History of cardiovascular disease at baseline in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial participants stratified by glomerular filtration rate (GFR). Normal or increased GFR is defined as greater than or equal to 90 mL /min per 1.73 m2 (n = 10 374); mild reduction in GFR, 60 to 89 mL /min per 1.73 m2 (n = 22 965); moderate reduction in GFR, 30 to 59 mL /min per 1.73 m2 (n = 6952); and severe reduction in GFR, less than or equal to 29 mL /min per 1.73 m2 (n = 223). MI/stroke indicates history of old or age-indeterminate myocardial infarction or stroke; CABG/angioplasty, history of coronary artery bypass graft, coronary angioplasty, or other revascularization procedure; CHD, coronary heart disease, defined as history of known prior MI, angina, primary cardiac arrest, coronary stenosis greater than 50%, reversible perfusion defect, or prior coronary revascularization procedure; ASCVD, arteriosclerotic cardiovascular disease, including history of intermittent claudication, gangrene, or ischemic ulcers, documented (eg, by means of angiography, intermittent claudication, ankle-brachial index, or Doppler imaging) atherosclerotic peripheral and/or carotid disease (≥50% occlusion, or ankle-arm index <0.9), or presence of abdominal aortic aneurysm or carotid bruits; and asterisk, P<.05 for differences between groups.
Baseline lipid profile in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial participants stratified by glomerular filtration rate (GFR). Definitions of different GFRs are given in the legend to Figure 1. To convert total, low-density lipoprotein (LDL), and high-density lipoprotein (HDL) cholesterol levels to millimoles per liter, multiply by 0.0259. Asterisk indicates P<.05 between groups.
Odds ratios for the prevalence of cardiovascular disease (CVD) and left ventricular hypertrophy on electrocardiography (ECG-LVH) stratified by deciles of glomerular filtration rate (GFR) and adjusted for age, race, sex, body mass index, baseline blood pressure, low- and high-density lipoprotein cholesterol levels, triglyceride level, diabetes, and smoking. CVD includes old or age-indeterminate myocardial infarction, stroke, coronary artery bypass graft, angioplasty, or other revascularization procedure, or major ST depression or T-wave inversion on baseline ECG or other documented CVD. Each GFR value indicates the cut point for the decile. The referent value for the odds ratio is the decile of GFR greater than 105 mL/min per 1.73 m2.
Rahman M, Brown CD, Coresh J, Davis BR, Eckfeldt JH, Kopyt N, Levey AS, Nwachuku C, Pressel S, Reisin E, Walworth C, . The Prevalence of Reduced Glomerular Filtration Rate in Older Hypertensive Patients and Its Association With Cardiovascular DiseaseA Report From the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial. Arch Intern Med. 2004;164(9):969-976. doi:10.1001/archinte.164.9.969
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
The prevalence of reduced glomerular filtration rate (GFR) in older hypertensive patients and the relationship between level of GFR and cardiovascular disease (CVD) and its risk factors are not well known.
We evaluated baseline renal function in 40 514 hypertensive patients 55 years or older who were enrolled in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). We used the simplified Modification of Diet in Renal Disease study equation to estimate GFR and examined the prevalence of CVD in patients with different levels of GFR.
Fifty-seven percent of patients had mild (60-89 mL/min per 1.73 m2), 17.2% had moderate (30-59 mL/min per 1.73 m2), and 0.6% had severe (≤29 mL/min per 1.73 m2) reductions in GFR. Compared with patients with normal or mildly reduced GFR, patients with moderate or severe reductions in GFR were more likely to have had a prior myocardial infarction or stroke (19.2% and 23.4% vs 28.7% and 26.9%, respectively), have ischemic changes on electrocardiography (ECG) (16.0% and 18.9% vs 24.6% and 34.1%, respectively), and have left ventricular hypertrophy on ECG (ECG-LVH) (3.9% and 4.2% vs 6.0% and 11.2%, respectively). A decrease in GFR of 10 mL/min per 1.73 m2 was independently associated with a 6% higher risk for CVD and 14% higher risk for ECG-LVH. The increase in risk was marked at a GFR of approximately 60 to 70 mL/min per 1.73 m2.
