Introduction

Hypertension is a chronic disease that is increasing in prevalence worldwide and spans race and gender lines. The worldwide prevalence of hypertension was estimated from a 2005 systematic review of published literature to be 26%with a projected 60% increase by 2025.1 In countries with established market economies, the prevalence of hypertension ranges between 22% and 55%.2 An analysis of data from the 19992004 National Health and Nutrition Examination Survey (NHANES) found that the prevalence of hypertension in the USA increased over these 5 years; prevalence rates were approximately 27% in 1999 and 29% in 2004.3 The 20042007 NHANES data revealed a high prevalence of hypertension in adults aged 55 years; hypertension affected 47% of individuals in this age group and increased to 56% in those aged 7584 years.4 NHANES data also show that the prevalence of chronic kidney disease (CKD) in the USA is increasing. The prevalence of stage 34 CKD (estimated glomerular filtration rate [eGFR] 1560 ml/min/1.73 m2) increased from ~5.5% in the 19881994 survey to >8% in the 19992004 survey.3 The prevalence of end-stage renal disease (ESRD) has also steadily increased, increasing by 18% between 2000 and 2007. In 2010, ESRD was estimated to affect 1,699 per 1 million US individuals.5 The worldwide prevalence of ESRD has mirrored the findings in the USA. In the UK, the incidence of ESRD continues to rise at an approximate rate of 58% per year.6 Correspondingly, health-care costs associated with ESRD have steadily increased. According to the US Renal Data System (USRDS) 2010 report, Medicare expenditures for ESRD in 2008 were US$26.8 billion in the USA, comprising 5.9% of the total Medicare budget.5 This estimation of costs, however, only accounts for care related to ESRD and does not account for all cardiovascular care for these individuals. As CKD is an independent risk factor for cardiovascular morbidity and mortality, contributing additional disease burden to the population worldwide, the inclusion of these additional costs would substantially increase the above figures. Uncontrolled hypertension is associated with increased risk of cardiovascular morbidity and mortality. The risk of cardiovascular mortality doubles for each 20 mmHg increase in systolic blood pressure and 10 mmHg increase in diastolic blood pressure above 'normal' blood pressure levels of 115/75 mmHg.7 Poorly controlled blood pressure in individuals with all forms of CKD is associated with disease progression and cardiovascular morbidity and mortality. Perry et al.8 demonstrated the intimate connection between blood pressure, cardiovascular disease risk and risk of CKD in their 15-year follow-up study of approximately 12,000 hypertensive individuals at multiple Veterans Affairs centers in the USA (Figure 1). Poorly controlled blood pressure was a clear risk factor for the development of CKD and cardiovascular complications and lowering of blood pressure was associated with a reduced risk of developing CKD. Specifically, the risk ratio for CKD was 2.8 for a pretreatment systolic blood pressure of 166180 mmHg and 7.6 for pretreatment blood pressure >180 mmHg. A decrease in systolic blood pressure of >2 mmHg after treatment was associated with a marked decrease in the relative risk of developing ESRD. The relative risk of developing ESRD was 0.65 for a decrease in systolic blood pressure level 215 mmHg, 0.56 for a decrease of 1520 mmHg and 0.39 for a decrease >20 mmHg.8 Figure 1 | The relationship between pretreatment systolic blood pressure level and the subsequent occurrence of ESRD.

KaplanMeier estimates of rates of ESRD over 15 years of follow-up in 12,000 individuals with hypertension from multiple Veterans Affairs centers in the USA. SBP >180 mmHg was associated with an increased risk of developing ESRD. Permission obtained from Wolters Kluwer Health Perry, H. M. et al. 25, 587594 (1995). Abbreviations: ESRD, end-stage renal disease; SBP systolic blood pressure.

y y y

Full size figure and legend (30 KB) Figures and tables index Download high-resolution PowerPoint slide (102 KB)

Even modest elevations in blood pressure seem to confer an increased risk of ESRD. The Multiple Risk Factor Intervention Trial (MRFIT), which investigated the development of various cardiovascular disease manifestations in a cohort of 12,000 men, examined the effect of blood pressure on the development of ESRD.9 Despite individuals with known baseline CKD (defined as a serum creatinine level >177 mol/l) being excluded from the trial, the investigators found that elevation of either systolic or diastolic blood pressure one standard deviation above the range in the lowest group was associated with a 1.7-fold increased risk of developing ESRD.9 Although this study was limited by inclusion of only men, a similar study performed in a cohort from Kaiser Permanente, CA, USA, included both men and women and reported similar results.10 Specifically, the risk of developing ESRD increased with each hypertension stage (as classified by the Fifth Joint National Committee Report on the Detection, Evaluation and Treatment of Hypertension).10 As in the MRFIT study, the investigators excluded individuals with baseline CKD (as defined by eGFR <60 ml/min/1.73 m2). Although hypertension was associated with an increased risk of developing ESRD, the overall rates of developing ESRD were low; approximately 15 cases per 100,000 person-years in the MRFIT study and 14 cases per 100,000 person-years in the Kaiser Permanente cohort.9, 10Nevertheless, these data highlight the association of hypertension with the development and progression of CKD. The purpose of this Review is to explore the epidemiology of hypertension associated with CKD. Although a complete discussion of the pathogenesis of hypertensive kidney disease is beyond the scope of this Review, we will briefly describe the pathogenic mechanisms of hypertension as a cause and consequence of CKD. Furthermore, we will outline the epidemiology of hypertensive CKD and ESRD owing to hypertension, explore important factors in the diagnosis of the disease and outline strategies to prevent its development and progression.
topof page

