Dose Response Effectiveness of Propranolol for the
Treatment of Angina Pectoris
By EDWIN L. ALDERMAN, M. D., RICHARD 0. DAVIES, M.D., JOHN J. CROtWLEY, PH.D.,
MARIO G. LOPES, M.D., JEFF Z. BROOKER, M.D., JOEL P. FRIEDMAN, M.D.,
ANTHONY F. GRAHAM, M.D., HARVEY J. MATLOF, M.D., AND DONALD C. HARRISON, M.D.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
SUMMARY
Seventeen patients received placebo medication during a 12-week run-in period, followed by four doubleblind study periods of six weeks each, during which time placebo, 80 mg, 160 mg and 320 mg propranolol
dosages were administered. Examination of the frequency of angina episodes and nonprophylactic
nitroglycerin consumption revealed significant beneficial clinical responses for both the 160 and 320 mg
dosages. Exercise testing also demonstrated increased exercise tolerance (320 mg dose) with a shift of the exercise end point from pain to fatigue in seven of 17 patients. The interrelationships between propranolol
daily dosage, clinical response assessed by percent reduction in anginal episodes, beta-adrenergic blockade
measured by percent reduction in exercise heart rate and serum levels were examined. In general, serum
levels of 30 ng/ml, when drawn 90 to 180 minutes following the last oral dose, were required to achieve a
25% or greater reduction in angina frequency. Serum levels above 30 ng/ml were similarly correlated with a
20% or greater reduction in exercise heart rate at equivalent levels of exercise. Detailed examination of
different patterns of clinical response with respect to beta-blockade, serum levels and oral doses are
presented.
Additional Indexing Words:
Beta-adrenergic blockade
Exercise testing
pROPRANOLOL has been the subject of many
investigations concerning its efficacy for the
treatment of angina pectoris. Previous studies using a
variety of protocols comparing active drug to placebo
have generally demonstrated that propranolol is an
effective agent.'14 The present study was designed to
assess the relationship between clinical response and
daily propranolol doses of 80, 160 and 320 mg, with
the goal of quantifying factors which influence this
relationship. Pharmacologic studies have previously
reported substantial patient-to-patient variability in
propranolol plasma levels for similar oral doses.4 6
Other studies have examined the relationship between
propranolol plasma levels and the degree of betaadrenergic blockade;6-8 however, clinical data were
not available. The present multi-dose trial provided
the opportunity to examine the interrelationships
between oral dosage, clinical response, propranolol
serum levels and degree of beta-adrenergic blockade
as assessed by exercise testing.
Methods
Clinical Features of Study Patients
Twenty-three patients entered this study; however, two
did not proceed beyond initial placebo periods because of
insufficient angina in one and development of unstable
angina in the other. All 21 patients who received
propranolol were men, with an age range of 41-70 years
(mean age 55). Three patients were taking oral
hypoglycemic drugs for diabetes mellitus, six patients had
mild hypertension (five of these were receiving diuretics);
seven patients were receiving anticoagulants, and 15
patients smoked cigarettes. The concomitant medications
were taken throughout the trial. Sixteen patients had
previously used long-acting oral nitrates, which were discontinued at the time of entry into the study. No other changes
in medications or dietary intake were made during the
course of the study. Coronary artery disease was
documented in all patients by prior myocardial infarction
(15 patients), by coronary arteriography (9 patients), or by
both. Angina pectoris had been present for one to 19 years,
and the severity of symptoms as indexed by the New York
Heart Association placed three patients in class II, 11
patients in class III, and seven patients in class IV. Two
From the Cardiology Division and the Division of Biostatistics,
Stanford University School of Medicine, Stanford, California.
Supported in part by a grant-in-aid from Ayerst Pharmaceuticals,
Inc., and NIH grants HL 5866 and Program Project grant NIH 1
P01 HL 15833. Dr. Lopes was supported by a grant from Instituto
de Alta Cultura, Portuguese Department of National Education,
Lisbon.
Dr. Lopes' present address is Hospital Santa Maria, Lisbon 4,
Portugal.
Presented in part at the American College of Cardiology
Meeting, New York, February 1974.
Address for reprints: Edwin L. Alderman, M.D., Cardiology
Division, Stanford University Medical Center, Stanford, California
94305.
Received November 4, 1974; revision accepted for publication
January 31, 1975.
964
Circulation, Volume 51, June 1975
965
PROPRANOLOL DOSE RESPONSE IN ANGINA
patients had mild congestive heart failure and two others
had moderate congestive heart failure, all four of whom
were controlled with chronic digitalis therapy. Thirteen
patients had significantly diminished pulse amplitude in the
lower extremities, with three patients having mild claudication and three having moderate claudication.
All patients had a stable level of angina pectoris for three
months prior to this trial, were ambulatory outpatients for
its duration, and had five or more clearly discernible angina
episodes per week. Informed consent for participation in
this study was obtained from each subject. Patients were instructed to be as active as their symptoms would allow and
to use sublingual nitroglycerin tablets for the relief of angina
episodes. Prophylactic nitroglycerin was used infrequently
by four of the study patients, and this consumption was
recorded.
Single-Blind Run-In Period
All patients initially entered
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
a single-blind placebo run-in
period of 12 weeks' duration. Four six-week study periods
were then utilized to determine the effects of the drug.
Every six weeks (during both run-in and study periods) a
new dose was initiated with one tablet daily, which was increased by one tablet every other day until a dose of one
tablet four times daily was reached. Patients were seen every
two weeks for history and physical examination. At each
visit, a new medication bottle was provided and the tablets
remaining in the previous bottle were counted. Patients
were asked to keep a written record of each angina attack in
a small printed diary. This booklet was collected every two
weeks and a new one supplied. At the conclusion of each
two-week period, the patient summarized his impressions of
the severity, frequency, and response of the angina to
nitroglycerin. At the end of each six-week dose period, electrocardiogram, chest X-ray, hematologic and biochemical
determinations were monitored. Treadmill exercise testing
was performed on three separate occasions during the run-in
period.
Eleven of the 21 patients completing the single-blind
period had previously participated in a study of practolol
efficacy for the treatment of angina.9 These 11 patients had
already participated in an identical 12-week single-blind
placebo run-in period and moved directly from the last
double-blind practolol dose period into the first six-week
dose period of the double-blind propranolol protocol. These
patients had stable angina and no more than 12 weeks
elapsed from the completion of their single-blind run-in
period until their initial double-blind propranolol period.
