Aggressive blood pressure reduction and renin angiotensin system blockade in chronic kidney disease: time for re-evaluation?

http://www.kidney-international.org & 2013 International Society of Nephrology Aggressive blood pressure reduction and renin angiotensin system blockade in chronic kidney disease: time for re-evaluation? Pantelis A. Sarafidis 1 and Luis M. Ruilope 2 1 Section of Nephrology and Hypertension, 1st Department of Medicine, AHEPA University Hospital, Thessaloniki, Greece and 2 Hypertension Unit, Hospital 12 de Octubre, Madrid, Spain Over the past decades, aggressive control of blood pressure (BP) and blockade of the renin angiotensin aldosterone system (RAAS) were considered the cornerstones of treatment against progression of chronic kidney disease (CKD), following important background and clinical evidence on the associations of hypertension and RAAS activation with renal injury. To this end, previous recommendations included a BP target of o130/80 mm Hg for all individuals with CKD (and possibly o125/75 mm Hg for those with proteinuria 41 g/day), as well as use of angiotensin-converting-enzyme inhibitors and angiotensin receptor blockers as first-line therapy for hypertension in all CKD patients. However, long-term extensions of relevant clinical trials support a low-bp goal only for patients with proteinuria, whereas recent cardiovascular trials questioned the benefits of low systolic BP for diabetic patients, leading to more individualized recommendations. Furthermore, our previous knowledge of the specific renoprotective properties of RAAS blockers in patients with proteinuric CKD is now extended with data on the use of these agents in patients with less advanced nephropathy and/or absence of proteinuria, deriving mostly from subanalyses of cardiovascular trials. This review discusses previous and recent clinical evidence on the issues of BP reduction and RAAS blockade by type and stage of renal damage, aiming to aid clinicians in their treatment decisions for patients with CKD. Kidney International (2014) 85, 536 546; doi:10.1038/ki.2013.355; published online 18 September 2013 KEYWORDS: blood pressure; chronic kidney disease; hypertension; renin angiotensin system Correspondence: Pantelis A. Sarafidis, Section of Nephrology and Hypertension, 1st Department of Medicine, AHEPA University Hospital, Aristotle University of Thessaloniki, St Kiriakidi 1, Thessaloniki 54006, Greece. E-mail: psarafidis11@yahoo.gr Received 23 April 2013; revised 26 May 2013; accepted 5 June 2013; published online 18 September 2013 Chronic kidney disease (CKD) is currently recognized as an independent risk factor for cardiovascular and all-cause mortality; 1 recent data further exemplify the importance of CKD for cardiovascular complications, in relation to established risk factors, such as diabetes. 2 Therefore, prevention of CKD or retardation of disease progression in affected individuals is proposed as another strategy toward cardiovascular protection. 3 On one hand, elevated blood pressure (BP) is a major risk factor for CKD, and on the other hand, kidney injury can cause hypertension. 4 6 Observational studies suggested a strong association between high BP and the risk for renal function decline or end-stage renal disease (ESRD), whereas in various clinical trials, patients with BP below conventional thresholds showed better preservation of renal function. 4,5 Thus, in the past decade, relevant guidelines recommended a BP target of o130/80 mm Hg for all CKD patients (and possibly o125/75 mm Hg for those with proteinuria 41g/day), 5,7 9 although evidence from trials with hard renal outcomes (that is, incidence of ESRD) randomizing patients to different BP targets was limited. 8 Recently, long-term cohort data of relevant trials supported a low-bp goal for patients with proteinuria, whereas cardiovascular trials questioned the beneficial effects of low BP for patients with diabetes, 10,11 leading to recommendations for less aggressive approaches to BP lowering in the latter. 6,12 Angiotensin-converting-enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) are recommended as first-line therapy for hypertension in patients with CKD, 5,7 9 following evidence ranging from background studies to major renal trials in proteinuric CKD that suggested these agents to slow nephropathy progression more effectively than other antihypertensive agents. 13 However, for patients with less advanced nephropathy or absence of proteinuria, previous reports suggested inhibitors of renin angiotensin aldosterone system (RAAS) to confer no additional renoprotective benefit, 14,15 and recent trials showed combined RAAS blockade to increase the risk of acute renal failure and related complications. 16,17 Overall, observations evolving over the past few years have made selection of appropriate BP targets and use of RAAS blockade a complicated issue for the average clinician. In this 536 Kidney International (2014) 85, 536 546

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD review review, we attempt to delineate past major studies and recent developments in these fields by analyzing available clinical evidence for the different types of individuals with CKD. TARGET BP IN PATIENTS WITH CKD Nondiabetic kidney disease Two clinical trials with hard renal end points have evaluated different BP targets in patients with nondiabetic CKD: the Modification of Diet in Renal Disease (MDRD) and the African-American Study on Kidney Disease (AASK). MDRD included two studies in patients with CKD (585 patients in study A (glomerular filtration rate (GFR) 25 55 ml/min per 1.73 m 2 ) and 255 in study B (GFR 13 24 ml/min per 1.73 m 2 )) with the rate of change in GFR (GFR slope) as primary outcome, and a mean follow-up of 2.2 years. 18 Diabetic patients on insulin treatment were excluded by protocol; thus, only 26 patients with diabetic nephropathy participated. In a 22 factorial design, patients were randomized to different levels of dietary protein intake and to a usual-bp goal (mean arterial pressure o107 mm Hg for patients p60 years old (roughly corresponding to o140/ 90 mm Hg) and o113 mm Hg for patients X61 years old) or a low goal (mean arterial pressure o92 mm Hg for patients p60 years old (corresponding to o125/75 mm Hg) and o98 mm Hg for patients X61 years old). Neither the projected GFR decline in 3 years (10.7 vs. 11.5 ml/min per 1.73 m 2 ) nor the risk of ESRD and death (0.85, 95% confidence interval (CI): 0.60 1.22 for low-bp arm) differed significantly between groups. 