Pathogenesis
Despite the widespread prevalence of hypertension and associated CKD, the pathogenesis of hypertensive nephrosclerosis remains unclear. Proposed pathogenic mechanisms include two, distinct initial injuries that lead to progressive injury. One potential pathogenic mechanism reflects the changes in extrarenal and renal vasculature associated with systemic hypertension. The breakdown of elastic fibers in the arterial circulation, replaced by progressive intimal thickening, is associated with normal aging and is exaggerated in the setting

of systemic hypertension. The downstream afferent arterioles lose their autoregulatory capacity owing to hyalinosis and dilatation. Ultimately, the glomeruli suffer from the diseased state of the vasculature, lose their ability to autoregulate and are therefore unable to attenuate hyperfiltration-mediated injury.11Moreover, in many cases, patients are treated with dihydropyridine calcium antagonists even in early stages of nephropathy, which results in a loss of renal autoregulation.12Loss of renal autoregulation alone does not mediate all cases of hypertensive nephrosclerosis. The presence of ischemic, rather than hypertrophic, glomeruli suggests an alternative process in which hypoxia and ischemia are the predominant mechanisms. The precise precipitant for hypoxia is unclear; however, hypertension-mediated endothelial injury is likely to be involved. Endothelial damage and the loss of endogenous vasodilators, such as nitric oxide, contributes to tissue hypoxia and, ultimately, to ischemic injury. The initiating factors that precipitate these changes within the kidney remain unclear. What is known is that loss of autoregulation and development of ischemia will eventually lead to renal parenchymal damage with all the associated manifestations of CKD, including progressive azotemia, abnormalities in bone and mineral metabolism, and impaired control of extracellular fluid volume. The 2009 discovery of an association between MYH9 polymorphisms and kidney disease13 added insight into the underlying etiology of nondiabetic CKD and, specifically, hypertensive nephrosclerosis. In an evaluation of 871 individuals with a variety of underlying nondiabetic causes of ESRD (696 of whom had ESRD attributable to hypertension), significant associations were found in 14 of 15 MYH9 polymorphisms in all 871 individuals. Specifically, presence of the E1 haplotype was significantly more common in individuals with hypertension-associated ESRD than in controls without kidney disease (odds ratio of 2.23; P = 4.52 1012). The study also identified alternative haplotypes that were significantly associated with an increased or decreased risk of hypertension-associated ESRD. The associations remained significant even after adjusting for age and gender.13 The product of MYH9, myosin-9, is a mechanoenzyme that is localized to the podocyte foot processes (which are responsible for moving actin filaments inside cells). The connection between myosin-9 and the development of kidney disease has not been definitively established. However, the current theory purports that MYH9polymorphisms result in products that disrupt normal podocyte function, causing podocyte injury, which ultimately leads to glomerulosclerosis. Importantly, not all individuals with the gene polymorphisms in the abovementioned study developed kidney disease.13 The strong association between MYH9 polymorphisms and ESRD, especially in individuals with ESRD attributed to hypertension, raised questions about the true association of hypertension with kidney disease. Although this discovery does not translate into the demise of hypertensive nephrosclerosis as a diagnosis, it does better explain the presence of severe hypertension and kidney disease and MYH9 polymorphisms could be a marker for risk of hypertension, especially in young, African Americans, in whom such polymorphisms are prevalent. The presence of genetic polymorphisms that lead to intrinsic kidney injury suggest that the failure of strict blood-pressure control to prevent CKD progression does not relate to blood pressure alone, but rather relates to factors intrinsic to the kidney that are yet to be defined. This hypothesis might partially explain why blood-pressure control alone cannot reverse existing kidney damage or stop progressive kidney damage, as was observed in the African American Study of Kidney Disease (AASK).14 Presence of stage 3 CKD or higher seems to independently aggravate hypertension. CKD is associated with impaired function of endogenous vasodilators, increased sympathetic activity and increased salt retention.

The ability of the normal kidney to handle variations in dietary sodium intake and sodium excretion is governed by the pressurenatriuresis curve: increased sodium intake is associated with increased blood pressure, increased glomerular filtration rate (GFR) and, ultimately, increased sodium excretion to match the increased intake of sodium. However, in the setting of advanced parenchymal kidney damage, the ability to achieve complete sodium balance is blunted by the decrease in GFR. Thus, the pressurenatriuresis curve shifts to the left and sodium balance is only achieved at the expense of an increase in blood pressure.15 Unfortunately, without adequate control of salt intake and appropriate antihypertensive therapy, the vicious cycle of pressure natriuresis continues and progressive renal injury and hypertension occur (Figure 2). The clinical picture is one of salt-sensitive hypertension, impaired normal variations in blood pressure (that is, the blunted or absent nocturnal 'dipping' of blood pressure) and systemic vascular disease that renders the control of blood pressure especially difficult and raises the risk of cardiovascular disease.11 Figure 2 | Mechanisms that contribute to salt sensitivity and hypertensive kidney injury.

