Abstract 232: Urinary Proteomics in Patients with Type 2 Diabetes at High Cardiovascular Risk

Background: We have previously demonstrated the value of a urinary proteomic classifier (CKD273) for diagnosis and prediction of diabetic nephropathy (DN). Whether CKD273 is only related to renal damage or also reflects generalised vascular damage in patients with diabetes remains unclear. Methods: We recruited 45 patients at high cardiovascular (CV) risk from secondary care: 15 normoalbuminuric; 15 with MA; and 15 with DN (urinary albumin:creatinine ratio: 1.1 (0-3.3), 7.7 (2.6-22.5) and 124.5 (0.8-412.6) mg/mmol; estimated glomerular filtration rate (eGFR): 74 (46-125), 69 (49-100) and 37 (6-60) mL/min/1.73m 2 ). Participants underwent pulse wave analysis for assessment of heart rate-corrected central augmentation index (AIx@75) and ultrasound for assessment of carotid intima-media thickness (c-IMT). Spot urine samples were analysed using capillary electrophoresis coupled to mass spectrometry. Results: There was no difference in age (61±8, 64±6 and 59±7 years; P =0.130), body mass index (34.4±6.2, 35.1±8.1 and 34.4±6.7 kg/m 2 ; P =0.955) and blood pressure (144±15/83±7, 149±20/83±10 and 148±16/82±12 mmHg; P =0.765/0.910) between groups. In keeping with inclusion criteria participants were at high CV risk (Framingham score: 30±11, 38±12 and 32±12; P =0.141; ASSIGN Score: 36±15, 43±15 and 39±17; P =0.415) and had subclinical vascular damage (AIx@75: 22 (7-38), 23 (13-21) and 25 (4-35)%; P =0.993; c-IMT: 0.723 (0.563-1.276), 0.760 (0.614-1.082) and 0.704 (0.581-0.986) mm; P =0.305) independent of eGFR ( r =0.259, P =0.086 for c-IMT; r =-0.082, P =0.598 for AIx@75). Despite similarities in CV risk and vascular phenotypes the CKD 273 classifier was significantly different between the groups (-0.169±0.373, 0.421±0.467 and 0.765±0.434; P =0.002) but not related to c-IMT ( r =-0.075, P =0.747) or AIx@75 ( r =-0.299, P =0.200). The previously defined CKD273 cut-off of 0.343 had 71% sensitivity for detecting MA or DN with specificity of 100%. Conclusions: CKD273 reliably distinguished normoalbuminuria from MA or DN in this high risk cohort independent of the vascular phenotype. Neither traditional renal markers nor a novel proteomic classifier appear to fully explain the vascular damage in our patients.