18 However, analyses of patients by baseline proteinuria showed that higher proteinuria was associated with steeper GFR decline and that low target BP had beneficial effects on GFR slope in patients with proteinuria 40.25 g/day in study A and 41 g/day in study B (Figure 1), and the results were not substantially altered after adjustment for 10 relevant covariates. 18,19 These findings indicated that a low target BP may be beneficial in proteinuric patients and was the basis of previous recommendations for BP o130/80 mm Hg in patients with CKD and o125/ 75 mm Hg in patients with proteinuria 41g/day. 5,7 9 A patient-level meta-analysis of trials on antihypertensive treatment with or without ACEIs in predominantly nondiabetic CKD confirmed the above: in patients with proteinuria 41 g/day, levels of systolic blood pressure (SBP) between 110 and 119 mm Hg were associated with similar risk of CKD progression with levels between 120 and 129 mm Hg and significantly lower risk compared with SBP X130 mm Hg; in contrast, in patients with proteinuria o1g/ day, a significant association of low BP with renoprotection was absent. 20 A subsequent analysis examined long-term outcomes considering the trial phase of MDRD (1989 1993) together with a follow-up cohort period (1993 2000) during which no specific target BP was recommended. 21 During a median follow-up of 10.7 years, low target BP was associated with a reduced risk for ESRD (adjusted hazard ratio (HR) 0.68; 95% CI: 0.57 0.82) and the composite of ESRD or death Rate of GFR decline (ml/min/year) 0 2 4 6 8 10 12 14 16 Study A Baseline urine protein (g/day) Study B (301) (119) (104) (54) (77) (59) (63) (32) 0 0.25 0.25 1.0 1.0 3.0 3.0 0 0.25 0.25 1.0 1.0 3.0 Figure 1 Effects of different blood pressure (BP) targets on glomerular filtration rate (GFR) in the Modification of Diet in Renal Disease (MDRD) studies by baseline proteinuria. Low target BP had beneficial effects on GFR slope in patients with proteinuria 40.25 g/day in study A and 41 g/day in study B. Black circles indicate usual-bp group and white circles indicate low-bp group; numbers in parentheses reflect patients in both BP groups with at least one follow-up GFR measurement (reprinted from Peterson et al. 19 ). (HR 0.77; 95% CI: 0.65 0.91) as compared with usual target BP. Again, in subgroup analyses, the benefits from low target BP for ESRD and the composite end point was significant only for proteinuria 41 g/day. The P-value for interaction of target BP with proteinuria was 0.09 for ESRD and 0.08 for the composite outcome. 21 AASK included 1094 African Americans with hypertensive CKD (GFR 20 65 ml/min per 1.73 m 2, mean proteinuria 0.6 g/day) randomized to goal mean arterial pressure 102 107 or p92 mm Hg, and to initial treatment with metoprolol, ramipril, or amlodipine in a 32 factorial design. The main outcomes were GFR slope and a composite of GFR reduction X50% (or X25 ml/min per 1.73 m 2 ), ESRD, or death. The mean achieved BP was 128/78 mm Hg in the low-bp group and 141/85 in the usual-bp group. After a median of 3.8 years, neither GFR slope ( 2.21±0.17 vs. 1.95±0.17 ml/ min per 1.73 m 2 per year; P ¼ 0.24) nor the composite outcome (risk reduction for low-bp group 2%; 95% CI: 22 to 21%) differed significantly between groups. 22 After the trial phase, B700 subjects were enrolled in an observational phase with a total follow-up of 8.8 12.2 years; target BP during the cohort phase was o130/80 mm Hg. A recent analysis including the two phases together showed no significant difference between the two groups in the risk of the composite outcome of doubling of serum creatinine (SCr), ESRD, or death (HR in low-bp group, 0.91; 95% CI: 0.77 1.08). However, there was a significant interaction with the baseline level of proteinuria (P ¼ 0.02); patients with urine protein-to-creatinine ratio of 40.22 (both measured in 24-h urine collections and expressed in mg/dl), which roughly equals a proteinuria of 320 mg/day, had lower risk of the primary outcome with intensive treatment (HR 0.73; 95% CI: 0.58 0.93; Figure 2). In those with urine 3.0 0 2 4 6 8 10 12 14 16 Kidney International (2014) 85, 536 546 537

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD Cumulative incidence (%) 100 80 60 40 20 Trial phase Standard control Intensive control Trial and cohort phases Cohort phase P/C ratio >0.22 P/C ratio 0.22 0 0 1 2 3 4 5 6 7 8 9 10 Follow-up year Figure 2 Effects of different blood pressure (BP) targets on the composite outcome of doubling of serum creatinine, end-stage renal disease (ESRD), or death in the trial and cohort phases of the African-American Study on Kidney Disease (AASK) (reprinted from Appel et al. 23 ). P/C, protein/creatinine. protein-to-creatinine ratio of p0.22, there was no betweengroup difference (HR 1.18; 95% CI: 0.93 1.50). 23 Taken together, the findings from MDRD and AASK indicate that a low-bp target is beneficial for long-term renal survival in patients with nondiabetic kidney disease and proteinuria 40.25 0.3 g/day (equivalent to urinary albumin excretion (UAE) of B0.15 g/day). Thus, a goal BP of o130/ 80 (that is, that of AASK cohort study) seems justifiable for proteinuria 40.3 g/day, whereas a lower BP target of o125/ 75 mm Hg may be applicable for patients with proteinuria 41 g/day (Table 1). It must be noted, however, that all available evidence derives from subgroup analyses or from combination of randomized phases with long-term observational follow-up phases and, still, no direct evidence from randomized trials on the effect of different BP targets on hard renal end points in proteinuric or nonproteinuric populations is available. Furthermore, both trials randomized to mean arterial pressure levels that correspond on average, but not for every patient, to specific levels of SBP and diastolic BP (DBP). Diabetic kidney disease To date, clinical trials with hard renal outcomes randomizing at different target BP in patients with diabetes are missing. Earlier studies in diabetic patients with varying levels of renal function and albumin excretion compared different BP goals with change in creatinine clearance (CrCl) as a primary outcome; they showed no difference in CrCl change between groups with intensive or moderate BP control but higher reductions of proteinuria and slower progression from microto macroalbuminuria with intensive BP. 24,25 Another report showed no difference in CrCl or proteinuria progression between intensive and standard BP reduction. 