Their clinical features and responses to propranolol were not
different from those observed for the rest of the patients.
period, but dropouts produced unequal numbers in each
group (see table 1).
Coded tablets were supplied which were identical in
appearance and taste for the double-blind and run-in
periods. Equivalence bioavailability of the active dose forms
had been demonstrated previously (Caron and Davies, unpublished observations). Pulse and blood pressure were
recorded by the nursing staff at each visit, and the numeric
results of current and prior visits were not known to the doctors interviewing and examining the patients. Blood samples
were obtained 90 to 180 min after the first morning dose
during the fourth and sixth weeks of each dose period.
Serum propranolol levels were measured using a modification of the fluorometric procedure of Shand et al.5
Of the 21 patients entering the double-blind protocol,
three were eliminated: one who died suddenly, one because
of acute myocardial infarction, and a third who developed
acute coronary insufficiency. Eighteen patients completed
the double-blind portion of the study; however, one patient
has been excluded from the data analysis because of the
marked atypicality of his chest pain and inconsistency of his
blood level measurements, suggesting failure to take
medication. It was necessary to shorten five of the 68
double-blind periods completed by these 17 patients from
six weeks to four weeks. One 80 mg double-blind period was
shortened because of severe back pain unrelated to medication; one 160 mg dose period was shortened because of a
missing bottle of medication; and three placebo periods
were shortened because of severe angina. Following completion of the 24-week double-blind protocol, patients were
given single-blind placebo tablets for an additional six
weeks, during which time procedures were followed identical to those during the single-blind and double-blind
protocols.
Exercise Testing
Exercise tolerance was measured by means of a symptomlimited exercise test performed between 15 and 45 min after
the blood was sampled for serum propranolol levels. A
multi-stage treadmill test was used, with a modification of
the procedure of Bruce and Hornsten'0 to introduce two
lesser exercise stages of 1 mph at 0% grade, and 1.2 mph at
5% grade. The speed and inclination of the treadmill were
increased every three minutes. Prior to entry into the study,
the initial exercise stage was selected for each patient such
Table 1
Treatment Sequence for Double Blind Periods
Double-blind periods*
Group
Double-Blind Periods
Following the 12-week single-blind run-in period,
patients entered the double-blind portion of the study only
if they had five or more angina episodes per week. Patients
were started on a new dose of medication every six weeks in
the same manner as during the run-in period. The same
observations of efficacy and safety parameters were made at
two-week intervals and at the end of the sixth week, as in the
single-blind periods. Placebo and three dose levels of
propranolol (80, 160 and 320 mg/day) were examined in
four sequential six-week double-blind periods (table 1). The
order of drug administration was selected by randomizing a
Latin square. Patients were randomly assigned to one of
four treatment sequences at the end of the single-blind
Circulation, Volume 51, June 1975
1
No.
A
4
PL
B
160
5
80
C
3
D
320
5
Duration of Treatment Periods
3
4
320
PL
160
80
80
320
160
80
320
PL
PL
160
Single-
Singleblind
run-in
Placebo
12 weeks
2
1
6 weeks
Double-blind periods
2
3
6 weeks
6 weeks
blind
4
F/U
6 weeks
6 weeks
*Propranolol doses are given as mg/day.
Abbreviations: F/U = follow-up; PL = placebo.
966
ALDERMAN ET AL.
that angina developed more than three minutes after the
start of exercise. Patients were instructed to stop walking at
a similar level of pain for each test. ECG leads II and V5
were continuously monitored. Systolic blood pressure was
measured at two-minute intervals and at exercise end point
by external cuff techniques. The end point of exercise
testing was chest pain for 16 patients and fatigue in one
patient. Significant flat or downsloping ST-segment depression (0.1 mV for 0.08 sec) was exhibited by 11 patients,
borderline ST-segment depression by four patients, and no
ST-segment depression by two patients. The maximum
heart rate at exercise end point ranged from 85 to 145 (mean
119), with ten patients limited at heart rates equal to or less
than 115.
Statistical Analysis
A multivariate approach to statistical analysis
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
was used,
and the effect of time, which should be minimal due to the
run-in period, was pooled into the over-all interaction.
Carry-over effects from drug doses given in a preceding
period were minimized by using only data from the last four
weeks of each six-week period for analysis. In order to
stabilize variance between patient groups (i.e., four treatment sequences), reduce the influence of a few large
numbers (range of angina episodes/two weeks = 0-149,
mean 15.9), and allow for the presence of zero angina
episodes, a square root transformation was employed. Computational analysis was done using the Biomed Series
Program X63.
Heart Rate and Blood Pressure Response
Resting heart rate following propranolol administration fell 16 to 27%, depending upon the
posture and dosage, as shown in table 2. Systolic and
diastolic blood pressures (table 2) tended to decrease
to
a
Clinical Response
of clinical response to propranolol
examined for statistical analysis: frequency of angina episodes, and nonprophylactic
nitroglycerin consumption. Examination of run-in,
single-blind placebo data, as shown in figure 1,
demonstrates that angina frequency and nitroglycerin
consumption fell significantly (P < 0.05) within eight
weeks of placebo therapy to two-thirds of their initial
level, and then tended to stabilize over the succeeding
four weeks. There was a substantial variability in individual patients from week to week, with some subjects exhibiting no placebo effect. A range of four to
twelve weeks for stabilization to occur was noted.
Figure 2 illustrates the results of angina frequency,
and figure 3 the results of nitroglycerin consumption
during the course of the study. Transformed square
root data are plotted, and P values for statistical comparisons are shown. Both figures demonstrate a linear
relationship betwen the logs of the three propranolol
dosages and reduction of both angina frequency and
nitroglycerin consumption.
Listed in table 3 are the degrees of freedom, F
ratios and P values for comparisons between placebo
periods and between placebo and propranolol drug
periods. Also shown is the statistical analysis of groupto-group variation (interaction) associated with each
main comparison between treatment periods. Since
main effects are averages over the four different treatment sequences, the presence of interactions would
Two
therapy
measures
were
lesser extent.
