|Home | About | Journals | Submit | Contact Us | Français|
The relationship between hypertension and chronic kidney disease (CKD) has long been the subject of controversy. The pathogenetic mechanisms of nephropathy in non-diabetic individuals with hypertension, as well as optimal hypertension treatment targets in populations with nephropathy remain important clinical concerns. This manuscript reviews breakthroughs in molecular genetics that have clarified the complex relationship between hypertension and kidney disease, answering the question of which factor comes first. An overview of the potential roles that hyperuricemia plays in the pathogenesis of hypertension and CKD and current blood pressure treatment guidelines in populations with CKD are discussed. The ongoing National Institutes of Health-sponsored Systolic Blood Pressure Intervention Trial (SPRINT) is underway to help answer these important questions. Enrollment of 9,250 hypertensive SPRINT participants will be completed in 2013; important results on ideal blood pressure control targets for reducing nephropathy progression, cardiovascular disease end-points, and preserving cognitive function are expected. As such, many of the controversial aspects of hypertension management will likely be clarified in the near future.
The interplay between hypertension and chronic kidney disease (CKD) has been the subject of intensive research efforts. This manuscript addresses three controversial aspects of this relationship: (a) genetic aspects of “hypertension-attributed” glomerulosclerosis, (b) the role of hyperuricemia in the pathogenesis of high blood pressure and CKD, and (c) optimal blood pressure (BP) treatment goals in patients with CKD.
Relationships between mild-moderate essential hypertension and the initiation of CKD have recently been clarified by molecular genetic analyses. Although “hypertension-attributed” nephropathy (HN; also called arteriolar nephrosclerosis or hypertensive nephrosclerosis) is diagnosed frequently, these terms are often applied indiscriminately and contribute to misunderstanding of the true pathogenesis of non-diabetic kidney.1–3 To understand the etiologies of non-diabetic nephropathy with low level proteinuria, it is critical that clinicians consider European-derived and African-derived population groups separately; different renal disorders are typically present.
After diabetes mellitus, HN is the second most commonly reported etiology of end-stage kidney disease (ESKD) requiring renal replacement therapy in the United States Renal Data System. Kidney biopsies are rarely performed in HN and many patients present to nephrologists late in their course with secondary hypertension related to the presence of longstanding CKD.4 Relative to European Americans, African Americans are more often diagnosed with HN and hypertension-attributed ESKD. However, few African Americans with HN had documented essential hypertension with normal indices of renal function and normal urinalyses prior to diagnosis.4 Instead, HN is typically a diagnosis of exclusion in non-diabetic African Americans with sub-nephrotic proteinuria and advanced CKD.2, 5 Nephrologists often demonstrate bias when applying this diagnosis. Physicians provided identical clinical histories of a patient with hypertension and CKD with low level proteinuria (except for differing ethnicities) diagnosed HN more often in African Americans and chronic glomerulonephritis more often in European Americans.6
Controlling systemic blood pressure often slows or arrests nephropathy progression in non-diabetic, hypertensive individuals of European ancestry who have reduced glomerular filtration rate (GFR) and sub-nephrotic proteinuria. This was observed in both the Hypertension Detection and Follow-up Program (HDFP) and Multiple Risk Factor Intervention Trial (MRFIT).7, 8 However, this was not the case in patients with African ancestry. In African Americans, nephropathy typically progressed despite usual (or aggressive) BP control, including with use of angiotensin-converting enzyme inhibitors (ACEi).7–9 Although the National Institute of Health-sponsored African American Study of Kidney Disease and Hypertension (AASK) Trial demonstrated a slight benefit in slowing the rate of decline in GFR among African American patients treated to low BP targets using ACEi (relative to usual BP targets with calcium channel blockers or beta blockers),10 long-term follow-up in the AASK Cohort revealed that nearly 60% reached a primary study end-point (doubling of serum creatinine concentration, renal replacement therapy or death) despite aggressive BP lowering using ACEi.9, 11
Histologic differences are seen in the kidney biopsies of subjects clinically diagnosed with HN based on ethnicity. African Americans have more solidified glomerulosclerosis and arteriolonephrosclerosis, relative to European Americans who have more obsolescent glomeruli with ischemia-related collapse.12 African Americans with HN often have close relatives with ESKD; many relatives have nephropathy attributed to diabetes, lupus, HIV-associated nephropathy (HIVAN) and chronic glomerular disorders.13–15 This familial clustering suggested the existence of an overarching kidney failure susceptibility gene independent from high BP.1 Together, these data strongly suggested the existence of ethnic-specific disorders that were apparently related to hypertension, but with a different pathogenesis in each population group.16 This hypothesis has now been proven in the laboratory.
