- Email: [email protected]
Validity of Plasma Aldosterone-to-Renin Activity Ratio in African American and White Subjects With Resistant Hypertension
AJH
2005; 18:805– 812
Validity of Plasma Aldosterone-to-Renin Activity Ratio in African American and White Subjects With Resistant Hypertension Mari K. Nishizaka, Monique Pratt-Ubunama, Mohammad A. Zaman, Stacey Cofield, and David A. Calhoun Background: Recent reports suggesting that primary aldosteronism (PA) is more common than historically thought have often relied on use of the plasma aldosterone concentration (PAC) to plasma renin activity (PRA) ratio (ARR) to identify patients with PA. Prior determinations of the validity of the ARR had been generally limited to subjects that could be withdrawn from antihypertensive therapy and to non–African American subjects. Methods and Results: The current study was designed to evaluate prospectively the diagnostic value of the ARR in treated African American and white subjects with resistant hypertension. Consecutive subjects referred to a university hypertension clinic for resistant hypertension were evaluated with an early morning ARR and a 24-h urinary aldosterone and sodium. The presence of PA was defined as a suppressed PRA (⬍1.0 ng/mL/h) and elevated urinary aldosterone excretion (⬎12 g/24 h) dur-
ing high dietary sodium ingestion (⬎200 mEq/24 h). In 58 subjects, PA was confirmed. The ARR was elevated (⬎20) in 45 of 58 subjects with PA and in 35 of the 207 patients without PA, resulting in a sensitivity of 78% and specificity of 83% with a corresponding positive predictive value of 56% and a negative predictive value of 93%. Among African American subjects, the ARR was less sensitive than in white subjects (75% v 80%), but it still had a high negative predictive value (92% v 94%). Conclusions: These data indicate that the ARR is valid as a screening test for PA in African American and white patients on stable antihypertensive treatments, but a high percentage of false-positive results precludes using it for accurate diagnosis of PA. Am J Hypertens 2005;18: 805– 812 © 2005 American Journal of Hypertension, Ltd. Key Words: Primary aldosteronism, aldosterone to renin ratio (ARR), African American, resistant hypertension.
ecent reports1–10 indicate that primary aldosteronism (PA) is common in both general and selected hypertensive populations, with reported prevalence rates ranging from 7% to 20%. One reason for the recognition of the higher prevalence of PA is that use of the plasma aldosterone concentration (PAC)/plasma renin activity (PRA) ratio (ARR) has facilitated broader screening of hypertensive subjects. Since Hiramatsu et al11 described the utility of the ARR to screen for PA, it has been increasingly used to identify subjects at high risk for PA.1–9 Primary aldosteronism is suggested by a high ARR. The cuff-off level is arbitrary, ranging from 20 to 100 (ng/ dl)/(ng/mL/h). Some recommendations on use of the ARR to screen for PA suggest an elevated ratio in conjunction with a minimally high PAC.4,5,9,12,13 Requiring a higher
R
ratio or a PAC of a certain level improves the specificity (fewer false-positive results) but lowers the sensitivity (more false-negative results) of the screening test. The validity of ARR has been assessed in multiple studies, using various methods to confirm the diagnosis of PA. It is generally agreed that failure to suppress aldosterone excretion during volume expansion remains the gold standard for confirmation of PA. The most commonly used methods to achieve volume expansion are fludrocortisone suppression testing (FST), intravenous saline infusion, or dietary oral salt loading. Prior reports of determination of ARR test characteristics have generally been done in patients with mild to moderate hypertension (allowing for discontinuation of therapy before evaluation). In that setting, an elevated
Received April 13, 2004. First decision January 12, 2005. Accepted January 18, 2005. From the Vascular Biology and Hypertension Program (MKN, MP-U, MAZ, DAC) and Department of Biostatistics (SC), University of Alabama at Birmingham, Birmingham, Alabama. This work was supported by American Heart Association postdoctoral
fellowship 0425389B (M.K.N.), American Heart Association Grant-in-Aid 0355302B (to DAC), National Heart, Lung, and Blood Institute Grants HL075614 (to DAC) and HL07457 (to MAZ and MP-U). Address correspondence and reprint requests to Dr. Mari K. Nishizaka, 115 CHSB, 933 South 19th Street, Birmingham, AL 352942041; e-mail: [email protected]
© 2005 by the American Journal of Hypertension, Ltd. Published by Elsevier Inc.
0895-7061/05/$30.00 doi:10.1016/j.amjhyper.2005.01.002
806
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
ARR has been reported to have a high sensitivity and specificity in identifying patients with PA.11,13–17 What remains unknown is the validity of the ARR to screen for PA in patients who are unable to discontinue antihypertensive therapy. Knowing the diagnostic value of ARR in the setting of ongoing antihypertensive treatment is particularly relevant as the prevalence of PA, as recently documented by this laboratory, is high in subjects with resistant hypertension.10,18 Individuals of black ethnicity (South-African individuals of black ethnicity or African Americans) are more likely to have low renin levels compared with persons of white ethnicity.6,19,20 Despite this tendency toward lower renin activity, perhaps suggesting a greater prevalence of aldosterone-associated hypertension, we have reported that in subjects referred to a specialty clinic for resistant hypertension, African Americans have a similar prevalence of PA compared with white subjects.10 What has not been previously determined is the validity of the ARR to screen for PA in African American subjects. With generally lower renin levels but similar rates of PA (at least in patients with resistant hypertension), it might be suspected that the ARR is less specific in African American versus white subjects. The current study was designed to determine prospectively the validity of an elevated ARR to screen for PA in African American and white subjects with resistant hypertension while these individuals were maintained on a stable antihypertensive regimen. Outpatient dietary salt loading was used to confirm the diagnosis of PA in all evaluated subjects.
AJH–June 2005–VOL. 18, NO. 6
outpatient basis (Fig. 1). Blood samples were collected using routine laboratory procedure including use of a tourniquet and vacutainer in subjects who had been ambulatory for at least 1 h. Hemolyzed serum was not used for analysis. Initial evaluation included an early morning PRA and PAC (when PRA tends to be lowest and PAC tends to be highest), and a 24-h urine collection for sodium and aldosterone during the subject’s ad libitum diet. If the urinary aldosterone was elevated (⬎12 g/24 h) but the urinary sodium was ⬍200 mEq/24 h, the 24-h urinary assessments were repeated after 3 days of dietary salt supplementation. A suppressed PRA (⬍1.0 ng/mL/h), and elevated urinary aldosterone (⬎12 g/24 h) in the setting of high dietary sodium ingestion (⬎200 mEq/24 h) confirmed the diagnosis of PA.12,21 To evaluate the validity of relying on high urinary sodium excretion during the subject’s ad libitum diet to indicate chronic dietary salt loading, we remeasured urinary aldosterone excretion after dietary salt supplementation (50 mEq NaCl/day) in 12 subjects whose ad libitum urinary sodium excretion was ⬎200 mEq/24 h and whose aldosterone excretion was high (⬎12 g/24 h). Laboratory Methods
Subjects
Levels of PAC, PRA, urinary aldosterone, and urinary sodium were measured by commercial laboratories using standard techniques. The PAC (Quest Diagnostics, Atlanta, GA) and PRA (Mayo Clinic Laboratories, Rochester, MN) levels were measured by radioimmunoassay. The reference range for PAC is 4.0 to 31.0 ng/dL. The reference range for an upright PRA is 1.31 to 3.95 ng/mL/h. The minimum reported PRA by this laboratory is 0.6 ng/mL/h. Urinary
We prospectively evaluated consecutive subjects referred for resistant hypertension to the University of Alabama at Birmingham Hypertension Clinic over a 44-month period (December 2000 to July 2004). The protocol was reviewed and approved by the Institutional Review Board and written informed consent was obtained from all participants. Resistant hypertension was defined as uncontrolled hypertension (⬎140/90 mm Hg) at two or more clinic visits, despite use of three or more antihypertensive medications at pharmacologically effective doses. All subjects were on a stable antihypertensive regimen for at least 4 weeks before biochemical evaluation. No medications were discontinued before evaluation except for spironolactone, triamterene, or amiloride, which were discontinued for at least 6 weeks before evaluation. Blood pressure was also measured by 24-h ambulatory blood pressure (BP) monitoring (SpaceLabs 90207, Redmond, WA). Serum potassium levels were corrected with oral supplementation to be ⬎3.5 mEq/L before evaluation, if necessary. Secondary hypertension, such as renovascular hypertension, pheochromocytoma, or Cushing syndrome, was excluded as clinically indicated. Subjects on chronic glucocorticoid therapy were not included in the analysis. Biochemical evaluation was done in all subjects on an
FIG. 1. Flow chart for the diagnostic evaluation for primary aldosteronism (PA). Na ⫽ sodium; PAC ⫽ plasma aldosterone concentration; PRA ⫽ plasma renin activity.
