血压控制目标21

Blood Pressure Control in Chronic Kidney Disease: Is Less Really More? Julia B. Lewis Division of Nephrology and Hypertension, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee “A mind is a terrible thing to change” brings up the problem of confirmation bias. In a recent analysis of nearly 8000 participants, people were twice as likely to seek information confirming what they already believed rather than con- sider evidence that would challenge those beliefs.1 Indeed, medicine is not immune to this bias; people reviewing the same data can reach dramatically dif- ferent conclusions on the basis of preex- isting beliefs. For many years, observa- tional studies suggested patientswhohad renal disease and were on erythropoie- tin-stimulating agents had better out- comes with higher hemoglobin levels. This belief that a normal hemoglobin level was a good thing persisted despite early evidence to the contrary and even- tually drove the formation of widely used guidelines.2 Did confirmation bias con- tribute to the sad fact that the guidelines did not change until after multiple clini- cal trials demonstrated harm? Now we have the BP problem. New BP goals significantly �140/90 mmHg may ormay not be good for patients with chronic kidney disease (CKD). What is unarguably true is that despite well-in- tended guidelines and passionate beliefs, no well-powered, randomized, inten- tion-to-treat clinical trials have demon- strated a clinically significant benefit of achieving a BP target of�130/80 mmHg in the setting of CKD. Multiple landmark studies demon- strated unequivocally that lowering BP in individuals who are of any age and havemoderate to severe hypertension re- duces risk for stroke and, less consis- tently, other cardiovascular events.3–12 Patients from these studies were ran- domly assigned to various antihyperten- sive regimens, resulting in different achievedBP.Of note, in all of those trials, patients in low-BP groups, on average, did not achieve mean BP �140/90 mmHg, and many showed benefit at considerably higher BP (Table 1). No ad- equately powered, randomized, con- trolled trials have demonstrated cardio- vascular benefit with BP goals �140/90 mmHg. Only epidemiologic observa- tions that patients who achieve lower BP have fewer cardiovascular events sup- ports the hypothesis that BP goals �140/90 mmHg may have cardiovascu- lar benefit. This hypothesis will be tested in the ongoing Systolic Blood Pressure INTervention (SPRINT) trial sponsored by the National Institutes of Health, which will randomly assign�10,000 pa- tients to one of two BP goals:�120/80 or �140/90 mmHg. Patients with CKD were systemati- cally excluded from the majority of early BP trials, and, consequently, few trials provided renal outcomes3–12; however, secondary analyses of these trials yielded some hypothesis-generating informa- tion. In the Hypertension Detection and Follow-up Program, patients who en- tered the trial with baseline serum creat- inine (SCr) level of 1.50 to 1.69 mg/dl benefited from lower BP in the stepped- care group with fewer hypercreatinine- mic events.13 In contrast, in the Euro- pean Working Party on High Blood Pressure in the Elderly trial, SCr in- creased significantly in the patients who achieved lowerBPcomparedwithplacebo.14 In the Multiple Risk Factor Intervention Trial (MRFIT), 32,544 men were en- rolled in a prevention trial to study the effects of BP control, lowering of serum cholesterol, and cessation of smoking on Published online ahead of print. Publication date available at www.jasn.org. Correspondence: Dr. Julia B. Lewis, Division of Ne- phrology and Hypertension, Department of Medicine, S-2332 Medical Center North, Vanderbilt University School of Medicine, Nashville, TN 37232-2372. Phone: 615-343-6105; Fax: 615-343-7156; E-mail: julia.lewis@vanderbilt.edu Copyright © 2010 by the American Society of Nephrology ABSTRACT BP control is critical in the treatment of patients with chronic kidney disease. Recent guidelines now recommend a BP goal of �130/80 mmHg. Clinical trials using a randomized, intention-to-treat design have not established the benefits of this goal; rather, observational data and secondary or subgroup analyses drove the development of the new guidelines. A variety of observations suggest potential adverse events associate with achieving too low a BP in patients with chronic kidney disease, and ongoing randomized trials will have to establish the benefits or risks of meeting this goal. J Am Soc Nephrol 21: 1086–1092, 2010. doi: 10.1681/ASN.2010030236 SPECIAL ARTICLE www.jasn.org 1086 ISSN : 1046-6673/2107-1086 J Am Soc Nephrol 21: 1086–1092, 2010 coronary artery disease. After 16 years of follow-up, there were 814 cases of ESRD, and elevated BP was a strong indepen- dent risk factor.15,16 The major cardio- vascular trials establishing the benefit of a systolic BP (SBP) goal of�140 mmHg on reducing cardiovascular outcomes did