Relative to European-derived populations, African Americans have significantly higher incidence rates of focal segmental glomerulosclerosis (FSGS) and HIVAN, a form of FSGS, collapsing variant. In contrast to HN, these disorders have a clearly defined renal histology and kidney biopsies are often performed. It is widely appreciated that FSGS can present with low level proteinuria; however, few investigators considered that sub-nephrotic FSGS could be the disorder present in many African Americans labeled with HN.17 In 2008, Kopp et al.18 and Kao et al.19 reported results of genetic admixture mapping experiments in African Americans with FSGS and several non-diabetic forms of ESKD. These reports detected strong association between markers in the non-muscle myosin heavy chain 9 gene (MYH9) on chromosome 22q with non-diabetic kidney disease. Strong association was also detected in African Americans labeled with hypertension-attributed ESKD,20 but not hypertension per se;21 therefore, it was likely that these patients had kidney diseases in the spectrum of FSGS but with low level proteinuria. Kidney biopsies in African Americans with HN also revealed high frequencies of focal global glomerulosclerosis (FGGS), with interstitial fibrosis and arteriolar nephrosclerosis.22
Two coding variants in the adjacent apolipoprotein L1 gene (APOL1) were subsequently found to be more strongly nephropathy-associated than MYH9; nephropathy-associated MYH9 variants appear to have been implicated since they are in strong linkage disequilibrium with APOL1 G1 and G2 coding variants.23, 24 MYH9 remains associated with nephropathy in Europeans and European Americans.25–27 Identification of the APOL1 association proved that HN in African Americans was a genetic disorder.5 APOL1 is strongly associated with several severe non-diabetic forms of CKD in those with African ancestry in an autosomal recessive fashion. Inheriting one APOL1 nephropathy risk variant (G1 or G2) protected from trypanosoma brucei rhodesiense infection, a cause of African sleeping sickness. This parasite is transmitted by the tse-tse fly in sub-Saharan Africa. Hence, mutations in APOL1 appear to have arisen relatively recently in sub-Saharan Africa (within the past ~5,000 years) due to selection, they are virtually absent in European and Asian populations. Individuals inheriting two APOL1 risk variants (G1G1, G2G2, or G1G2) face markedly increased risk for non-diabetic nephropathy. Odds ratios (ORs) for APOL1 association with kidney disease rank among the highest in complex human disease, 29 for HIVAN, 17 for FSGS, and 7.3 for hypertension-attributed ESKD.23, 28 Mechanisms whereby APOL1 mutations cause glomerulosclerosis, interstitial, and vascular injury remain under intense study.29, 30 Novel treatments for this spectrum of FSGS-related kidney disease are likely to result from unraveling the pathogenesis of APOL1-associated nephropathy. Although controlling systemic BPs with ACEi remain important for reducing cardiovascular disease (CVD) risk, these treatments had minimal effects on slowing nephropathy progression in hypertensive African Americans with CKD.9
It is now apparent that FGGS in African Americans resides within the FSGS disease spectrum and is poorly responsive to systemic BP reduction. We believe that lesions in the intima and media of small, intra-renal blood vessels (true arteriolar nephrosclerosis) are directly caused by mild-moderate hypertension (as well as smoking and hyperlipidemia) in non-diabetic patients with CKD and low-level proteinuria of European-ancestry.16 These individuals truly have the systemic disorder hypertensive nephrosclerosis and nephropathy progression slows with control of systemic BP. African-ancestry patients more often have primary renal diseases with glomerular involvement (FGGS and/or FSGS), with interstitial and vascular changes.
Analyses in AASK prove this theory. DNA was available in 675 AASK participants and 618 African American non-nephropathy controls. APOL1 was impressively associated with nephropathy attributed to hypertension in all AASK cases (OR=2.6), even stronger association was present in the subset with baseline proteinuria (OR=6.3 with urine protein:creatinine ratio >0.6 g/g) or progressive nephropathy (OR=4.6 with serum creatinine concentration rising to ≥3 mg/dl during the study).31 AASK cases had been carefully phenotyped to ensure that their clinical histories were consistent with HN. Not only did their kidney disease progress despite aggressive anti-hypertensive therapy, cases proved to be genetically similar to FSGS, HIVAN and hypertension-attributed ESRD in other African American samples.