IuuMUi
w
Table 2
Heart Rate and Blood Pressure Changes after
Propranolol4
U,)
0
ANGINA EPISODES
3: 90k
0
Dose
(mg/day)
No.
Supine
Position
Standing
A. Heart rate changes as percentage of double-blind
placebo value
80
17
-20.1% 1tt
-15.9%t
160
17
- 18.3%t
-24.0%tt
320
17
- 21.9%t
-27.4%tt
B. Systolic pressure changes as percentage of double-blind
placebo value
80
17
-3.0%
-3.5%*
160
1]7
-2.5°7%
-3.7%*
320
17
-3.0%
-5.2%*
C. Diastolic pressure changes as percentage of double-blind
placebo value
80
17
- 10.1%
0.0%
160
17
- 9.9%
-3.7%
320
17
- 5.7 U
-10.3%
11
-1
/
*p < 0.05.
tP < 0.01.
tAll data are results of averages of three separate determinations during each double-blind period.
3:
80k
-J
z
70
0
W
(D
ct
60k
z
W
* NON-PROPHYLACTIC
NITROGLYCERIN
c-)
wr rU
1-2
3-4
5-6
7-8
9-10
WEEKS of PLACEBO
11-12
* p<.05
Figure 1
The frequency of angina episodes and nonprophylactic
nitroglycerin consumption is illustrated as a percentage of the initial
two weeks for each two week segment of the twelve week duration
of the single-blind run-in placebo period. Both indices of anginal
severity fell within eight weeks of placebo therapy by one-third and
tended to level off during the succeeding four weeks.
Circulation, Volume 51, June 1975
PROPRANOLOL DOSE RESPONSE IN ANGINA
invalidate these comparisons. The interactions
between the double-blind placebo period and the
drug periods for angina frequency (table 3A) are large
enough to indicate that caution must be used in the
interpretation of the- corresponding main effects. The
cause of the interaction is the high level of symptoms
in the double-blind placebo period for some patients.
Figures 2 and 3 similarly emphasize that the levels of
angina frequency and nitroglycerin consumption during double-blind placebo periods were more comparable to the levels of angina at the time of initial entry into the study than the level of angina at the end of
the run-in period. For 13 of the 17 patients (groups B,
C and D in table 1) the double-blind placebo period
followed the therapeutically effective 160 mg dose.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
SINGLE
BLIND
PLACEBO
967
These same patients (groups B, C and D) exhibited a
significant difference (P < 0.05) between the doubleblind placebo period and the end of the run-in period
(table 3A). These facts suggest that patients, upon
cessation of propranolol therapy, return to a level of
angina similar to that experienced upon initial entry
into the study, rather than back to the level attained
after prolonged placebo therapy.
Because of the differences between placebo periods
and the possible presence of interactions, each of the
propranolol dose periods has been compared with
both the double-blind and the run-in placebo periods,
the results of which are shown in table 3 (main comparisons) and figures 2 and 3. Comparisons of
propranolol periods with double-blind placebo data
DOUBLE BLIND t
PROPRANOLOL
(daily dosage)
4 Last 4
5.0First
5.0 -Weeks Weeks
m
W
W
-80mg
4.5
W
W04.0-
0
~~~~~160 mg
1.
3g53>
z o
z
[;
0
320mg
43.0-
t all data averaged over last 4 weeks in each
* p( .01
*p<.05
Figure 2
period
The results of angina frequency during the first and last four weeks
of the placebo periods and the last four weeks of each of the doubleblind periods are shown as the square root of the number of angina
episodes per two weeks. Frequency of angina episodes during the
double-blind placebo period more closely approximated frequency
upon initial entry into the study (first four weeks), than during the
last four weeks of the single-blind placebo period. A linear
relationship between the logs of the three propranolol dosages and
reduction in angina frequency are exhibited. Statistical data, shown
in the bottom half of the illustration, are obtained from table 3.
Circulation, Volume 51, June 1975
t all dota averaged over last 4 weeks in eoch period
* p<.OI
Pp<.05
Figure 3
The results of nonprophylactic nitroglycerin consumption are
illustrated in the same manner as angina frequency in figure 2.
Nitroglycerin consumption during the double-blind placebo period
more closely approximated consumption upon initial entry into the
study than during the last four weeks of the twelve week singleblind placebo period. A linear relationship between the log
propranolol dosage and nitroglycerin consumption is exhibited.
Statistical significances are illustrated in the lower half of the graph
and appropriate F test statistics are listed in table 3.
ALDERMAN ET AL.
968
Table 3-A
Statistical Analysis-Total Angina Episodes (square root transformation)
Period A
Plac
Plac
Plac
Plac
RI, SB
RI, SB
RI, SB
DB
Comparison of periods
Compared to
First 4
Last 4
Last 4
Plac
Plac
80
160
320
80
160
320
Last 4
Period B
RI, SB
DF
DB
DB
DB
DB
Last 4
Last 4
Last 4
Last 4
Last 4
DB
DB
DB
Last 4
Last 4
Last 4
1, 13
l, 13
1, 13
1, 13
1, 13
1, 13
1, 13
1, 13
Main comparison
F
Group-to-group
variation (interaction)
DF
F
P
P
2.35
5.53
0.15.
4.06
9.47
0.05*
NS
0.07
0.05
7.41
21.55
39.56
0.05
0.01
0.01
NS
3, 13
2, 13
1.04
0.56
NS
NS*
3, 13
1.30
NS
3, 13
(2, 13
3.20
4.42
0.07
0.05)*
Table 3-B
Statistical Analysis-Non-prophylactic Nitroglycerin Consumption (square root transformation)
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Comparison of periods
Plac
Plac
Plac
Plac
RI, SB
RI, SB
RI, SB
DB
Period B
Compared to
Period A
First 4
Last 4
Last 4
Last 4
Plac
Plac
80
160
320
80
160
320
RI, SB
DF
1,
1,
1,
1,
Group-to-group
Main comparison
F
P
DB
DB
DB
DB
Last 4
Last 4
Last 4
Last 4
Last 4
13
13
13
13
1,13
12.68
2.73
0.22
3.47
6.30
DB
DB
DB
Last 4
Last 4
Last 4
1, 13
5.87
1, 13
1, 13
13.12
19.01
variation (interaction)
F
P
DF
0.01
NS*
3, 13
2, 13
1.84
1.11
NS
NS
0.10
3, 13
1.17
NS
0.05 J
0.05 )
0.01
0.01
3, 13
(2, 13
1.81
2.57
NS
NS*
NS)*
*Analysis of Groups B, C, and D only (i. e., includes only those patients for whom placebo followed the 160 mg dose).
