Strict glycaemic control in diabetic patients with CKD or ESRD: beneficial or deadly?

Several large randomized controlled trials including the Diabetes Control and Complications Trial in type 1 diabetes [1] and the UK Prospective Diabetes Study [2] as well as the Kumamoto Study [3] in type 2 diabetes indicate that good glycaemic control can reduce the risk of nephropathy. More recently, the ADVANCE trial [4] documented in subjects with type 2 diabetes (T2DM) that strict glycaemic control (mean HbA1c: 6.5%) in comparison with the standard control (mean HbA1c: 7.3%) is associated with a significant reduction in renal events, including onset of or worsening of nephropathy [hazard ratio (HR) 0.79; P = 0.006], new-onset microalbuminuria (HR 0.91; P = 0.02) and, in particular, development of macroalbuminuria (HR 0.70; P < 0.001). Moreover, patients with macroalbuminuria and chronic kidney disease (CKD; eGFR 60 mL/min/1.73 m2 at baseline) had a 3.2-fold higher risk for cardiovascular (CV) events and a 22.2-fold higher risk for renal events compared with patients who had neither risk factor. In a study on kidney plus pancreas transplantation in type 1 diabetic patients, Morath et al. [5] had shown that metabolic control improves long-term renal allograft and patient survival. A recent Austrian cohort study [6] of 798 first renal allograft recipients with a diagnosis of diabetes mellitus prior to transplantation also showed that glucose control was significantly associated with patient and graft survival after renal transplantation. Can one extrapolate that these very positive renoprotective effects of intensive glucose control are seen in all diabetic patients with CKD or end-stage renal disease (ESRD)? Unfortunately, results of prospective randomized controlled intervention studies are not yet available to answer this clinically relevant question, despite the fact that diabetes is present in 25–50% of patients with ESRD. Information about the effect of strict glycaemic control on outcome in diabetic patients with CKD is very limited. In the PROactive study (Prospective Pioglitazone Clinical Trial in Macrovascular Events), patients with type 2 diabetes and CKD (n = 597; 11.6% of the 5154 patients) had a particularly high CV risk and cardiovascular disease (CVD) [7]. The incidence of the combined end point (non-fatal myocardial infarction, stroke and death) was 18.3% in patients with CKD compared with 11.5% in patients without CKD (HR 1.65; P < 0.0001). In addition, all-cause mortality was 10.9% compared to 5.9% (HR 1.86) in those without CKD. Remarkably, patients who had CKD and were treated with pioglitazone were less likely to reach the combined end point (HR 0.66; P < 0.001) independently of the severity of renal impairment [7]. Since at study end the HbA1c difference between pioglitazone and placebo was only 0.5% (6.9 vs 7.4%), presumably a variety of well-documented anti-atherogenic effects of pioglitazone [8] were responsible for CV protection in diabetic patients presenting with CVD. It is well known that the 5-year survival is much lower in diabetic vs non-diabetic patients on HD [9]. During the last decade, several observational studies [10–14] were published indicating that the survival of diabetic patients on haemodialysis (HD) is influenced by the glycaemic control. Morioka et al. [10] evaluated the impact of glycaemic control on survival in 150 diabetic subjects with ESRD starting HD treatment. During the short follow-up period of 2.8 years, 76% of the patients died; however, in those with good control (HbA1c < 7.5%), mortality was lower than in those with poor glycaemic control (HbA1c ≥ 7.5%). Oomichi et al. [11] found in a 7-year observational study of 114 diabetic patients on HD that mortality was similar in patients with good HbA1c <6.5% and fair glycaemic control (HbA1c > 6.5% < 8.0%), but mortality was significantly higher (HR 2.89; P = 0.01) in those with poor glycaemic control (HbA1c > 8.0%). Hayashino et al. [12] analysed mortality in the large Japanese Dialysis Outcomes and Practice Pattern Study on 1569 HD patients with diabetes and 3342 patients without diabetes. Mortality was significantly higher in diabetic patients on HD whose HbA1c was in the fifth quintile (HbA1c ≥ 7.3%), but was not different in the four quintiles with HbA1c ranging from 5.0 to 7.2%. In contrast to these studies in Asian dialysis populations, no correlation between HbA1c and survival was found at 12 months in 24 875 US diabetic dialysis patients [15]. The failure to see a beneficial effect of glycaemic control may be explained by the underlying differences in the tools used to measure glycaemia. Kalantar-Zadeh et al. [13] evaluated 23 618 US diabetic HD patients and evaluated survival as a function of HbA1c. He observed that higher HbA1c values were incrementally associated with higher mortality. Compared with patients with HbA1c in the range of 5–6%, patients with HbA1c >10% had higher HRs of adjusted all-cause and CV death: 1.41 and 1.73 (P < 0.001), respectively. Remarkably, this relationship was only seen in patients without anaemia (Hb > 11.0 g/day; n = 19 316) and not in patients with anaemia or malnutrition. Recently, Drechsler et al. [14] investigated the impact of glycaemic control on cardiac and vascular outcomes in 1255 diabetic HD patients in the German Diabetes and Dialysis Study (4D Study). During 4 years of follow-up, patients with HbA1c >8.0% or HbA1c from >6 to <8% had an increased risk of sudden death (HR 1.85 and 2.26, respectively; P < 0.003) compared with patients with HbA1c <6.0%. In contrast, the risk of myocardial infarction and all-cause mortality did not differ between the three groups. Unfortunately, HbA1c is not an ideal control parameter to assess glycaemic control in patients with advanced kidney disease. As early as 1979, HbA1 levels were measured in patients on HD or with CKD using ion exchange chromatography; even then, the accuracy of this method in patients with advanced CKD had been questioned [16]. In chronic renal failure, the lifespan of erythr