220-OR: Plasma Adrenomedullin and Allelic Variation in the ADM Gene and Risk for Lower Extremity Amputation in People with Type 2 Diabetes

Patients with diabetes have an increased risk for lower limb amputation (LLA) but biomarker to assess risk of LLA is lacking. Adrenomedullin (ADM), a vasodilator peptide, has been shown to be associated with cardiovascular and renal risk in diabetes. The aim of this study is to assess the association between plasma MR-proADM (a surrogate of ADM), allelic variation in the ADM gene and the risk for LLA in cohorts of patients with type 2 diabetes. We studied 4532 participants of the DIABHYCAR and SURDIAGENE cohorts (men 70%, mean age 66, mean duration of diabetes 13 years, A1c 8,0%). MR-proADM was measured by immunofluorescence. Five SNPs were chosen in the haplotype block containing the ADM gene (chr11p15.4): rs4399321, rs11042725, rs7944706, rs2957692 and rs2957717. The mean follow-up was 5.3 years. Incidence of LLA by tertiles of MR-proADM (T1, T2, T3) was 1.0% (T1), 1.9% (T2) et 4.7% (T3) (p<0,001). In a Cox multivariable analysis model (adjusted for cohort, age, sex, diabetes duration, hypertension, A1c, cholesterol, history of peripheral arterial disease and glomerular filtration rate eGFR), HR for LLA was 4.21 (2.18-8.66; p<0.001) for T3 vs. T1, 2.46 (1.50-4.12; p=0.003) for T3 vs. T2 and 1.82 (1.25-2.67; p=0.002) for 1 unit of log [ADM]. All SNPs were significantly associated with baseline plasma ADM. Four were significantly associated with LLA: HR adjusted for sex, age, duration of diabetes, A1c, eGFR and cohort: 1.40 (1.046 - 1.84; p=0.02) for the G-allele of rs4399321 in a codominant model; 1.71 (1.13 - 2.54; p=0.01) for the AA vs. CX genotype of rs11042725; 1.61 (1.09 - 2.35; p=0.02, for the GG vs. AX genotype of rs7944706; 1.37 (1.06 - 1.77; p=0.04) for the G-allele of rs2957692 in a codominant model. We observed an association of plasma ADM with LLA and of allelic variations in the AMP locus with plasma ADM and with LLA in a large population of people with type 2 diabetes. This pattern of mendelian randomisation supports the causality of the association of plasma ADM with LLA.