Additional APOL1-associated kidney diseases include sickle cell nephropathy32 and severe lupus nephritis with ESKD (personal communication B.I. Freedman). In one study, collapsing FSGS was more often seen in HIV positive patients with two APOL1 risk variants, as opposed to immune complex disease in those with fewer than two risk variants.33 However, a subsequent report suggests that APOL1 genotypes failed to predict renal histology or the clinical course in patients with HIVAN.34 In addition, kidneys donated for transplantation by African Americans are known to function for shorter time periods than kidneys donated by European Americans; APOL1 risk variants appear to underlie this observation.35 African American deceased donor kidneys with two APOL1 risk variants functioned for significantly shorter intervals than those with zero or one risk variant. Kidneys from African Americans with fewer than two APOL1 risk variants functioned for similar durations as European American donated kidneys. In contrast, recipient genotypes did not impact graft survival.36
The association of APOL1 with a variety of severe forms of nephropathy (and weaker association with milder kidney disease) suggests that APOL1 leads to nephropathy progression.37, 38 In support of this observation, African Americans with two APOL1 risk variants initiate dialysis at an earlier age than those with less than two risk variants.39, 40 An additional second hit, either a gene-gene or gene-environment interaction, appears necessary to trigger development of kidney disease.34, 41 These second hits will be very important since they may lead to novel therapies to treat this refractory family of non-diabetic kidney diseases in individuals with African-ancestry, including the kidney disease long attributed to hypertension, but now proven to be a variant of FSGS.
As stated above, true arteriolar nephrosclerosis, more frequent in those of European ancestry, is thought to be caused by systemic hypertension and related metabolic derangements that contribute to vascular disease. However, some researchers suggest that hyperuricemia may mediate both hypertension and CKD by producing renal vascular pathology. The controversial question is whether uric acid is simply a by-product or a causal agent in these disorders.
Purine catabolism results in the generation of uric acid, present in the plasma in its soluble form urate. In most mammals, uric acid is converted to the more soluble allantoin, which is ultimately eliminated by the kidneys. Humans lack the urate oxidase enzyme (uricase) and therefore must excrete uric acid in their urine.42 It remains unclear as to what evolutionary advantage was gained in humans by loss of expression mutations in the uricase gene. Uric acid has been shown to have antioxidant properties in plasma, potentially conferring benefit.43 It has also been proposed that uric acid may provide the advantage of adequate BP maintenance under low-salt dietary conditions via renin-dependent mechanisms, a trait necessary for survival prior to the modern high sodium diet.42
However, plasma uric acid concentrations must remain in a tight physiologic range, above which deleterious effects may result. Gout represents one harmful consequence of hyperuricemia. It remains unclear whether hyperuricemia represents a response to oxidative stress resulting from other metabolic derangements or whether elevations in uric acid concentrations independently contribute to the development of CKD and hypertension; both processes may co-exist. Here, we review the epidemiological, laboratory, and clinical studies surrounding this controversial relationship.
Several large epidemiological studies have detected association between high serum uric acid levels and the development of hypertension. Among normotensive men in the Osaka Health Survey, increasing quintiles of serum uric acid were associated with higher risk for new-onset hypertension (defined as a BP >160/95 mmHg).44 Baseline hyperuricemia was also associated with an increased risk for incident hypertension (BP >140/90 mmHg) in 4400 men and women from Okinawa.45 Among more than 3300 men and women initially free of hypertension, myocardial infarction, heart failure, CKD, or gout in the Framingham Heart Study, each standard deviation increase in serum uric acid was associated with an increased risk for new-onset hypertension.46 Uric acid levels independently predicted incident hypertension in adult men from the Normative Aging Study and MRFIT47; and analysis of the effects of hyperuricemia across ethnic strata in more than 9000 male and female Atherosclerosis Risk in Communities (ARIC) cohort participants revealed an increased risk for hypertension with each standard deviation increase in serum uric acid, an effect that was more pronounced in African Americans than European Americans.48 Despite relatively modest risk increases, these large studies strongly support a role for hyperuricemia in the initiation of hypertension.