The prevalence of reduced GFR is high in older hypertensive patients. Patients with moderate or severe reduction in GFR are more likely to have a history of CVD and ECG-LVH. Even modest reductions in GFR are independently associated with a higher prevalence of CVD and ECG-LVH.
Approximately 8 million adults (4.5% of the US population) are estimated to have moderate or severe reductions in glomerular filtration rate (GFR; ≤59 mL/min per 1.73 m2).1 However, few large-scale epidemiological studies have been performed to determine the prevalence of decreased GFR in older patients, racial/ethnic minorities, and patients who are at high risk for cardiovascular disease (CVD). Accurate estimation of the prevalence of decreased GFR in previous studies has been limited by the variability of serum creatinine measurements between laboratories and the lack of sensitivity of serum creatinine level used as a sole measure of renal function.2 The National Kidney Foundation Kidney Disease Outcome Quality Initiative (K/DOQI) Clinical Practice Guidelines on Chronic Kidney Disease (CKD) now recommend the use of equations, such as those derived from the Modification of Diet in Renal Disease (MDRD) study, that incorporate age, race, and sex in addition to serum creatinine level for estimation of GFR.3,4 Early identification of patients with decreased GFR has important clinical and research implications. In addition, it is well recognized that the prevalence of CVD is very high in patients with end-stage renal disease (ESRD).5 However, the epidemiology of CVD in patients with earlier stages of chronic renal disease and the relationship of GFR to CVD are not well documented. It is also not clear whether the increased risk for CVD seen in patients with renal disease is due to the coexistence of multiple traditional cardiovascular risk factors (eg, hypertension, diabetes, or smoking) or independently related to the decrement in GFR.
We evaluated baseline renal function in patients enrolled in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). The ALLHAT is a randomized, double-blind, active-controlled clinical trial designed to determine whether the incidence of the primary outcome, fatal coronary heart disease (CHD) or nonfatal myocardial infarction (MI), differs between treatment initiation with a diuretic vs each of 3 other antihypertensive drugs for a mean follow-up of 5 years (range, 4-8 years).6 The participation of more than 40 000 patients and the measurement of serum creatinine levels in a single central laboratory presented a good opportunity to study the prevalence of decreased GFR and the relationship between GFR and CVD and its risk factors in this population.
The purpose of this report is to (1) determine the prevalence of decreased GFR at baseline in ALLHAT participants, (2) describe the association between GFR and history of CVD, and (3) describe the association between GFR and left ventricular hypertrophy on electrocardiography (ECG-LVH).
The double-blind, randomized, active-controlled design of ALLHAT has been described in detail previously.6 In brief, ALLHAT has 2 components. The antihypertensive component is a randomized double-blind trial designed to determine whether the combined incidence of fatal CHD and nonfatal MI differs between diuretic treatment (chlorthalidone, 12.5-25 mg/d) and each of the following 3 alternative antihypertensive therapies: a calcium channel blocker (amlodipine besylate, 2.5-10 mg), an angiotensin-converting enzyme inhibitor (lisinopril, 10-40 mg), and an α-adrenergic blocker (doxazosin mesylate, 2-8 mg). The component for lowering of lipid levels is a randomized, open-label trial designed to determine in 10 356 moderately hypercholesterolemic adults (a subset of the hypertension trial) whether lowering serum cholesterol levels with a hydroxymethylglutaryl–coenzyme A reductase inhibitor (pravastatin sodium) and a cholesterol-lowering diet (National Cholesterol Education Program Step I Diet) will reduce all-cause mortality compared with a control group receiving dietary counseling and usual care. The participants in ALLHAT are high-risk hypertensive patients 55 years or older, recruited at 623 clinical sites in the United States, Canada, Puerto Rico, and the US Virgin Islands between February 14, 1994, and January 31, 1998.7 Patient recruitment techniques used in the study have been published in detail previously.8 All participants gave written informed consent, and all centers obtained institutional review board approval. The trial ended in March 2002, after a mean follow-up of 5 years. Eligibility for participation in ALLHAT required the participant to have at least 1 other risk factor for CVD in addition to hypertension. These include the presence of underlying atherosclerosis, as evidenced by prior MI, stroke, vascular surgical repair or bypass, abnormal ECG, or vascular stenosis as detected by ultrasound; a history of diabetes; current smoking; ECG-LVH or LVH confirmed by echocardiography; or 2 high-density lipoprotein (HDL) cholesterol values less than 35 mg/dL (<0.9 mmol/L) on 2 separate readings within a 5-year period. Because there was limited acceptance of equations to estimate GFR when the study was designed, a serum creatinine level greater than 2 mg/dL (>176.8 µmol/L) as reported by the investigator was used as the criterion to exclude patients with advanced renal dysfunction. However, if the serum creatinine level drawn at the time of randomization was found to exceed 2 mg/dL (>176.8 µmol/L), these patients were maintained in the trial and followed up according to the study protocol.