Methods
AJH–June 2005–VOL. 18, NO. 6
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
aldosterone was measured by radioimmunoassay (Mayo Clinic Laboratories, Rochester, MN). The reference range for urinary aldosterone is 2 to 16 g/24 h. Statistical Analysis Values are expressed as mean ⫾ SD. A PRA value ⬍0.6 ng/ mL/h was treated as 0.6 ng/mL/h. Values between groups were compared by the Student t test, Mann-Whitney test, and 2 test. A P value ⬍ .05 was considered significant. Receiver operator characteristic (ROC) analysis was used to determine the test characteristics of the ARR.
Results A total of 265 subjects (115 African American and 150 white) with resistant hypertension were evaluated. Demographic values of the evaluated subjects listed in Table 1. Of the total subjects, 58% had low-renin hypertension, defined as a PRA ⬍1.0 ng/mL/h. The majority of the subjects were receiving an angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB), a -blocker, a calcium channel blocker (CCB), and/or a diuretic. Compared with their white counterparts, African American subjects were younger (53 ⫾ 12 v 59 ⫾ 11 years), heavier (body mass index 34.4 ⫾ 8.5 v 30.8 ⫾ 6.4 kg/m2), and had higher clinic (91 ⫾ 18 v 87 ⫾ 15 mm Hg) and mean 24-h ambulatory diastolic BP levels (88 ⫾ 13 v 82 ⫾ 13 mm Hg). Systolic BP values were not different, either in the clinic or during ambulatory BP monitoring, in African American and white subjects (Table 1). African Americans were less likely to be on a -blocker and more likely to be on a calcium channel
807
blocker than were white subjects, but the total number of antihypertensive medications was not different between the two groups (3.9 ⫾ 1.1 v 3.8 ⫾ 1.3). The mean PRA was similar in the two ethnic groups, but the mean PAC tended to be lower in African American subjects (12.5 ⫾ 8.1 v 14.5 ⫾ 10.2 ng/dL). The mean ARR was not different in African American versus white subjects. The prevalence of the low-renin hypertension (PRA ⬍1.0 mg/mL/h) was also similar in African American and white subjects (62% v 55%) (Table 2). Of the 265 subjects, 58 (28 African American and 30 white) were confirmed to have PA. Of these, 23 of the 58 PA (40%) and 31 of the 207 non-PA subjects (15%) were hypokalemic (⬍3.6 mEq/L) on presentation to our clinic. In 31 of the 58 subjects confirmed to have PA, diagnoses were made based on a high urinary aldosterone (range 13 to 55 g/24 h) and sodium (range 202 to 363 mEq/24 h) excretion during their normal diet. In the remaining 27 subjects, diagnoses were made after subjects had repeated the 24-h urinary collection after 3 days of dietary salt supplementation. Six patients were excluded from having PA after dietary sodium loading. Serum potassium levels were monitored in all subjects who underwent dietary salt loading. If the subject become hypokalemic with salt loading and the urinary aldosterone level was not elevated (⬎12 g/ 24 h), the urinary aldosterone assessment was repeated with potassium supplementation sufficient to maintain normokalemia. It is still possible, however, that reductions in serum potassium to less than hypokalemic range may have affected aldosterone secretion. In 12 subjects with high dietary sodium (259 ⫾ 46 mEq/ 24 h) and aldosterone excretion (21.3 ⫾ 6.2 g/24 h)
Table 1. Baseline characteristics of subjects during ad libitum diet Characteristic
Total
African American
White
N Age (y) Male/female Body mass index (kg/m2) Clinic BP (mm Hg) Systolic Diastolic ABPM, 24-h Systolic Diastolic Number of antihypertensive medications Medication use (% of subjects) ACE inhibitor ARB Diuretic -Blocker CCB
265 56 ⫾ 12 116/149 32.4 ⫾ 7.4
115 53 ⫾ 12 45/70 34.4 ⫾ 8.5
150 59 ⫾ 11* 71/79 30.8 ⫾ 6.4*
160 ⫾ 27 89 ⫾ 16
160 ⫾ 27 91 ⫾ 18
160 ⫾ 27 87 ⫾ 15*
143 ⫾ 18 85 ⫾ 13
145 ⫾ 20 88 ⫾ 13
142 ⫾ 16 82 ⫾ 13*
3.8 ⫾ 1.2
3.9 ⫾ 1.1
3.8 ⫾ 1.3
55% 47% 82% 64% 68%
55% 41% 85% 57% 76%
55% 52% 79% 70%* 63%*
ABPM ⫽ ambulatory blood pressure monitor; ACE ⫽ angiotensin converting enzyme; ARB ⫽ angiotensin receptor blocker; BP ⫽ blood pressure; CCB ⫽ calcium channel blocker. Values are mean ⫾ SD. * Significantly different from African American subjects, P ⬍ .05.
808
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
AJH–June 2005–VOL. 18, NO. 6
Table 2. Biochemical parameters and test characteristics of an elevated plasma aldosterone/plasma renin activity ratio (ARR) Characteristic N PAC (ng/dL) PRA (ng/mL/h) ARR Serum potassium (mEq/L) Urinary sodium (mEq/24-h) Urinary aldosterone (g/24-h) Primary aldosteronism (%) Test characteristics of ARR ⬎20 Sensitivity (%) Specificity (%) ⫹PV (%) ⫺PV (%)
Total 13.6 3.5 15.9 4.0 184 11.5
265 ⫾ 9.3 (1–49) ⫾ 6.6 (0.6–55.5) ⫾ 15.7 (0.1–78.3) ⫾ 0.5 (2.4–5.5) ⫾ 79 (46–442) ⫾ 9.2 (1.0–78) 58 (22%) 78 83 56 93
African American
White
115 12.5 ⫾ 8.1 3.7 ⫾ 9.2 14.0 ⫾ 13.2 3.9 ⫾ 0.5 172 ⫾ 85 12.0 ⫾ 10.3 28 (24%)
150 14.5 ⫾ 10.2 3.5 ⫾ 6.6 15.9 ⫾ 15.7 4.0 ⫾ 0.5 194 ⫾ 74 11.1 ⫾ 8.3 30 (20%)
75 87 66 92
80 80 50 94
PAC ⫽ plasma aldosterone concentration; PRA ⫽ plasma renin activity; ⫹PV ⫽ positive predictive value; ⫺PV ⫽ negative predictive value. Values are mean ⫾ SD (range).
during their ad libitum diet, the 24-h urine collection was repeated after 3 days of additional dietary salt supplementation (50 mEq/day). In all 12 subjects, aldosterone excretion remained elevated (21.8 ⫾ 4.7 g/24 h) despite higher sodium excretion (353 ⫾ 31 mEq/24 h). This prevalence of PA was similar among African American and white subjects (24 v 20%, respectively). Thirteen of the 58 subjects with confirmed PA had a false-negative ARR (⬍20). Seven of these subjects were African American and six were white. Of the 207 hypertensive subjects without PA, 35 had an elevated ARR (⬎20). Accordingly, an ARR ⬎20 had an overall sensitivity of 78% and a specificity of 83%. This corresponds to a positive predictive value of 56% and a negative predictive value of 93%. The PRA was suppressed (⬍1.0 mg/ mL/h) in all subjects with an ARR ⬎20. An ARR⬎20 was less sensitive (75 v 80% respectively) but more specific (87 v 80%, respectively) in African American compared with white subjects (Table 2). Despite this difference in sensitivity, the negative predictive values of an elevated ARR were similar in African American and white subjects, indicating that in this clinical setting, an ARR ⬎20 is an equally effective screening test for African American and white subjects. We determined the test characteristics of different cut-off levels of the ARR (Fig. 2). Using a higher ARR as a screening test provided better specificity, but lowered the sensitivity. Requiring a high ARR (⬎50) as a positive screen resulted in a very high specificity and improved the positive predictive value, but at the expense of a substantially lower sensitivity and negative predictive value. Using ROC analysis, in this cohort of subjects being treated for resistant hypertension, an ARR of 16.3 provided the highest sensitivity and the highest negative predictive value, resulting in the greatest value as a screening test (Fig. 3). We also compared biochemical values in the presence or absence of different classes of antihypertensive agents (Table 3). Urinary aldosterone excretion was not different
in relation to use of the any of the evaluated classes of agents. The PRA was higher with use of diuretics and tended to be so with use of ACE inhibitors. The PRA and PAC were not different in relation to calcium channel blocker (CCB) use. The PRA tended to be lower in the setting of -blocker use resulting in significant increases in the ARR. As a consequence, the ARR in the setting of -blocker use was associated with a higher false positive rate (22% for -blockers v 14 to 17% for other classes of agents). However, the areas under the ROC curves were similar in the presence or absence of -blockers (84% v 92%, respectively) as were the best ARR cut-off values (17.0 v 16.3, respectively).
FIG. 2. Test characteristics of different levels of plasma aldosterone/plasma renin activity ratio (ARR) to identify primary aldosteronism in subjects being treated for resistant hypertension. ⫹PV ⫽ positive predictive value; ⫺PV ⫽ negative predictive value.