The pressurenatriuresis curve governs how the kidney handles variations in dietary sodium intake and sodium excretion so that increased sodium intake is ultimately matched by increased sodium excretion. In patients with kidney disease, the balance of sodium intake and excretion is altered by the decrease in GFR such that sodium balance is only achieved with an increase in blood pressure. Without adequate control of salt intake or antihypertensive therapy, continued renal injury and hypertension can occur. Abbreviations: GFR, glomerular filtration rate; RAS, reninangiotensin system.

y y y

Full size figure and legend (54 KB) Figures and tables index Download high-resolution PowerPoint slide (125 KB)

topof page

Epidemiology
ESRD owing to hypertension Estimating the prevalence of CKD attributable to hypertension alone is difficult. Individuals with presumed hypertensive nephrosclerosis are rarely biopsied, making the ability to determine a definitive diagnosis and precisely estimate prevalence challenging. Adherence to diagnostic criteria for hypertensive nephrosclerosis does not always occur in clinical practice. Thus, the best estimates of hypertension-associated CKD can only be made from the examination of large-scale databases such as the USRDS, NHANES and the Kidney Early Evaluation Program (KEEP). Hypertension remains the second most common primary diagnosis for patients with prevalent ESRD, accounting for ~125,000 of individuals who receive dialysis in the USA. Hypertension also remains the second leading cause of incident ESRD, with approximately one-third of patients on incident dialysis having hypertension reported as their primary diagnosisapproximately 30,000 individuals in the USA are diagnosed with hypertension-associated ESRD yearlyand there continues to be a steady increase in the number of patients with incident ESRD as a primary diagnosis. The rate of ESRD in the USA attributable to

hypertension increased by 8% from 2000 to 2007. As the authors of the USRDS annual 2010 data report point out, this finding does not necessarily reflect a true increase in percentage cases of hypertensive nephropathy that lead to ESRD.5Rather, this finding might simply reflect the continued increase in the prevalence of hypertension and the absence of an identified alternative diagnosis for hypertensive nephropathy. In addition, this finding could also reflect the aging population of patients with incident ESRD, in whom hypertension is also increasingly common.5 The limitation of using USRDS data to investigate the epidemiology of ESRD secondary to hypertension was highlighted by Zarif et al.,16 who assessed the accuracy of data recorded by the US Health Care Financing Administration (HCFA) by comparing HCFA-listed cases of hypertensive nephrosclerosis against the prevalence of 'true' hypertensive nephrosclerosis according to two sets of strict clinical criteria. The first, established by Schlessinger and colleagues,17 included family history of hypertension in a first-degree relative, left ventricular hypertrophy (identified on an electrocardiogram or echocardiogram), proteinuria of <500 mg per 24 h or a score of <2+ on urinary dipstick and hypertension preceding any evidence of renal dysfunction, any congenital or intrinsic renal disease or systemic illness associated with renal damage or exposure to nephrotoxins. The second criteria were based on criteria used by the AASK trial investigators18 and included age 1870 years, urine protein levels <2.5 g per day and no evidence of immune complex disease or diabetes mellitus. Of a random sample of 100 individuals identified as having hypertensive nephrosclerosis according to the HCFA, only four individuals satisfied the Schlessinger criteria for hypertensive nephrosclerosis and only 28 of 91 African-American individuals satisfied the AASK criteria for hypertensive nephrosclerosis. However, the predominant reason as to why 96 of 100 individuals did not satisfy the Schlessinger criteria was a lack of available information, specifically information on the presence of hypertension before evidence of renal dysfunction. In 14 individuals, alternative diagnoses were identifiedvarying from IgA nephropathy to myeloma kidney.16 In a similar study, Zucchelli and Zuccala19 reviewed patients with ESRD who were diagnosed with hypertensive nephrosclerosis according to clinical criteria and underwent renal biopsy. Of 56 patients examined, only 26 patients had nephrosclerosis as confirmed by biopsy. The second most common diagnosis was atheromatous renovascular disease, which affected 19 of the patients. The remainder of individuals had various diseases ranging from analgesic nephropathy to immunotactoid glomerulonephritis.19 The above studies highlight the challenges in estimating the precise prevalence of ESRD secondary to hypertension alone. Individuals with presumed hypertensive nephrosclerosis are rarely biopsied and strict clinical criteria, while established, are not routinely followed to make a diagnosis. The reported incidence and prevalence of ESRD owing to hypertension is, therefore, likely an overestimate because of alternative diseases that are undiagnosed or overlooked. Use of standardized criteriaclinical and, when available, histologicalwould potentially improve the estimates of ESRD related to hypertensive nephrosclerosis. The level of blood-pressure control that slows the progression and time course for the development of CKD remains unclear. The best available data arise from studies of blood-pressure intervention and its effect on CKD progression. The AASK study20followed ~1,100 individuals with baseline CKD (eGFR 2060 ml/min/1.73 m2) and evaluated the effect of 'strict' blood-pressure lowering on disease progression. The intervention group was treated to target mean arterial pressure (MAP) levels of <92 mmHg whereas the control group was treated to a target MAP of 102107 mmHg. Importantly, individuals with substantial