At baseline, participants underwent measurement of serum potassium, glucose, creatinine, total cholesterol, HDL cholesterol, triglyceride, and alanine aminotransferase levels from fasting blood samples. An ECG was performed if there was no existing tracing within the past year, and read in a central ECG laboratory. The clinical center study coordinator was instructed to complete a questionnaire listing the inclusion criteria, checking all conditions that were known and documented to apply to the participant. In addition, the questionnaire included items about race, ethnicity, sex, years of education, current estrogen use, current regular aspirin use, cigarette smoking (past or current), and the presence of CHD. The presence or absence of CHD was ascertained as a single yes/no question and defined as a history of MI (including silent MI), primary cardiac arrest, coronary revascularization, angina, angiographically defined coronary stenosis greater than 50%, or reversible coronary perfusion defect on results of noninvasive cardiac testing. In addition, history of CVD was also defined as 1 or more of the following: (1) old (>6 months) or age-indeterminate MI or stroke; (2) history of a revascularization procedure, angina pectoris, or transient ischemic attack; and (3) other documented atherosclerotic CVD, which included history of intermittent claudication, gangrene, ischemic ulcers, documented (eg, by angiography, intermittent claudication, ankle-brachial index, or Doppler imaging) atherosclerotic peripheral and/or carotid disease (≥50% occlusion, or ankle-arm index <0.9), or presence of abdominal aortic aneurysm or carotid bruits. A detailed operations manual standardizing these definitions was provided to all the sites. The LVH reported herein was defined by Minnesota code 3-1 or 3-3 with ST/T-wave changes (Minnesota codes 4 and 5) as determined by the central ECG laboratory.
Serum creatinine level was measured at a single central laboratory using the VITROS chemistry system (Ortho-Clinical Diagnostics, Rochester, NY) with a coefficient of variation of approximately 2%. The following simplified MDRD study equation9 incorporating age, race, and sex in addition to serum creatinine level was used to estimate GFR: Estimated GFR (mL/min per 1.73 m2) = 186.3 × (Serum Creatinine)−1.154× (Age, y)−0.203× 1.212 (if Black) × 0.742 (if Female).
Since minor variations in the measurement of serum creatinine level can have a significant impact on estimated GFR, it is recommended that the creatinine assays be compared with those performed at the MDRD study laboratory to determine whether an additional correction is needed to use the abbreviated MDRD study equation.2,10 This was performed in our study by means of an indirect calibration based on analyzing 148 frozen ALLHAT samples at the Third National Health and Nutrition Examination Survey (NHANES III) White Sands Laboratory, White Sands, NM, which has recently been compared with The Cleveland Clinic Laboratory, Cleveland, Ohio, using 212 frozen MDRD study samples and 342 frozen NHANES III samples.10 Analyses of these data indicated that the ALLHAT serum creatinine measurements averaged 0.02 mg/dL (SE, 0.02 mg/dL) (1.76 µmol /L [SE, 1.76 µmol /L]) higher than the measurements during the MDRD study. Thus, serum creatinine values measured in ALLHAT are comparable to measurements during the MDRD study and justify the use of the MDRD study equation without additional calibration. Patients were classified into 4 baseline GFR categories, using the following definitions for GFR ranges: normal or increased (≥90 mL /min per 1.73 m2), mild reduction (60-89 mL /min per 1.73 m2), moderate reduction (30-59 mL /min per 1.73 m2), and severe reduction (≤29 mL /min per 1.73 m2). The GFR cutoffs for these categories are consistent with the recommendations in the National Kidney Foundation K /DOQI Clinical Practice Guidelines on CKD.2 The guidelines define GFR of less than 60 mL /min per 1.73 m2 for longer than 3 months as CKD. Individuals with higher levels of GFR and a marker of kidney damage, such as proteinuria, for longer than 3 months are also defined as having CKD. Urine protein excretion was not ascertained in ALLHAT; therefore, some of the study participants with a GFR greater than or equal to 60 mL/min per 1.73 m2 may also have CKD.2
This report examines the relationship between GFR at baseline and prevalent CVD and ECG-LVH at the time of entry into the study. The level of GFR or control of risk factors in the years before entry into the study is not known. The relationship of a history of CVD and ECG-LVH with GFR was initially examined by the use of contingency tables (4 categories of GFR by the yes/no outcome of the dependent variable). We used analysis of variance to compare continuous variables between GFR categories, and χ2 tests for categorical variables. Logistic regression was used to assess the relationship of history of CVD (as defined by the presence of old or age-indeterminate MI or stroke; coronary artery bypass grafting, angioplasty, and/or another revascularization procedure; major ST depression or T-wave inversion on baseline ECG; or other documented CVD) and the presence of ECG-LVH with GFR after adjusting for other risk factors. The independent variables in the logistic regression were age, race, sex, body mass index, baseline blood pressure, low-density lipoprotein (LDL) and HDL cholesterol and triglyceride levels, diabetes, and smoking.
