AimsThis study aimed to evaluate the association between baseline bilirubin (TBiL) and follow-up TBiL changes for diabetic kidney disease (DKD) incidence and progression based on a 5?years' cohort study.MethodsThis cohort study was conducted in Beijing between 2009 and 2013. The subjects were consisted of 5342male diabetic patients with baseline retinopathy. Cox proportional risk model was used to calculate hazards ratio (HR).ResultsThe mean age of the 5342 diabetic patients was 78.68?±?8.40 (65–102?yrs). The total five year incidence was 8.7% (95%CI: 7.9%–9.4%) for DKD and 10.5% (95%CI: 9.7%–11.3%) for eGFR decrease. The HR of baseline TBiL showed a decreasing trend for both DKD incidence and eGFR decrease. The HRs of baseline TBiL (per μmol/L increase) for DKD and eGFR decrease were 0.967(95%CI: 0.946–0.988) and 0.955(95%CI: 0.936–0.975) respectively. For follow-up TBiL changes, after adjusted for related co-variables and baseline TBiL levels (as continuous variable) in the model, the HRs (per μmol/L of follow-up TBiL changes) for DKD and eGFR decrease were 0.973(95%CI: 0.952–0.995) and 0.991(95%CI: 0.974–0.998) respectively. The results were similar when baseline TBiL and follow-up TBiL changes were used as tertiary variable.ConclusionNot only baseline TBiL, but also follow-up changes were significantly associated with DKD incidence and progression. |