Risk of microalbuminuria in relatives of subjects with diabetic nephropathy: a predictive model based on multivariable dimensionality reduction approach.

Aim: To evaluate a predictive model of microalbuminuria by using anthropometric, clinical and genetic variables in relatives of subjects with diabetic nephropathy.

Methods: Eligible subjects, aged 18-63 years with body mass index<35 kg/m2, and first degree relatives of patients with type 2 diabetes and diabetic nephropathy were enrolled in a cross-sectional study. A total of 70 individuals with microalbuminuria were compared with 60 individuals without microalbuminuria. Based on a morning urinary sample, microalbuminuria was defined as albumin≥30<300 mg/dL. Genotyping of single nucleotide polymorphisms (SNPs) G-174C of IL6 (rs1800795), G-308A of TNF (rs1800629), and Pro12Ala of PPARγ2 (rs1801282) genes were determined. The multivariable dimensionality reduction analysis was performed using the software multifactor dimensionality reduction package.

Results: The multivariable dimensionality reduction analysis showed that obesity and SNP G-308A of TNF gene exhibited main effects with 1.10 and 1.98% of information gain (IG), respectively. The IL6 showed synergy (interaction) with HDL-c (IG 1.27%) and sex (IG 1.02%); also high-sensitivity C-reactive protein and triglycerides levels showed synergy (IG 1.08%). The consistency of the cross-validation for this model was 0.6836, with sensitivity and specificity of 0.58 and 0.76 (odds ratio 4.64; 95% CI 4.0-10.0, p<0.0001).

Conclusion: Our results indicate that obesity and/or high blood pressure, in synergism with high-sensitivity C-reactive protein and high-density lipoprotein cholesterol levels, is the main predictive risk factor of diabetic nephropathy in healthy subjects, relatives of patients with type 2 diabetes and diabetic nephropathy.