It is important to note that variable statistical adjustments were applied for relevant confounders in the aforementioned epidemiological studies in adult participants.49 In addition to hypertension, hyperuricemia commonly associates with obesity, dyslipidemia, glucose intolerance, and decreased GFR; these factors could potentially confound study results.50 Healthy children and adolescents lacking these co-morbidities may be a better population in which to assess this relationship. Feig and Johnson evaluated 125 children aged 6 to 18 years who were referred to nephrologists for initial evaluation of hypertension and 40 normotensive controls seen in the nephrology clinic. Among hypertensive children, 63 were diagnosed with primary (essential) hypertension and 56 (89%) of them had a serum uric acid >5.5 mg/dL. In contrast, none of the 40 normotensive controls had a serum uric acid >5.5 mg/dL. The two groups had a similar GFR; however, the hypertensive children had a significantly higher body mass index percentile (BMI%); although the correlation between BP and uric acid remained highly significant after adjusting for BMI%.51 The National Health and Nutrition Survey (NHANES) examined more than 6000 American adolescents aged 12 to 17 years. In those with a normal GFR, 209 (3.3%) had an elevated BP. The mean uric acid level was 5.6 mg/dL in the hypertensive group, relative to 5.0 mg/dL in normotensive adolescents; the OR for an elevated BP was 1.38 (95% confidence interval [CI] 1.16–1.65 per 0.1 mg/dL increase in serum uric acid, adjusted for age, sex, ethnicity, and BMI%). Adolescents with elevated BP had a significantly higher BMI%.52
Epidemiological studies also examined relationships between serum uric acid concentration and the subsequent development of CKD. The large number of potential confounders makes these studies more difficult to interpret; nonetheless, intriguing results were seen in adjusted analyses. Iseki et al., examined data from over 6400 subjects in the Okinawa General Health Maintenance Association and detected an increased risk of developing an elevated serum creatinine concentration (>1.2 mg/dL in women, >1.4 mg/dL in men) for those with a serum uric acid >8.0 mg/dL, relative to <5.0 mg/dL (relative risk [RR] 2.9 men, 10.39 women).53 Pooled data from more than 13,000 ARIC and Cardiovascular Health Study participants with a mean baseline GFR of 90 ml/min/1.73m2 revealed a modestly increased risk for declining GFR below 60 ml/min/1.73m2 for each 1 mg/dL increase in serum uric acid.54 In more than 21,000 healthy Vienna Health Screening Project participants, uric acid levels of 7.0–8.9 mg/dL and ≥9.0 mg/dL were associated with significant increases in risk for incident GFR <60 ml/min/1.73m2 (RR 1.74 and 3.12, respectively).55 In a carefully selected cohort including more than 800 healthy adults aged 20–65 years who were followed for five years, higher baseline serum uric acid levels were associated with GFR declines of >10 ml/min/1.73m2.56 Associations between high uric acid levels and incident CKD have also been demonstrated in populations enriched for diabetes, hypertension, and coronary artery disease (CAD).57 Collectively, these studies raise the possibility that uric acid may be an independent risk factor for subsequent CKD.
In support of these reports, experiments in the laboratory have examined the effects of hyperuricemia on BP, vascular and kidney histology. A rodent model created by inhibiting the uricase enzyme with oxonic acid allows for development of hyperuricemia. On low sodium diets, normouricemic rats maintained normal BPs (or had drops in BP), whereas hyperuricemic rats had increases in BP that correlated with uric acid levels. Rising BPs were attenuated by lowering uric acid concentrations with allopurinol. Furthermore, the rise in BP appeared to be mediated by the renin-angiotensin system.58 Hyperuricemic rats developed pathologic changes in the kidney, with afferent arteriolopathy and mild interstitial injury without uric acid crystal deposition. Administration of allopurinol prevented hyperuricemia, hypertension, and vasculopathy. Enalapril and losartan both prevented the hypertension and arteriolopathy; however, hydrochlorothiazide failed to prevent arteriolopathy despite lowering systemic BP.58, 59 In the rat 5/6 nephrectomy remnant kidney (RK) model of CKD, inducing hyperuricemia led to renal hypertrophy with increased glomerulosclerosis and more severe interstitial fibrosis.60 Relative to RK rats without hyperuricemia, hyperuricemic RK rats developed significantly more severe afferent arteriolopathy with smooth muscle cell proliferation. Treating hyperuricemic RK rats with allopurinol decreased uric acid levels and prevented these renal histologic changes. Mediation of the renal injury via the renin-angiotensin system was demonstrated by increased renin and COX-2 expression.60 The chronic effects of uric acid induced pre-glomerular arteriolar lesions were subsequently tested by inducing hyperuricemia in low-salt diet fed rats over seven weeks. Hyperuricemia was allowed to resolve after two weeks and presence of persistent afferent arteriolopathy was confirmed. RK and control rats were then fed a high salt or a low salt diet. Chronic increases in BP were only observed in previously hyperuricemic rats fed a high salt diet, suggesting that hyperuricemia-induced chronic vascular changes led to salt sensitivity.42
Additional experiments provide evidence that hyperuricemia induces vascular injury. In vitro, uric acid promoted proliferation of rat aorta vascular smooth muscle cells61 and up-regulated intracellular pro-inflammatory pathways; evidenced by increased levels of monocyte chemoattractant protein-1 (MCP-1).62 Low nitric oxide (NO) levels, commonly associated with endothelial dysfunction, are seen in hyperuricemic rats and were reversible upon reducing serum uric acid with allopurinol.63 Hyperuricemia increased the proliferation of human vascular smooth muscle cells (HVSMC) in vitro, which appears to be mediated by increased C-reactive protein (CRP) expression.64 Collectively, this large body of evidence supports the possibility that uric acid directly contributes to the development of hypertension and CKD by producing non-crystal associated renal vascular injury.
