3 prevalent diab etic subjects, respectively. Associations were stronger among diabetic indi viduals in the epsilon 3/4 group: the ORs were 3.0 (1.1-8.8) and 4.1 (1.9-8 .7) for epsilon 3/4 newly-diagnosed and epsilon 3/4prevalent diabetic subje cts, respectively. Despite associations whose pattern and magnitude suggest ed interaction between apo E genotype and diabetes on PAD prevalence among never-smokers, formal testing of this interaction did not reach statistical significance. Our finding of an apo E-PAD association among never-smokers may result from the effects of an apo epsilon 4-related atherogenic lipid p rofile on peripheral arteries. Further studies are needed to clarify the po tential mediating role of diabetes on the apo E-PAD association. Genet. Epi demiol. 19:52-63, 2000. (C) 2000 Wiley-Liss, Inc.