AJH–June 2005–VOL. 18, NO. 6
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
FIG. 3. Receiver operator characteristic analysis of plasma aldosterone/plasma renin activity ratio (ARR) to identify primary aldosteronism in subjects being treated for resistant hypertension.
Discussion This is the first prospective study to determine the validity of the ARR to screen for PA in a resistant hypertensive population consisting of a large percentage of African American subjects. In this study population of patients evaluated on a stable antihypertensive regimen, we found an ARR ⬎20 to have sufficient sensitivity and negative predictive value to be an effective screen for, but lacking the specificity to reliably confirm the diagnosis of PA. This was true of both African American and white subjects. In the current study cohort, to have relied on an elevated
809
ARR only as indication of PA would have resulted in 35 false-positive diagnoses among 265 evaluated subjects. Prior reports7,11,13–16,22–24 have evaluated the validity of the ARR. These studies have mostly reported an elevated ARR to have both a high sensitivity and specificity (test characteristics varying according to the cut-off level of ARR applied). In the current analysis, the sensitivity and specificity of the ARR was less than described in these earlier reports, leading us to conclude that in the setting of ongoing antihypertensive therapy, an elevated ARR (⬎20) is not sufficient to diagnose PA and must be confirmed with suppression testing. Reported differences in ARR sensitivity and specificity may be related to differences in study design. In the current study, the validity of an elevated ARR was compared with lack of suppression of 24-h urinary aldosterone excretion during high dietary salt ingestion. In other studies, fludrocortisone suppression (FST), saline infusion, and captopril suppression of plasma aldosterone were used as confirmatory testing. It is generally accepted that FST is the most accurate method of confirming the diagnosis PA.3,4,7 However, wider clinical use of FST is limited in that it generally requires a hospitalization of 4 to 5 days for both safety and accuracy reasons. Because dietary salt loading is more amenable to outpatient use, it is being increasingly used for confirmatory testing of PA, particularly within the United States. While intravenous saline infusion has been compared with FST,25,26 sometimes favorably, sometimes not, there has not been a systematic comparison of dietary salt loading to FST. We recognize this as a limitation of our study. Depending on the results of such a comparison, the relative test characteristics of the ARR with oral salt loading versus FST may differ.
Table 3. Biochemical values in relation to use of different classes of antihypertensive agents in total group of subjects Subgroup (n) ACE inhibitor use (⫹) (145) (⫺) (120) ARB use (⫹) (125) (⫺) (140) Diuretic use (⫹) (216) (⫺) (49) -Blocker use (⫹) (170) (⫺) (95) CCB use (⫹) (181) (⫺) (84)
PAC
PRA
ARR
Aldo
14.0 ⫾ 9.7 13.1 ⫾ 8.9
4.1 ⫾ 8.9 3.0 ⫾ 6.2
15.6 ⫾ 14.6 14.4 ⫾ 14.8
11.5 ⫾ 9.3 11.5 ⫾ 9.1
13.2 ⫾ 8.9 14.0 ⫾ 9.7
3.8 ⫾ 7.8 3.4 ⫾ 7.9
15.1 ⫾ 15.2 15.0 ⫾ 14.3
11.9 ⫾ 9.8 11.1 ⫾ 8.7
13.9 ⫾ 9.5 12.6 ⫾ 8.7
4.0 ⫾ 8.5* 1.6 ⫾ 2.7
14.4 ⫾ 14.4 17.4 ⫾ 15.7
11.7 ⫾ 9.6 10.7 ⫾ 7.4
13.6 ⫾ 8.8 13.6 ⫾ 10.3
2.9 ⫾ 7.1 4.8 ⫾ 8.9
16.7 ⫾ 14.8* 12.1 ⫾ 14.5
11.9 ⫾ 10.2 10.7 ⫾ 7.1
13.9 ⫾ 9.8 13.0 ⫾ 8.4
3.7 ⫾ 7.3 3.5 ⫾ 8.9
14.2 ⫾ 14.8 16.9 ⫾ 14.4
11.7 ⫾ 8.6 11.1 ⫾ 10.4
ACE ⫽ angiotensin converting enzyme; Aldo ⫽ urinary aldosterone (g/24-h); ARB ⫽ angiotensin receptor blocker; ARR ⫽ PAC to PRA ratio; CCB ⫽ calcium channel blocker; PAC ⫽ plasma aldosterone concentration (ng/dL); PRA ⫽ plasma renin activity (ng/mL/h). Values are mean ⫾ SD. * Significantly different from subjects who are not taking the medication, P ⬍ .05.
810
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
In most prior studies of the diagnostic value of the ARR, the study population had mild-to-moderate hypertension or included normotensive controls, allowing for withdrawal of medications before evaluation. Because some classes of agents affect PRA and/or aldosterone excretion, the diagnostic value of the ARR undoubtedly differs according to whether medications have been discontinued. Although it is admittedly preferable to evaluate each subject after withdrawal from antihypertensive therapy, in subjects with poorly controlled hypertension (as in the current study), withdrawal of therapy is often not possible. Accordingly, it is important clinically to know the test characteristics of the ARR in the setting of continued medication use. As has been shown by other investigators, diuretic use in the current study was associated with greater PRA and a reduced ARR. Aldosterone excretion was not different with or without diuretic use. Accordingly, a diuretic-related increase in PRA ⬎1.0 ng/mL/h (our limit for diagnosis of PA) in subjects with true PA would have resulted in a false-negative evaluation. If assessment of the 24-h urinary aldosterone excretion in this setting was normal (⬍12 g/24 h by our criteria), PA would be excluded regardless of the PRA value. However, observation of a high PRA during diuretic use in the setting of high urinary aldosterone excretion could represent a falsely high PRA value, and repetition of the biochemical evaluation after cessation of the diuretic would be necessary to exclude PA. Use of ACE inhibitors was associated with higher PRA values. Urinary aldosterone excretion was not different in relation to ACE inhibitor use, suggesting that, as with diuretics, a normal urinary aldosterone excretion would exclude PA, but exclusion of PA in the setting of a high PRA and urinary aldosterone would require repeat testing after withdrawal of the ACE inhibitor. Biochemical parameters were not different with or without use of CCB, suggesting that chronic use of such agents does not preclude assessment of PA. The PRA was lower and the ARR higher in subjects receiving -blockers, whereas urinary aldosterone excretion was not different. Accordingly, -blocker use was associated with a higher ARR false-positive rate. A high rate of -blocker use in the current study cohort (64%) undoubtedly increased the overall ARR false-positive rate, emphasizing the need to confirm the diagnosis of PA with suppression testing, particularly in the setting of chronic -blocker use. It is also possible that in the current study greater use of -blockers in white subjects minimized ethnic differences in ARR test characteristics. Multiple other reports1,13,27,28 suggest that continuation of antihypertensive medications (except spironolactone) does not preclude accurate assessment for PA as long as suppression testing is done to confirm the diagnosis. Muratero et al27 recently evaluated single drug effects on PRA and PAC in subjects with high baseline PAC values. Fosinopril, doxazosin, and amlodipine increased PRA while significantly reducing the ARR. Atenolol suppressed
AJH–June 2005–VOL. 18, NO. 6
PRA while increasing the ARR. All of the drugs induced modest suppression of PAC. Despite these divergent effects on the PRA and ARR, all of the agents were associated with a low rate of false-negative ARR, whereas atenolol was associated with a higher rate of false-positive ARR. These results are consistent with our findings in subjects with resistant hypertension in that a negative ARR had a high predictive value and in that -blockers are more often associated with falsely high ARR, increasing the need for confirmatory testing. Our results differ in that we found in general a high false-positive ARR rate. This is undoubtedly a consequence of using a lower cut-off for ARR (20 v 50), but it also may reflect a high proportion of our subjects receiving -blockers or a higher proportion of low-renin hypertension in subjects with resistant hypertension. Importantly, the current study design does not exclude the possibility that concomitant use of two or more medications affected renin-aldosterone values in ways opposite to what would be observed with single medicine use. For this reason, the validity of the ARR in the setting of multidrug regimens is not known, and its interpretation must be made cautiously. The current study was done in an effort to help with that interpretation, as analysis of single medication use would be difficult to accomplish in subjects with resistant hypertension. If the goal of screening for a disease is a high sensitivity and high negative predictive value (ie, a negative result reliably rules out the presence of the disease), then in a cohort of subjects with resistant hypertension, an ARR cut-off of 20 was appropriate (in the context of a minimum PRA of 0.6 mg/mL/h). Higher cut-off levels of ARR increased the specificity of the test but at the expense of lowering the sensitivity. The higher the ARR cut-off, the less likely a disease negative patient will have a falsepositive screening result; but the higher cut-off will result in a greater number of patients being falsely labeled as disease free. Requiring a high ARR (⬎20) in conjunction with a minimal PAC (⬎15 ng/dL) had a similar overall effect. Specificity was improved but with loss of sensitivity (Fig. 2). The current results indicate that an elevated ARR is equally valid as a screening test in both African American and white subjects. Before this evaluation, no study had specifically determined the test characteristics of the ARR in a population that included a large number of African American subjects. This is an important determination, as we have recently reported that PA is as common in African Americans with resistant hypertension as in individuals of white ethnicity. The sensitivity of an elevated ARR was somewhat less in African American compared with white subjects, but this small difference in sensitivity did not alter the negative predictive value; so, as a screening test, a high ARR was equally valid regardless of ethnicity. Our approach in diagnosing PA is to rely on a sodium excretion of ⬎200 mEq/24 h during the subject’s ad libitum diet to indicate chronic dietary salt loading. This approach assumes that additional dietary salt loading in