14 An analysis of mean achieved SBP versus estimated GFR (egfr) decline in controlled trials of diabetic CKD suggested that lowering SBP down to 130 mm Hg may be associated with reduction of egfr loss down to 2 ml/min per 1.73 m 2 per year. 5 A post hoc analysis of the RENAAL (Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan) study, which included 1513 patients with type 2 diabetes, hypertension, and macroalbuminuria (mean SCr 1.9 mg/dl and median albumin-to-creatinine ratio (ACR) 1237 mg/g) and compared the effects of losartan or placebo on CKD progression, showed that baseline SBP of 140 159 mm Hg increased the risk for ESRD or death by 38% (P ¼ 0.05) as compared with SBP o130 mm Hg. In a multivariate model, every 10-mm Hg rise in baseline SBP increased the risk for ESRD or death by 6.7% (P ¼ 0.007), whereas the same rise in DBP decreased the risk by 10.9% (P ¼ 0.01). The authors concluded that SBP is a stronger predictor of renal outcomes than DBP and those with the highest baseline PP have the highest risk for nephropathy progression. 26 Subsequently, a post hoc analysis of achieved BP and renal outcomes from the Irbesartan Diabetic Nephropathy Trial (IDNT; which included 1590 patients with type 2 diabetes, hypertension, and proteinuria 4900 mg/day, to compare the effects of irbesartan, amlodipine; and placebo) showed that SBP 4149 mm Hg was associated with a 2.2-fold increase in the risk for doubling SCr or ESRD compared with SBP o134 mm Hg, and follow-up achieved SBP most strongly predicted renal outcomes (Figure 3); moreover, progressive lowering of SBP down to 120 mm Hg improved renal and patient survival, but below 120 mm Hg, all-cause mortality increased. 27 Based on evidence like that above and as progression of renal injury is believed to follow the same pathways once proteinuria develops, it has been argued that in patients with proteinuric diabetic nephropathy (particularly 41 g/day), the above BP targets for nondiabetic CKD should apply. 5 This argument is difficult to dispute, but the National Health and Nutrition Examination Survey (NHANES) 1999 2000 in the United States suggest that the prevalence of macroalbuminuria (ACR X300 mg/g, equivalent to proteinuria 40.5 g/day) in adult patients with diabetes is only B10%, whereas another 20% have microalbuminuria (ACR 30 300 mg/g), and 70% normoalbuminuria. Simultaneously, B11% of diabetics (rising to 26% of diabetics 465 years old) have egfr o60 ml/min per 1.73 m 2. 28 Thus, an important proportion of diabetic patients (especially the elderly type 2 with concomitant hypertension) have CKD stage 3 or higher without proteinuria. For these individuals, BP o130/ 80 mm Hg may not be required for renoprotection. However, a lower BP target could also be warranted for cardioprotection and benefits in mortality, as discussed below. All clinical practice guidelines in previous decade recommended target BP o130/80 mm Hg for diabetic patients. The first evidence pointing toward that target derived from the UK Prospective Diabetes Study (UKPDS) 38 and Hypertension Optimal Treatment (HOT) trials. 29,30 In UKPDS 38, 1148 hypertensive type 2 diabetic patients were randomized to BP of o150/85 or o180/105 mm Hg (and achieved mean BP values of 144/82 and 154/87 during 8.4 years); the tight control group had significant reductions in diabetes-related deaths and complications. 29 The HOT study randomized 18,790 hypertensives to DBP targets of p90, 538 Kidney International (2014) 85, 536 546

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD review % Of patients with doubling of Scr or ESRD 80 60 40 20 Follow-up SBP, mm Hg > 149 141 149 134 140 < 134 0 0 6 12 18 24 30 36 42 48 54 Follow-up time (months) Figure 3 Cumulative proportions of patients who reached a renal end point (doubling of serum creatinine or end-stage renal disease (ESRD)) in the Irbesartan Diabetic Nephropathy Trial (IDNT) by quartile of achieved systolic blood pressure (SBP) during follow-up; patients with SBP o134 mm Hg had the longest renal survival (reprinted from Pohl et al. 27 ). p85, or p80 mm Hg; no difference between groups was observed for the total study population, but a 51% reduction in cardiovascular events between the p80 and p90 mm Hg target was evident for the subgroup of 1501 diabetic patients. 30 The problem in recommending the o130/80 mm Hg target in diabetes was that of SBP. Observational data supported that SBP o130 mm Hg would confer cardiovascular benefits 31 but the tight control arms in UKPDS and HOT achieved mean SBP values of 144 and 141 mm Hg. 29,30 Furthermore, a target of SBP o130 mm Hg is very difficult to achieve in clinical practice, even within well-structured clinical environments 32 and will often require 43 antihypertensive drugs and many clinic visits. Further trials attempted to provide evidence for this important clinical question. The ADVANCE (Action in Diabetes and Vascular disease preterax and diamicron MR Controlled Evaluation) randomized 11,140 type 2 diabetics to fixed perindopril indapamide combination or placebo, on top of background therapy. After 4.3 years, the average BP was 135/74 versus 140/76 in the two groups. Differences of 9% in major macrovascular or microvascular events, 18% in cardiovascular death, and 14% in all-cause mortality favoring the active treatment were noted. 