Abbreviations: Plac = placebo; RI = run-in: SB = single-blind; DB = double-blind; First 4 = first 4 weeks; Last 4 = last
4 weeks; F = F statistic; DF = degrees of freedom; P = probability value; NS = not significant.
statistically demonstrate propranolol efficacy for all
three dose levels. This result is, however, favorably
biased by the relative severity of symptoms during the
double-blind control period. A more conservative
comparison using placebo data at the end of the runin period documents propranolol's efficacy at the 160
mg (P < 0.07) and 320 mg (P < 0.05) dose levels.
At the end of each two-week visit, patients were
asked to comment concerning the severity and duration of pain and how they felt in relation to the
previous visit. Using subjective patient data, each
double-blind period was assigned a global rating of
excellent, good, fair or poor, which were assigned
scores of 4, 3, 2 and 1, respectively. Figure 4 illustrates
the number of patients in each clinical response
category for placebo and the three propranolol dose
periods. Although a favorable trend was observed for
patients receiving 80 mg per day, only the 160 mg and
320 mg daily dosages achieved statistically significant
results.
Using the criteria of 25% improvement of angina
frequency compared to the last four weeks of the runin placebo period, six of 17 patients (35%) showed
benefit at the 80 mg dose, ten of 17 patients (59%)
showed benefit at the 160 mg dose, and 13 of 17
patients (76%) showed benefit at the 320 mg dose
(table 4).
For 11 of the 13 patients responding at the 320 mg
dose, the increase from 160 mg to 320 mg improved
angina suppression by an additional 14 to 53%. Four
patients failed to demonstrate 25% or greater reduction in angina frequency at any propranolol dose
when compared to run-in placebo data; however two
individuals (patients 3 and 4) showed a dosecorrelated response compared to double-blind placebo
values. The other two individuals had significant
arteriographically-proven coronary disease and positive exercise tests. Although these four patients did
not show significant suppression of angina episodes,
all exhibited some clinical response based on strong
dose-correlated changes in both physical activity and
subjective ratings of the double-blind periods.
Exercise Tolerance
The results of exercise testing for each of the
double-blind periods are summarized in figure 5. The
Circulation, Volume 51, June 1975
969
PROPRANOLOL DOSE RESPONSE IN ANGINA
limitation after exercise end point during
the placebo double-blind period was pain for 16 of the
17 patients. With increasing doses of propranolol, exercise tolerance progressively lengthened from 325 sec
to 410 sec, reaching statistical significance for the 320
mg dose period (P < 0.01). In addition, symptom
limitation at exercise end point was shifted from pain
to fatigue for seven of the 17 patients.
Listed in table 5 are the results of exercise-induced
changes in heart rate, systolic blood pressure and the
double product (systolic arterial pressure X heart
rate). All three measures of propranolol effect were
significantly suppressed at rest and at the exercise end
point in a dose-related fashion. The increase in
double-product with exercise from baseline was also
attentuated progressively as the dose increased.
exhibited a linear relationship between
the log serum propranolol concentration and log dose.
Individual patients were noted who had serum levels
which were consistently in the upper or lower extremes for all doses. For each subject, the variation
between samples taken on the fourth and sixth weeks
of each double-blind period was small compared to
the differences between results from the three
propranolol dosages.
Figure 7 illustrates the relation between individual
clinical responses and log propranolol serum levels for
those 13 individuals (patients 5 through 17, table 4)
exhibiting a 25% or greater reduction in angina
episodes at one or more propranolol doses (r = 0.52;
P < 0.001). This figure shows that a serum level in excess of 30 ng/ml, when drawn 90 to 180 min following
the last oral dose, was generally required to achieve
25% or greater reduction in angina frequency.
The degree of beta-adrenergic blockade for each
propranolol dose was assessed during exercise testing
by measuring the maximum heart rate at maximum
exercise while patients were receiving placebo
therapy. Heart rates were noted at the same levels of
exercise during each of the three drug periods as had
been maximally attained during placebo treatment.
The measured heart rates were then recorded as a
percentage of the placebo heart rate (table 4).
Figure 8 illustrates the relationship between the
most patients
symptom
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Correlation of Propranolol Serum Levels with
Clinical Response and Degree of Beta-Blockade
Averages of two propranolol serum determinations,
obtained during each of the 68 double-blind periods
are listed in table 4 and shown in figure 6. The mean
serum propranolol level was 24 ng/ml for the 80 mg/day dose, 67 ng/ml for the 160 mg/day dose and 173
ng/ml for the 320 mg/day dose. All determinations
made during the placebo period yielded insignificant
serum levels. A ten-fold range of serum levels was
seen in different patients from an individual dose, and
Table 4
Correlation Between Angina Suppression, Seruim Drug Levels and Beta Blockade
Percent reduction in
anginal frequency
compared to SB PL
Number of anginal episodes during
last 4 weeks of each period
Propranolol serum
level (ng/ml)
90-180 min samples
Percent reduction in
exercise heart rate
from PL at same
exercise level
Propranolol dose (mg/day)
DB
PL
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Mean
SEM
39
49
156
45
37
21
24
34
63
10
68
81
230
165
7
30
12
63.0
15.2
SB
PL
15
17
45
24
31
11
23
14
79
19
46
58
109
256
80
160
320
80
160
320
80
160
320
80
23
36
40
97
24
27
9
39
8
49
3
31
15
82
172
1
0
3
37.4
10.7
32
52
101
23
13
8
0
0
0
0
0
0
17
0
0
0
9
52
28
28
43
90
40
18.1
6.3
0
0
0
0
13
i8
0
43
38
84
33
74
25
33
86
100
70
36.3
8.4
0
0
0
0
38
27
53
57
39
100
11
10
27
31
6
18
3
64
40
66
35
25
50
19
22
24
64
81
170
102
104
147
105
133
76
68
100
138
681
79
118
19
2
24
5
16
13
13
17
3
33
9
23
26
334
21
207
190
24
57
136
43
30
12
19
18
94
21
42
28
78
7
29
10
46.6
14.7
18a
4
3
6
47.0
12.2
15
6
48
0
8
2
53
120
0
0
1
28.9
8.7
61
97
51
53
100
66
90
50.1
8.6
17
30
32
38
61
30
19
25
23.8
3.3
60
193
75
68
140
53
50
65
67.5
10.2
Abbreviations: DB = double-blind; SEM = standard error of mean; SB = single-blind; PL
Circulation, Volume 51, June 1975
182
173
35.4
=
placebo.