AJH–June 2005–VOL. 18, NO. 6
this setting would not suppress aldosterone excretion sufficient to change the diagnosis of PA in some subjects. Such an assumption was validated in showing that aldosterone excretion remained high in 12 subjects, even with additional salt loading. Clinically, we think that this is important, as this method allows maximizing patient safety and convenience by minimizing the number of patients subjected to additional salt loading. Overall, the current results suggest that the ARR can be used to screen for PA during chronic administration of the most commonly used classes of antihypertensive agents, in that a low ratio has a high probability of excluding the diagnosis. A high ARR in the setting of diuretic, ACE inhibitor, and/or ARB use increases suspicion of aldosterone excess, as these medications tend to increase PRA, and thereby lower the ARR. -Blockers can falsely increase the ARR, necessitating confirmatory testing to exclude PA. The current results also emphasize that a high ratio is not reliably diagnostic and that suppression testing is necessary to confirm PA. Such confirmatory testing is best done in the absence of any medications; but if that is not possible, as is often the case in subjects with resistant hypertension, use of a regimen (such as a combination of verapamil, hydralazine, and prazosin) that is unlikely to have a significant effect on renin and/or aldosterone levels is preferred. It should be emphasized that the current analysis included only subjects with resistant hypertension. In patients with less severe hypertension, screening for PA is best done after withdrawal of antihypertensive medications to maximize sensitivity and specificity. With increasing pressure on clinicians to reduce costs, particularly through reduced hospitalizations, an effective outpatient strategy for the diagnosis of PA is needed. Based on the current results, use of a screening ambulatory ARR and confirmatory testing with assessment of 24-h urinary aldosterone excretion during high dietary salt ingestion represents one such approach. We believe that confirming the diagnosis of PA is appropriate as opposed to simply adding an aldosterone antagonist, as it allows consideration of surgical treatment with adrenalectomy and, when medical therapy is used, guides titration of the aldosterone antagonist. That is, we more often titrate to a higher dose of an aldosterone antagonist in subjects with confirmed PA compared with treating resistant hypertension in the absence of PA.29
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19.
References 1.
2.
3.
Gallay BJ, Ahmad S, Xu L, Toivola B, Davidson RC: Screening for primary aldosteronism without discontinuing hypertensive medications: plasma aldosterone-renin ratio. Am J Kidney Dis 2001;37: 699 –705. Gordon RD, Stowasser M, Tunny TJ, Klemm SA, Rutherford JC: High incidence of primary aldosteronism in 199 patients referred with hypertension. Clin Exp Pharmacol Physiol 1994;21:315–318. Gordon RD, Ziesak MD, Tunny TJ, Stowasser M, Klemm SA: Evidence that primary aldosteronism may not be uncommon: 12%
20.
21. 22.
23.
811
incidence among antihypertensive drug trail volunteers. Clin Exp Pharmacol Physiol 1993;20:296 –298. Fardella CE, Mosso L, Gómez-Sánchez C, Cortés P, Pinto M, Huete A, Oestreicher E, Foradori A, Montero J: Primary hyperaldosteronism in essential hyperensives: prevalence, biochemical profile, and molecular biology. J Clin Endocrinol Metab 2000;85:1863–1867. Rayner BL, Opie LH, Davidson JS: The aldosterone/renin ratio as a screening test for primary aldosteronism. S Afr Med J 2000;90:394 – 400. Rayner BL, Myers JE, Opie LH, Trinder YA, Davidson JS: Screening for primary aldosteronism—normal ranges for aldosterone and renin in three South African population groups. S Afr Med J 2001;91:594 –599. Lim PO, Dow E, Brennan G, Jung RT, MacDonald TM: High prevalence of primary aldosteronism in the Tayside hypertension clinic population. J Hum Hypertens 2000;14:311–315. Lim PO, Rodgers P, Cardale K, Watson AD, MacDonald TM: Potentially high prevalence of primary aldosteronism in a primarycare population (letter). Lancet 1999;353:40. Loh K-C, Koay ES, Khaw M-C, Emmanuel SC, Young WF Jr: Prevalence of primary aldosteronism among Asian hypertensive patients in Singapore. J Clin Endocrinol Metab 2000;85:2854 – 2859. Calhoun DA, Nishizaka MK, Zaman MA, Thakkar RB, Weissman P: Hyperaldosteronism among black and white subjects with resistant hypertension. Hypertension 2003;40:892– 896. Hiramatsu K, Yamada T, Yukimura Y, Komiya I, Ichikawa K, Ishihara M, Nagata H, Izumiyama T: A screening test to identify aldosterone-producing adenoma by measuring plasma renin activity. Arch Intern Med 1981;141:1589 –1593. Young WF Jr: Minireview: Primary aldosteronism– changing concepts in diagnosis and treatment. Endocrinology 2003;144:2208 – 2213. Weinberger MH, Fineberg NS: The diagnosis of primary aldosternism and separation of two major subtypes. Arch Intern Med 1993; 153:2125–2129. Trenkel S, Seifarth C, Schobel H, Hahn EG, Hensen J: Ratio of serum aldosterone to plasma renin concentration in essential hypertension and primary aldosteronism. Exp Clin Endocrinol Diabetes 2002;110:80 – 85. Soule S, Davidson JS, Rayner BL: The evaluation of primary aldosteronism. S Afr Med J 2000;90:387–394. McKenna TJ, Sequiera SJ: Diagnosis under random conditions of all disorders of renin-angiotensin-aldosterone axis, including primary aldosteronism. J Clin Endocrinol Metab 1991;73:952–957. Stowasser M, Gordon RD, Gunasekera TG, Cowley DC, Ward G, Archibald C, Smithers BM: High rate of detection of primary aldosteronism, including surgically treatable forms, after ‘non-selective’ screening of hypertensive patients. J Hypertens 2003;21: 2149 –2157. Mosso L, Carvajal C, González A, Barraza A, Avila F, Montero J, Huete A, Gederlini A, Fardella CE: Primary aldosteronism and hypertensive disease. Hypertension 2003;42:161–165. Falkner B: Differences in blacks and whites with essential hypertension: biochemistry and endocrine. State of the art lecture. Hypertension 1990;15:681– 686. Pratt JH, Rebhun JF, Zhou L, Abbrosius WT, Newman SA, GomezSanchez CE, Mayes DF: Levels of mineralocorticoids in whites and blacks. Hypertension 1999;34:315–319. Thakkar RB, Oparil S: Primary aldosteronism: a practical approach to diagnosis and treatment. J Clin Hypertens 2001;3:189 –195. Ignatowska-S´witalska H, Chodakowska J: Evaluation of plasma aldosterone to plasma renin activity ratio in patients with primary aldosteronism. J Hum Hypertens 1997;11:373–378. Lyons DF, Kem DC: Single dose captoril as a diagnostic test for primary aldosteronism. J Clin Endocrinol Metab 1983;57:892– 896.
812
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
24. Rossi E, Regolisti G, Negro A, Sani C, Davoli S, Perazzoli F: High prevalence of primary aldosteronism using postcaptopril plasma aldosterone to renin ratio as a screening test among Italian hypertensives. Am J Hypertens 2002;15:896 –902. 25. Streeten DHP, Tomycz N, Anderson GH: Reliability of screening methods for the diagnosis of primary aldosteronism. Am J Med 1979;67:403– 413. 26. Holland OB, Brown H, Kuhnert L, Fairchild C, Risk M, GomezSanchez G: Further evaluation of saline infusion for the diagnosis of primary aldosteronism. Hypertension 1984;6:717–723.
AJH–June 2005–VOL. 18, NO. 6
27. Muratero P, Rabbia F, Milan A, Paglieri C, Morello F, Chiandussi L, Veglio F: Drug effect on aldosterone/plasma renin activity ratio in primary aldosteronism. Hypertension 2002;40:897–902. 28. Seifarth C, Trenkel S, Schobel H, Hahn EG, Hensen J: Influence of antihypertensive medication on aldosterone and renin concentration in the differential diagnosis of essential hypertension and primary aldosteronism. Clin Endoclinol 2002;57:457– 465. 29. Nishizaka MK, Zaman MA, Calhoun DA: Efficacy of low-dose spironolactone in subjects with resistant hypertension. Am J Hypertens 2003;16:925–930.