33 These results indicated a favorable effect of further SBP lowering, but should be interpreted with caution, as the actual comparison involved a drug intervention. In contrast, a subgroup analysis of 6400 patients with diabetes, hypertension, and coronary artery disease from the International VerapamilSR Trandolapril Study pointed toward no additional benefit with SBP o130 mm Hg. 34 It showed a higher risk for cardiovascular events in patients having SBP 4140 mm Hg during follow-up compared with those with SBP between 130 and 140 mm Hg (HR 1.46; 95% CI: 1.25 1.71). However, patients with SBP o130 mm Hg had similar risk with those achieving conventional control, whereas SBP o115 mm Hg suggested an additional hazard. The ACCORD-BP (Action to Control CardiOvascular risk in Diabetes-BP) trial randomized 4733 high-risk patients with type 2 diabetes to target SBP o120 or o140 mm Hg. 35 After 4.7 years of follow-up, no difference between groups was observed in the primary outcome of myocardial infarction, stroke, or cardiovascular death (HR 0.88; 95% CI: 0.73 1.06) or all-cause mortality (HR 1.07; 95% CI: 0.85 1.35). Tight SBP control was associated with a reduction in the risk of stroke (HR 0.59; 95% CI: 0.39 0.89), but with more serious adverse events (3.3% vs. 1.3%, Po0.001). These findings were considered by many as conclusive evidence against the 130/80 BP target. However, a careful interpretation of ACCORD-BP suggests differently for several reasons: first, the intensive SBP goal in ACCORD-BP was o120 and not o130 mm Hg; second, after 1 year and until the study end, the mean SBP values were 119.3 and 133.5 mm Hg, respectively; third, because of factorial design and achieved BP values, the power of the study was reduced and the event rate was much lower than expected. Thus, the main conclusion of ACCORD- BP was that SBP target o120 mm Hg is rather not justified as it confers benefit only for strokes and is associated with more side effects. However, the issue of the optimal SBP goal in diabetes remained unresolved, 10,11 as it cannot be excluded that the SBP target of o125, o130, or o135 mm Hg reduces cardiovascular risk compared with a target of o140 mm Hg. Recent meta-analyses attempted to provide further light on this important issue. One meta-analysis 36 included randomized studies of patients with type 2 diabetes or impaired fasting glucose/impaired glucose tolerance with X1-year follow-up. Thirteen trials with 37,736 participants examining either different BP targets or drug interventions were included, and reductions of 10% in all-cause mortality and 17% in stroke with intensive BP control were noted. Meta-regression suggested continued risk reduction for stroke to SBP o120 mm Hg, but for levels o130 mm Hg, there was also a 40% increase in serious adverse events with no other benefits. 36 Another meta-analysis included 31 drug or BP level comparing trials with 73,913 diabetics. 37 Allocation to moretight BP control reduced the risk of stroke by 39%, but not that of myocardial infarction. However, this analysis did not distinguish between different target levels; thus, study arms with similar BP targets from different studies were grouped to more-tight or less-tight groups depending on the BP of the comparator. A third meta-analysis included only 5 studies with 7,312 participants that had compared prespecified BP targets and assessed at least one of the outcomes of mortality, myocardial infarction, or stroke. 38 Intensive BP targets (p130/80 mm Hg) were associated with insignificant decreases in mortality (relative risk (RR) 0.78; 95% CI: 0.55 1.05) and myocardial infarction (RR 0.93; 95% CI: 0.80 1.08), and significant decrease in stroke (RR 0.65; 95% CI: 0.48 0.86), compared with standard targets (p140 160/ 85 100 mm Hg). The strength was the inclusion of only trials that compared interventions with different BP goals, but 4 of Kidney International (2014) 85, 536 546 539

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD Table 1 Possible blood pressure targets for patients with kidney disease according to available evidence from trials with renal and/or cardiovascular primary end points Protein excretion o0.3 g/day (albumin excretion o150 mg/day) Protein excretion 0.3 1 g/day (albumin excretion 150 500 mg/day) Protein excretion 41g/day (albumin excretion 4500 mg/day) Nondiabetic CKD o140/90 mm Hg o130/80 mm Hg o125/75 mm Hg a Diabetic CKD SBP o130 140 mm Hg b DBP o80 mm Hg b o130/80 mm Hg c o130/80 mm Hg c (o125/75 mm Hg c for young patients with heavy proteinuria) Abbreviations: CKD, chronic kidney disease; DBP, diastolic blood pressure; SBP, systolic blood pressure. a As evident from Modification of Diet in Renal Disease (MDRD) study B trial phase and MDRD long-term study. b From cardiovascular outcome trials. c Following extrapolation of data from non-diabetic CKD studies and post hoc or observational analyses in diabetic CKD. 5 studies (with the exception of ACCORD, which conferred 65% of the total populations) randomized patients to DBP targets, and 2 studies included normotensives at baseline. As the authors acknowledge, this meta-analysis could not draw conclusions regarding any specific target, but only comment on the comparative effectiveness of intensive versus standard BP-lowering strategies. Furthermore, the mix-up of the target with actually achieved BP values in standard arms included leads to problems similar to that of ACCORD-BP. What is of major importance is that in both the ACCORD-BP and the relevant meta-analyses, even with the above limitations, RRs of all outcomes studied pointed strongly toward benefit with the intensive target; thus, the possibility that higher power could have led to significant differences in favor of the intensive targets cannot be excluded. Overall, based on all available data, the DBP target of o80 mm Hg seems justified for all patients with diabetes (Tables 1 and 2) for reasons of reduction in cardiovascular end points and mortality. On the other hand, the critical question of whether the o130 mm Hg SBP target in patients with diabetes is justified remains unanswered. This could be conclusively answered by an adequately powered trial comparing o130 with o140 mm Hg SBP goals and ensuring relevance of achieved BP to target BP levels. Until such evidence appears, caution in data interpretation and individualization of treatment seems the best way forward. For patients with CKD, this has been exemplified by reports noting that in all major renal outcome studies, the proportion of participants 470 years old was particularly small, and such evidence is of little relevance to the elderly diabetics, which is the faster growing patient group. 39 When it comes to BP target, a level of o125/75 mm Hg may be easily tolerated and confer nephroprotection in young individuals; however, the same level in the elderly diabetics could be associated with frequent episodes of hypotension and acute renal failure (especially with concomitant aggressive RAAS blockade or diuretic use, and presence of atherosclerotic renal artery lesions), leading to loss of renal function much faster than anticipated. 40 RAAS BLOCKADE IN PATIENTS WITH CKD Proteinuric kidney disease The first outcome trial to evaluate the effect of RAAS blockers on renal disease was the Collaborative Study, which randomized 409 type 1 diabetics with protein excretion 40.5 g/day and SCr p2.5 mg/dl to captopril or placebo. 