proteinuria (>2.5 g per day), diabetes or an alternative cause of CKD were excluded. The investigators found no difference in CKD progression between the two groups during the study period20 (up to ~6 years) or in the longer-term cohort study (up to 10 years).14 The event rate for the development of the combined end point of doubling of serum creatinine level or the development of ESRD was 6.0 per 1,000 person-years of follow-up. The event rate for development of ESRD alone was approximately 0.85 per 1,000 person-years of follow-up.14, 20 The AASK population, which excluded individuals with marked proteinuria, diabetes or alternative etiologies of renal dysfunction, might provide the best representation of the natural history of hypertensive nephrosclerosis. Level of blood pressure seems to contribute to CKD progression. However, lowering blood pressure to a MAP <92 mmHg (approximately 120/80 mmHg) failed to demonstrate noteworthy slowing of CKD progression. AASK and other randomized studies of patients with normoalbuminuria or microalbuminuria demonstrate that lowering blood pressure to <130/80 mmHg has no benefit.21 By contrast, long-term follow-up studies of patients with advanced stage 3 nephropathy and macroalbuminuria (that is, >300 mg per day) clearly show an advantage of lowering blood pressure to <130/80 mmHg for slowing the progression of nephropathy.22 Hypertension in predialysis CKD The epidemiology of hypertension associated with predialysis CKD depends largely on the definition of CKD used. The Kidney Dialysis Outcomes Quality Initiative (KDOQI) classification provides a step-wise progressive definition of CKD ranging from albuminuria with preserved GFR (stage 1) to ESRD with GFR <15 ml/min/1.73 m2 (stage 5). Whether or not early-stage kidney disease (for instance, stages 1 and 2) are included in a particular study will vastly affect the overall prevalence estimates of CKD and hypertensionassociated CKD. The intimate connection between the definition of CKD and its effect on estimation of prevalence is highlighted by the evaluation of the NHANES population by Crews and colleagues.23 In an evaluation of 17,794 patients, the investigators estimated the overall prevalence of CKD to be 20%. Furthermore, the prevalence of CKD in individuals with normal blood pressure was 13.4%. Despite the use of modified KDOQI criteria for the classification of CKD, these findings are alarming because the prevalence of CKD was found to be much higher than previously thought. However, as highlighted by Agarwal in an accompanying commentary, prevalence estimates would be markedly different if the researchers had included only individuals with stage 3 CKD or greater in their analysis.24 The presence of a single measurement of microalbuminuria does not confirm the presence of glomerular disease or parenchymal kidney injury. Estimates of CKD are probably overestimated even by the measurement of persistent microalbuminuria.24 A 2010 debate questioned the relevance of microalbuminuria measurements in the diagnosis of kidney disease and controversy regarding its use as a marker of CKD still exists.25, 26, 27 The results of the Avoiding Cardiovascular Events through Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) study highlights the limitations of using microalbuminuria as a marker of kidney disease and predictor of CKD progression. The randomized controlled trial comparing two antihypertensive regimensbenazepril and hydrochlorothiazide versus benazepril and amlodipine found that although the reduction of microalbuminuria was greater in the group receiving benazepril and hydrochlorothiazide than in the group receiving benazepril and amlodipine; a doubling of serum creatinine level and progression to ESRD in those aged >65 years was less frequent in the latter group.28

The findings of Crews et al.23 mentioned earlier also highlight the limitations of using criteria that have been established for the classification of disease severity for establishing a diagnosis. Early (stage 1 and stage 2) CKD is included in the KDOQI classification to incorporate individuals with obvious renal parenchymal damage (for example, proliferative lupus nephritis) but preserved GFR. Without further clinical information, it is impossible to identify which individuals classified as having stage 1 CKD in the NHANES database have true parenchymal injury and should actually be included in the analysis of CKD prevalence. Thus, use of the KDOQI classification of CKD and, specifically, use of a single measurement of kidney function such as microalbuminuria, might not be the best method for identifying CKD prevalence in the general population. Estimating the prevalence of hypertension associated with predialysis CKD is difficult in that no mandatory reporting of etiology is required for patients with predialysis CKD. In the USA, physicians caring for patients with ESRD are required to identify an etiology for the disease, as denoted on HCFA medical evidence forms. However, physicians are not required to identify an etiology for individuals with predialysis CKD. Accurately determining the prevalence and incidence of predialysis CKD attributable to hypertension is, therefore, particularly difficult. To date, the available literature exploring the epidemiology of CKD overall has not classified individuals according to the underlying etiology. Rather, the available population cohort data gathered via NHANES and the National Kidney Foundation's KEEP have provided information on comorbid conditions associated with CKD, including hypertension as opposed to providing accurate estimates of CKD resulting from hypertension; as a result, only estimates of CKD associated with hypertension can be provided. The 2010 USRDS report summarizes the NHANES data on CKD and comorbid conditions.5In comparison to individuals without CKD, individuals with any stage of CKD (stages 15) have higher rates of hypertension. The 19992006 NHANES data noted poor blood-pressure control in those with and without CKD. The proportion of individuals with uncontrolled blood pressure were greater in those with CKD than in those without CKD, 68.8% and 51.7%, respectively (according to disease-specific definitions).29 With each progressive stage of CKD, the rates of hypertension increased (approximately 36% in individuals with stage 1 CKD, increasing to 84% of individuals with stage 4 or 5 CKD). At stage 3 CKD and above, individuals were more likely to have hypertension than not. The propensity for individuals with stage 3 and more severe CKD to have prevalence estimates of hypertension >50% held true for all demographic subgroups analyzed individuals aged >60 years, males and females, non-Hispanic white individuals, African Americans and other races. Individuals aged >60 years (irrespective of race), non-Hispanic white individuals, and African Americans seem to be particularly affected, with the prevalence of hypertension in individuals with stage 2 CKD being >50%, 58% and 63% in these groups, respectively.5 The strong association of hypertension with CKD is confirmed by data from KEEP.30However, the KEEP data suggest that hypertension is more prevalent at early stages of CKD and that individuals with CKD have even higher rates of hypertension than NHANES data suggest. The KEEP investigators found that the overall prevalence of hypertension exceeded 50% in all individuals with CKD, regardless of eGFR, and in those with 'normal' renal function (eGFR >100 ml/min/1.73 m2) the prevalence of hypertension was ~57%.30 The two databases, however, represent different methods of sampling individuals and, overall, have varied demographics. The KEEP population is targeted and includes more African Americans than the NHANES data, making the finding of higher rates of hypertension in early-stage CKD more difficult to interpret and,

perhaps, overestimating the prevalence for the general population.30 Nevertheless, these two sets of largescale population data highlight the strong association between CKD and hypertension and the vigilance necessary on the part of clinicians to recognize the association and guide therapy accordingly. Although both NHANES and KEEP data provide estimates of the prevalence of CKD with concomitant hypertension, the available data remain limited without information on the underlying etiology of CKD (for example, diabetic nephropathy or hypertensive nephrosclerosis). Furthermore, data on self-reported or physician-reported etiologies of CKD are limited by the same absence of established criteria as the definition of CKD itself. As mentioned earlier, the study by Zarif et al.16 used two, different criteria to establish a diagnosis. Although the two criteria share the common feature of not including individuals with diabetes or those with alternative underlying diseases associated with kidney injury, the acceptable ranges of proteinuria are markedly disparate (urine protein <500 mg per day in the Schlessinger criteria versus <2.5 g per day in the AASK study criteria). Estimates of disease prevalence could therefore change substantially depending on the criteria used by caregivers to identify hypertensive nephrosclerosis. An estimation of the prevalence of predialysis hypertensive CKD is not currently available, but the epidemiology of this entity is an area for future investigation.
topof page