34
14
17.4
2.3
160
21
17
24
21
16
16
17
33
21
37
32
18
21
30
37
42
20
24.9
2.1
320
22
26
30
22
23
21
25
3,5
22
38
34
28
29
26
34
49
27
28.9
1.8
ALDERMAN ET AL.
970
600-
500 -
*
o
0
N=17
o
@00
0
~
0 0
TIME TO
EXERCISE 400
ENDPOINT 40
ENDPOINTS
* PAIN
°
a FATIGUE
0
@0
°°
~
0
0
(sec)
300 -
NUMBER
PERSONS
T
0+o
4
12-
IN EACH
200100
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
0PLACEBO 80mg
160mg
PROPRANOLOL
320mg
DAILY DOSAGE
Figure 4
The number of patients in each of four clinical response categories
are illustrated for double-blind placebo and each of the propranolol
doses. Of the 17 patients,
fi;ve
exhibited a good to excellent clinical
-
N.S. P( 01
N.S.
PLACEBO 80mg 160mg 320mg
PROPRANOLOL DAILY DOSAGE
Figure 5
The time to exercise end point is illustrated for each of the doubleblind study periods. The duration of exercise tolerance increased
significantly with the 320 mg propranolol dose compared to doubleblind placebo. Chest pain as an end point for exercise tolerance is illustrated by the solid circles and was present for 16 of 17 patients
during placebo periods. Fatigue as an end point of exercise
tolerance is illustrated by the open circles and at the highest
propranolol dosage became the exercise limitation for eight of the
1 7 patients.
response while receiving placebo. The number of persons in these
increased
categories
propranolol
hibited
a
significantly
with
the
160
good
or
excellent clinical
320 mg
patients ex-
and
doses, such that at the 320 mg dose 16 of 17
response.
degree of beta-blockade exhibited during exercise, the
log propranolol serum levels, and the percent reduction in angina frequency for the same 13 patients
whose data are illustrated in figure 7. The correlation
coefficient relating propranolol serum levels to the
percent reduction in exercise heart rates was 0.51. In
addition, the data in this figure suggest that a serum
propranolol level above 30 ng/ml or a heart rate
reduction at equivalent levels of exercise of 20% or
more yields a satisfactory clinical response. It appears
that patients receiving lower propranolol dose levels
who did not exhibit a 25% reduction in angina frequency are somewhat better discriminated by their inadequate degree of beta-blockade than by their
propranolol serum level. Figure 9 confirms in these
same 13 patients the relationship between the degree
Table 5
Effects of Exercise on Heart Rate and Blood Pressure
160 mg/day
80 mg/day
Placebo (DB)
Mean
SEM
Mean
SEM
80.1
138
10994
3.2
4.6
431
62.3
129
8040
2.8
400
0.01
0.0o
0.01
119.3
3.9
5.2
997
101.8
153
15516
3.1
4.9
587
0.01
NS
0.01
P
320 mg/day
SEM
Mean
SEM
P
Mean
58.8
126
7385
2.4
4.2
373
0.01
0.01
0.01
54.7
129
6942
2.4
4.2
466
0.01
0.05
0.01
94.6
3.0
5.1
661
0.01
0.01
0.01
89.9
143
12830
3.0
4.0
577
0.01
0.Oi
0.01
P
Rest
Heart rate
Blood pressure (systolic)
Heart rate X systolic
blood pressure
3.a
Exercise end point
Heart rate
Blood pressure (systolic)
Heart rate X systolic
blood pressure
Abbreviations: Ns
=
162
19380
145
13759
not significant; DB = double-blind; SEM= standard error of mean; P = probability value.
Circulation,
Volume 51, June
1975
971
PROPRANOLOL DOSE RESPONSE IN ANGINA
of beta-blockade measured by exercise heart rate
reduction and the clinical response evaluated as
reduction in angina frequency (r = 0.69).
The relationship between propranolol dose, serum
level, beta-blockade and clinical response is illustrated
in figure 10 in the same manner as in figure 8 for three
selected patients. Patient A (patient 7, table 4) ex-
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
hibited a clinical response while receiving 320 mg per
day; however this patient did not show a clinical
response while receiving 80 or 160 mg doses. The data
in figure 10 suggest that at both lower dosages, inadequate serum levels to produce significant reduction in exercise heart rate were attained. Three of the
13 who exhibited 25% or greater reductions in anginal
frequency demonstrated this pattern (patients 5, 6 and
7). Patient B (patient 11, table 4) exhibited inadequate
serum levels and little beta-blockade with the 80 mg
dose, but with higher dosages demonstrated high
serum levels and increased beta-blockade. Four
patients (patients 8, 9, 10 and 11) exhibited this
pattern. The most common response pattern (patients
12, 13, 14, 15, 16 and 17) is typified by patient C
(patient 15, table 4), who exhibited beneficial clinical
responses to propranolol at all three dose levels. This
individual, while receiving 80 mg per day, achieved a
partial degree of exercise blockade and a partial
therapeutic response. With larger propranolol doses, a
maximum level of beta-blockade was reached, along
with further reduction in angina frequency.
Adverse Reactions (table 6)
Every two weeks each patient was asked to
volunteer drug side effects and was then specifically
questioned concerning 20 possible adverse reactions.
The most common responses were fatigue and
lightheadedness, which occurred with sufficient frequency during both placebo and propranolol treatment periods to make these particular symptoms
difficult to evaluate. Only those symptoms which
could be directly attributable to propranolol administration during double-blind treatment periods,
1000
CO
A
500
1000 r
A
0
500
200-
ZQ
A
0
_i
100-
200
c
0'
W
0
50-
_i100
0
C
_
A
it
50
LLJ
E
2
20DOSAGE
10-
n 20 _E-
0 80 mg.