not consistently suggest renal bene- fit. These trials were not designed to en- roll patients with CKD or carefully mea- sure renal outcomes. Numerous observational studies also reported an association between higher levels of BP and ESRD.17–20 More re- cently, the Action inDiabetes andVascu- lar Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) trial, a study of intensive blood sugar control—not BP control— was conducted of 11,140 patients with type 2 diabetes. In a secondary analysis of BP in these patients, there were fewer re- nal events in patients with lower SBP over a range of 110 to 170 mmHg.21,22 Although observational data sets sup- port the hypothesis that there is a contin- uous benefit of reducing BP to lower and lower levels, they are problematic: Pa- tients were not randomly assigned to BP goals in these studies. As a result, there is the potential for confounding factors to account for the observed association of lower BP and better renal outcomes rather than a cause-and-effect relation- ship. As renal disease worsens, BP rises and is more difficult to control, so the observation in a study in which patients with worse renal function may on aver- age have higher BP and those with less severe renal disease have lower BP and patients are not randomly assigned to different BP goals may not be due to a lower BP slowing the rate of loss of renal function. One cannot distinguish whether patientswith less severe renal disease have lower BP because they have less severe renal disease or lower BP leads to fewer adverse renal outcomes. There may also be common factors that both worsen re- nal disease and cause higher BP. Again there might be a strong relationship be- tween higher BP and worse kidney dis- ease, but this would not be due to a cause-and-effect relationship. Last, drugs used to lower BP, such as inhibitors of the renin-angiotensin sys- tem,may benefit the kidney beyond low- ering BP23–25; that is, lower BP would as- sociate with better renal outcomes, but it would not be the cause of better renal outcomes. The only way to exclude these confounding factors is to conduct a trial in which patients are randomly assigned to different BP goals and the renal event rate is examined. Thus, studies that sim- ply observe an association between achieved BP and renal events are open to interpretation and irresolvable concern. The classic study by Parving et al.26 was one of the first to demonstrate that early antihypertensive treatment re- duced the rate of declining renal function in 12 patients who had diabetes and served as their own controls. Several other small studies similarly revealed a benefit to randomly assigning patients to lower BP goals but also had too few pa- tients to be definitive and used decreas- ing albuminuria as an outcome rather than a change in estimated GFR.27–30 More recently, the Ramipril Efficacy in Nephropathy 2 (REIN-2) study ran- domly assigned 338 patients to either a diastolic BP (DBP) of�90 mmHg (con- ventional group) or BP�130/80 mmHg (intensified group) and reported no de- crease in ESRD events in the intensified group.31 The Modification of Diet in Renal Disease (MDRD) study and the African American Study of Kidney Disease and Hypertension (AASK) are two large, well-controlled clinical trials that exam- ined the hypothesis that being randomly assigned to a mean arterial BP (MAP) of �92 versus 107 mmHg would slow the rate of decline in renal function as mea- sured by I125-iothalamate GFR.32MDRD bears careful review because the results of that trial, along with other observa- tional studies, drove the current guide- line recommendation for a BP goal of �130/80 mmHg in patients who have CKD and proteinuria�1 g/24 h. InMDRD, 840 patients with a GFR of either 25 to 55 ml/min (study A) or 13 to 24 ml/min (study B) were randomly as- signed to a MAP goal of �92 or �107 mmHg. Of note, because MAP is calcu- lated as DBP� one third pulse pressure, the patientswhowere randomly assigned to aMAP of�92mmHg (roughly equiv- alent to 125/75mmHg) had SBPbetween 98 and 154 mmHg, and the usual BP group with a MAP of 107 mmHg (roughly equivalent to 140/90 mmHg) had SBP between 103 and 163 mmHg. As can be seen in Figure 1, by the intention-to-treat design, there was no significant benefit on slowing the rate of decline of GFR in the groupof patientswhowere randomly assigned to a MAP of �92 mmHg.32 Analysis of other outcomes such as time to doubling of SCr or ESRD similarly showed no benefit of the low BP goal; however, a significant benefit of the low BP goal was seen in 54 patients who en- tered the study with proteinuria�3 g/24 h. Although the 104 patients with pro- teinuria of 1 to 3 g/24 h did not benefit as a group from the low BP goal, when grouped with patients with proteinuria �3 g/24 h, there was benefit. Further- more, no significant benefit was associ- ated with the low BP goal in any group of patients in study B (GFR 13 to 24 ml/ min; Figure 2).32,33 There were significant interactions in Table 1. Placebo-controlled hypertension trials Trial N Low BP Achieved Comparison Arms Primary Results SHEP4 4736 143/68 Diuretic-based regimen versus placebo 36% RRR in stroke STOP6 1627 166/85 � blocker and diuretic regimen versus placebo 29% RRR in stroke Syst-Eur3 4695 153/77 CCB-based regimen versus placebo 42% RRR in stroke Syst-China5 2394 150/81 CCB-based regimen versus placebo 38% RRR in stroke CCB, calcium channel blocker; RRR, relative risk reduction. SPECIAL ARTICLEwww.jasn.org J Am Soc Nephrol 21: 1086–1092, 2010 BP Control in CKD 1087 MDRD between BP interventions and baseline proteinuria, and there was a re- duction in proteinuria in patients who were randomly assigned to a MAP of �92mmHg in both study A and study B; however, a major confounding factor was the more frequent use of renin-an- giotensin inhibitors, knownmodifiers of renal progression, in the low-BP group.33 The primary result of the BP study in the MDRD was negative: There was no ben- efit in slowing the rate of decline of renal function. Only a secondary analysis in the subgroup with proteinuria supports the low BP goal and is confounded by imbalance in the use of the renin-angio- tensin inhibitors. The MDRD results and the results of the observational studies noted formed the basis for the Kidney Disease Outcomes Quality Initiative (KDOQI) guideline of �130/80 mmHg for patients with CKD and proteinuria �1 g/24 h.34 Guideline committees tend to use available evi- dence to make best practical recommen- dations, but a careful review of the data confirms that real uncertainty remains on appropriate BP goals for patients with CKD. Another quirk of the KDOQI guidelines is that they were largely based on the MDRD study, in which the low-BP group was randomly assigned to a MAP of �92 mmHg. A MAP goal al- lows a wide range of SBP and DBP com- binations. Although never published, the actual mean follow-up systolic and dia- stolic BP in the usual group were 132.7 (�1.0 to 14.9)/80.2 (�1.0 to 7.5) mmHg and in the low-BP group were 125.6 (�1.0 to 13.8)/76.7 (�1.0 to 6.8) mmHg (Tom Greene, PhD, personal communi- cation, October 2009). The low-MAP group had a wide range of SBP, approxi- mately 98 to 154 mmHg. Achieving SBP in this range is a very different interven- tion than targeting all patients’ SBP to �130 mmHg (Figure 3). Other evidence has become available since the appearance of these KDOQI guidelines. At the completion of the MDRD study, the patients were returned to the care of their usual physician, and there was no longer any BP separation between the two randomly assigned groups. After 7 years in routine care, fewer of those patients who were origi- nally randomly assigned to a MAP �92 versus�107mmHg reached ESRDor the composite end point of ESRD or death.35 The second large study, the AASK, was completed by randomly assigning 1094 black patients with hypertension and a urine protein-creatinine ratio of �2.5 to a MAP of�92 versus 102 to 107 mmHg. There was no significant benefit of being randomly assigned to the low-BP group for the full cohort.36 In D ec lin e in g lo m er ul ar fi ltr at io n ra te (m l/m m) Month after randomization Low blood pressure group B3 15 12 9 6 3 0 F4 F12 F20 F28 F36 Usual blood pressure group Figure 1. Estimated mean � SE decline in the GFR from baseline (B) to selected follow-up times (F) in MDRD. The patients who were assigned to the usual-BP group (dashed line) are compared with those assigned to the low-BP group (solid line). Adapted from Klahr et al.,32 with permission. <1 1–<3 3 Base-line urinary protein (g/day) M ea n ra te o f G FR d ec lin e (m l/m in/ yr) <1 1–<3 3 n = 420 n = 104 n = 54 n = 136 n = 63 n = 3212 8 4 0 12 8 4 0 Figure 2. Decline in the GFR according to baseline urinary protein excretion and BP group in MDRD. The mean � SE rate of decline per year in the GFR from baseline to 3 years, for study A and study B categorized by baseline urine protein, is projected. F, patients with usual BP; E, patients with low BP. The number at the bottom of each panel indicates the total number of patients with follow-up GFR measurements in the two BP groups combined. Adapted from Klahr et al.,32 with permission. 