A number of underpowered studies attempted to examine the effects of lowering uric acid levels on BP and CKD. In an unblinded pilot study, five hypertensive children aged 14 to 17 years with a mean serum uric acid level of 6.9 mg/dL were treated with allopurinol for one month. The mean uric acid concentration was reduced to 3.3 mg/dL and systolic BP (SBP) fell from 140±3.6 mmHg to 131.2±6.1 mmHg (p=0.017). Four of the five children were normotensive after one month of therapy and all five had rebounds in BP upon discontinuing allopurinol.65 In a follow-up, double blind, placebo controlled crossover trial, thirty adolescents aged 11 to 17 years with newly diagnosed (untreated) hypertension and serum uric acid ≥6 mg/dL were randomized to receive allopurinol 200 mg twice daily or placebo. A significantly larger reduction in BP was seen with allopurinol; SBP fell 6.9 mmHg (95% CI −4.5 to −9.3 mmHg) versus a 2.0 mmHg fall with placebo (95% CI 0.3 to −4.3 mmHg; p=0.009); and DBP fell 5.1 mm Hg with allopurinol (95% CI −2.5 to −7.8 mmHg) versus a 2.4 mmHg fall with placebo (95% CI 0.2 to −4.1 mmHg; p=0.05). Twenty of thirty patients (67%) achieved a normal BP with allopurinol versus only one of thirty (3%) with placebo.66 There was concern that trials attempting BP reduction with allopurinol in adults could be less successful due to irreversible vascular damage and resultant salt sensitivity in chronically hyperuricemic patients, as in hyperuricemic salt sensitive rats.42, 65 This may explain the observation that the association between uric acid and hypertension is less robust in older adults.67, 68
The effect of reducing uric acid concentrations on kidney function was assessed in 54 hyperuricemic adult patients with CKD randomized to allopurinol or usual therapy for twelve months. The primary end-point was the composite of death, dialysis, or >40% increase in baseline serum creatinine concentration. Sixteen percent (4/25) of patients in the allopurinol treatment group and 46% (12/26) of usual therapy controls reached a study end-point (p=0.015).69 Another trial randomized thirty asymptomatic hyperuricemic patients to allopurinol 300 mg/day, thirty seven patients to no treatment, and those results were compared with an additional thirty normouricemic controls. Allopurinol treatment resulted in decreases in serum uric acid and SBP, with increases in flow mediated vasodilation and GFR, relative to a lack of significant change in the other two groups.70 Stevens-Johnson syndrome, a potentially fatal desquamative skin disorder associated with allopurinol must be considered when weighing the impact of these studies. The potential risks and benefits of prescribing allopurinol to large numbers of patients must be included in the decision analysis. However, the limited data available suggests that pharmacological reduction of uric acid has a beneficial effect on blood pressure in children and kidney function in adults.
