Poor Glycemic Control Is Associated With More Rapid Kidney Function Decline After the Onset of Diabetic Kidney Disease

Abstract

Background: The role of glycemic control and its variability on the rate of kidney function decline after the onset of diabetic kidney disease (DKD) remains unclear.

Methods: The association between baseline glycated hemoglobin (HbA1c) and rates of estimated glomerular filtration rate (eGFR) loss during follow-up was examined by mixed-effects linear regression in 530 individuals with type 1 diabetes and early-to-moderate DKD from the Preventing Early Renal Loss (PERL) trial and 2378 individuals with type 2 diabetes and established DKD from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. The benefit of intensive vs standard glycemic control in slowing eGFR decline was examined in ACCORD. The associations between continuous glucose monitoring-derived short-term glycemic variability indices and rate of eGFR decline were also evaluated in PERL.

Results: A higher baseline HbA1c was associated with a more negative eGFR slope in both PERL and ACCORD (-0.87 and -0.27 mL/min/1.73 m2/year per Hba1c unit increment, P < .0001 and P = .0002, respectively). In both studies, the strength of this association progressively increased with increasing levels of albuminuria (P for interaction <.05). Consistent with this, the benefit of intensive glycemic control on eGFR decline was greater in ACCORD participants with severe rather than moderate albuminuria (+1.13 vs + 0.26 mL/min/1.73 m2/year, P = .01). No independent associations were found in PERL between short-term glycemic variability indices and rate of eGFR decline.

Conclusion: In both type 1 and type 2 diabetes, poor glycemic control is associated with a more rapid rate of glomerular filtration rate decline after DKD onset, especially in persons with severe albuminuria.

Keywords: diabetic kidney disease; glycemic control; kidney function decline.

Figure 1.

Interaction of glycemic control and albuminuria on rate of GFR decline after onset of diabetic kidney disease. Y-axes represent rate of GFR decline (mL/min/1.73  m²/year). Data points represent beta estimates (±SE) from mixed-effects linear regression models. GFR at each time-point, including baseline, was the dependent variable; time from randomization to GFR assessment was the fixed effect; study subjects were random effects. P-trends correspond to the P-values for the interaction between time × glycemic control variable; P-interactions correspond to the P-values for the 3-way interaction terms including time × AER or ACR × glycemic control. (A) Rate of eGFR decline by AER and HbA1c strata in PERL. (B) Rate of iGFR decline by AER and HbA1c strata in PERL. (C) Rate of eGFR decline by ACR and HbA1c strata in ACCORD. Within each AER/ACR stratum in (A-C), data points on the left represent those with HbA1c ≤ 7.5%, those in the center represent HbA1c 7.5%-8.5%, and those on the right represent HbA1c ≥ 8.5%. (D) Rate of eGFR decline by ACR and glycemic control groups in ACCORD. Within each ACR stratum, data points on the left represent intensive glycemic control and those on the right represent standard glycemic control. Abbreviations: ACCORD, Action to Control Cardiovascular Risk in Diabetes trial; ACR, urinary albumin-creatinine ratio, mg/mmol; AER, albumin excretion rate, mg/24 hours; eGFR, estimated glomerular filtration rate based on creatinine-derived Chronic Kidney Disease Epidemiology Collaboration 2009 equation; GFR, glomerular filtration rate; HbA1c, glycated hemoglobin; iGFR, iohexol-based glomerular filtration rate; PERL, Preventing Early Renal Loss trial.

Figure 2.

Rate of eGFR decline by CV and HbA1c strata in PERL. Y-axes represent the rate of GFR decline (mL/min/1.73 m²/year). X-axes represent eGFR slopes by strata of baseline HbA1c and CV. Data points represent beta estimates (±SE) from mixed-effects linear regression models. GFR at each time-point, including baseline, was the dependent variable; time from randomization to GFR assessment was the fixed effect; study subjects were random effects. P-trends correspond to the P-values for the interaction between time × glycemic control variable; P-interaction corresponds to the P-value from 3-way interaction term including time × CV × HbA1c. Within each stratum, data points on the left represent those with CV ≤ 41%, those in the center represent CV 41%-48%, and those on the right represent CV ≥ 48%. Abbreviations: CV, glucose coefficient of variation obtained from continuous glucose monitoring devices; eGFR, estimated glomerular filtration rate based on creatinine-derived Chronic Kidney Disease Epidemiology Collaboration 2009 equation; HbA1C, glycated hemoglobin; PERL, Preventing Early Renal Loss trial.