Lower Haemoglobin Level and Subsequent Decline in Kidney Function in Type 2 Diabetic Adults Without Clinical Albuminuria

Babazano T, Hanai K, Suzuki K, et al. Diabetologia. 2006;49:1387-1393.

Several studies have demonstrated that anemia is indicative of an increased risk of kidney disease progression in type 2 diabetes patients with overt nephropathy characterized by albuminuria. This study investigates the relationship between baseline hemoglobin concentrations and the rate of glomerular filtration rate (GFR) decline in type 2 diabetes patients with urinary albumin in the normal to microalbuminuric range.

This study is an observational prospective cohort study of 464 participants (149 women and 315 men, aged 55±13 years) referred to the Tokyo Women’s Medical University Hospital in Tokyo, Japan in 1999. Participants were individuals with type 2 diabetes, modest renal impairment (mean estimated GFR = 74.8±14.5 mL•min-1•1.73m-2), baseline serum creatinine concentration <177 mmol/L (mean±SD = 74.8±14.5 mmol/L), and urinary albumin:creatinine ratio (ACR) <300 mg/g creatinine (geometric mean [95% confidence interval] = 11.5 mg/g creatinine [10.5-12.8]) in the first morning urine. Exclusion criteria included previous treatment with human recombinant erythropoietin or <2.5 years of follow-up observation since the first visit. The mean observation time was 5.0±0.9 years. Most baseline hemoglobin concentrations were normal (mean±SD = 144±15 g/L). The primary outcome of the study was rate of change of estimated GFR. Univariate and multiple regression analyses were performed to determine relationships between hemoglobin levels and GFR changes. Covariates for multiple regression analysis included: age, diabetes duration, diabetes and hypertension medications, body-mass index (BMI), systolic and diastolic blood pressure (BP), hemoglobin A1C (A1C), logarithmically transformed urinary ACR, and GFR.

Univariate correlation analysis indicated that lower baseline hemoglobin concentration, higher baseline ACR, and higher baseline GFR were associated with faster GFR decline overall (P = .001), although hemoglobin concentration did not correlate with change in GFR among women (P = .177). Other variables associated with changes in GFR were age, BMI, diastolic BP, and serum creatinine. Multiple regression analyses confirmed the results of the univariate analyses with the exception that hemoglobin concentration did correlate with GFR changes among women (P = .022). The relationship between baseline hemoglobin concentration and change in GFR was also considered across hemoglobin quartiles after data were adjusted for covariates. Rates of GFR decrease were significantly lower for the third and fourth quartiles, defined by the highest hemoglobin levels (>136 g/L for women and >150 g/L for men), compared to the first quartile, defined by the lowest hemoglobin level (<126 g/L for women and <141 g/L for men) (see Table 1).

Table 1. Comparison of adjusted rate of GFR change by baseline hemoglobin concentration quartiles

The difference in the rate of decline between the lowest and highest hemoglobin quartiles was modest but significant, demonstrating that lower hemoglobin levels can predict more rapid decline in GFR among diabetic patients who have not developed albuminuria or anemia. The study does have limitations. For example, the study population may be too homogeneous to generalize results to all type 2 diabetes patients. Although it is not designed to demonstrate whether correcting anemia in early stages of DN attenuates the rate of kidney function decline, it may provide insight regarding the timing of treatments, such as erythropoietin, that do.