41 After 3 years, treatment with captopril led to 43% reduction in the risk of the primary end point of doubling of SCr, 50% reduction in the combined end point of death, need for dialysis, and transplantation, and 30% reduction in UAE compared with placebo. Small differences in BP favoring the captopril group were noted, but did not affect outcomes. In type 2 diabetes, this effect of RAAS blockade was tested in the RENAAL 42 and IDNT 43 trials. In RENAAL, losartan treatment was associated with a 16% reduction in doubling of SCr, ESRD, or death, a 35% reduction in ACR, and a 15% decrease in the rate of CrCl reduction. 42 In IDNT, irbesartan resulted in a 20% reduction as compared with placebo and a 23% reduction as compared with amlodipine in doubling of SCr, ESRD, or death; proteinuria decreased by 33% with irbesartan versus 6% with amlodipine and 10% with placebo. 43 Careful meta-analyses in patients with diabetic nephropathy confirmed these effects of ACEIs and ARBs in reduction of the risk of ESRD and SCr doubling. 44,45 The post hoc analyses of the above trials investigated the relationship of baseline proteinuria and reductions in proteinuria during follow-up with the outcome. In RENAAL, baseline proteinuria had a nearly linear relationship with the risk for the primary outcome. Furthermore, for every 50% reduction in albuminuria in the first 6 months, there was a 36% reduction of the primary end point and a 45% reduction for ESRD at study end. The authors attributed the renoprotection to the antiproteinuric effect of losartan and not to BP. 46 Similarly, in IDNT, every twofold increase in baseline proteinuria doubled the risk of the primary end point. Irrespective of treatment group, this risk was cut in half with every 50% reduction in proteinuria at 1 year. 47 These results clearly supported the value of proteinuria as an intermediate outcome in patients with overt diabetic nephropathy. Studies in nondiabetic CKD also support the use of ACEIs. In the REIN-2 (Ramipril Efficacy In Nephropathy-2) study, patients with mean SCr of 2.4 mg/dl and proteinuria of 43 g/day were randomized to ramipril or placebo; ramipril showed significant reductions in proteinuria, GFR decline, and the risk of SCr doubling or ESRD compared with placebo, even after adjustment for changes in BP. 48 In the AASK trial, ramipril showed a 36% reduction in the composite of 50% reduction of GFR, ESRD, or death 540 Kidney International (2014) 85, 536 546

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD review Table 2 Preferred first-line agents for hypertensive patients with kidney disease, according to available evidence from trials with renal and/or cardiovascular primary end points Normoalbuminuria (o30 mg/day or ACR o30 mg/g) Microalbuminuria (30 300 mg/day or ACR 30 300 mg/g) Macroalbuminuria (4300 mg/day or ACR 4300 mg/g) or clinical proteinuria (40.5 g/day) Nondiabetic kidney None preferred None preferred ACEI or ARB a disease Diabetic kidney disease ACEI or ARB b ACEI or ARB b ACEI or ARB a Abbreviations: ACEI, angiotensin-converting-enzyme inhibitor; ACR, albumin-to-creatinine ratio; ARB, angiotensin receptor blocker. a Based on short-term controlled trials with proteinuria as primary outcome, ultra-high doses of ACEIs or ARBs or dual blockade with conventional dosing of ACEIs, ARBs, aliskiren, or spironolactone may be tried with caution by experienced physicians in selected individuals with high levels of proteinuria and low risk of complications from aggressive renin angiotensin aldosterone system (RAAS) blockade (see text). b RAAS blockers have been shown to reduce progression of typical diabetic nephropathy from normo- to microalbuminuria and from micro- to macroalbuminuria. However, no specific agents are indicated in patients with diabetes and possible other causes of reduced estimated glomerular filtration rate (egfr; especially in the elderly). compared with amlodipine, and a 22% reduction compared with metoprolol. 22 Another study randomized 224 patients with more advanced CKD (SCr 3.1 5.0 mg/dl and mean proteinuria 1.6 g/day) to benazepril or placebo on top of conventional antihypertensive therapy; 49 benazepril showed a 43% reduction in doubling of SCr, ESRD, or death, a 23% decrease in the rate of GFR decline, and 2.5 times greater reduction in proteinuria after an average of 3.4 years. A previous meta-analysis in nondiabetic CKD showed that regimens including an ACEI were associated with a 31% reduction in ESRD incidence and a 30% reduction in the composite of SCr doubling or ESRD. 50 The renoprotective properties of RAAS blockers in nondiabetic patients are also more prominent in those with heavier proteinuria. In REIN-2, higher baseline proteinuria was associated with larger differences in GFR decline and the percentage of patients reaching the primary end point between the two groups. 48 In a post hoc analysis of AASK (where only 33% of participants had proteinuria at baseline), a reduction in proteinuria at 6 months was associated with less progression to ESRD at 5 years. 51 In another study of benazepril in patients with CKD of various causes, the reduction in doubling of SCr or need for dialysis was greater in those with proteinuria 41 g/day. 52 Similarly, the aforementioned meta-analysis showed a better renoprotective action of ACEIs in patients with higher proteinuria, whereas in protein excretion o0.5 g/day no benefit of ACEIs compared with other antihypertensives was found. 50 There are currently no outcome data supporting differences in renoprotection between ACEIs and ARBs. This was exemplified in the DETAIL (Diabetics Exposed to Telmisartan And enalapril) study that compared the effects of enalapril and telmisartan in 250 patients with type 2 diabetes, hypertension, and UAE between 11 and 999 mg/min. After 2 years, the two agents had similar effects on GFR change, SCr, UAE, BP, and the rates of ESRD, cardiovascular events, and all-cause mortality. 53 Some authors call for the use of ARBs in place of ACE inhibitors because they are generally better tolerated, have a lower incidence of hyperkalemia and cough, and are less frequently associated with angioedema. 54 The higher cost of ARBs makes this argument untenable for some, but ARBs have been proven to be cost effective in the management of nephropathy in various settings. 