Awareness and control of hypertension

Despite widespread awareness of the complications associated with hypertension, especially in those with CKD, the rates of hypertension, the rates of ESRD attributable to hypertension and the rates of CKD continue to rise in the USA and worldwide.29Furthermore, cardiovascular outcomes of patients with CKD, often related to hypertension, continue to be poor. Although control of hypertension requires physician input, adherence to prescribed medications and dietary modifications ultimately relies on patient effort. Awareness of the disease process is necessary to optimize adherence to therapy and patient outcomes.31 Despite widespread awareness, the NHANES 19992006 data suggest that awareness of hypertension remains suboptimal.29 Among individuals with hypertension, more than one-quarter of patients both with and without CKD were unaware of their disease. Approximately 31% of individuals without CKD were unaware of their diagnosis of hypertension. Individuals with both early and late-stage CKD were almost equally unaware of their diagnosis; 36% of individuals with stage 1 or stage 2 CKD and 26% of individuals with stage 3 CKD or higher did not know that they had hypertension.5 As mentioned earlier, the KEEP database provides a targeted view of individuals with CKD and comorbid hypertension. Awareness rates of hypertension in the KEEP population were better than in the NHANES population. Overall, 80% of individuals with CKD were aware of their diagnosis of hypertension.32 Rates of awareness of hypertension increased with advancing stages of CKD. In individuals with stage 1 CKD, awareness rates were ~70% and in those with more advanced CKD (stages 4 or 5) awareness rates were >90%.32 The Chronic Renal Insufficiency Cohort (CRIC) study represented another targeted study population for the assessment of comorbid conditions associated with CKD. Designed as a prospective cohort to study CKD risk progression and cardiovascular disease outcomes, the CRIC study enrolled individuals from seven clinical centers in the USA and represents a formalized assessment of individuals with CKD. Accordingly, the rates of awareness of hypertension in the CRIC cohort are even higher than in the KEEP population, with >98% of individuals with hypertension reporting disease awareness.33Importantly, the CRIC

cohort represents a self-selected population rather than a randomly sampled population (as in the NHANES and KEEP cohorts) as participants of the CRIC cohort actively sought study enrollment. Thus, it is not surprising that the rates of disease awareness are better than in individuals in the NHANES and KEEP cohorts. Nevertheless, the increased rates of awareness with advancing stages of CKD as well as in more targeted cohorts represent opportunities for increased rates of awareness in all populations. Unfortunately, data from these varied surveys suggest that in the general population, awareness of hypertension is poor. Only after CKD develops or becomes more advanced does awareness improve. The CRIC cohort represents the study population with the best awareness of hypertension; however, the largest group of individuals within this cohort have stage 3 CKD or higher. Although targeting individuals with advanced CKD is important, the failure to generate increased awareness of hypertension among patients with early stage CKD represents missed opportunities for education and treatment. The reason for these discrepancies in awareness between study populations remains unclear. Increased input from specialty physicians, such as nephrologists, or those seen at academic medical centers might have a role, but this approach remains to be elucidated. The increased awareness of hypertension in targeted populations and in those with advanced stage CKD has not, unfortunately, translated to improved blood-pressure control. In the NHANES population, rates of blood-pressure control followed a pattern opposite to that of hypertension, with increased rates of uncontrolled hypertension associated with advancing stages of CKD. Despite treatment, 58.6% of individuals without CKD had uncontrolled hypertension compared with 69.2% of individuals with stage 3 or 4 CKD.29 Data from KEEP illustrate similar findings with regard to overall blood-pressure control. Despite an increased awareness of hypertension, target blood-pressure control (<130/80 mmHg) remained poor and was only achieved by 11% of individuals.32 In contrast to NHANES data, Sarafidis and colleagues32 found that advanced CKD was associated with improved blood-pressure control. Rates of blood-pressure control in patients with stage 4 or 5 and stage 3 CKD were 21% and 20%, respectively, compared with 13% and 11% for patients with stages 2 and 1 CKD, respectively.32 The CRIC study data also demonstrated that a targeted (and self-selected) population seems to fare better at translating hypertension awareness to control of blood pressure than a randomly selected population.33 Rates of blood-pressure control remained suboptimal, but were improved compared with the NHANES and KEEP populations. 67% of individuals achieved a blood pressure target <140/90 mmHg, and 46% of individuals achieved a target <130/80 mmHg. Importantly, these findings represent baseline characteristics of the cohort. No study-based interventions were performed. Given that the NHANES and KEEP databases represent randomly sampled cohorts, these data best represent the general population of individuals with CKD while the CRIC population represents individuals who were aware of their CKD status and were referred for participation in a study. Nevertheless, data from the CRIC population suggest that motivated individuals with appropriate medical follow-up can achieve improved rates of adherence and successfully reach blood-pressure goals in the setting of CKD. How individuals involved in the CRIC study achieved better rates of blood-pressure control is unclear. Certainly, both demographic and clinical variables have a role. Identifying factors associated with improved blood-pressure control across populations might help increase rates of control in the general population. The CRIC cohort, as noted above, represents a self-selected population and both awareness of disease and motivation to improve the state of