2005; 18:805– 812
Validity of Plasma Aldosterone-to-Renin Activity Ratio in African American and White Subjects With Resistant Hypertension Mari K. Nishizaka, Monique Pratt-Ubunama, Mohammad A. Zaman, Stacey Cofield, and David A. Calhoun Background: Recent reports suggesting that primary aldosteronism (PA) is more common than historically thought have often relied on use of the plasma aldosterone concentration (PAC) to plasma renin activity (PRA) ratio (ARR) to identify patients with PA. Prior determinations of the validity of the ARR had been generally limited to subjects that could be withdrawn from antihypertensive therapy and to non–African American subjects. Methods and Results: The current study was designed to evaluate prospectively the diagnostic value of the ARR in treated African American and white subjects with resistant hypertension. Consecutive subjects referred to a university hypertension clinic for resistant hypertension were evaluated with an early morning ARR and a 24-h urinary aldosterone and sodium. The presence of PA was defined as a suppressed PRA (⬍1.0 ng/mL/h) and elevated urinary aldosterone excretion (⬎12 g/24 h) dur-
ing high dietary sodium ingestion (⬎200 mEq/24 h). In 58 subjects, PA was confirmed. The ARR was elevated (⬎20) in 45 of 58 subjects with PA and in 35 of the 207 patients without PA, resulting in a sensitivity of 78% and specificity of 83% with a corresponding positive predictive value of 56% and a negative predictive value of 93%. Among African American subjects, the ARR was less sensitive than in white subjects (75% v 80%), but it still had a high negative predictive value (92% v 94%). Conclusions: These data indicate that the ARR is valid as a screening test for PA in African American and white patients on stable antihypertensive treatments, but a high percentage of false-positive results precludes using it for accurate diagnosis of PA. Am J Hypertens 2005;18: 805– 812 © 2005 American Journal of Hypertension, Ltd. Key Words: Primary aldosteronism, aldosterone to renin ratio (ARR), African American, resistant hypertension.
ecent reports1–10 indicate that primary aldosteronism (PA) is common in both general and selected hypertensive populations, with reported prevalence rates ranging from 7% to 20%. One reason for the recognition of the higher prevalence of PA is that use of the plasma aldosterone concentration (PAC)/plasma renin activity (PRA) ratio (ARR) has facilitated broader screening of hypertensive subjects. Since Hiramatsu et al11 described the utility of the ARR to screen for PA, it has been increasingly used to identify subjects at high risk for PA.1–9 Primary aldosteronism is suggested by a high ARR. The cuff-off level is arbitrary, ranging from 20 to 100 (ng/ dl)/(ng/mL/h). Some recommendations on use of the ARR to screen for PA suggest an elevated ratio in conjunction with a minimally high PAC.4,5,9,12,13 Requiring a higher
R
ratio or a PAC of a certain level improves the specificity (fewer false-positive results) but lowers the sensitivity (more false-negative results) of the screening test. The validity of ARR has been assessed in multiple studies, using various methods to confirm the diagnosis of PA. It is generally agreed that failure to suppress aldosterone excretion during volume expansion remains the gold standard for confirmation of PA. The most commonly used methods to achieve volume expansion are fludrocortisone suppression testing (FST), intravenous saline infusion, or dietary oral salt loading. Prior reports of determination of ARR test characteristics have generally been done in patients with mild to moderate hypertension (allowing for discontinuation of therapy before evaluation). In that setting, an elevated
Received April 13, 2004. First decision January 12, 2005. Accepted January 18, 2005. From the Vascular Biology and Hypertension Program (MKN, MP-U, MAZ, DAC) and Department of Biostatistics (SC), University of Alabama at Birmingham, Birmingham, Alabama. This work was supported by American Heart Association postdoctoral
fellowship 0425389B (M.K.N.), American Heart Association Grant-in-Aid 0355302B (to DAC), National Heart, Lung, and Blood Institute Grants HL075614 (to DAC) and HL07457 (to MAZ and MP-U). Address correspondence and reprint requests to Dr. Mari K. Nishizaka, 115 CHSB, 933 South 19th Street, Birmingham, AL 352942041; e-mail: [email protected]
© 2005 by the American Journal of Hypertension, Ltd. Published by Elsevier Inc.
0895-7061/05/$30.00 doi:10.1016/j.amjhyper.2005.01.002
806
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
ARR has been reported to have a high sensitivity and specificity in identifying patients with PA.11,13–17 What remains unknown is the validity of the ARR to screen for PA in patients who are unable to discontinue antihypertensive therapy. Knowing the diagnostic value of ARR in the setting of ongoing antihypertensive treatment is particularly relevant as the prevalence of PA, as recently documented by this laboratory, is high in subjects with resistant hypertension.10,18 Individuals of black ethnicity (South-African individuals of black ethnicity or African Americans) are more likely to have low renin levels compared with persons of white ethnicity.6,19,20 Despite this tendency toward lower renin activity, perhaps suggesting a greater prevalence of aldosterone-associated hypertension, we have reported that in subjects referred to a specialty clinic for resistant hypertension, African Americans have a similar prevalence of PA compared with white subjects.10 What has not been previously determined is the validity of the ARR to screen for PA in African American subjects. With generally lower renin levels but similar rates of PA (at least in patients with resistant hypertension), it might be suspected that the ARR is less specific in African American versus white subjects. The current study was designed to determine prospectively the validity of an elevated ARR to screen for PA in African American and white subjects with resistant hypertension while these individuals were maintained on a stable antihypertensive regimen. Outpatient dietary salt loading was used to confirm the diagnosis of PA in all evaluated subjects.
AJH–June 2005–VOL. 18, NO. 6
outpatient basis (Fig. 1). Blood samples were collected using routine laboratory procedure including use of a tourniquet and vacutainer in subjects who had been ambulatory for at least 1 h. Hemolyzed serum was not used for analysis. Initial evaluation included an early morning PRA and PAC (when PRA tends to be lowest and PAC tends to be highest), and a 24-h urine collection for sodium and aldosterone during the subject’s ad libitum diet. If the urinary aldosterone was elevated (⬎12 g/24 h) but the urinary sodium was ⬍200 mEq/24 h, the 24-h urinary assessments were repeated after 3 days of dietary salt supplementation. A suppressed PRA (⬍1.0 ng/mL/h), and elevated urinary aldosterone (⬎12 g/24 h) in the setting of high dietary sodium ingestion (⬎200 mEq/24 h) confirmed the diagnosis of PA.12,21 To evaluate the validity of relying on high urinary sodium excretion during the subject’s ad libitum diet to indicate chronic dietary salt loading, we remeasured urinary aldosterone excretion after dietary salt supplementation (50 mEq NaCl/day) in 12 subjects whose ad libitum urinary sodium excretion was ⬎200 mEq/24 h and whose aldosterone excretion was high (⬎12 g/24 h). Laboratory Methods
Subjects
Levels of PAC, PRA, urinary aldosterone, and urinary sodium were measured by commercial laboratories using standard techniques. The PAC (Quest Diagnostics, Atlanta, GA) and PRA (Mayo Clinic Laboratories, Rochester, MN) levels were measured by radioimmunoassay. The reference range for PAC is 4.0 to 31.0 ng/dL. The reference range for an upright PRA is 1.31 to 3.95 ng/mL/h. The minimum reported PRA by this laboratory is 0.6 ng/mL/h. Urinary
We prospectively evaluated consecutive subjects referred for resistant hypertension to the University of Alabama at Birmingham Hypertension Clinic over a 44-month period (December 2000 to July 2004). The protocol was reviewed and approved by the Institutional Review Board and written informed consent was obtained from all participants. Resistant hypertension was defined as uncontrolled hypertension (⬎140/90 mm Hg) at two or more clinic visits, despite use of three or more antihypertensive medications at pharmacologically effective doses. All subjects were on a stable antihypertensive regimen for at least 4 weeks before biochemical evaluation. No medications were discontinued before evaluation except for spironolactone, triamterene, or amiloride, which were discontinued for at least 6 weeks before evaluation. Blood pressure was also measured by 24-h ambulatory blood pressure (BP) monitoring (SpaceLabs 90207, Redmond, WA). Serum potassium levels were corrected with oral supplementation to be ⬎3.5 mEq/L before evaluation, if necessary. Secondary hypertension, such as renovascular hypertension, pheochromocytoma, or Cushing syndrome, was excluded as clinically indicated. Subjects on chronic glucocorticoid therapy were not included in the analysis. Biochemical evaluation was done in all subjects on an
FIG. 1. Flow chart for the diagnostic evaluation for primary aldosteronism (PA). Na ⫽ sodium; PAC ⫽ plasma aldosterone concentration; PRA ⫽ plasma renin activity.