55 Overall, it seems reasonable to use the two classes interchangeably in patients with proteinuria. Nonproteinuric kidney disease Although the beneficial actions of RAAS blockers in patients with proteinuria or CKD with established natural course (that is, diabetic nephropathy) are based on solid background and clinical evidence, no outcome trial has specifically investigated their effects on hypertensive subjects with early CKD or those with reduced egfr in the absence of proteinuria. This issue is of major clinical importance, as with the existing CKD definition, 40% of adults 470 years old have egfr o60 ml/min per 1.73 m 2, but only 5% have macroalbuminuria. Among hypertensive patients, B15% have egfr o60 ml/min per 1.73 m 2 (reaching to 30% in those 465 years old), but again 45% have macroalbuminuria. 28,39 An important systematic review suggested that guidelines on ACEI and ARB use toward renoprotection have limited relevance to patients 470 years old, as 75% of studies on which contemporary guidelines were based did not include patients of this age group, and only one (the Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)) included an important proportion of elderly subjects. 39 In recent years, indirect data on renal outcomes in such populations were also available, mainly from subanalyses of cardiovascular trials. The first trial with population of early CKD was the ABCD (Appropriate Blood Pressure Control in Diabetes), 14 that included 470 hypertensives with type 2 diabetes, of whom only 18% had macroalbuminuria. Baseline CrCl was 85 ml/min per 1.73 m 2 overall and 75 ml/min per 1.73 m 2 in the subgroup of macroalbuminuric patients. Participants were randomized to nisoldipine or enalapril and intensive or moderate BP control in a 22 design. There was no difference in CrCl between the two groups over 5.3 years, although enalapril significantly lowered UAE. However, the most definite end point of incident ESRD was not recorded. Furthermore, it was unknown whether this benefit of enalapril on proteinuria would translate in retardation of ESRD if follow-up was longer. Kidney International (2014) 85, 536 546 541

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD ALLHAT randomized 433,000 patients with hypertension and 41 cardiovascular risk factor aged 455 years to chlorthalidone, amlodipine, or lisinopril with a primary cardiovascular outcome. With respect to renal function, with exclusion criteria of SCr 42.0 mg/dl and ACEI use for underlying CKD, the mean baseline egfr was 78 ml/min per 1.73 m 2. Measurements of urine protein were not included, but proteinuric patients should have been a minority. At study end, egfr was significantly higher in the amlodipine group than in the chlorthalidone and lisinopril groups (75 vs. 70 and 71 ml/min per 1.73 m 2, respectively). 56 In post hoc analyses there were no differences in the incidence of ESRD or a X50% decrement in egfr among the three groups in total and in patients with egfr of 60 89 or o60 ml/min per 1.73 m 2. 15 These results are in direct contrast with the aforementioned findings of the IDNT, where amlodipine has accelerated renal function decline; this, however, is rather directly related to different populations under study. 13 The authors of ALLHAT suggested that presumably participants with decreased renal function were patients with ischemic renal disease, for which an overwhelming renoprotective effect of ACEIs is not expected. 15 The absence of renoprotective effects of lisinopril (and possibly the beneficial actions of amlodipine) in a population with mean age 67 years and mean egfr 78 ml/min per 1.73 m 2, with the few individuals at risk for CKD progression suffering mainly from ischemic renal disease, seems by all means reasonable. A previous meta-analysis 57 also challenged the renoprotective action of ACEIs and ARBs. In placebo-controlled trials, these agents showed benefits on all renal outcomes, but the studies also had reductions in BP in active treatment groups. Studies comparing ACEIs or ARBs with other antihypertensives yielded the RR of 0.71 (95% CI: 0.49 1.04) for SCr doubling and 0.87 for ESRD (95% CI: 0.75 0.99). Among studies including only diabetic patients and comparing RAAS blockers with placebo, RAAS blockers were associated with significant reductions of 21 and 22% in the risk of ESRD and SCr doubling; however, in studies using active treatment, no renal benefit with RAAS blockade was evident. 57 This meta-analysis met criticism for several methodological issues such as substantial heterogeneity across studies, domination by ALLHAT (which may have overridden any effect in patients with true diabetic nephropathy), ignorance of the issues of proteinuria and CKD staging, equal attention to soft and hard renal end points, absence of patient-level data, and division of studies into those using placebo or active treatment, although in all placebo-controlled studies, placebo was added on standard therapy. 13,58,59 Overall, these findings seemed weak to controvert the clear results of studies on RAAS blockade in proteinuric CKD; however, they raised attention to the issue of renoprotection in early stages of CKD or nonproteinuric disease. The renal sub-study of the ACCOMPLISH (Avoiding Cardiovascular Events through Combination Therapy in Patients Living with Systolic Hypertension) provided further data for nonproteinuric patients. This study randomized 11,506 patients with hypertension and high cardiovascular risk to benazepril plus amlodipine or benazepril plus hydrochlorothiazide, and was terminated early because of benefit of the former in the primary cardiovascular outcome. Among study participants, about 85% were 465 years old and 60% were diabetic; however, mean egfr was 79 ml/min per 1.73 m 2 (only 10% of the patients had egfr o60 ml/min per 1.73 m 2 ) and baseline micro- and macroalbuminuria were at 19% and 5% of the population, respectively. 60 The benazepril and amlodipine groups had a slower egfr decline ( 0.88 vs. 4.22 ml/min per 1.73 m 2 per year), although it reduced ACR less effectively than benazepril and hydrochlorothiazide. Furthermore, although the number of ESRD events was small, the benazepril and amlodipine combination was associated with a 48% reduction in SCr doubling, egfr o15 ml/min per 1.73 m 2, and dialysis (HR 0.52; 95% CI: 0.41 0.65), and a 27% reduction in SCr doubling, ESRD, and death (HR 0.73; 95% CI: 0.64 0.84). 