Mean ± SD change from baseline in systolic and diastolic blood pressure (BP) over 24 weeks of treatment. With respect to systolic BP reduction, P<.001 for the noninferiority hypothesis, indicating strong certainty that systolic BP reduction with eplerenone lies within the 6–mm Hg margin of systolic BP reduction with amlodipine. Furthermore, P = .83 for the customary comparison of 2 groups, consistent with the hypothesis of similar systolic BP reductions.
Change in generalized symptom distress presented as mean ± SEM. The difference in response to amlodipine, which showed a negative change, and eplerenone, which showed a net positive (better) change was statistically significant (P = .03). Note that the average erosion associated with amlodipine was almost 0.1 SD negative, and the average improvement associated with eplerenone was almost 0.1 SD.
Symptom distress change for individual symptoms associated with amlodipine and eplerenone treatment. Note the positive (rightward) changes reflect reduced or improved distress. Leftward changes represent increased distress. The positive treatment effects favor eplerenone.
Mean changes in psychosocial measures assessed by SF-36 Health Survey in groups of responses defined by confirmed reports of change in symptom distress. The point is subtended by a 95% confidence interval.
Hollenberg NK, Williams GH, Anderson R, Akhras KS, Bittman RM, Krause SL. Symptoms and the Distress They CauseComparison of an Aldosterone Antagonist and a Calcium Channel Blocking Agent in Patients With Systolic Hypertension. Arch Intern Med. 2003;163(13):1543-1548. doi:10.1001/archinte.163.13.1543
Although there has been increasing recognition that a substantial part of the cardiovascular, central nervous system, and renal conditions induced by renin-angiotensin-aldosterone system activation reflects an action of aldosterone, the potential influence of therapy designed to block aldosterone has been limited by the fact that spironolactone (until recently the only aldosterone antagonist available) exerts a substantial array of adverse effects. We sought to compare the magnitude of the distress induced by a widely used calcium channel blocking agent, amlodipine, and a new aldosterone antagonist, eplerenone, in patients treated for systolic hypertension.
A total of 269 patients older than 50 years with systolic hypertension were randomized to either eplerenone, 50 mg/d, or amlodipine, 2.5 mg/d, and titrated to a maximum 200-mg eplerenone dose or 10-mg amlodipine dose. Patients were followed up for 24 weeks. Quality-of-life questionnaires (SF-36 Health Survey) and a validated instrument for assessing symptom distress (Symptom Distress Index) were administered at randomization and 24 weeks after starting treatment.
The systolic blood pressure response to eplerenone and amlodipine did not differ (eplerenone = −20.5 mm Hg and amlodipine = −20.1 mm Hg). For the quality-of-life analysis, 119 patients were randomized to eplerenone and 122 to amlodipine. No significant treatment group differences in the Symptom Distress Index were detected at baseline. There was an overall significant treatment effect on symptom distress in favor of eplerenone (P = .03). Indeed, Symptom Distress Index showed significant worsening distress in 36 of 71 symptoms in the amlodipine arm and none in the eplerenone arm. Significant treatment effect in favor of eplerenone was observed in 5 symptoms: ankle swelling, weight gain, nocturia, increased urination, and shortness of breath. Patients with symptom distress also showed an erosion of psychosocial measures of quality of life (P<.001).
The aldosterone antagonist eplerenone is substantially better tolerated than the widely used calcium-channel blocking agent amlodipine, with comparable reductions in systolic blood pressure. This feature should improve therapeutics in patients in whom blockade of aldosterone's effect would be helpful.
BLOCKADE OF the renin-angiotensin-aldosterone system with angiotensin-converting enzyme inhibitors and angiotensin II antagonists has become a major element in the treatment of a wide range of cardiovascular and renal conditions. Over the past decade there has been growing recognition that aldosterone contributes to the pathologic conditions induced by activation of this system.1- 6 In animal models, blockade of the effects of aldosterone at the receptor level have had a substantial positive influence on natural history.1- 4 Clinical development has been limited because spironolactone, the only aldosterone antagonist that was available until recently, is poorly tolerated.7- 9 The recent development of eplerenone, an aldosterone receptor antagonist with greater specificity for the mineralocorticoid receptor than spironolactone, has created the possibility that improved therapeutics will follow.1- 4
We have validated an instrument, the Symptom Distress Index (SDI), for assessing not only symptom frequency, a straightforward measure, but also the severity of symptoms in patients treated for hypertension.10,11 Among the antihypertensive agents assessed was amlodipine, a calcium channel blocking agent that is widely used and has a well-defined series of annoying but not life-threatening adverse effects.10 We used amlodipine as a positive control in comparison with eplerenone, an aldosterone receptor antagonist. We explored a range of symptoms associated with a wide variety of antihypertensive agents with special emphasis on adverse effects recognized to occur with amlodipine, and those associated with spironolactone use.7- 9
One of the fundamental tenets of medicine is the Hippocratic injunction that we avoid doing harm. In the case of medical therapy, the most common harm that we wish to avoid is symptoms due to our choice of treatment. When adverse events are sufficiently severe, they lead to discontinuation of treatment, with consequent loss of the benefits that treatment might provide.11,12 Clinical trials predict inadequately withdrawal from treatment in clinical practice.13 Thus, attempts to quantify the frequency of symptoms induced by treatment, and their severity, is a fundamental in therapeutics.14,15 The primary objective of this study was to evaluate the symptom distress associated with eplerenone compared with amlodipine at comparable levels of efficacy.
