Factor analysis of diabetic nephropathy in Chinese patients

ObjectiveTo investigate factors related to albuminuria in diabetic nephropathy (DN).MethodsClinical data of 873 Chinese patients with type 2 diabetes were gathered retrospectively. Urinary albumin–creatinine-ratio was tested thrice for each patient before calculating its mean value for DN diagnosing and staging. Inter-relationship among variables was studied using factor analysis, and factor scores were compared and used as independent variables in regression models to identify the parameter clusters that predict the development and/or progression of DN.ResultsFactor analysis reduced 21 original variables to eight unique factors, representing obesity, glycemic, C peptide, lipids, time, renal, blood pressure and metabolism status. Logistic regression revealed that presence of hyperuricemia/gout (OR = 1.821, 95%CI 1.224–2.707), time factor (OR = 1.404, 95%CI 1.194–1.651) and blood pressure factor (OR = 1.424, 95%CI 1.216–1.668) were positively associated with DN, while C peptide factor (OR = 0.816, 95%CI 0.691–0.963) was negatively associated with DN. Ordinal regression revealed another positively related lipid factor (OR = 1.237, 95%CI 1.060–1.445) besides those determined in Logistic regression.ConclusionHyperuricemia/gout, time, blood pressure and lipid factors are predictors of DN, while C peptide factor is negatively associated with the development and/or progression of DN.

► Factor analysis was used to eliminate multicollinearity among inter-related variables in regression models.
► Factor analysis reduced 21 original variables into eight unique factors.
► Hyperuricemia/gout, time, blood pressure and lipid factors are predictors of diabetic nephropathy.
► C peptide factor is negatively associated with the development and/or progression of diabetic nephropathy.