Considering “hyperuricemia” as a disease state begs the question, what is the underlying cause? In adults, several etiologies exist including impaired kidney function and use of diuretics. However, it is important to consider potential causes of hyperuricemia in otherwise healthy children. Children born with low nephron number may be predisposed to hyperuricemia through increased proximal tubular reabsorption associated with hyperfiltration.65 The human urate transporter 1 regulates proximal tubular urate reabsorption. Gene polymorphisms in this transporter leading to reduced urate excretion have been proposed.71 Potential dietary etiologies include fructose consumption, a source that is receiving increasing attention. Fructose ingestion, through sucrose (50% fructose bound to glucose) and high fructose corn syrup (HFCS) (55% free fructose) has increased markedly in the U.S. over the past thirty years, in part through consumption of sweetened beverages.72 Fructose uptake lacks a negative feedback mechanism and is metabolized primarily in the liver after absorption in the jejunum. Hepatic fructokinase, the initial enzyme in fructose metabolism, has a high affinity for its substrate and will continue to phosphorylate fructose as long as it is available. This results in transient intracellular depletion of adenosine triphosphate (ATP) to adenosine monophosphate (AMP). AMP is then metabolized to uric acid. In contrast, glucose metabolism is controlled by a feedback loop which prevents ATP depletion.72 As such, serum uric acid levels correlate with ingestion of added sugars or sugar sweetened beverages.73, 74 The effect of a low fructose diet was tested in 28 adults with stage 2–3 CKD who were placed on a low fructose (12 g/24 h) diet for 6 weeks followed by resumption of their normal diet. During the low fructose diet, BP improved in the overall group, with significant reduction in SBP and DBP in “dippers” (n=20); there was also a non-significant decrease in serum uric acid from 7.1 to 6.6 mg/dL (p<0.1).75 Epidemiological studies reveal parallel increases in fructose ingestion and metabolic syndrome traits.72 Given these studies and the association of hyperuricemia with obesity, dietary fructose ingestion will likely be the subject of future research.
The epidemiological data associating hyperuricemia with incident hypertension and CKD are supported by laboratory studies providing biologic plausibility for a direct causal relationship. Consumption of excess sugar in childhood likely leads to obesity and also hyperuricemia, which may begin the process of vascular injury in the kidney. We agree with Dr. Johnson and colleagues76 that this process offers a credible explanation for the global rise in the burden of hypertension and CKD seen in recent decades.
Hypertension is a significant risk factor for CVD and cerebrovascular disease, contributing to approximately 47% of ischemic heart disease events and 54% of strokes worldwide.77 Pharmacologic treatment of hypertension in general populations reduces these risks.78 Hypertension clearly contributes to progression of CKD79 and both CKD and ESKD are independent risk factors for CVD.80 Recent studies have sought to define BP targets in both diabetic and non-diabetic kidney disease which safely reduce mortality, CVD risk, and rates of nephropathy progression. While reduction of SBP to <140–150 mmHg had benefit in most CKD populations, the answer was less clear when reducing SBP below 120–130 mmHg. Here, we review hypertension trials that included participants with CKD, along with the current BP target recommendations, and controversies surrounding targeting lower BPs.
Studies investigating BP targets in patients with diabetes have included participants with diabetic nephropathy. The United Kingdom Prospective Diabetes Study (UKPDS) randomized 1148 patients with type 2 diabetes to intensive BP control (goal BP <150/<85 mmHg) or usual control (goal <180/<105 mmHg). Approximately 17% of the sample had albuminuria. Baseline mean BP was 159/94 mmHg and achieved BPs were 144/82 mmHg (intensive) and 154/87 mmHg (usual). Results were dramatic with reductions in microvascular diabetes complications including nephropathy (12% reduction per 10 mmHg decrease in SBP), deaths related to diabetes (17% reduction per 10 mmHg decrease in SBP) and all-cause mortality (12% reduction per 10 mmHg decrease in SBP) down to a SBP <120 mmHg.81 The Appropriate Blood Pressure Control in NIDDM - Hypertensive (ABCD) trial randomized 480 patients with hypertension to intensive BP control (goal DBP <75 mmHg) or moderate control (goal DBP <80–89 mmHg) using either nisoldipine or enalapril. Patients with diabetic nephropathy were included and 18.5% had albuminuria >200 ug/min. Initial mean BP was 156/98 mmHg and achieved BPs were 132/78 mmHg (intensive) and 138/86 mmHg (moderate).82 All-cause mortality was lower in the intensive treatment group than the moderate treatment group (5.5% vs. 10.7%, p=0.037). Interestingly, progression of albuminuria was not significantly different between groups.83 The Irbesartan Diabetic Nephropathy Trial (IDNT) included only diabetic patients with nephropathy and randomized 1715 patients to treatment with irbesartan, amlodipine, or placebo to a goal BP <135/<85 mmHg. Initial mean BP was 159/87 mmHg.84 The study found a favorable effect of irbersartan on the composite endpoint of death, doubling of serum creatinine or ESKD. A post-hoc analysis of achieved BPs demonstrated that a SBP down to 120 mmHg was associated with decreased CVD mortality and congestive heart failure. However, an achieved SBP <120 mmHg was associated with an increased risk of CVD death and congestive heart failure. Achieved DBPs below 85 mmHg were associated with an increased risk of all-cause mortality and myocardial infarction, but decreased risk for stroke.85 These findings suggested a J-shaped relationship in achieved BP control (discussed below). The Action to Control Cardiovascular Risk in Diabetes (ACCORD BP) trial sought to clarify whether lowering SBP <135–140 mmHg was beneficial in 4733 patients with diabetes. ACCORD participants included approximately 38.5% with albuminuria, but excluded patients with serum creatinine >1.5 mg/dL. Participants were randomized to intensive goal SBP <120 mmHg or standard goal SBP <140 mmHg. Starting mean BP was 139.2/76.0 mmHg and achieved BPs were 119.3/64.4 mmHg (intensive) and 133.5/70.5 mmHg (standard). There was no difference in the primary composite outcome of nonfatal MI, nonfatal stroke, or CVD death between groups. There was also no difference in progression to ESKD between groups, but development of macroalbuminuria was reduced in the intensive treatment group.86 Among secondary outcomes, there was a significant reduction in stroke with a higher rate of adverse events in the intensive BP treatment group.