60 Again, it can be argued that better preservation of renal function when adding on a RAAS blocker, a dihydropiridine versus a thiazide, is not unreasonable in groups of patients with mean age 465 years and egfr well above 60 ml/min per 1.73 m 2. The absence of specific benefit of a dihydropiridine on UAE would not be of particular importance in elderly patients with normoalbuminuria who are at much higher risk for renal injury through dehydration and hypotension. Use of high doses or combined treatment with RAAS blockers Because of the absence of specific therapies for proteinuric nephropathies, several groups have evaluated the effects of aggressive RAAS blockade toward renoprotection. 13 Shortterm randomized studies in patients with micro- or macroalbuminuria have shown that use of a single RAAS blocker in ultra-high dose (that is, 2 to 3 times the maximum dose for hypertension) could reduce UAE more effectively than conventional doses without important side effects. 61 64 Similarly, combination treatment of ACEIs and ARBs reduced proteinuria more than single blockade at maximum doses. 65,66 On this basis, the results of the Combination- Treatment of ARB and ACEI in Nondiabetic Renal Disease (COOPERATE) trial in 2003, showing that a combination of trandolapril and losartan in patients with proteinuric nondiabetic CKD resulted in a 60% reduction in SCr doubling or ESRD compared with either drug alone, 67 were considered perfectly reasonable and led to inclusion of combination treatment in clinical practice; however, they were followed by great embarrassment for the nephrology community when the whole trial was found to be a fraud. 68 The Ongoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial (ONTARGET) randomized 25,620 patients with previous cardiovascular event to maximum doses of ramipril, telmisartan, or their combination, and was anticipated to provide the definite answer on the value of double RAAS blockade. 69 With regard to renoprotection, the study provided very useful information but not for patients with proteinuric disease because of the 542 Kidney International (2014) 85, 536 546

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD review Table 3 Characteristics of outcome trials evaluating the effects of combined RAAS blockade on cardiovascular and renal end points ONTARGET 16,69 ALTITUDE 17 N 25,620 8561 Study groups Ramipril 10 mg, telmisartan 80 mg, combination of both Aliskiren 300 mg versus placebo on top of ACEI or ARB Primary composite outcome Death from cardiovascular causes, myocardial infarction, stroke, or hospitalization for heart failure Cardiovascular death or first occurrence of cardiac arrest with resuscitation; myocardial infarction; stroke; unplanned hospitalization for heart failure; ESRD, death attributable to kidney failure, or need for RRT with no dialysis or transplantation available or initiated; or doubling of SCr Renal composite outcome Any dialysis, doubling of SCr, or death ESRD, death attributable to kidney failure, or the need for renal replacement therapy with no dialysis or transplantation available or initiated; or doubling of SCr Median follow-up 4.7 Years 2.8 Years Population characteristics Mean age 66.4 Years 64.5 Years Gender (%) Male 73% 68% Female 27% 32% BMI (kg/m 2 ) 28.1 29.9 Smoking (%) 12.5% 13% Diabetes (%) 37% 100% Hypertension (%) 69% 94.5% Mean baseline BP (mm Hg) 142/82 137/74 Baseline egfr (ml/min per 1.73 m 2 ) 73.6 57.0 egfr o60 ml/min per 1.73 m 2 24% 67.5% Urine albumin excretion Normoalbuminuria 82.9% 14.5% Microalbuminuria a 13.1% 25.6% Macroalbuminuria b 4% 58.4% Main findings Primary composite outcome: telmisartan was not inferior to ramipril; combination treatment was not superior to ramipril, and was associated with higher incidence of hypotension, hyperkalemia, and syncope Renal outcomes: telmisartan had similar effects with ramipril; combination treatment increased by 2.2-fold the risk of dialysis for ARF compared with ramipril. Terminated early because of hypotension, hyperkalemia, and ARF in the aliskiren group Cardiovascular composite: 11% increase with aliskiren (HR 1.11 (CI: 0.99 1.25), P ¼ 0.09), driven by 2.4-fold higher risk of resuscitated cardiac arrest. Renal composite: no difference between groups (HR 1.03 (CI: 0.87 1.23), P ¼ 0.74) Abbreviations: ACEI, angiotensin-converting-enzyme inhibitor; ALTITUDE, Aliskiren Trial in Type 2 Diabetes Using Cardiorenal Endpoints; ARB, angiotensin receptor blocker; ARF, acute renal failure; BMI, body mass index; BP, blood pressure; CI, confidence interval; egfr, estimated glomerular filtration rate; ESRD, end-stage renal disease; HR, hazard ratio; ONTARGET, Ongoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial; RAAS, renin angiotensin aldosterone system; RRT, renal replacement therapy; SCr, serum creatinine. a Defined as urinary albumin-to-creatinine ratio X3.4 to o33.9 mg/mmol (roughly equivalent to X30 to o300 mg/g) in ONTARGET and X20 to o200 mg/g in ALTITUDE. b Defined as urinary albumin-to-creatinine ratio X33.9 mg/mmol in ONTARGET and X200 mg/g in ALTITUDE. population characteristics (Table 3); 69% of participants had hypertension, 37% diabetes, and 24% egfr o60 ml/min per 1.73 m 2, but only 13% had microalbuminuria and 4% macroalbuminuria. Along with higher incidence of hypotension and hyperkalemia with combination treatment, the renal outcome of dialysis and doubling of SCr was higher with combination versus ramipril (HR 1.24; 95% CI: 1.01 1.51). UAE rose continuously during follow-up in all groups, with combination treatment displaying the lowest rate. 16 These results were considered by many as conclusive evidence against the use of double RAAS blockade in CKD; it was also argued that proteinuria should not be used as an intermediate renal outcome. However, such conclusions cannot be drawn from cohorts where the vast majority of participants have normoalbuminuria. Differences in the renal outcome derived from significant differences only in dialysis for acute renal failure, which was included in the composite, whereas the risks of SCr doubling and chronic dialysis were not different. Thus, ONTARGET was another example of the risks of aggressive RAAS blockade in susceptible individuals (that is, elderly normotensives, with reduced egfr, most probably having ischemic renal disease). 70,71 The introduction of aliskiren in clinical practice added another option for RAAS blockade. The Aliskiren in the Kidney International (2014) 85, 536 546 543