(
a)
0
< I O
_~
cn
S
0~
C]160 mg.
0
5-
A 320 mg.
0
CL
W
0
2-
2
.
,
320
160
80
DAILY PROPRANOLOL DOSE
* 90 to 180 min samples
Figure 6
The log of the propranolol serum level attained 90 to 180 min after
the last morning dose is plotted in ng/ml against the log of the daily
propranolol dosage. A generally linear relationship was exhibited,
with some patients exhibiting strikingly high or low serum level
responses at the same oral dosage.
Circulation, Volume 51, June 1975
0-25
26-50
51-75
76-100
angina episodes
compared to placebo run-in period
Figure 7
per cent reduction in
The log of the propranolol serum level obtained 90 to 180 min after
the first morning oral dosage is plotted against the percent reduction in angina episodes as compared to the placebo run-in period.
Data for only those 13 patients who had a 25% or greater reduction
in angina frequency at one or more dosages of propranolol are illustrated. In general, a serum level in excess of 30 ng/ml was required for significant clinical response.
ALDERMAN ET AL.
972
1000
50 _
0
v-
500
5o
* ***
40
0*-
x
0h0
X~~~~
°X
-
0
_
200
0
E
*
m
C_
c
0
0*
100
0
°
0
c.a
~0-
° IO1
A
tA
0
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
CL
A
0
Percent reduction
in angina frequency
o 0-25
* 26-75
A 76-100
0
5
10
20
30
40
rate at same exercise level
Figure 8
The log of the propranolol serum level in ng/ml obtained 90 to 180
min following the last morning dose is plotted against the percentage reduction of exercise heart rate as measured at the same exercise level. Drug dosages which yielded less than 25% reduction in
angina frequency generally were associated with serum levels below
50 ng/ml or less than 25% reduction in exercise heart rate. Patients
exhibiting the most striking clinical responses (76 to 100% reduction
in angina frequency) demonstrated 30% or greater reduction in exercise heart rate, with quite variable propranolol serum levels
=
0.51;
y
=
14.1
e
l
60
40
100
80
S0
Percent reduction in excercise heart
(r
I
20
The percent reduction in exercise heart rate at the same exercise
level is plotted against the percent reduction in angina episodes
compared to the last four weeks of the placebo run-in period. A
linear relationship is exhibited with a P value of 0.001. These data
demonstrate the correlation between the degree of beta-blockade
exerted by propranolol and its clinical effectiveness in reducing
angina frequency (r = 0.68; y = 0.20 x + 15.7).
0
0
0
Percent reduction in angina episodes
compared to placebo run -in period
Figure 9
0
a-
101
H .
0
0
1100
0
o0
20
*
_~~~
o
c 0
m
.5
W
~~~~
00
w
-_-
50
')o
*
m0 20
4
0
~0
056x)
as compared to placebo periods, are listed in table 6.
Adverse reactions were observed in seven of the 21
patients, including the 17 patients who completed the
trial and the four patients who were treated for shorter
periods. Drug-related side effects were mild, occurred
generally in patients who had pre-existing symptoms
and did not lead to drug cessation. Congestive heart
failure was accentuated in two individuals, both of
whom had pre-existing symptoms. Increased claudication was noted in two individuals, both of whom had
pre-existing symptomatic peripheral vascular disease.
Serum levels obtained during propranolol treatment
periods, during which adverse reactions were noted,
were in excess of 50 ng/ml for 12 of the 13 episodes.
Adverse reactions were not more common for those individuals who exhibited very high serum levels.
Comparison of angina frequency during run-in and
double-blind placebo periods has already suggested
that following discontinuation of propranolol patients
return to the same level of angina which they had exhibited at the time of initial entry into the study. Four
of the 13 patients who had abrupt transitions from the
160 mg daily propranolol dose to placebo experienced
not only immediate return of angina, but also the
onset of rest and nocturnal angina episodes (unstable
angina) which had not previously been their pattern.
One of these four patients developed frequent venTable 6
Adverse Reactions (21 Patients at Risk)
Lowest dose at which
No. reaction was encountered
Reaction
Fatigue
Heart failure
Extreme bradyeardia (<45)
Claudication
Postural symptoms
Gastrointestinal
Total episodes in seven patients
4
2
2
2
2
1
13
160 (1),
80 (1),
160 (1),
160 (1),
320 (2)
320 (1)
Circulation, Volume
320 (3)
320 (1)
320 (1)
320 (1)
51, June
1975
PROPRANOLOL DOSE RESPONSE IN ANGINA
973
therapy was reinstituted. The only other major cardiovascular event during this study was acute cor320 mg
onary insufficiency, which occurred while the patient
was receiving propranolol. A detailed description of
500 F
these cases, along with three additional case reports,
has been presented elsewhere.1
Discussion
k 320 mg
200 160
7' \This study, which was designed to circumvent the
deficiencies of other clinical trials evaluating
propranolol efficacy, confirms that propranolol, when
1oogiven in appropriate doses to patients with moderate
:0 mg \
to severe angina pectoris, reduces the angina episodes
0'
and improves exercise tolerance. Conservative
,s160 mg
1-* 50160
mg
statistical analysis of our data demonstrated
C
propranolol efficacy for both the 160 (P < 0.07) and
E
)mg
the 320 mg (P < 0.01) daily doses. These data are consistent with previous propranolol trials which have
0
uniformly documented efficacy with daily doses of 160
E 20k
0
mg or more.", 3 The very prominent linear relationship
Percent reduction
between clinical response and the log of the
in angina frequency
A
a0 0- 25
propranolol dosage, which we noted, has previously
@ 26- 75
been observed by Prichard and Gillam.12
Several previous multi-dose trials of propranolol12-15
A
76-100
demonstrated
a favorable response at daily doses of 80
5Fmg, whereas another similarly designed trial reported
efficacy only when the 160 mg dose was reached.16
O80mg
These trials used a double-blind crossover design
causing half of the placebo periods to follow drug
2L
2
40
0
periods which may have accentuated control levels of
10
20
so
50
angina frequency, as occurred in our trial. Although
Percent reducttion in exercise
over-all statistical analysis of the data in this study did
heart rate at sa me exercise level
not suggest efficacy for the 80 mg dose, it was apFigure 10
parent that some individuals reached a sufficient
The data on three selected patients are illustrated in order t
blood level and sufficient degree of beta-blockade to
serum llevels,
emphasize the relationship between prc)pranolol
,pranolo
sem
exhibit a partial therapeutic effect. In six such patients
degree of beta-blockade measured by the piercent reduction in exercise heart rate, the daily propranolol dosa, ge and clinical response
a significant gain in clinical response was noted at
I00r
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
'OH
measured by percent reduction in angis rsa frequency (table 4).