Three major studies have examined hypertension treatment goals in non-diabetic kidney disease. The Modification of Diet in Renal Disease (MDRD) studied two groups of participants, moderate CKD (GFR 25–55 ml/min/1.73m2; n=585) and severe CKD (GFR 13–24 ml/min/1.73m2; n=255). Groups were randomized to two different levels of dietary protein intake and subsequently to a low BP target (mean arterial pressure [MAP] <93 mmHg) or a usual BP target (MAP <107 mmHg). The primary outcome was rate of GFR decline. Baseline mean BPs were 131/81 mmHg and achieved MAPs were 90 mmHg (~126/77 mmHg) and 94 mmHg (~133/80 mmHg), respectively. After three years there was no difference in the primary outcome between usual and low BP targets in either CKD group.87, 88 A post-hoc analysis found that patients in both groups with baseline proteinuria >1 gm/day had a significantly reduced rate of GFR decline when treated to the lower BP target. This was most pronounced in participants with >3 gm/day of proteinuria.89 An analysis of MDRD participants from 1993–2000 after the trial ended revealed reduced nephropathy progression in the low BP target group (Hazard Ratio [HR] for kidney failure 0.68, p=0.001; HR for composite kidney failure or all-cause mortality 0.77, p=0.0024).90 We previously discussed the AASK Trial, where 1094 African American patients with hypertension attributed non-diabetic CKD were randomized to intensive (MAP <92 mmHg) or standard (MAP 102–107 mmHg) BP control. Baseline BPs were 152/96 mmHg and 149/95 mmHg, respectively; and achieved BPs were 128/78 mmHg (intensive) and 141/85 mmHg (standard). There was no difference in the composite outcome of doubling of serum creatinine, ESKD, or death in the entire cohort after 10 year follow-up. However, when stratifying participants by baseline proteinuria, a reduction in composite risk was observed for those with a urine protein:creatinine ratio >0.22 g/g compared to <0.22 g/g (HR 0.73, p=0.01).9 In the Blood-Pressure Control for Renoprotection in Patients with Non-Diabetic Chronic Renal Disease (REIN-2) trial, 338 participants with proteinuria were randomized to intensified (<130/80 mmHg) or conventional (DBP <90 mmHg) BP control. All participants were treated with ramipril and other anti-hypertensive medications were adjusted to achieve goal BPs. Baseline BPs were 137.0/84.3 mmHg and 136.4/83.9 mmHg, respectively; and achieved blood pressures were 129.6/79.5 mmHg (intensive) and 133.7/82.3 mmHg (conventional). The study was terminated early, after the first interim analysis revealed no differences in the cumulative incidence of ESKD, rate of GFR decline, or residual proteinuria between the study arms.91
The aforementioned trials demonstrate the controversy as to whether lowering SBP <130 mmHg with pharmacologic treatments leads to a favorable risk/benefit ratio in patients with CKD. In the latter three studies, achieving lower BP targets was associated with a slightly higher rate of adverse events but appeared to confer benefit in the subset of patients with proteinuria.88 Current BP guidelines vary with regard to the treatment of patients with CKD based on the presence and amount of proteinuria. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7), the National Kidney Foundation Disease Outcomes Quality Initiative (NKF K/DOQI), and the Canadian Society of Nephrology (CSN) guidelines for hypertension currently recommend a goal BP <130/<80 mmHg in patients with CKD with and without proteinuria.92–94 The Caring for Australasians with Renal Impairment (CARI) guidelines recommend BP <130/<80 mmHg in patients with CKD, including those with proteinuria <1 gm/day and <125/<75 mmHg for those with >1 gm/day (www.cari.org.au). In 2011, the United Kingdom National Institute for Health and Clinical Excellence (NICE) recommended a target BP <140/<90 mmHg for all patients below 80 years of age and a target <150/<90 mmHg for all patients >80 years old (www.nice.org.uk).
