Analysis of Metabolic Parameters as Predictors of Risk in the RENAAL Study

The RENAAL study design, inclusion and exclusion criteria, treatment regimen, and baseline patient data have been reported elsewhere (3,7). Patients were stratified according to baseline proteinuria and, after a 6-week screening period, randomized to receive losartan (50 mg titrated to 100 mg once daily) or placebo. Open-label antihypertensive medication (calcium channel blockers, diuretics, α- and β-blockers, and centrally acting agents) were added if trough sitting blood pressure did not reach the goal of <140/90 mmHg. The study population consisted of 1,513 patients of both sexes, aged 31–70 years.

Clinical and laboratory evaluations were conducted at baseline and every 3 months. Laboratory measurements included (but were not limited to) HbA_1c, total, LDL, and HDL cholesterol, triglycerides, uric acid, and serum potassium. Lipoprotein a [Lp(a)] levels were measured only at baseline and at 1 year in 100% of the patients in the U.S., 25% of patients in Europe, and 15% of patients in Asia. Chemistry tests were performed after an overnight fast of ∼8 h. A central laboratory analyzed blood and urinary specimens (albumin/creatinine levels).

Selected baseline characteristics were compared between the placebo and losartan groups using a χ² test (for discrete variables) or a t test (for continuous variables). The relationship between baseline metabolic variables and the primary composite end point or ESRD alone was examined by pooling the data from both arms of the study. The composite end point was examined using the time-to-first-event principle, and only the first event was counted in the analysis. A univariate analysis was performed on each baseline variable using a Cox regression model that was adjusted for region. The hazard ratio for each baseline variable and its 95% CI were calculated. A hazard ratio >1 indicates increasing risk of reaching the end point with increasing values of the variable. A P value <0.05 was considered statistically significant. All statistical tests were two sided. For the variables found to be significant in the univariate analysis, the hazard ratios were examined graphically by quartile. For this analysis, the first quartile served as the base of comparison. Serum triglycerides were log transformed for the hazard ratio analysis.

At baseline, no differences were observed between the losartan and placebo groups for age, BMI, sex, or smoking (3). At baseline, similar percentages of patients in the losartan and placebo groups were receiving insulin (61.4 vs. 58.9%, P = NS), oral antidiabetic agents (48.1 vs. 50.0%, P = NS), and lipid-lowering agents (36.5 vs. 36.1%, P = NS). Most patients in both treatment groups had poor glycemic control, with ∼81% of patients in both groups having HbA_1c levels >7% (8) (Table 1). About one-third of patients had baseline total cholesterol levels within the recommended range of <200 mg/dl, whereas one-third of patients had baseline total cholesterol levels >240 mg/dl (Table 1). Only 13.8% of patients taking losartan and 12.5% of patients on placebo had HDL cholesterol levels >60 mg/dl, as recommended by the Adult Treatment Panel III (9) (Table 1). Approximately half of the patients had high LDL cholesterol (>130 mg/dl; losartan 49.9% and placebo 50.7%) and one-third triglyceride levels >200 mg/dl (losartan 35.6% and placebo 38.7%) (Table 1).

Table 2 shows the mean levels of HbA_1c, lipids, serum uric acid, and serum potassium over the course of the RENAAL study. Neither treatment group showed significant changes in HbA_1c values. In both groups, total cholesterol levels were elevated at baseline and decreased at the last follow-up. At 12 and 36 months, total cholesterol was significantly lower in the losartan group compared with the placebo group despite equivalent use of lipid-lowering medication (Table 2). Although LDL cholesterol levels decreased in both treatment groups, the losartan group showed greater numerical decreases and the difference between groups was statistically significant at 12 months (losartan 126.07 mg/dl vs. placebo 134.01 mg/dl, P < 0.05). At baseline, HDL cholesterol levels were almost identical for the losartan and placebo groups, and modest but similar decreases were observed in both groups (Table 2). Baseline triglyceride levels were also similar between the two groups, and no significant change was observed over time (Table 2). Baseline Lp(a) levels were significantly higher in the losartan group than in the placebo group (P < 0.05), and there was no significant change in the levels of Lp(a) in either group at the final 1-year analysis (Table 2). Although losartan was associated with increased serum potassium at all time points, the mean rise never exceeded 0.3 mEq/l. Increased serum potassium levels led to similar discontinuation rates (losartan 1.1% vs. placebo 0.5%, P = NS). Baseline uric acid levels were similar in the two groups, and, in the losartan group, uric acid tended to be lower at all measured points and was significantly lower at 24 months (losartan vs. placebo, P < 0.05).

Total cholesterol (risk increase 67% per 100 mg/dl, P < 0.001), LDL cholesterol (risk increase 32% per 50 mg/dl, P < 0.001), and triglycerides (risk increase 47% per log-transformed mg/dl, P = 0.011) were associated with increased risk of developing the primary composite outcome (Fig. 1A). Hazard ratios for the primary composite end point showed no significant relationship with baseline serum potassium, HbA_1c, and HDL cholesterol. Baseline total cholesterol levels >220 mg/dl (third and fourth quartiles) were associated with an increased risk of reaching the primary composite end point compared with levels in the first quartile (Fig. 1B). Similarly, an increased risk of reaching the primary composite end point was related to baseline LDL cholesterol >167 mg/dl (fourth quartile) (Fig. 1C) and baseline triglyceride levels >245 mg/dl (fourth quartile) (Fig. 1D).