100 Pa tie nt s (% ) 110 120 130 140 Guideline BP MDRD low MAP group Figure 3. The solid line is the hypothetical representation of the average mean SBP and 1 SD on either side of that mean found in the MDRD study low-MAP group. Dashed line is the hypothetical representation of distribution of average SBP if guidelines target all SBP at �130 mmHg. SPECIAL ARTICLE www.jasn.org 1088 Journal of the American Society of Nephrology J Am Soc Nephrol 21: 1086–1092, 2010 contrast to the MDRD study, there also was no benefit in any subgroup, includ- ing those with proteinuria �1 g/24 h.36 There was, however, a reduction in pro- teinuria in the low-BP group of the full cohort. After the AASK was completed, the majority of participants entered an ob- servational cohort study. Ironically, al- though no benefit of the low BP goal was demonstrated in these same patients, the data safety committee asked all partici- pants in the AASK cohort study to have a BP goal of �130/80 mmHg to comply with current guidelines. Seven years after the completion of the AASK, during which time there was no difference in BP between the two originally randomly as- signed groups, those who originally were randomly assigned in the trial to a MAP of �92 mmHg had fewer ESRD events but only in the prespecified subgroup of patients, approximately 33%of those en- rolledwith urine protein-creatinine ratio �0.22.37 It is fair to ask whether these results from two trials are compelling evidence to support a low BP goal when the bene- fits of that goal are not apparent at the time BP is actually lowered but are ap- parent only in a subgroup of patients 7 years after there is no longer any differ- ence in the BP. Although there is indirect evidence to support the hypothesis that BP �130/80 mmHg may benefit the kidney, uncertainty abounds. Well-designed stud- ieswereunable todemonstrate that benefit in their full cohorts.32,36 Of note, in 385 children in the Effect of Strict Blood Pressure Control and ACE Inhibition on the Progression of CRF in Pediatric Pa- tients (ESCAPE) trial, those who were randomly assigned to the MAP goal of �50th percentile had 50% fewer events (50% decreases in GFR or ESRD) as compared with those who were ran- domly assigned to a MAP in the 50th to 95th percentiles.38Although it supports lower BP as renoprotective, this trial in children is difficult to align with treat- ment goals in adults. Not all hypotheses need be tested in clinical trials. For example, it would not be prudent to test the hypothesis that it is safer to jump out of an airplane with a parachute than without one.39 Although a few poor souls have survived the jump without an open chute, common sense will always prevail; however, if an inter- vention has the potential for a harmful effect rather than beneficial or is costly or a burden to patients, then a valid clinical trial is desirable, appropriate, and re- quired. Lowering the SBP goal from 140 to 130 mmHg will require our nation of patients with hypertension to take one or possibly two additional antihypertensive medications with a considerable increase in cost and potential for medication-as- sociated adverse effects. Furthermore, in the case of low BP goals, there is evidence they may be associated with clinical harm. Observational studies have suggested the risk for stroke and death in patients withCKDmay actually be greater among those with SBP levels�120mmHg com- pared with higher levels.40 Of concern, individuals with CKD have a J-shaped relationship with stroke outcomes, such that those with SBP �120 mmHg are at significantly increased risk compared with those with CKD and SBP of 120 to 129 mmHg; the risk for stroke increases, of course, for SBP �130 mmHg. This J shape was not seen for patients without CKD.41 This latter finding suggests the possibility that if studies targeting lower BP goals are conducted on populations of patients who do not include patients with CKD, then the results may be mis- leading, because the risk–benefit rela- tionship is different for patients with CKD. As is the case with observational studies supporting lower BP goals, these observational studies regarding risk are limited, because patients with more se- vere overall illness have lower BP, or drugs used to achieve lower BP levels may independently increase the risk for stroke. The few BP-lowering trials that have analyzed the effects of BP lower- ing in the subgroup of patients with CKD reported mixed results: Some showed benefit in reducing cardiovas- cular outcomes, whereas others were unable to show benefit. In the Irbesartan in Diabetic Nephrop- athy Trial (IDNT), patients who had CKD and achieved SBP �120 mmHg had a higher risk for all-causemortality than any other BP level, and this mortality risk was comparable to that seen in those with SBP �180mmHg.42 A careful analysis of all co- morbidities was unable to account for the increased mortality in patients who achieved SBP �120 mmHg. Further- more, a recentmeta-analysis of trials that examined the effects of regimens target- ing low BP goals in patients with CKD demonstrated no clear benefit for car- diovascular or CKD outcomes.43 The re- cently reported ONgoing Telmisartan Alone and in combination with Ramipril Global EndpoinT (ONTARGET) trial also shows that negative outcomes asso- ciate with very low SBP. In ONTARGET, 25,620 patients (9595 of whom had dia- betes and 6157 of whom had stage 3 or 4 CKD) were randomly assigned to receive ramipril 10 mg/d, telmisartan 80 mg/d, or both.44 Telmisartan was equivalent to ramipril with respect to the primary out- come, but in the combination group, there were