We used SAS software (SAS Institute Inc, Cary, NC) for all statistical analyses, and a P value of less than .05 was considered significant. For some variables (BMI, education, total cholesterol, LDL cholesterol, HDL cholesterol, fasting triglycerides, and lipid-lowering treatment at baseline), there were small numbers of missing values at baseline. Therefore, the denominators for those percentages, means, and standard deviations, are slightly less than the total number of participants in the estimated GFR groups.
Baseline serum creatinine levels were available in 40 514 of 42 418 ALLHAT participants. The prevalence of decreased GFR in ALLHAT participants at baseline is shown in Table 1. Of the participants with available creatinine levels, 22 965 (56.7%) had a mild reduction in GFR (60-89 mL/min per 1.73 m2), 6952 (17.2%) had a moderate reduction (30-59 mL/min per 1.73 m2), and 223 (0.6%) had a severe reduction (≤29 mL/min per 1.73 m2). The mean estimated GFR and serum creatinine level in each group is listed in Table 1.
Patients with moderate and severe reductions in GFR were older and included a higher proportion of women compared with patients with a mild reduction in or a normal GFR (P<.05; Table 2). The majority of patients with a moderate reduction in GFR were white non-Hispanic (57.6%), whereas black non-Hispanic patients were the largest racial/ethnic group among the patients with a severe reduction of GFR (42.2%) and with a normal or a high GFR (43.0%). Level of education was low in the study population as a whole, with more than 70% of patients reporting 12 or fewer years of education. The mean BMI was lower in patients with severe and moderate reductions in GFR (P<.05). Approximately 90% of patients in all groups were taking antihypertensive medications, and mean baseline systolic blood pressure was higher in patients with a severe reduction in GFR (Table 2).
Patients with moderate (28.7%) and severe (26.9%) reductions in GFR were more likely to have a history of old or age-indeterminate MI or stroke than patients with a mild reduction (23.4%) or a normal or increased GFR (19.2%; P<.05) (Figure 1). A higher proportion of patients with moderate (17.2%) and severe (14.4%) reductions of GFR had undergone coronary artery bypass grafting, coronary angioplasty, or another revascularization procedure compared with patients with a mild reduction in GFR or with a normal or a high GFR (13.6% and 9.2% respectively; P<.05). More patients with moderate (31.3%) and severe (28.7%) reductions in GFR had a history of CHD compared with patients with a mild reduction or a normal or high GFR (26.4% and 21.2%, respectively; P<.05). Patients with moderate (24.6%) and severe (34.1%) reductions in GFR were more likely to have major ST-segment depression or T-wave inversion on the baseline ECG compared with the other groups (18.9% and 16.0%, respectively; P<.05). In addition, the prevalence of ECG-LVH was higher in patients with moderate (6.0%) and severe (11.2%) reductions in GFR than those with a normal GFR (3.9%) or a mild reduction (4.2%; P<.05 for differences between groups).
