Lack of interaction of beta-cell-function-associated variants with hypertension on change in fasting glucose and diabetes risk: the Framingham Offspring Study.

OBJECTIVE: To test whether pancreatic beta-cell genetic frailty and hypertension (HTN) interact in their associations with change over time in fasting glucose (ΔFG) or type 2 diabetes mellitus (T2D) risk. METHODS AND RESULTS: We pooled data from 3471 Framingham Offspring Study participants into six ∼4-year periods (15 852 person-examinations; mean age 52; 54% women). We defined two genetic exposures reflecting beta-cell genetic risk burden: single nucleotide polymorphism (SNP) score counts of fasting glucose-associated and T2D-associated risk alleles at 16 and 33 putative beta-cell loci, respectively; and three HTN exposures: HTN versus no-HTN; treated versus untreated HTN; and five mutually exclusive antihypertensive categories (beta-blockers, thiazides, renin-angiotensin system agents, combinations, others) versus untreated HTN. We tested ∼4-year mean ΔFG or odds of T2D by per-risk allele score change and HTN category, seeking genetic score-by-HTN interaction. Genetic scores increased ∼4-year ΔFG (0.6  mg/dl per-risk allele; P = 8.9 × 10(-16)) and T2D-risk (∼17% per-risk allele; P = 2.1 × 10(-7)). As compared to no-HTN, HTN conferred higher ΔFG (2.6 versus 1.7 mg/dl; P < 0.0001) and T2D-risk [odds ratio (OR) = 2.9, 95% confidence interval (CI) 2.8-3.0; P < 0.0001]. As compared to untreated HTN, treated HTN conferred higher ΔFG (3.4 versus 3.0 mg/dl; P < 0.0001) and T2D-risk (OR = 1.4, 95% CI 1.3-1.5; P = 0.02). Beta-blockers (OR = 1.6, 95% CI 1.1-2.4), combinations (OR = 1.6, 95% CI 1.1-2.5), and others (OR = 2.0, 95% CI 1.4-2.9) increased T2D-risk (all P < 0.02). In joint models including interaction terms, all genetic score-by-HTN interaction terms were P value greater than 0.05. In joint models without interaction, fasting glucose-SNP or T2D-SNP genetic scores (both P < 0.001) and HTN (P < 0.0001) independently increased ΔFG or T2D-risk. CONCLUSION: HTN, HTN treatment, and common fasting glucose-SNP genetic score/T2D-SNP genetic score independently predicted ΔFG and T2D incidence, but did not modify each other's association with ΔFG or T2D risk.

Duke Scholars

Author Michael J Pencina Biostatistics & Bioinformatics, Division of Biostatistics