Similar to the relationships between reaching the primary end point and baseline levels of metabolic factors, the risk of reaching ESRD was strongly related to baseline total cholesterol (risk increase 96% per 100 mg/dl, P < 0.001) and LDL cholesterol (risk increase 47% per 50 mg/dl, P < 0.001) (Fig. 2A). No significant relationship was observed for HbA_1c or HDL cholesterol, and triglycerides were borderline significant. Hazard ratios by quartile of baseline total cholesterol showed a significant relationship with the risk of developing ESRD at the two highest quartiles (>220 mg/dl) (Fig. 2B). Baseline LDL cholesterol in the fourth quartile (>167 mg/dl) was also strongly associated with the risk of developing ESRD (Fig. 2C).

Diabetic nephropathy has reached near-epidemic proportions worldwide, and its incidence continues to rise. It is the leading cause of ESRD in the U.S., and ∼50% of new patients starting dialysis have type 2 diabetes (1,2). Among patients with ESRD, those with diabetes have the highest rates of morbidity and mortality. Given these data, any treatment that reduces the progression of diabetic nephropathy to ESRD should have significant effects on morbidity, mortality, and the cost of ESRD. In addition, the importance of strict maintenance of glycemic control and serum lipid profile necessitates that these parameters should not be adversely affected by the treatment. Recently, the RENAAL trial documented the role of losartan, an angiotensin II blocker, in the reduction of progression to renal failure in patients with type 2 diabetes and clinical nephropathy (3). This post hoc analysis demonstrates that losartan is not associated with adverse effects on metabolic parameters such as glycemic control and lipid profile.

Achieving optimal glycemic control in patients with type 2 diabetes is often challenging. In the current study, with usual care from the patients’ physicians, baseline HbA_1c levels were well above the guidelines recommended by the American Diabetes Association (8). These levels did not decrease significantly throughout the study despite consistent monitoring by a clinical management committee. It is possible, and even likely, that tighter glycemic control may have benefitted the patients long term. In contrast to prior studies (6), baseline HbA_1c levels did not correlate with the primary composite outcome in this population. This may simply reflect the difficulty of maintaining optimal glycemic control in the RENAAL patient population. Importantly, unlike ACE inhibitors, for which there is conflicting data (10–12), losartan was not associated with adverse effects on glycemic control or with any significant difference in the incidence of hyper- or hypoglycemia in patients in the RENAAL study. Similar findings have been reported for irbesartan (5).

Few studies have dealt with the relationship between hyperlipidemia and progression to renal failure. In the RENAAL study, total and LDL cholesterol were elevated at baseline and a strong correlation for total and LDL cholesterol and triglycerides was observed for the primary composite end point. In addition, both total and LDL cholesterol were associated with an increased risk of developing ESRD. The association between lipids and renal disease has attracted significant interest, especially because half of the deaths in dialysis patients are of cardiovascular origin (13). Several studies in experimental animal models support a causal relationship between elevated lipid levels and the development of glomerular damage (14); in addition, pharmacologic lowering of lipids by a variety of classes of antihyperlipidemic medications has been reported to ameliorate renal damage (15–18). In humans, high serum cholesterol in conjunction with obesity has been reported to induce structural glomerular changes (19), and some prospective studies in humans have shown that hyperlipidemia may exacerbate progressive renal disease (6,20–24). Elevated cholesterol levels and low HDL cholesterol levels have been identified as independent risk factors for progressive renal disease (6), and smaller trials support elevations of LDL cholesterol or apolipoprotein B as predictors of renal progression (20–22). The role of individual lipid fractions in promoting renal dysfunction in humans is unclear; however, the data presented here suggest that elevated total and LDL cholesterol are important predictors of the development of ESRD in patients with type 2 diabetic nephropathy and are in agreement with prior studies in patients with impaired renal function with or without diabetes (6,20–24). Whether this is due to a primary role of hyperlipidemia or a correlation with proteinuria and other risk factors for renal progression remains to be proven. As such, further prospective studies are required; however, aggressive lipid lowering should clearly be an important consideration in treating patients with type 2 diabetes and clinical nephropathy.

Hyperkalemia is a clinically relevant adverse event associated with agents that block the renin-angiotensin system. In six clinical trials involving over 1,500 patients with renal insufficiency, increases in serum potassium (mean 0.3–0.6 mEq/l) occurred in patients randomized to ACE inhibitors (25–30). Additionally, it has been reported that in patients with moderate renal impairment, less hyperkalemia was observed in the AIIA-treated group when compared with patients receiving an ACE inhibitor (31). In the current study, serum potassium levels were elevated at all study points in the losartan group; however, the mean increase never exceeded 0.3 mEq/l. A small number of patients, 1.1% in the losartan group and 0.5% in the placebo group, had to discontinue therapy because of hyperkalemia, which demonstrates that hyperkalemia associated with losartan is clinically manageable in this patient population.

It should be noted that the analyses described herein are post hoc, and as such the data should be interpreted with this limitation in mind. Further analysis of the RENAAL database is required to examine the role of baseline variables of predictors of outcome in this trial.

In the RENAAL study, losartan therapy in patients with type 2 diabetes and nephropathy provided renal protection by delaying the time to the composite end point of doubling of serum creatinine, ESRD, or death. In this post hoc analysis, losartan had no adverse effect on glycemic control, lipid profile, or serum uric acid. Overall, losartan was generally well tolerated in these patients.