Given the higher prevalence of CVD in patients with moderate and severe reductions in GFR, additional analyses focused on traditional risk factors for atherosclerosis in these patients (Table 3). Patients with moderate and severe reductions in GFR had significantly higher total and LDL cholesterol levels, higher triglyceride levels, and lower HDL cholesterol levels (P<.05 for differences between groups). In addition, they were more likely to have high total and LDL cholesterol profiles and low HDL cholesterol profiles as defined by the current National Cholesterol Education Program guidelines (Figure 2).11 The prevalence of type 2 diabetes was 44.0% in patients with a severely reduced GFR, 32.4% in patients with a moderately reduced GFR, 32.8% in patients with a mildly reduced GFR, and 45.2% in patients with a normal or high GFR (Table 3). A significant number of patients with moderate and severe reductions in GFR were current smokers (17.3% and 24.2%, respectively), as were those with a mild reduction in GFR and a normal GFR (20.9% and 27.1%, respectively; P<.05). Finally, the prevalence of obesity was lower in patients with moderate and severe reductions in GFR.
Multiple logistic regression analyses were used to determine the independent effects of estimated GFR on prevalent CVD. After adjustment for age, race, sex, body mass index, baseline blood pressure, LDL and HDL cholesterol and triglyceride levels, diabetes, and smoking, estimated GFR was independently associated with CVD. A 10-mL /min per 1.73 m2 decrease in GFR was associated with a 6% higher risk for CVD (odds ratio [OR], 1.06 per 10 mL
/min per 1.73 m2; P<.001). When GFR was categorized according to the cut points previously defined, patients with severe reduction in GFR were almost twice as likely to have CVD at baseline (OR, 1.99; P<.001), compared with the reference group (normal or increased GFR). Patients with a mild reduction in GFR had a slightly increased risk (OR, 1.09; P = .004). When GFR was categorized into deciles, patients in the lowest decile of GFR (<53.6 mL /min per 1.73 m2) had the highest likelihood of having prevalent CVD (OR, 1.55; P<.001; Figure 3) compared with patients in the highest decile of GFR. The increased risk appeared to begin at GFR values approximating 60 mL /min per 1.73 m2.
Similar logistic regression models were used to determine the relationship between estimated GFR and the presence of ECG-LVH. After adjustment for the covariates listed in the preceding paragraph, estimated GFR was independently associated with the presence of ECG-LVH. A 10-mL /min per 1.73 m2 decrease in GFR was associated with a 14% higher risk of LVH (OR, 1.14 per 10 mL /min per 1.73 m2; P<.001). When GFR was categorized according to the cut points previously defined, patients with a severe reduction in GFR had more than a 3-fold increase (OR, 3.19; P<.001) in the prevalence of ECG-LVH, compared with the reference group (normal or increased GFR). Patients with a mild reduction in GFR also had a significantly increased risk (OR, 1.26; P = .001). When GFR was categorized into deciles, patients in the lowest decile of GFR had the highest likelihood of having ECG-LVH (OR, 2.51; P<.001; Figure 3) compared with patients in the highest decile of GFR. The likelihood of ECG-LVH seemed to increase at a GFR below approximately 70 mL /min per 1.73 m2 (Figure 3). Since ECG may be less accurate in the diagnosis of LVH in black patients,12 separate regression models were run stratified by race, and qualitatively similar results were obtained (data not presented).
This report demonstrates a high prevalence of reduced GFR in older hypertensive patients. Patients with moderate or severe reduction in GFR in this population are more likely to have a history of CVD and ECG-LVH. Although most traditional risk factors for CVD are common, dyslipidemia is particularly severe in patients with moderate and severe reduction in GFR. After adjustment for several known risk factors for CVD, even modest reductions in GFR are independently associated with a higher prevalence of clinical CVD and ECG-LVH.
Few studies have documented the prevalence of reduced GFR in large populations, and recent published estimates have been derived from analyses of the NHANES database.13,14 Since we documented that measurement of serum creatinine level in the NHANES and ALLHAT laboratories was comparable, use of the modified MDRD equation allows for valid comparisons between the 2 populations. Although patients with known serum creatinine levels greater than 2 mg/dL (>176.8 µmol /L) were excluded, the prevalence of reduced GFR (<60 mL /min per 1.73 m2, defined as CKD by the National Kidney Foundation K/DOQI guidelines) is much higher in ALLHAT participants (approximately 18%) compared with the general adult population (4.6%).2 This is likely due to the fact that the ALLHAT population was older and had multiple risk factors for chronic renal disease.