their disease might be much higher than in a randomly sampled population. In the KEEP and CRIC populations, African Americans and patients who did not have diabetes were less likely than white individuals and patients with diabetes to achieve target levels of blood-pressure control.32, 33 In addition, data from the CRIC database show that individuals with an annual household income >$100,000 were more likely to achieve adequate blood-pressure control than those with a household income of <$100,000.33 Presence of albuminuria at all ranges >30 mg per day was also associated with poor rates of blood-pressure control. Individual study data on predictors of blood-pressure control are helpful in improving understanding of the barriers to achieving adequate blood-pressure control, but these data do not explain the differences in rates of blood-pressure control across study populations. Furthermore, predictors of blood-pressure control identified by these studies, such as race and socioeconomic status, do not suggest obvious targets for intervention.
topof page

Diagnosis
Biomarkers Hypertension-associated renal dysfunction is manifested primarily by increases in serum creatinine level. However, persistent increases in serum creatinine levels reflect substantial renal parenchymal damage and some degree of irreversible kidney dysfunction. Identifying individuals with early signs of CKD might help the targeting of therapies to more effectively prevent disease progression and associated complications. Identifying an appropriate marker of early renal dysfunction, however, remains challenging and depends on the underlying etiology of kidney disease. Hypertension in the setting of CKD from glomerular disease can be associated with obvious markers of renal parenchymal disease, such as proteinuria. However, in the absence of overt glomerular disease, such as in hypertensive nephrosclerosis or early diabetic nephropathy, evidence of early kidney injury remains elusive. Microalbuminuria has been suggested as a potential marker of early kidney dysfunction but, as noted earlier, is clinically relevant only when increases into the macroalbuminuric range (>300 mg per day) occur in the presence of appropriate blood-pressure control (that is, <140/90 mmHg). The ReninAngiotensin System Study highlights the limitations of using microalbuminuria as a surrogate marker for early CKD or as a therapeutic target.34 This study examined the histological and clinical progression of individuals with type I diabetes, normoalbuminuria, and normal blood pressure who received either angiotensin-converting-enzyme (ACE) inhibitors, angiotensin-receptor blockers, or placebo. The investigators found that histological progression of diabetic nephropathy persisted despite maintenance of normal blood pressure and microalbuminuria.34 The disparity between reduction of microalbuminuria and CKD progression reported by the ACCOMPLISH study further illustrates the limitations of using microalbuminuria as a surrogate marker for CKD and a prognostic tool for CKD progression.28 Thus, microalbuminuria might accurately represent vascular dysfunction and serve as a marker of cardiovascular risk rather than CKD progression.28 Conversely, macroalbuminuria represents true renal parenchymal damage and should continue to serve as a prognostic marker of CKD progression and a therapeutic target in CKD treatment.

Novel markers of kidney injury are currently under investigation and could be used to identify the first signs of hypertension-associated renal injury. Serum levels of cystatin C have demonstrated promise as an early marker of hypertension-associated kidney dysfunction and may accurately reflect eGFR in various populations.35 Investigators of the Heart and Soul Study examined the effect of baseline systolic blood pressure by assessing serum cystatin C level and serum creatinine-based estimations of GFR. In individuals with an eGFR of >60 ml/min/1.73 m2, serum cystatin C level better correlated with systolic blood pressure than did serum creatinine level.36 Despite the association of cystatin C level with systolic blood pressure, its ability to detect early kidney injury remains limited in its inherent connection with GFR. Like serum creatinine level, the serum level of cystatin C will only increase (indicating renal dysfunction) in the presence of a fall in GFR. The propensity of the kidney to hyperfilter and preserve GFR in the setting of early kidney injury limits the ability of any GFR-based estimation of kidney injury. In addition, novel biomarkers that reflect tubular damage (for example, neutrophil gelatinase-associated lipocalin and interleukin 18) could serve as early markers of hypertension-associated kidney injury. Further investigation into biomarkers of hypertension-associated CKD is ongoing and will hopefully be fruitful. Blood-pressure monitoringwhere and when? The associations of hypertension with CKD was initially identified in observational and experimental studies in which blood pressure was measured in the setting of a specific health-care encounter in the clinic. The cumulative effect of elevated blood pressure on end organ damage, however, might best be demonstrated via blood-pressure monitoring outside of the health-care setting: that is, with home and ambulatory monitoring of blood-pressure levels, as shown by Agarwal and colleagues.37 Home and ambulatory blood-pressure monitoring is important because of its ability to identify white-coat hypertension, masked hypertension and the failure of circadian variations in blood pressure. White-coat hypertension is defined as having persistently elevated blood-pressure readings (>140/90 mmHg) in a clinical setting but normal 24 h average blood-pressure levels (<130/80 mmHg).38 The presence of whitecoat hypertension was previously thought to be a benign finding, but is now thought to increase the risk of cardiovascular complications similarly to essential hypertension. In an observational study of 420 patients with newly diagnosed hypertension, 76 individuals were identified as having white-coat hypertension after undergoing 24 h ambulatory blood-pressure monitoring. After a 10-year follow-up, the rates of cardiovascular eventsboth fatal and nonfatalwere similar between the two groups (18.4% in the whitecoat hypertension group, 16.3% in those without white-coat hypertension), which was higher than the rate reported in nonhypertensive, healthy controls (event rate of 6.8%).39 Masked hypertension, the opposite of white-coat hypertension, carries a similarly increased risk of cardiovascular complications. The SHEAF (Self-Measurement of Blood Pressure at Home in the Elderly: Assessment and Follow-up) study assessed the history of cardiovascular complications in individuals with isolated elevated in-office blood-pressure levels versus those with isolated elevated at-home blood-pressure levels and those with uncontrolled hypertension. Rates of historical cardiovascular complications, specifically coronary artery disease, peripheral vascular disease and stroke were greater in individuals with masked hypertension than in those with white-coat and controlled hypertension. No difference in rates of cardiovascular complications was observed between individuals with masked hypertension and those with uncontrolled hypertension.40