In this study, 241 patients with systolic hypertension who were 50 years or older were randomized to either eplerenone or amlodipine. The entry criteria in this study involved an untreated seated systolic blood pressure between 140 and 190 mm Hg. After a placebo run-in, patients were randomized to treatment. The initial doses were 2.5 mg of amlodipine daily, or 50 mg of eplerenone daily. Each agent was then titrated to a maximum of 200 mg of eplerenone or 10 mg of amlodipine, to achieve a systolic blood pressure below 140 mm Hg.
Baseline blood pressure was measured at visit 0, at the end of the single-blind placebo treatment interval. The single-blind placebo treatment interval lasted 2 to 4 weeks. Blood pressure was measured at trough, about 24 hours after the last dose with a mercury manometer. The patients were to have been sitting for at least 5 minutes. Thereafter, over 5 minutes, blood pressure was measured 3 times in one arm. The first measurement was discarded and the blood pressure recorded was the average of the second and third measurements.
After the placebo run-in and randomization, a dose escalation format was adopted. The initial dose of amlodipine was 2.5 mg, and of eplerenone was 50 mg. At 2-week intervals, office visits provided the opportunity to increase the amlodipine dose to 5 mg, and then to 10 mg if goal blood pressure (140/90 mm Hg) was not achieved. In a similar fashion, eplerenone doses could be increased to 100 and 200 mg/d.
The patients were followed up for 24 weeks and quality-of-life (QOL) questionnaires were administered at randomization and 14 and 24 weeks after randomization. The approach to QOL assessment used in this study was similar to our previous studies.16 At each study center a specifically trained office employee proctored the battery of patient self-administered questionnaires. The battery included the SF-36 Health Survey instrument, primarily to assess several psychosocial measures described below, the Symptom Distress Index (SDI), and Cantril's ladder. The SDI consisted of 73 items. The questionnaire presented a list of symptoms associated with hypertension or side effects from all of the drugs used. Each question asked the respondent, "Have you had this symptom at all in the past month?" The respondent was to circle a number between 0 (didn't have it at all) through 5 ("yes, and it has bothered me extremely").
In addition, to assess the global QOL at baseline, we used "Cantril's ladder."17 In this global QOL instrument, a visual analog scale is used as a vertical 10-rung "ladder." The participant is asked to grade his or her current QOL on the ladder. Standing on the ground (0) represents the worst QOL that one could imagine. Standing on the top rung of the ladder (10) represents the best QOL that can be imagined, and can only be sustained for a short time. Standing on the middle rung (5) represents
"the average QOL in your community"—without further definition. Because baseline QOL can act as a determinant of QOL during treatment,18 we administered Cantril's ladder at baseline to provide an alternative to the instruments used for therapeutic evaluation to block for baseline effects.
Finally, the SF-36 Health Survey, widely used in QOL studies, assessed 8 aspects of health-related QOL labeled as physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health.
Sample size was designed to accommodate the efficacy assessment, providing at least 90% power to establish noninferiority of eplerenone vs amlodipine by a margin of 6 mm Hg change in systolic blood pressure from baseline. The efficacy analysis was based on an intention to treat with missing values being handled by means of the "last observation carried forward" method. Analysis of covariance was used with baseline value, treatment, and center included in the model.
To be included in the QOL analysis, the patient had to have a valid baseline and end point QOL assessment. Treatment effects on baseline to end point change in QOL were measured by multivariate 1-way analysis of variance (MANOVA). Pillai's trace statistic was used as an exact test of the multivariate significance in each analysis.19 If the main MANOVA detected an overall multivariate effect, we planned to examine the associated univariate effects. The intended analysis had too few degrees of freedom: with 73 variables and only about 60 patients in each of the 4 cells (amlodipine males and females, eplerenone males and females). For that reason, we summarized the data in accordance with the O'Brien rank sum procedure20; we also reduced the dimensionality of the criterion variable space by means of a principal components factor analysis. The O'Brien rank sum and the factor score associated with the first principal component provide parallel overall summaries of each patient's reported symptom distress. We also used the Wilcoxon rank sum test for nonparametric analyses. All QOL analyses were implemented with SPSS statistical software.21
Table 1, Table 2, and Table 3 compare the treatment groups with respect to demographic and baseline characteristics. There were no significant differences in sex, ethnicity, employment status, initial QOL, or baseline blood pressure.