It is important for primary care physicians to appreciate that a significant percentage of their older hypertensive patients may have at least a moderate reduction in GFR. Rather than relying on serum creatinine measurements alone, use of the MDRD study equation allows for recognition of these patients. The clinician can then intervene on manifestations of CKD such as anemia and bone disease, which are often established in the GFR range of 30 to 60 mL/min per 1.73 m2.3 In addition, better blood pressure and glycemic control and inhibition of the renin-angiotensin axis, particularly in proteinuric patients, should be instituted to slow the rate of decline of GFR.3
A second important finding in our study was the high prevalence of CVD in patients with moderate and severe reduction in GFR. Forty percent of US patients undergoing hemodialysis have coronary artery disease, and CVD mortality is approximately 15 times greater in dialysis patients than in the general population.15 The fact that significant morbidity and mortality due to CVD occur soon after the onset of ESRD suggests that the atherosclerotic vascular disease begins during the earlier stages of chronic renal disease.5 Despite the plethora of evidence in patients with ESRD,5 the epidemiology of CVD in patients with earlier stages of chronic renal disease is not well delineated. It is in this often hard-to-capture population that this report makes important observations about the prevalence of CVD and its risk factors. We confirm that patients with reduced GFR were more likely to have a history of CVD and ischemic changes on ECG. We used information about the history of CVD provided by the clinician at the time of study entry, which was validated by review of source documentation at site visits. In addition, the more objective findings of ischemic changes on ECG read in a central laboratory showed similar results. Thus, these data demonstrate the high prevalence of CVD in patients with reduced GFR and underscore the need for early recognition and aggressive risk factor management in these patients.
The reason why patients with reduced GFR are at such high risk for CVD is an area of active investigation5 and is thought to be related to 2 issues. First, the prevalence of traditional risk factors for CVD such as smoking, diabetes, and hypertension may be higher in patients with reduced GFR. Second, there may be factors unique to the altered homeostatic milieu in patients with CKD that increase risk for CVD. This report provides important data supporting both concepts. Hypertension and diabetes were common in patients with reduced GFR. Hyperlipidemia in particular was more severe in patients with reduced GFR. Patients with moderate and severe reduction in GFR had lower HDL cholesterol levels and higher total and LDL cholesterol levels. However, patients with reduced GFR were less likely to be current smokers and less likely to be obese than patients with preserved GFR. Although it is possible that the level of control of cardiovascular risk factors in the years before entry in the study may have differed between the GFR groups, data to confirm or refute this possibility are not available in our study participants.
The fact that GFR remained independently associated with prevalence of CVD in our study following adjustment for traditional risk factors lends support to the concept that there may be other factors that predispose patients with reduced GFR to atherosclerotic disease. For example, altered homocysteine levels,16 vascular calcifications associated with renal osteodystrophy,17 poor nutrition,18 and increased levels of inflammatory markers19 are some of the factors thought to play a role in this process. Additional research is clearly needed to define the role of these and other factors that increase risk of CVD in patients with reduced GFR. The newly launched National Institutes of Health–sponsored prospective cohort study of chronic renal insufficiency will be an important resource in this area of investigation. In addition, the relationship of baseline levels of GFR to incidence of CVD and ECG-LVH will be the subject of another report.
The third important finding in our study is the high prevalence of ECG-LVH in patients with moderate and severe reductions in GFR. Left ventricular hypertrophy is well documented as an independent marker for CVD in the general population.20 In a study of Framingham participants, Culleton et al21 reported that the prevalence of ECG-LVH in patients with mild CKD (as defined by a serum creatinine level of 1.4-3.0 mg/dL [123.8-265.2 µmol /L]) was significantly higher compared with the rest of the population. Echocardiographic LVH has also been shown to be higher in patients with creatinine clearance less than or equal to 30 mL /min (38%) than in those with a creatinine clearance greater than 30 mL /min (16%).22 Our findings in this large and representative sample of patients confirm the high prevalence of ECG-LVH in patients with reduced GFR. In fact, fairly modest reductions in GFR (less than approximately 70 mL/min per 1.73 m2) are associated with a significant increase in the prevalence of LVH.
