Nutritional biomarkers and heart failure requiring hospitalization in patients with type 2 diabetes: the SURDIAGENE cohort

Heart failure (HF) is a growing complication and one of the leading causes of mortality in people living with type 2 diabetes (T2D). Among the possible causes, the excess of red meat and the insufficiency of vegetables consumption are suspected. Such an alimentation is associated with nutritional biomarkers, including trimethylamine N-oxide (TMAO) and its precursors. Here, we aimed to study these biomarkers as potential prognostic factors for HF in patients with T2D. We used the SURDIAGENE (SURvival DIAbetes and GENEtics) study, a large, prospective, monocentric cohort study including 1468 patients with T2D between 2001 and 2012. TMAO and its precursors (trimethylamine [TMA], betaine, choline, and carnitine) as well as thio-amino-acids (cysteine, homocysteine and methionine) were measured by liquid chromatography-tandem mass spectrometry. The main outcome was HF requiring Hospitalization (HFrH) defined as the first occurrence of acute HF leading to hospitalization and/or death, established by an adjudication committee, based on hospital records until 31st December 2015. The secondary outcomes were the composite event HFrH and/or cardiovascular death and all-cause death. The association between the biomarkers and the outcomes was studied using cause-specific hazard-models, adjusted for age, sex, history of coronary artery disease, NT-proBNP, CKD-EPI-derived eGFR and the urine albumin/creatinine ratio. Hazard-ratios (HR) are expressed for one standard deviation. The data of interest were available for 1349/1468 of SURDIAGENE participants (91.9%), including 569 (42.2%) women, with a mean age of 64.3 ± 10.7 years and a median follow-up of 7.3 years [25th-75th percentile, 4.7-10.8]. HFrH was reported in 209 patients (15.5%), HFrH and/or cardiovascular death in 341 (25.3%) and all-cause death in 447 (33.1%). In unadjusted hazard-models, carnitine (HR = 1.20, 95% CI [1.05; 1.37]), betaine (HR = 1.34, [1.20; 1.50]), choline (HR = 1.35, [1.20; 1.52]), TMAO (HR = 1.32, [1.16; 1.50]), cysteine (HR = 1.38, [1.21; 1.58]) and homocysteine (HR = 1.28, [1.17; 1.39]) were associated with HFrH, but not TMA and methionine. In the fully adjusted models, none of these associations was significant, neither for HFrH nor for HFrH and/or CV death, when homocysteine only was positively associated with all-cause death (HR = 1.16, [1.06; 1.27]). TMAO and its precursors do not appear to be substantial prognosis factors for HFrH, beyond usual cardiac- and kidney-related risk factors, w