The systolic blood pressure response to the 2 drugs was essentially identical (Figure 1). Amlodipine induced a reduction of systolic blood pressure of −20.1 ± 1.13 mm Hg, and eplerenone reduced systolic blood pressure by −20.5 ± 1.14 mm Hg. Changes in diastolic blood pressure were substantially smaller: −6.9 ± 0.3 mm Hg for amlodipine and −4.5 ± 0.2 mm Hg for eplerenone (P = .01). At week 24, the numbers of patients receiving the first dose level of eplerenone (50 mg) and amlodipine (2.5 mg) were 25 (20%) and 18 (9%), respectively. The numbers of patients receiving the second dose level of eplerenone (100 mg) and amlodipine (5 mg) were 19 (15%) and 35 (29%), respectively. At week 24, most patients were receiving the highest dose level of eplerenone (200 mg) and amlodipine (10 mg): 81 (65%) and 68 (56%), respectively. At the end of the 24-week double-blind treatment period, the mean daily dose of eplerenone was 154 mg and the mean dose of amlodipine was 7.4 mg.
During active treatment, premature discontinuation was uncommon, and did not differ by treatment group, although the dropout rate was 50% greater for the amlodipine-treated patients (n = 30, 24.8%) compared with eplerenone-treated patients (n = 19, 16%).
For the 73 symptoms for which symptom distress was assessed in this study, there was a striking difference identified between the 2 treatments (Figure 2). The generalized summary measure (Figure 2) revealed an average erosion in symptoms distress approached 0.1 SD negative for amlodipine, and approached 0.1 SD positive for eplerenone. This treatment effect that was statistically significant (P = .03). If we take 0.1 SD of difference as clinically meaningful measure,16 then 36 of the 73 measures favored eplerenone, and 1 favored amlodipine (Figure 3; P<.001). The symptoms associated with amlodipine included those that are commonly recognized, including ankle swelling, headache, flushing, and constipation. Others that are not widely recognized include dryness of mouth and loss of taste, sleep disturbances, which included both difficulty getting to sleep and getting back to sleep. The eplerenone-treated group specifically did not develop breast pain or tenderness, breast enlargement, changes in menstrual periods, gynecomastia, or loss of libido. Adverse events of specific interest are summarized in Table 4.
The 8 scales of the SF-36 Health Survey showed no significant treatment effect, or difference between the 2 treatments overall.
To clarify this relationship, the symptom distress responses were collapsed into 3 categories. Under this scheme, there were 5 possible distress level transitions: none to much (2 steps worse), 1 step worse, no change, 1 step better, and much to none (2 steps better). In 46 608 concurrent reports of SF-36 Health Survey and symptom distress changes since the preceding visit, only 57 of these events reported a 2-step worsening of symptom distress, and only 70 events showed a 2-step improvement—far too few for a stable analysis. There were almost 3000 1-step symptom distress change events in each direction. For each SF-36 scale, Figure 4 displays a pair of 95% confidence intervals for the mean scale score, separately for patients with 1-step improved ("better") symptom distress and for those with 1-step deteriorated ("worse") symptom distress. The SF-36 subscale data are presented in units of unadjusted pooled sample standard deviations of the raw index.
For all 8 SF-36 scales, the "scale score" improved on average where it was associated with reports of symptom distress improvement. For 7 scales, the mean score also deteriorated in association with worsening symptom distress (P<.01). This erosion in QOL reflected primarily the negative effects of amlodipine. In magnitude, moreover, the typical change in SF-36 scores lies very much in the same range (0.1 or 0.2 of a raw score standard deviation) that we observed10,11,16 in calibrating the subscales of the Rand Mental Health Index by a similar procedure.
The SDI is a sensitive technique for assessing the influence of drug treatment on QOL. The symptoms associated with amlodipine use that were statistically significant in this study are those that we identified in an earlier study,10,11 and included especially ankle swelling, headache, facial flushing, constipation, and pronounced heartbeat. These instruments, as anticipated, have a sensitivity adequate to identify physical symptoms when they occur. Thus, the absence of any statistically significant symptoms associated with eplerenone, including absence of breast tenderness or gynecomastia, loss of libido, and menstrual irregularities—all symptoms associated with spironolactone use7- 9—must be considered significant.