The reasons why patients with CKD have a higher prevalence of LVH have not been completely elucidated. Clearly, as we demonstrate in this study, patients with CKD often have several traditional risk factors for LVH (eg, hypertension, diabetes, and older age). However, the fact that estimated GFR remained independently associated with LVH after adjustment for several traditional risk factors suggests that the altered homeostatic milieu in CKD may also predispose to the development of LVH. For example, the effect of anemia on LVH has been extensively studied, and a strong association between hemoglobin levels and progressive left ventricular growth has been demonstrated in patients with CKD.23- 25 In addition, extracellular fluid volume expansion and abnormalities of calcium/phosphate homeostasis may promote development of LVH in patients with CKD.26 Therefore, early recognition of LVH in patients with CKD may provide an opportunity for intervention with correction of anemia, control of blood pressure, optimization of fluid status, and balance of calcium phosphate. Future prospective studies will be needed to document the impact of these interventions on left ventricular geometry and cardiovascular outcomes.
A strength of this study is its large and diverse patient population, drawn from across the United States, Canada, and Puerto Rico from a variety of health care providers, allowing for generalizability of the results. Although the level of health care before entry into the study may have differed between the sites, it is unlikely to have an impact on the relationship between GFR and CVD/LVH reported herein. In addition, the use of a single central laboratory and careful calibration with the MDRD study assay enhances the validity of the findings. A limitation, however, was our inability to identify patients at the earliest stages of chronic renal disease, based on markers of kidney damage such as proteinuria. According to the National Kidney Foundation K/DOQI Clinical Practice Guidelines on CKD,2 patients with a normal or increased GFR (≥90 mL /min per 1.73 m2) or a mild reduction in GFR (60-89 mL /min per 1.73 m2) are defined as having CKD if they also have a marker of kidney damage. Therefore, it is possible that some of the ALLHAT participants with a GFR greater than or equal to 60 mL/min per 1.73 m2 may also have CKD, and that our data may have underestimated the prevalence of CKD in this population.
Finally, this cohort of more than 7000 well-characterized patients with GFR less than 60 mL/min per 1.73 m2 provides an important opportunity to study the progression of renal insufficiency in high-risk hypertensive patients. Although the primary aim of the antihypertensive arm of ALLHAT is to determine the incidence of fatal CHD and nonfatal MI, change in renal function and development of ESRD are prespecified secondary outcomes of the study. Serum creatinine measurements are obtained periodically during follow-up, and development of ESRD is captured as a study event. Future analyses of these data can provide hypothesis-generating information about comparative effects on renal outcomes in each of the randomized antihypertensive and lipid-lowering drug groups. In addition, ALLHAT can be an important resource for epidemiological data. For example, despite the increased prevalence of CVD, there have been conflicting reports from the NHANES data whether renal insufficiency has an independent influence on cardiovascular mortality.27,28 Long-term follow-up of the ALLHAT participants may provide additional understanding about the effects of baseline renal insufficiency on cardiovascular outcomes.
We have demonstrated that the prevalence of reduced GFR in older hypertensive patients is high. Patients with a moderate or a severe reduction in GFR in this population are more likely to have a history of CVD and ECG-LVH. Even modestly low GFR levels are independently associated with a higher prevalence of clinical CVD and ECG-LVH.
Corresponding author and reprints: Mahboob Rahman, MD, MS, Divisions of Hypertension and Nephrology, Case Western Reserve University/University Hospitals of Cleveland, 11100 Euclid Ave, Cleveland, OH 44106 (e-mail: firstname.lastname@example.org).
Accepted for publication June 24, 2003.
This study is supported by a contract with the National Heart, Lung, and Blood Institute (NHLBI), Bethesda, Md. The ALLHAT investigators acknowledge contributions of study medications supplied by Pfizer Inc, New York, NY (amlodipine and doxazosin); AstraZeneca, Wayne, Pa (atenolol and lisinopril); and Bristol-Myers Squibb, Princeton, NJ (pravastatin), and financial support provided by Pfizer Inc to the NHLBI. Dr Levey is supported by grant RO1 DK 53869 from the National Institutes of Health, Bethesda, Md.
This study was presented in part at the 17th Annual Meeting of the American Society of Hypertension; May 15, 2002; New York, NY.