Methods
AJH–June 2005–VOL. 18, NO. 6
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
aldosterone was measured by radioimmunoassay (Mayo Clinic Laboratories, Rochester, MN). The reference range for urinary aldosterone is 2 to 16 g/24 h. Statistical Analysis Values are expressed as mean ⫾ SD. A PRA value ⬍0.6 ng/ mL/h was treated as 0.6 ng/mL/h. Values between groups were compared by the Student t test, Mann-Whitney test, and 2 test. A P value ⬍ .05 was considered significant. Receiver operator characteristic (ROC) analysis was used to determine the test characteristics of the ARR.
Results A total of 265 subjects (115 African American and 150 white) with resistant hypertension were evaluated. Demographic values of the evaluated subjects listed in Table 1. Of the total subjects, 58% had low-renin hypertension, defined as a PRA ⬍1.0 ng/mL/h. The majority of the subjects were receiving an angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB), a -blocker, a calcium channel blocker (CCB), and/or a diuretic. Compared with their white counterparts, African American subjects were younger (53 ⫾ 12 v 59 ⫾ 11 years), heavier (body mass index 34.4 ⫾ 8.5 v 30.8 ⫾ 6.4 kg/m2), and had higher clinic (91 ⫾ 18 v 87 ⫾ 15 mm Hg) and mean 24-h ambulatory diastolic BP levels (88 ⫾ 13 v 82 ⫾ 13 mm Hg). Systolic BP values were not different, either in the clinic or during ambulatory BP monitoring, in African American and white subjects (Table 1). African Americans were less likely to be on a -blocker and more likely to be on a calcium channel
807
blocker than were white subjects, but the total number of antihypertensive medications was not different between the two groups (3.9 ⫾ 1.1 v 3.8 ⫾ 1.3). The mean PRA was similar in the two ethnic groups, but the mean PAC tended to be lower in African American subjects (12.5 ⫾ 8.1 v 14.5 ⫾ 10.2 ng/dL). The mean ARR was not different in African American versus white subjects. The prevalence of the low-renin hypertension (PRA ⬍1.0 mg/mL/h) was also similar in African American and white subjects (62% v 55%) (Table 2). Of the 265 subjects, 58 (28 African American and 30 white) were confirmed to have PA. Of these, 23 of the 58 PA (40%) and 31 of the 207 non-PA subjects (15%) were hypokalemic (⬍3.6 mEq/L) on presentation to our clinic. In 31 of the 58 subjects confirmed to have PA, diagnoses were made based on a high urinary aldosterone (range 13 to 55 g/24 h) and sodium (range 202 to 363 mEq/24 h) excretion during their normal diet. In the remaining 27 subjects, diagnoses were made after subjects had repeated the 24-h urinary collection after 3 days of dietary salt supplementation. Six patients were excluded from having PA after dietary sodium loading. Serum potassium levels were monitored in all subjects who underwent dietary salt loading. If the subject become hypokalemic with salt loading and the urinary aldosterone level was not elevated (⬎12 g/ 24 h), the urinary aldosterone assessment was repeated with potassium supplementation sufficient to maintain normokalemia. It is still possible, however, that reductions in serum potassium to less than hypokalemic range may have affected aldosterone secretion. In 12 subjects with high dietary sodium (259 ⫾ 46 mEq/ 24 h) and aldosterone excretion (21.3 ⫾ 6.2 g/24 h)
Table 1. Baseline characteristics of subjects during ad libitum diet Characteristic
Total
African American
White
N Age (y) Male/female Body mass index (kg/m2) Clinic BP (mm Hg) Systolic Diastolic ABPM, 24-h Systolic Diastolic Number of antihypertensive medications Medication use (% of subjects) ACE inhibitor ARB Diuretic -Blocker CCB
265 56 ⫾ 12 116/149 32.4 ⫾ 7.4
115 53 ⫾ 12 45/70 34.4 ⫾ 8.5
150 59 ⫾ 11* 71/79 30.8 ⫾ 6.4*
160 ⫾ 27 89 ⫾ 16
160 ⫾ 27 91 ⫾ 18
160 ⫾ 27 87 ⫾ 15*
143 ⫾ 18 85 ⫾ 13
145 ⫾ 20 88 ⫾ 13
142 ⫾ 16 82 ⫾ 13*
3.8 ⫾ 1.2
3.9 ⫾ 1.1
3.8 ⫾ 1.3
55% 47% 82% 64% 68%
55% 41% 85% 57% 76%
55% 52% 79% 70%* 63%*
ABPM ⫽ ambulatory blood pressure monitor; ACE ⫽ angiotensin converting enzyme; ARB ⫽ angiotensin receptor blocker; BP ⫽ blood pressure; CCB ⫽ calcium channel blocker. Values are mean ⫾ SD. * Significantly different from African American subjects, P ⬍ .05.
808
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
AJH–June 2005–VOL. 18, NO. 6
Table 2. Biochemical parameters and test characteristics of an elevated plasma aldosterone/plasma renin activity ratio (ARR) Characteristic N PAC (ng/dL) PRA (ng/mL/h) ARR Serum potassium (mEq/L) Urinary sodium (mEq/24-h) Urinary aldosterone (g/24-h) Primary aldosteronism (%) Test characteristics of ARR ⬎20 Sensitivity (%) Specificity (%) ⫹PV (%) ⫺PV (%)
Total 13.6 3.5 15.9 4.0 184 11.5
265 ⫾ 9.3 (1–49) ⫾ 6.6 (0.6–55.5) ⫾ 15.7 (0.1–78.3) ⫾ 0.5 (2.4–5.5) ⫾ 79 (46–442) ⫾ 9.2 (1.0–78) 58 (22%) 78 83 56 93
African American
White
115 12.5 ⫾ 8.1 3.7 ⫾ 9.2 14.0 ⫾ 13.2 3.9 ⫾ 0.5 172 ⫾ 85 12.0 ⫾ 10.3 28 (24%)
150 14.5 ⫾ 10.2 3.5 ⫾ 6.6 15.9 ⫾ 15.7 4.0 ⫾ 0.5 194 ⫾ 74 11.1 ⫾ 8.3 30 (20%)
75 87 66 92
80 80 50 94
PAC ⫽ plasma aldosterone concentration; PRA ⫽ plasma renin activity; ⫹PV ⫽ positive predictive value; ⫺PV ⫽ negative predictive value. Values are mean ⫾ SD (range).
during their ad libitum diet, the 24-h urine collection was repeated after 3 days of additional dietary salt supplementation (50 mEq/day). In all 12 subjects, aldosterone excretion remained elevated (21.8 ⫾ 4.7 g/24 h) despite higher sodium excretion (353 ⫾ 31 mEq/24 h). This prevalence of PA was similar among African American and white subjects (24 v 20%, respectively). Thirteen of the 58 subjects with confirmed PA had a false-negative ARR (⬍20). Seven of these subjects were African American and six were white. Of the 207 hypertensive subjects without PA, 35 had an elevated ARR (⬎20). Accordingly, an ARR ⬎20 had an overall sensitivity of 78% and a specificity of 83%. This corresponds to a positive predictive value of 56% and a negative predictive value of 93%. The PRA was suppressed (⬍1.0 mg/ mL/h) in all subjects with an ARR ⬎20. An ARR⬎20 was less sensitive (75 v 80% respectively) but more specific (87 v 80%, respectively) in African American compared with white subjects (Table 2). Despite this difference in sensitivity, the negative predictive values of an elevated ARR were similar in African American and white subjects, indicating that in this clinical setting, an ARR ⬎20 is an equally effective screening test for African American and white subjects. We determined the test characteristics of different cut-off levels of the ARR (Fig. 2). Using a higher ARR as a screening test provided better specificity, but lowered the sensitivity. Requiring a high ARR (⬎50) as a positive screen resulted in a very high specificity and improved the positive predictive value, but at the expense of a substantially lower sensitivity and negative predictive value. Using ROC analysis, in this cohort of subjects being treated for resistant hypertension, an ARR of 16.3 provided the highest sensitivity and the highest negative predictive value, resulting in the greatest value as a screening test (Fig. 3). We also compared biochemical values in the presence or absence of different classes of antihypertensive agents (Table 3). Urinary aldosterone excretion was not different
in relation to use of the any of the evaluated classes of agents. The PRA was higher with use of diuretics and tended to be so with use of ACE inhibitors. The PRA and PAC were not different in relation to calcium channel blocker (CCB) use. The PRA tended to be lower in the setting of -blocker use resulting in significant increases in the ARR. As a consequence, the ARR in the setting of -blocker use was associated with a higher false positive rate (22% for -blockers v 14 to 17% for other classes of agents). However, the areas under the ROC curves were similar in the presence or absence of -blockers (84% v 92%, respectively) as were the best ARR cut-off values (17.0 v 16.3, respectively).
FIG. 2. Test characteristics of different levels of plasma aldosterone/plasma renin activity ratio (ARR) to identify primary aldosteronism in subjects being treated for resistant hypertension. ⫹PV ⫽ positive predictive value; ⫺PV ⫽ negative predictive value.