In addition to diagnosing white-coat and masked hypertension, variation in blood pressure between home and office recordings has diagnostic and prognostic importance for individuals with CKD. Agarwal and Andersen41 evaluated the relative predictive value of home blood-pressure readings versus office bloodpressure readings on progression of CKD and death. Elevated home blood-pressure readings were the best predictors of patient prognosis and home blood-pressure readings had a stronger association with the composite outcome (ESRD or death) and each individual outcome than office blood-pressure readings.41 In the setting of ESRD, home blood-pressure monitoring also seems to accurately reflect changes in volume status in patients with ESRD and may be an effective surrogate for the gold standard of blood-pressure measurementambulatory blood-pressure monitoring. Agarwal et al.42 performed home 24 h ambulatory blood-pressure monitoring three-times daily over the course of a week, and home blood-pressure recordings before and after each hemodialysis session. Home blood-pressure recordings, averaged over a week, correlated closely with ambulatory blood-pressure monitoring, with the area under a receiver operating characteristic curve of 0.89 for the diagnosis of hypertension.42 In a separate study by Agarwal and colleagues,43 home blood-pressure monitoring correlated more closely with ambulatory blood-pressure monitoring than did predialysis or postdialysis blood-pressure measurements when patients on hemodialysis underwent protocol-guided lowering of their dry weight in an effort to achieve blood-pressure control. Perhaps the most important value of ambulatory blood-pressure monitoring, especially in patients with CKD, is the evaluation of circadian variations in blood pressure. During sleep, blood pressure should decrease by at least 10%, a concept called nocturnal dipping. The absence of this decrease, referred to as nondipping, is associated with increased severity of kidney disease and cardiovascular disease in patients with CKD. In an evaluation of 184 individuals with CKD and 152 individuals without CKD, Agarwal and Light44 assessed the relationship between the degree of circadian variation in blood pressure level with eGFR and proteinuria. Individuals with proteinuria were more likely to have blunted circadian variations in blood pressure than individuals without proteinuria. The degree of proteinuria was a stronger predictor of absence of blood pressure variation than was CKD stage based on eGFR.44 Given the limitations in serum creatinine measurements as a marker of kidney damage, these findings suggest that variations in blood pressure might provide additional information regarding severity of kidney disease in individuals with hypertension and proteinuria. Absence of nocturnal dipping also seems to be associated with an increased risk of cardiovascular disease in individuals with and without CKD. In a prospective study of 5,682 patients, Kikuya et al.45 identified ambulatory blood pressure targets based on 10-year risk equivalents to in-office blood pressure measurements and outcomes. The 10-year risk equivalent of night-time blood pressure to optimal blood pressure (<120/80 mmHg) was 100/65 mmHg. High blood pressure readings at night, even within the 'normal' range, were associated with an increased risk of cardiovascular events.45Similarly, in patients on hemodialysis, absence of nocturnal dipping was associated with increased cardiovascular morbidity and mortality. Liu and colleagues46 assessed the development of further cardiovascular disease in 80 patients on hemodialysis who had undergone coronary artery catheterization. Individuals were categorized as 'nocturnal dippers' or 'nondippers' according to ambulatory blood pressure recordings. The 56 individuals classed as nondippers had higher rates of coronary artery disease, left-ventricular dyssynchrony, nonfatal cardiovascular events and cardiovascular death than those who were classed as nocturnal dippers.46

The findings of these studies highlight the valuable diagnostic and prognostic information gathered from outof-office blood-pressure monitoring, especially for individuals with hypertension and CKD. Although the best way of managing nondippers remains unclear, the prognostic information that out-of-office monitoring enables physicians to aggressively address known modifiable cardiovascular risk factors. One caveat exists, however. Agarwal and Light47 found that ambulatory blood-pressure monitoring, itself, was associated with decreased sleep time and qualitypotential mediators of the nondipping phenomenon seen in patients with CKD. The findings certainly do not negate the importance of ambulatory blood-pressure monitoring and the prognostic value of nondipping, but do suggest that the findings of nondipping on an isolated ambulatory blood-pressure recording should be placed in the context of sleep quality to fully interpret their prognostic value.
topof page