Patient A typifies those individuals who jfail to exhibit a clinical
response while receiving 80 or 160 mg dalily dosages, because in-,
ckd
adequate serum levels and inadequate d
were achieved. Patient B typifies the indivi,
with aninadequate
clinical response while receiving 80 mg/clay, but upon receiving
160 mg or 320 mg/day achieved adequate? serum levels with adequate degrees of beta-blockade. Some patiients (e.g. patient C) exhibited responses at all three oral dosage'levels, largely because
effective degrees of beta-blockade we
propranolol dosages.
sglreeso
rdual
beta-bl
Xreattained.atloe
tricular premature beats, includLing trigeminy and
bigeminy, which were documenited on ambulatory
tape monitoring and disappeared immediately upon
reinstitution of propranolol therap)y. One patient with
unstable angina during placetbo therapy died
suddenly, and another patient in t:he same clinical setting had a myocardial infarction. Jnstable angina persisted in the fourth individual until propranolol
Circulation, Volume 51, June 1975
higher doses.
As is the rule with all trials of drug efficacy for the
treatment of angina pectoris, protocol design influences the conclusions which can be drawn. This
fixed multi-dose trial was designed with a randomly
selected, but fixed sequence of a Latin square for
propranolol dosages. The level of symptoms during
dosages.
during
the double-blind placebo periods was intensified for
13 of the 17 patients following changeover from the
preceding therapeutically effective 160 mg dose
period. Discontinuance of this therapeutic propranolol dose had the effect of restoring the angina
level of many patients back to that observed upon initial entry into the study, rather than to that level obtained after prolonged placebo treatment, when some
patients had probably curtailed activity in order to
lessen the frequency and severity of pain. During drug
periods, they tended to increase activities when this
974
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
could be accomplished without experiencing significant angina. Increased physical activity carried over
from the therapeutically effective 160 mg dose period
into a placebo dose period could account for this
adverse transition effect. In order to partially
eliminate this carryover effect, only data from the last
four weeks of each six-week period were employed.
The data obtained during the 12-week placebo runin period pointed out that symptom reduction due to
placebo effect was prominent for only the initial six to
eight weeks. This is in contrast with earlier reports
which suggested that an 11-week rapport period is
necessary for stabilization.'7 The use of the last four
weeks of the run-in placebo period as control data
provided the most conservative estimate of propranolol efficacy.
The characterization of patients selected for participation in this study influences the extent to which
extrapolations to the general population can be made.
The requirement of five or more angina episodes per
week at the end of a prolonged three-month placebo
period strictly limited inclusion to those patients with
stable, reproducible, severe angina and coronary
artery disease. Fourteen of the 17 patients in our study
had previously experienced myocardial infarctions,
their first in all cases; however, only four had congestive heart failure, all well compensated. The data
strongly suggest that patients with prior infarction but
without uncompensated heart failure respond well to
propranolol therapy.
The adverse reactions observed in this study were
consistent in frequency and type with those observed
in other studies.1' ' Severe angina symptoms occurring
during placebo treatment periods following
propranolol therapy have previously been observed by
Mizgala15 and Zsoter.'6 The episodes of acute myocardial infarction and sudden death which we observed
following abrupt propranolol cessation are similar to
previous cases reported elsewhere.",16, 18-20
A wide range of serum levels was encountered
between individual patients receiving the same dose.
Shand5 noted a seven-fold range in peak plasma levels
(36-212 ng/ml) following a single 80 mg oral dose.
Zacest6 reports plasma levels ranging from 20 to 80
ng/ml following chronic daily dosages of 1.8 to 2.5
mg/kg (160 mg/day), a range which is similar to the
20 to 84 ng/ml range observed for most of our patients
receiving similar dosages. Two individual patients exhibited strikingly high serum levels at all dosages of
propranolol, the explanation for which may be a
metabolic defect in the hydroxylation of propranolol
as suggested by Zacest.6
The data from our study suggest that, for patients
responsive to propranolol therapy, a threshold serum
level in excess of 30 ng/ml is required for a clinical
ALDERMAN ET AL.
response. This clinical response is associated with a
level of beta-adrenergic blockade which produces a
20% or greater reduction in the heart rate attained at
the same exercise level reached at the point of symptom limitation without beta-blockade. The observed
correlation between clinical response and pharmacologic effect as assessed by exercise testing was, in
general, better than the correlation of clinical
response with propranolol serum levels. The fact that
serum levels were obtained at a time somewhat
different from that of exercise testing (maximum 45
min) is probably not significant. Data from Evans and
Shand21 suggest that for blood samples drawn 90 to
180 min following repeated oral doses, a variation of
approximately 20% in plasma levels is found. This
residual variability between serum levels and receptor
effect is probably not due to the pharmacologic effects
of an active metabolite, in view of the work of
Cleaveland and Shand,22 which shows that the effect
of chronic oral administration of propranolol can be
attributed entirely to circulating levels of the parent
drug. An additional explanation for this variability
between clinical response and serum levels may be
differences of propranolol kinetics between the intravascular compartment and the extravascular sites
where the drug elicits its pharmacologic effect.
Differences in propranolol distribution between body
compartments has been suggested by analysis of
propranolol data by Perrier and Gibaldi."