AJH–June 2005–VOL. 18, NO. 6
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
FIG. 3. Receiver operator characteristic analysis of plasma aldosterone/plasma renin activity ratio (ARR) to identify primary aldosteronism in subjects being treated for resistant hypertension.
Discussion This is the first prospective study to determine the validity of the ARR to screen for PA in a resistant hypertensive population consisting of a large percentage of African American subjects. In this study population of patients evaluated on a stable antihypertensive regimen, we found an ARR ⬎20 to have sufficient sensitivity and negative predictive value to be an effective screen for, but lacking the specificity to reliably confirm the diagnosis of PA. This was true of both African American and white subjects. In the current study cohort, to have relied on an elevated
809
ARR only as indication of PA would have resulted in 35 false-positive diagnoses among 265 evaluated subjects. Prior reports7,11,13–16,22–24 have evaluated the validity of the ARR. These studies have mostly reported an elevated ARR to have both a high sensitivity and specificity (test characteristics varying according to the cut-off level of ARR applied). In the current analysis, the sensitivity and specificity of the ARR was less than described in these earlier reports, leading us to conclude that in the setting of ongoing antihypertensive therapy, an elevated ARR (⬎20) is not sufficient to diagnose PA and must be confirmed with suppression testing. Reported differences in ARR sensitivity and specificity may be related to differences in study design. In the current study, the validity of an elevated ARR was compared with lack of suppression of 24-h urinary aldosterone excretion during high dietary salt ingestion. In other studies, fludrocortisone suppression (FST), saline infusion, and captopril suppression of plasma aldosterone were used as confirmatory testing. It is generally accepted that FST is the most accurate method of confirming the diagnosis PA.3,4,7 However, wider clinical use of FST is limited in that it generally requires a hospitalization of 4 to 5 days for both safety and accuracy reasons. Because dietary salt loading is more amenable to outpatient use, it is being increasingly used for confirmatory testing of PA, particularly within the United States. While intravenous saline infusion has been compared with FST,25,26 sometimes favorably, sometimes not, there has not been a systematic comparison of dietary salt loading to FST. We recognize this as a limitation of our study. Depending on the results of such a comparison, the relative test characteristics of the ARR with oral salt loading versus FST may differ.
Table 3. Biochemical values in relation to use of different classes of antihypertensive agents in total group of subjects Subgroup (n) ACE inhibitor use (⫹) (145) (⫺) (120) ARB use (⫹) (125) (⫺) (140) Diuretic use (⫹) (216) (⫺) (49) -Blocker use (⫹) (170) (⫺) (95) CCB use (⫹) (181) (⫺) (84)
PAC
PRA
ARR
Aldo
14.0 ⫾ 9.7 13.1 ⫾ 8.9
4.1 ⫾ 8.9 3.0 ⫾ 6.2
15.6 ⫾ 14.6 14.4 ⫾ 14.8
11.5 ⫾ 9.3 11.5 ⫾ 9.1
13.2 ⫾ 8.9 14.0 ⫾ 9.7
3.8 ⫾ 7.8 3.4 ⫾ 7.9
15.1 ⫾ 15.2 15.0 ⫾ 14.3
11.9 ⫾ 9.8 11.1 ⫾ 8.7
13.9 ⫾ 9.5 12.6 ⫾ 8.7
4.0 ⫾ 8.5* 1.6 ⫾ 2.7
14.4 ⫾ 14.4 17.4 ⫾ 15.7
11.7 ⫾ 9.6 10.7 ⫾ 7.4
13.6 ⫾ 8.8 13.6 ⫾ 10.3
2.9 ⫾ 7.1 4.8 ⫾ 8.9
16.7 ⫾ 14.8* 12.1 ⫾ 14.5
11.9 ⫾ 10.2 10.7 ⫾ 7.1
13.9 ⫾ 9.8 13.0 ⫾ 8.4
3.7 ⫾ 7.3 3.5 ⫾ 8.9
14.2 ⫾ 14.8 16.9 ⫾ 14.4
11.7 ⫾ 8.6 11.1 ⫾ 10.4
ACE ⫽ angiotensin converting enzyme; Aldo ⫽ urinary aldosterone (g/24-h); ARB ⫽ angiotensin receptor blocker; ARR ⫽ PAC to PRA ratio; CCB ⫽ calcium channel blocker; PAC ⫽ plasma aldosterone concentration (ng/dL); PRA ⫽ plasma renin activity (ng/mL/h). Values are mean ⫾ SD. * Significantly different from subjects who are not taking the medication, P ⬍ .05.
810
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
In most prior studies of the diagnostic value of the ARR, the study population had mild-to-moderate hypertension or included normotensive controls, allowing for withdrawal of medications before evaluation. Because some classes of agents affect PRA and/or aldosterone excretion, the diagnostic value of the ARR undoubtedly differs according to whether medications have been discontinued. Although it is admittedly preferable to evaluate each subject after withdrawal from antihypertensive therapy, in subjects with poorly controlled hypertension (as in the current study), withdrawal of therapy is often not possible. Accordingly, it is important clinically to know the test characteristics of the ARR in the setting of continued medication use. As has been shown by other investigators, diuretic use in the current study was associated with greater PRA and a reduced ARR. Aldosterone excretion was not different with or without diuretic use. Accordingly, a diuretic-related increase in PRA ⬎1.0 ng/mL/h (our limit for diagnosis of PA) in subjects with true PA would have resulted in a false-negative evaluation. If assessment of the 24-h urinary aldosterone excretion in this setting was normal (⬍12 g/24 h by our criteria), PA would be excluded regardless of the PRA value. However, observation of a high PRA during diuretic use in the setting of high urinary aldosterone excretion could represent a falsely high PRA value, and repetition of the biochemical evaluation after cessation of the diuretic would be necessary to exclude PA. Use of ACE inhibitors was associated with higher PRA values. Urinary aldosterone excretion was not different in relation to ACE inhibitor use, suggesting that, as with diuretics, a normal urinary aldosterone excretion would exclude PA, but exclusion of PA in the setting of a high PRA and urinary aldosterone would require repeat testing after withdrawal of the ACE inhibitor. Biochemical parameters were not different with or without use of CCB, suggesting that chronic use of such agents does not preclude assessment of PA. The PRA was lower and the ARR higher in subjects receiving -blockers, whereas urinary aldosterone excretion was not different. Accordingly, -blocker use was associated with a higher ARR false-positive rate. A high rate of -blocker use in the current study cohort (64%) undoubtedly increased the overall ARR false-positive rate, emphasizing the need to confirm the diagnosis of PA with suppression testing, particularly in the setting of chronic -blocker use. It is also possible that in the current study greater use of -blockers in white subjects minimized ethnic differences in ARR test characteristics. Multiple other reports1,13,27,28 suggest that continuation of antihypertensive medications (except spironolactone) does not preclude accurate assessment for PA as long as suppression testing is done to confirm the diagnosis. Muratero et al27 recently evaluated single drug effects on PRA and PAC in subjects with high baseline PAC values. Fosinopril, doxazosin, and amlodipine increased PRA while significantly reducing the ARR. Atenolol suppressed
AJH–June 2005–VOL. 18, NO. 6
PRA while increasing the ARR. All of the drugs induced modest suppression of PAC. Despite these divergent effects on the PRA and ARR, all of the agents were associated with a low rate of false-negative ARR, whereas atenolol was associated with a higher rate of false-positive ARR. These results are consistent with our findings in subjects with resistant hypertension in that a negative ARR had a high predictive value and in that -blockers are more often associated with falsely high ARR, increasing the need for confirmatory testing. Our results differ in that we found in general a high false-positive ARR rate. This is undoubtedly a consequence of using a lower cut-off for ARR (20 v 50), but it also may reflect a high proportion of our subjects receiving -blockers or a higher proportion of low-renin hypertension in subjects with resistant hypertension. Importantly, the current study design does not exclude the possibility that concomitant use of two or more medications affected renin-aldosterone values in ways opposite to what would be observed with single medicine use. For this reason, the validity of the ARR in the setting of multidrug regimens is not known, and its interpretation must be made cautiously. The current study was done in an effort to help with that interpretation, as analysis of single medication use would be difficult to accomplish in subjects with resistant hypertension. If the goal of screening for a disease is a high sensitivity and high negative predictive value (ie, a negative result reliably rules out the presence of the disease), then in a cohort of subjects with resistant hypertension, an ARR cut-off of 20 was appropriate (in the context of a minimum PRA of 0.6 mg/mL/h). Higher cut-off levels of ARR increased the specificity of the test but at the expense of lowering the sensitivity. The higher the ARR cut-off, the less likely a disease negative patient will have a falsepositive screening result; but the higher cut-off will result in a greater number of patients being falsely labeled as disease free. Requiring a high ARR (⬎20) in conjunction with a minimal PAC (⬎15 ng/dL) had a similar overall effect. Specificity was improved but with loss of sensitivity (Fig. 2). The current results indicate that an elevated ARR is equally valid as a screening test in both African American and white subjects. Before this evaluation, no study had specifically determined the test characteristics of the ARR in a population that included a large number of African American subjects. This is an important determination, as we have recently reported that PA is as common in African Americans with resistant hypertension as in individuals of white ethnicity. The sensitivity of an elevated ARR was somewhat less in African American compared with white subjects, but this small difference in sensitivity did not alter the negative predictive value; so, as a screening test, a high ARR was equally valid regardless of ethnicity. Our approach in diagnosing PA is to rely on a sodium excretion of ⬎200 mEq/24 h during the subject’s ad libitum diet to indicate chronic dietary salt loading. This approach assumes that additional dietary salt loading in
AJH–June 2005–VOL. 18, NO. 6
this setting would not suppress aldosterone excretion sufficient to change the diagnosis of PA in some subjects. Such an assumption was validated in showing that aldosterone excretion remained high in 12 subjects, even with additional salt loading. Clinically, we think that this is important, as this method allows maximizing patient safety and convenience by minimizing the number of patients subjected to additional salt loading. Overall, the current results suggest that the ARR can be used to screen for PA during chronic administration of the most commonly used classes of antihypertensive agents, in that a low ratio has a high probability of excluding the diagnosis. A high ARR in the setting of diuretic, ACE inhibitor, and/or ARB use increases suspicion of aldosterone excess, as these medications tend to increase PRA, and thereby lower the ARR. -Blockers can falsely increase the ARR, necessitating confirmatory testing to exclude PA. The current results also emphasize that a high ratio is not reliably diagnostic and that suppression testing is necessary to confirm PA. Such confirmatory testing is best done in the absence of any medications; but if that is not possible, as is often the case in subjects with resistant hypertension, use of a regimen (such as a combination of verapamil, hydralazine, and prazosin) that is unlikely to have a significant effect on renin and/or aldosterone levels is preferred. It should be emphasized that the current analysis included only subjects with resistant hypertension. In patients with less severe hypertension, screening for PA is best done after withdrawal of antihypertensive medications to maximize sensitivity and specificity. With increasing pressure on clinicians to reduce costs, particularly through reduced hospitalizations, an effective outpatient strategy for the diagnosis of PA is needed. Based on the current results, use of a screening ambulatory ARR and confirmatory testing with assessment of 24-h urinary aldosterone excretion during high dietary salt ingestion represents one such approach. We believe that confirming the diagnosis of PA is appropriate as opposed to simply adding an aldosterone antagonist, as it allows consideration of surgical treatment with adrenalectomy and, when medical therapy is used, guides titration of the aldosterone antagonist. That is, we more often titrate to a higher dose of an aldosterone antagonist in subjects with confirmed PA compared with treating resistant hypertension in the absence of PA.29
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19.