Whether to target a lower BP in populations with CKD remains important. Some studies have suggested that excessive lowering of the DBP beyond a certain threshold may be detrimental. The existence of a J-shaped curve for risk in BP control remains controversial; some studies observe it and others do not. A prospective cohort study of 7830 National Health and Nutrition Examination Survey II (NHANES II) participants without prior CVD examined associations between baseline SBP, DBP, and pulse pressures and mortality. In patients >65 years (n=2055) a linear relationship was seen between SBP and mortality. However, a J-shaped relationship was seen for DBP, with an increase in risk below 80 mmHg.95 A post-hoc analysis of IDNT limited to patients with diabetic kidney disease found an increased risk of all-cause mortality for achieved DBP <85 mmHg.85 The International Verapamil-Trandolapril Study (INVEST) randomized more than 22,000 hypertensive patients with CAD to treatment targets <140/<90 mmHg or <130/<85 mmHg.96 A secondary analysis found a J-shaped relationship between mean treated DBP and all-cause mortality, with increased risk for DBP <~80 mmHg. However, this relationship was attenuated in a multivariate model including history of MI, baseline DBP, and heart failure.97
In the Hypertension Optimal Treatment Trial (HOT), more than 18,000 patients with hypertension and DBPs 100–115 mmHg were randomized to treatment goals of DBP ≤90, ≤85, and ≤80 mmHg. Achieved DBPs were 85.2, 83.2, and 81.1 mmHg, respectively. In the goal ≤80 mmHg group, there was a non-statistically significant increase in risk for myocardial infarction (RR 1.37, 95% CI 0.99–1.91, p=0.05), but not for all-cause mortality or CVD mortality. In the diabetic subgroup with goal DBP ≤80 mmHg, there was an increased risk for CVD (RR 3.0, CI 1.8–7.08, p=0.016) and a trend toward all-cause mortality, (RR 1.77, CI 0.98–3.21, p=0.068).98 The MRFIT included 5362 men who had suffered a previous myocardial infarction. In this cohort, a J-shaped relationship for DBP was observed within the first 2 years. During that interval, those with DBPs <70 mmHg had the highest risk of death, compared to 70–79, 80–89, 90–99, and ≥100 mm Hg. After the initial two years, the risk of CVD death was lowest in the DBP 70–79 and 80–89 groups, similar in <70 and 90–99 groups, and highest in the ≥100 mmHg group.99 In the Systolic Hypertension in the Elderly study (SHEP), 4736 patients over the age of 60 years with isolated systolic hypertension were randomized to treatment or placebo with the goal of reducing the SBP <160 mmHg or by 20 mmHg, whichever was greater. The baseline DBP was 76.6 mmHg. Despite a mean reduction in DBP of 9 mmHg in the treatment group and 4–5 mmHg in the placebo group, no J-shaped association was observed for mortality.100 Except for IDNT, the above studies were not specifically designed to address CKD populations. However, the possibility of a J-shaped relationship between BP and mortality must be considered in CKD patients because of current recommendations for lower BP targets in this population.
To answer questions regarding the benefits and safety of targeting lower BPs in “at risk” populations, the National Institute of Health recently launched a large clinical study, Systolic Blood Pressure Intervention Trial (SPRINT). This trial plans to enroll 9250 hypertensive adults over the age of 50 years with a history of CVD, high risk for heart disease based on at least one additional CVD risk factor (smoking or high blood cholesterol levels), or CKD. The goal is for 45% of the study population to have CKD. Participants will be randomized to a SBP goal <140 mmHg or <120 mmHg and monitored for the endpoints of CVD events and mortality (www.nih.gov). This important study is expected to provide evidence for the optimal hypertension treatment goal in patients with stage 3–4 CKD.
We conclude that current data support lowering the SBP to below 140 mmHg in patients with CKD and between 120–130 mmHg in those with diabetic kidney disease or overt albuminuria (≥300 mg/day). We propose monitoring DBP closely when lowering SBP, in order to prevent marked drops. Avoiding large reductions in DBP may prevent achievement of SBP targets, but are likely to create more favorable risk/benefit ratios.
In summary, the complex relationship between hypertension and CKD continues to produce controversies in clinical care. The three topics that we have addressed promise to be the subject of future research endeavors. Solutions to these clinical issues will optimize the outcomes in patients who have high blood pressure and CKD.