The fall in systolic blood pressure was substantial, exceeding 20 mm Hg for both amlodipine and eplerenone groups. The drop in diastolic blood pressure was much smaller, the response to amlodipine exceeding the response to eplerenone modestly, but to a statistically significant degree. This pattern of response, the fall in systolic blood pressure exceeding substantially the fall in diastolic blood pressure, is probably appropriate to a patient population recruited because of systolic hypertension.
The explanation for the amlodipine-induced drop in systolic blood pressure is straightforward: amlodipine is a vasodilator, which accounts not only for its efficacy but for the adverse effects associated with amlodipine use, including edema, headache, and facial flushing. Why a specific aldosterone receptor antagonist should induce an essentially identical fall in systolic blood pressure requires explanation. Perhaps aldosterone contributes to inducing the vasoconstriction that is reversed by amlodipine.
In studying treatment effects on QOL, both the distress specifically associated with symptoms and the change in psychosocial factors should be evaluated. The distress associated with specific symptoms escapes the ambiguity that is implicit in the assessment of QOL.22 Overall, the psychosocial domains examined by the SF-36 Health Survey showed no difference at baseline or at the end of treatment for either drug group overall. As in our earlier study, however, there was a clear influence of symptom distress on the associated psychosocial measures.10 For all 8 SF-36 scales, the domain score improved on average when it was associated with reports of symptom distress improvement. Conversely, 7 of the 8 scales showed deterioration in association with worsening symptom distress. The influence of symptom distress, as previously, goes beyond the specific symptoms.10,11 Quality of life, in general, was worsened with the increased symptoms.
As in several earlier studies, the scale showing the largest change is vitality.10,11,16 The explanation for that phenomenon requires further investigation.23 The only scale not showing a difference with symptom distress was physical functioning. Again, the relationship is consistent both in character and direction with the relationship that we observed earlier.10,11,16 In our earlier study we used the full Rand Mental Health Index, whereas in this study we used the SF-36 Health Survey, which is a brief questionnaire derived from the Rand Mental Health Index. The results of this study suggest that the SF-36 Health Survey has adequate power for such studies.
In the case of amlodipine, the symptom distress is associated with a cluster of annoying but not life-threatening side effects. In the case of eplerenone, side effects related to an action on steroid receptors other than mineralocorticoid—as is the case with spironolactone—were absent, along with any other symptoms. Presumably, the side effect profile is influenced by a relatively greater affinity for mineralocorticoid than other steroid receptors.1- 4 This difference translates into a better tolerated agent, which should have implications for therapeutics.
A number of specific adverse events merit specific discussion, beyond the SDI findings. Hyperkalemia is a potential adverse effect of blocking of the effect of aldosterone on the kidney, and occurred in 2 eplerenone-treated patients and in 1 amlodipine-treated patient. In contrast, hypokalemia was found in 2 amlodipine-treated patients, and in no eplerenone-treated patients. Erectile dysfunction was reported in 2 of 61 eplerenone-treated men, and in no amlodipine-treated men. On the other hand, there was no complaint of gynecomastia, tender breasts, or menstrual irregularities. The anticipated difference in edema was found, present in 25.2% of the amlodipine-treated patients and 4.5% of the eplerenone-treated patients.
In an earlier study in which we attempted to calibrate QOL instruments, including the Mental Health Index, through the prospective and systematic assessment of life events, we made an observation immediately relevant to the findings in this study.16 The threshold change in QOL that appeared to be clinically significant was evidenced in a change in measured QOL of about 0.1 SD—the same measure as the change assessed with a single step in physical symptom distress defined in this study. In our earlier study this level of erosion of QOL approximated that which occurred with concern about one's job. Larger changes, from 0.2 to 0.4 SD, were associated with a larger and more important life event changes such as loss of one's employment, divorce, or death of a spouse. Whether or not the
"none to extreme" change in physical symptom distress is associated with a QOL change that truly matches the impact on one's QOL of divorce or death of a spouse cannot be answered from the data in this study. Given the language in the instruments, it would come as no surprise that they lack precision at the extreme.
A remarkable series of studies in a wide range of animal models of renal, cardiac, and central nervous system injury have suggested a crucial pathogenetic role for aldosterone. The availability of a new, specific, effective, and well-tolerated aldosterone receptor antagonists will facilitate the exploration of the clinical implications of these findings. The results of this study suggest that eplerenone may be such an agent.
Corresponding author and reprints: Norman K. Hollenberg, MD, PhD, Brigham and Women's Hospital, 15 Francis St, Boston, MA 02115.
Accepted for publication September 27, 2002.
This study was supported by a grant from Pharmacia Corporation, New York, NY.
We are grateful to Diana Capone for her assistance in the preparation and submission of the manuscript.