References 1.
2.
3.
Gallay BJ, Ahmad S, Xu L, Toivola B, Davidson RC: Screening for primary aldosteronism without discontinuing hypertensive medications: plasma aldosterone-renin ratio. Am J Kidney Dis 2001;37: 699 –705. Gordon RD, Stowasser M, Tunny TJ, Klemm SA, Rutherford JC: High incidence of primary aldosteronism in 199 patients referred with hypertension. Clin Exp Pharmacol Physiol 1994;21:315–318. Gordon RD, Ziesak MD, Tunny TJ, Stowasser M, Klemm SA: Evidence that primary aldosteronism may not be uncommon: 12%
20.
21. 22.
23.
811
incidence among antihypertensive drug trail volunteers. Clin Exp Pharmacol Physiol 1993;20:296 –298. Fardella CE, Mosso L, Gómez-Sánchez C, Cortés P, Pinto M, Huete A, Oestreicher E, Foradori A, Montero J: Primary hyperaldosteronism in essential hyperensives: prevalence, biochemical profile, and molecular biology. J Clin Endocrinol Metab 2000;85:1863–1867. Rayner BL, Opie LH, Davidson JS: The aldosterone/renin ratio as a screening test for primary aldosteronism. S Afr Med J 2000;90:394 – 400. Rayner BL, Myers JE, Opie LH, Trinder YA, Davidson JS: Screening for primary aldosteronism—normal ranges for aldosterone and renin in three South African population groups. S Afr Med J 2001;91:594 –599. Lim PO, Dow E, Brennan G, Jung RT, MacDonald TM: High prevalence of primary aldosteronism in the Tayside hypertension clinic population. J Hum Hypertens 2000;14:311–315. Lim PO, Rodgers P, Cardale K, Watson AD, MacDonald TM: Potentially high prevalence of primary aldosteronism in a primarycare population (letter). Lancet 1999;353:40. Loh K-C, Koay ES, Khaw M-C, Emmanuel SC, Young WF Jr: Prevalence of primary aldosteronism among Asian hypertensive patients in Singapore. J Clin Endocrinol Metab 2000;85:2854 – 2859. Calhoun DA, Nishizaka MK, Zaman MA, Thakkar RB, Weissman P: Hyperaldosteronism among black and white subjects with resistant hypertension. Hypertension 2003;40:892– 896. Hiramatsu K, Yamada T, Yukimura Y, Komiya I, Ichikawa K, Ishihara M, Nagata H, Izumiyama T: A screening test to identify aldosterone-producing adenoma by measuring plasma renin activity. Arch Intern Med 1981;141:1589 –1593. Young WF Jr: Minireview: Primary aldosteronism– changing concepts in diagnosis and treatment. Endocrinology 2003;144:2208 – 2213. Weinberger MH, Fineberg NS: The diagnosis of primary aldosternism and separation of two major subtypes. Arch Intern Med 1993; 153:2125–2129. Trenkel S, Seifarth C, Schobel H, Hahn EG, Hensen J: Ratio of serum aldosterone to plasma renin concentration in essential hypertension and primary aldosteronism. Exp Clin Endocrinol Diabetes 2002;110:80 – 85. Soule S, Davidson JS, Rayner BL: The evaluation of primary aldosteronism. S Afr Med J 2000;90:387–394. McKenna TJ, Sequiera SJ: Diagnosis under random conditions of all disorders of renin-angiotensin-aldosterone axis, including primary aldosteronism. J Clin Endocrinol Metab 1991;73:952–957. Stowasser M, Gordon RD, Gunasekera TG, Cowley DC, Ward G, Archibald C, Smithers BM: High rate of detection of primary aldosteronism, including surgically treatable forms, after ‘non-selective’ screening of hypertensive patients. J Hypertens 2003;21: 2149 –2157. Mosso L, Carvajal C, González A, Barraza A, Avila F, Montero J, Huete A, Gederlini A, Fardella CE: Primary aldosteronism and hypertensive disease. Hypertension 2003;42:161–165. Falkner B: Differences in blacks and whites with essential hypertension: biochemistry and endocrine. State of the art lecture. Hypertension 1990;15:681– 686. Pratt JH, Rebhun JF, Zhou L, Abbrosius WT, Newman SA, GomezSanchez CE, Mayes DF: Levels of mineralocorticoids in whites and blacks. Hypertension 1999;34:315–319. Thakkar RB, Oparil S: Primary aldosteronism: a practical approach to diagnosis and treatment. J Clin Hypertens 2001;3:189 –195. Ignatowska-S´witalska H, Chodakowska J: Evaluation of plasma aldosterone to plasma renin activity ratio in patients with primary aldosteronism. J Hum Hypertens 1997;11:373–378. Lyons DF, Kem DC: Single dose captoril as a diagnostic test for primary aldosteronism. J Clin Endocrinol Metab 1983;57:892– 896.
812
VALIDITY OF ALDOSTERONE-TO-RENIN RATIO
24. Rossi E, Regolisti G, Negro A, Sani C, Davoli S, Perazzoli F: High prevalence of primary aldosteronism using postcaptopril plasma aldosterone to renin ratio as a screening test among Italian hypertensives. Am J Hypertens 2002;15:896 –902. 25. Streeten DHP, Tomycz N, Anderson GH: Reliability of screening methods for the diagnosis of primary aldosteronism. Am J Med 1979;67:403– 413. 26. Holland OB, Brown H, Kuhnert L, Fairchild C, Risk M, GomezSanchez G: Further evaluation of saline infusion for the diagnosis of primary aldosteronism. Hypertension 1984;6:717–723.
AJH–June 2005–VOL. 18, NO. 6
27. Muratero P, Rabbia F, Milan A, Paglieri C, Morello F, Chiandussi L, Veglio F: Drug effect on aldosterone/plasma renin activity ratio in primary aldosteronism. Hypertension 2002;40:897–902. 28. Seifarth C, Trenkel S, Schobel H, Hahn EG, Hensen J: Influence of antihypertensive medication on aldosterone and renin concentration in the differential diagnosis of essential hypertension and primary aldosteronism. Clin Endoclinol 2002;57:457– 465. 29. Nishizaka MK, Zaman MA, Calhoun DA: Efficacy of low-dose spironolactone in subjects with resistant hypertension. Am J Hypertens 2003;16:925–930.