Treatment of hypertensive CKD

A detailed discussion of the treatment of hypertension-associated CKD is beyond the scope of this Review; however, some points regarding antihypertensive therapy are required. First, achieving blood-pressure control to target levels of <140/90 mmHg clearly slows progression of nephropathy owing to hypertension regardless of whether or not albuminuria is present. Second, lowering blood pressure to <130/80 mmHg in patients with stage 3 or higher nephropathy who have proteinuric kidney disease (>300 mg per day) is of additional benefit. Lastly, data that show a beneficial effect of blockers of the reninangiotensinaldosterone system (RAAS) on nephropathy progression are from the subgroup of hypertensive nephropathy patients with late stage 3 CKD and >300 mg per day of proteinuria; no data exist to support the use of these agents in elderly patients or in those with early stage nephropathy with normoalbuminuria or microalbuminuria.48, 49 Interestingly, data from the ACCOMPLISH trial demonstrated that use of an ACE inhibitor in combination with a calcium antagonist was associated with a reduced requirement for dialysis than use of an ACE inhibitor and a diuretic in individuals >65 years of age.28 For most people, blood-pressure control with agents that are well tolerated will ensure the progression of nephropathy is slowed. In addition, as discussed earlier, CKD is associated with increased salt retention. Excess salt retention probably has a major contributory role in the pathophysiology of difficult-to-control hypertension in CKD. The importance of volume overload is highlighted by the impact of volume removal on blood pressure in patients with ESRD. Observational studies have previously demonstrated the marked, albeit delayed, decrease in blood pressure in patients on hemodialysis who have additional fluid removal. Furthermore, Agarwal et al.50 demonstrated in a randomized controlled trial of additional ultrafiltration in patients on maintenance hemodialysis with hypertension that additional fluid removal led to improved blood-pressure control. How these findings translate to the treatment of patients with predialysis CKD and hypertension requires further investigation. A small, single-center study in 12 individuals with hypertension and CKD demonstrated the ability of loop diuretic to effectively reduce blood pressure. However, whether the individuals in this study were taking diuretics before study enrollment was not reported.51 Nevertheless, these findings together with data from the dialysis population highlight the importance of expanded extracellular-fluid volume control in the treatment of hypertension associated with CKD. Agents that inhibit the RAAS can effectively reduce blood pressure, proteinuria, and CKD progression as well as cardiovascular events associated with hypertension, diabetes and vascular disease.52, 53, 54 As mentioned

earlier, however, these findings are limited to individuals with stage 3 CKD or higher and proteinuria of >300 mg per day. In spite of these findings and guideline recommendations to use such agents in this subgroup of patients, their use is low owing to concerns about increasing serum creatinine levels and risk of hyperkalemia. The risk of hyperkalemia in patients with stage 3 CKD or higher with use of RAAS blockers is higher than in the general population; however, it is still relatively low. The AASK investigators found that the event rate for hyperkalemia in individuals on ramipril was 2.45 per 100 patient-years. Furthermore, use of a diuretic reduced the risk of hyperkalemia by almost 60%.55 These findings suggest that although hyperkalemia occurs more commonly in individuals with stage 3 CKD or higher on an ACE inhibitor, the risk of hyperkalemia should not preclude their use. Moreover, a 2009 study in patients with stage 2 and stage 3 CKD identified risk factors associated with developing hyperkalemia after the addition of an aldosterone antagonist for additional blood-pressure control. Before study enrollment patients were on maximum doses of an ACE inhibitor or angiotensin-receptor blocker. The investigators discovered that baseline serum potassium levels >4.5 mmol/l, baseline eGFR <45 ml/min/1.73 m2, a decrease in systolic blood pressure of >15 mmHg or a decrease in eGFR of >30% after treatment were associated with the development of hyperkalemia (potassium level >5.5 mmol/l).56However, only 17.3% of patients developed hyperkalemia and only one of 46 patients developed serum potassium levels >6 mmol/l. These findings suggest that with careful monitoring, especially of high-risk individuals, RAAS blockade can be safely used in individuals with CKD. Increases in serum creatinine level remain another concern of RAAS blockade. Increases in serum levels of creatinine by up to 30% should not be a reason for discontinuation of RAAS blockers. In a review of 12 randomized controlled trials of ACE inhibitor use, an initial rise in serum creatinine level of up to 30% was associated with long-term stabilization of renal function. After this initial increase, serum creatinine level should stabilize in the first 2 months of treatment.57 The finding of increased creatinine levels after initiation of an ACE inhibitor is similar to the effect of blood-pressure lowering on GFR in individuals with hypertension-associated CKD. In a prospective study of 40 individuals with nondiabetic CKD given antihypertensive therapy with an ACE inhibitor (enalapril) or -blocker (atenolol), the greater initial decreases in GFR after blood-pressure lowering were associated with improved long-term GFR outcomes.58 Use of RAAS inhibitors should be considered in patients with CKD and hypertension given the beneficial effects of these agents on CKD progression and cardiovascular risk reduction. In addition, guidelines for blood-pressure management clearly state that a 30% increase in serum creatinine level over a period of 3 months in the absence of hyperkalemia is associated with a long-term slowing of the progression of nephropathy.59, 60, 61 Thus, patients who will receive the greatest benefits from these agents are those studied in the trials discussed earlier in this Reviewthat is, those with eGFR <40 ml/min/1.73 m2 and proteinuria >300 mg per day. Kidney function, blood pressure and serum potassium level should be monitored closely after starting or adjusting the dose of RAAS blockers, usually within 1 month if not earlier. Only persistently increasing levels of serum creatinine after 2 months of treatment with stable blood pressure or refractory hyperkalemia (that is, persistence above 5.7 mmol/l with appropriate dietary and pharmacological interventions) should be indications for discontinuation.
topof page

Conclusions
Hypertension remains one of the most common chronic diseases in the developed world and, with increasing industrialization, is rapidly growing into a worldwide epidemic. Hypertension can lead to kidney disease or exist as a comorbid condition of kidney disease and can contribute to disease progression. The rates of hypertension-associated CKD and ESRD continue to rise and have a substantial influence on public health and health-care financing. Despite increasing rates of disease and evidence supporting the role of hypertension in CKD progression, patient awareness and control of hypertension remain suboptimal. Lifestyle and pharmacological interventions have both been demonstrated to be effective in lowering blood pressure. Continued vigilance by primary care physicians, nephrologists and other physicians that routinely care for these individuals is required to increase the rates of individuals with CKD and hypertension achieve target blood pressure levels.