PA Sarafidis and LM Ruilope: BP reduction and RAAS blockade in CKD Evaluation of Proteinuria in Diabetes (AVOID) study randomized 599 patients with type 2 diabetes, hypertension, and macroalbuminuria to aliskiren (150 mg titrated to 300 mg/day) or placebo on top of losartan 100 mg and optimal antihypertensive therapy. After 6 months, a reduction of 20% in ACR ratio was noted with aliskiren, 72 an effect that was independent of BP. 73 Based on the above data, the Aliskiren Trial in Type 2 Diabetes Using Cardiorenal Endpoints (ALTITUDE) study compared the effects of a combination of aliskiren and ACEI or ARB with ACE or ARB alone on cardiovascular and renal outcomes in 8561 patients with type 2 diabetes (Table 3). ALTITUDE was prematurely terminated at 69% of events because of renal complications, hyperkalemia, and hypotension in the aliskiren group; 74 this received great attention and was considered again as the end of double RAAS blockade. However, the recently published full report clarified a lot of issues. 17 All components of primary outcome in ALTITUDE did not differ between groups, with the exception of resuscitated cardiac arrest; SCr doubling was practically similar, and the end point of ESRD, dialysis, or death because of kidney failure also did not differ between groups. In the aliskiren group, BP was lower by 1.3/ 0.6 mm Hg, and reduction of ACR by 14% greater than placebo. Thus, the main differences between groups was the proportion of patients with hyperkalemia (11.2% vs. 7.2%), and that of patients with reported hypotension (12.1% vs. 8.3%; Po0.001 for both). 17 One can argue that following the inclusion criteria of (1) macroalbuminuria or (2) egfr 30 60 ml/min per 1.73 m 2 and microalbuminuria or (3) egfr 30 60 ml/min per 1.73 m 2 and a history of cardiovascular events, the baseline population of ALTITUDE resembled that of cardiovascular and not renal outcome trials and was much different from AVOID; mean age was at 65 years, 42% had cardiovascular disease, 67% had egfr o45 ml/min per 1.73 m 2, and only 58% had macroalbuminuria. Included patients also had to have BP o135/85 mm Hg, or BP 135/85 170/110 mm Hg at early visits if treated with at least three antihypertensives; thus, mean BP at baseline was 137/74 mm Hg and 69% of participants were receiving diuretics along with ACEI or ARB. According to the above, a large proportion of participants seemed susceptible to complications from BP lowering (indeed, hypotension was more frequent in the elderly and those receiving loop diuretics), much more from RAS blockade. Again, the important side effects leading to premature termination can be directly attributed to potent RAS blockade in susceptible individuals. Furthermore, in subgroup analyses, the risk of the primary outcome was significantly higher only in patients with baseline hyperkalemia (X6 mmol/l) and this could directly affect the outcome. In the AVOID study, higher frequency of hyperkalemia with aliskiren was observed only in subjects with egfr o60 ml/min per 1.73 m 2. 75 Thus, it is still not known whether dual RAAS blockade may be beneficial for renoprotection in selected patient groups, that is, the young patient with proteinuria, preserved egfr, no vascular disease, and high compliance to dietary potassium restrictions. 76 The ongoing VA NEPHRON-D trial 77 that randomized 1850 patients with diabetes and overt proteinuria to a combination of losartan and lisinopril versus losartan alone is expected to answer this important question. Addition of an aldosterone receptor antagonist to patients already on ACEI or ARB is suggested as another option for renoprotection, as plasma aldosterone is elevated in CKD and may independently contribute to renal injury, 78 whereas use of ACEIs or ARBs does not necessarily result in maintained reductions in plasma aldosterone. 79 Addition of spironolactone in proteinuric patients receiving ACEIs or ARBs reduced proteinuria in several pilot studies. 80 83 Similarly, eplerenone added to an ACEI further reduced UAE in patients with hypertension and left ventricular hypertrophy. 84 A detailed study randomized 81 patients with diabetes, hypertension, and macroalbuminuria already on lisinopril 80 mg to losartan 100 mg, spironolactone 25 mg, or placebo for 48 weeks. 85 Compared with placebo, ACR decreased by 34.0% with spironolactone and by 16.8% with losartan; ambulatory BP, CrCl, sodium and protein intake, and glycemic control did not differ between groups. Serum potassium was higher with either spironolactone or losartan (yet on average o5.2 meq/l in all groups). Based on the above, the use of low-dose spironolactone is highly promising, but as hyperkalemia is always a cause of concern for practicing nephrologists, 86 outcome trials with hard renal end points are still necessary to determine the benefits and risks of this approach. CONCLUSIONS Selection of BP targets for renoprotection should be based on the type of CKD and the degree of proteinuria. According to recent observations, in patients with nondiabetic CKD, a goal BP of o130/80 mm Hg seems justifiable for those with protein excretion 40.25 0.3 g/day (or albumin excretion 40.15 g/day); a lower BP target of o125/75 may be applicable for patients with proteinuria 41 g/day. In contrast, individuals with normoalbuminuria should be treated to the conventional goal of o140/90 mm Hg, as further BP lowering does not confer renal benefits. For patients with diabetic CKD, data from observational analyses and surrogate outcomes (and extrapolation of nondiabetic CKD findings) rather suggest a BP of o130/80 mm Hg when protein excretion is 40.3 g/day. For nonproteinuric patients with diabetes, cardioprotection must be the main determinant of BP goals; DBP o80 mm Hg is rather warranted, whereas the optimal SBP target lies between 130 and 140 mm Hg and should be decided on individual basis, according to anticipated benefits of stroke reduction and possible risks of hypotension and acute renal failure. 12 A similar approach is advisable when selecting the degree of RAAS blockade in CKD. Major trials have established that ACEIs and ARBs can decrease the risk of ESRD in patients with diabetic or nondiabetic proteinuric CKD, and to retard the progression from micro- to macroalbuminuria in kidney diseases with predictable natural course (that is, diabetic 544 Kidney International (2014) 85, 536 546