For most patients, little increase in beta-blockade
was observed between the 160 mg and 320 mg daily
dosages, suggesting that a relative maximum had
been reached. Clinical response was, however, considerably enhanced at the higher dose, probably
reflecting the fact that the degree of beta-blockade
was assessed at a time of peak propranolol serum
levels, whereas clinical response is assessed over a 4week interval. The propranolol serum level at the
point of relative maximum beta-blockade was
generally in excess of 50 ng/ml, which is consistent
with the levels noted by Bodem et al.7 at the time of
maximum reduction in exercise tachycardia. Coltart
and Shand8 similarly observed plasma levels of 40 to
75 ng/ml at the time of maximum beta-blockade
following single propranolol doses. Our data,
therefore, suggest that substantial beta-blockade, as
measured by exercise heart rate suppression, is required for a clinical response. Thus, the specific
propranolol dosage required for successful therapy
depends both on the blood drug level itself and upon
the extent of beta-adrenergic blockade which results.
Acknowledgment
We gratefully acknowledge the contribution of Ms. Pamela Lewis
in organizing and tabulating outpatient visit data. Messrs. Marvin
Quan, James Ryan and William Northeroft provided assistance in
Circulation, Volume 51, June 1975
PROPRANOLOL DOSE RESPONSE IN ANGINA
the organizational and computational aspects of this study. We are
appreciative of the outpatient nursing assistance provided by Mary
Nivison and Amelia Turczinski, the comments provided by Dr.
Peter Meffin and the editorial assistance provided by Mrs. Dorothy
McCain.
References
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
1. ALDERMAN EL, HARRISON DC: Beta-adrenergic blockade in the
management of angina pectoris. In Circulatory Effects and
Clinical Uses of Beta-Adrenergic Blocking Drugs, edited by
HARRISON DC. Amsterdam, Excerpta Medica, 1971, p 67
2. SHARMA B, TAYLOR SH: A critical review of the symptomatic
electrocardiographic and circulatory effects of adrenergic
beta-receptor antagonists in angina pectoris. In New
Perspectives in Beta-Blockade edited by BURLEY DM, FRIER
JH, RANDEL RK, TAYLOR SH. Horsham, England, CIBA
Laboratories, 1973, p 129
3. GIANELLY RE, GOLDMAN RH, TREISTER B, HARRISON DC:
Propranolol in patients with angina pectoris. Ann Intern
Med 67: 1216, 1967
4. GRANT RHE, KEELAN P, KER.NOHAN RJ, LEONARD JC,
NANCEKIEVILL L, SINCLAIR K: Multicenter trial of
propranolol in angina pectoris. Am J Cardiol 18: 361, 1966
5. SHArND DG, NUCKOLLS EM, OATES JA: Plasma propranolol
levels in adults, with observations in four children. Clin
Pharmacol Ther 11: 112, 1970
6. ZACEST R, KOCH-WESER J: Relation of propranolol plasma level
to beta-blockade during oral therapy. Pharmacology 7: 178,
1972
7. BODEM G, BROM MELL HL, WEIL JV, CHIDSEY CA:
Pharmacodynamic studies of beta-adrenergic antagonism induced in man by propranolol and practolol. J Clin Invest 52:
747, 1973
8. COLTART DJ, SHAND DG: Plasma propranolol levels in the
quantitative assessment of beta-adrenergic blockade in man.
Br Med J 3: 731, 1970
9. ALDERNIAN EL, DAVIES RO, FRIEDMAN JP, GRAHAM AF,
MATLOF HJ, HARRISON DC: Practolol in patients with angina
pectoris. Clin Pharmacol Ther 14: 175, 1973
Circulation, Volume 51, June 1975
975
10. BRUCE RA, HORNSTEN TR: Exercise stress testing in evaluation
of patients with ischemic heart disease. Prog Cardiovasc Dis
11: 371, 1969
11. ALDERMAN EL, COLTART DJ, WETTACH GE, HARRISON DC:
Coronary artery syndromes after sudden propranolol
withdrawal. Ann Intern Med 81: 625, 1974
12. PRICHARD BNC, GILLAM PMS: Assessment of propranolol in
angina pectoris. Clinical dose response curve and effect on
electrocardiogram at rest and on exercise. Br Heart j 33: 473,
1971
13. GILLAM PMS, PRICHARD BNC: Propranolol in the therapy of
angina pectoris. Am J Cardiol 18: 366, 1966
14. HEBB AR, GODWIN TF, GLNTON RW: New beta-adrenergic
blocking agent. Propranolol in the treatment of angina pectoris. Can Med Assoc J 98: 246, 1968
15. MIZCALA HF, KHAN AS, DAVIES RO: Propranolol in the
prophylactic treatment of angina pectoris. Can Med Assoc J
100: 756, 1960
16. ZSOTER TT, BEANLANDS DS: Propranolol in angina pectoris.
Arch Intern Med 124: 594, 1969
17. COLE SL, KAYE H, GRIFFITH GC: Assay of anti-anginal drugs.
The rapport period. JAMA 168: 275, 1958
18. SLOME R: Withdrawal of propranolol and myocardial
infarction. Lancet 1: 156, 1973
19. DIAZ RG, SOMBERG JC, FREEMAN E, LEVITT B: Withdrawal of
propranolol and myocardial infarction. Lancet 1: 1068, 1973
20. NELLEN M: Trial of two beta-blockade drugs, CIBA 39089 and
ICI 50172, in angina pectoris. S Afr Med J 6 (suppl): 15,
1969
21. EXANS GH, SHAND DG: Disposition of propranolol. V. Drug
accumulation and steady-state concentrations during
chronic oral administration in man. Clin Pharmacol Ther 14:
487, 1973
22. CLEAV'ELAND CR, SHAND DG: Effect of route of administration
on the relationship between beta-adrenergic blockade and
plasma propranolol level. Clin Pharmacol Ther 13: 181, 1972
23. PERRIER D, GIBALDI M: Drug concentrations in the plasma as an
index of pharmacologic effect. J Clin Pharmacol 14: 415,
1974
Dose response effectiveness of propranolol for the treatment of angina pectoris.
E L Alderman, R O Davies, J J Crowley, M G Lopes, J Z Brooker, J P Friedman, A F
Graham, H J Matlof and D C Harrison
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Circulation. 1975;51:964-975
doi: 10.1161/01.CIR.51.6.964
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1975 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/51/6/964
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not
the Editorial Office. Once the online version of the published article for which permission is being
requested is located, click Request Permissions in the middle column of the Web page under Services.
Further information about this process is available in the Permissions and Rights Question and Answer
document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/