Download Prognostic Significance of Baseline Heart Rate and Its

Original article
Prognostic Significance of Baseline Heart Rate and
Its Interaction With Beta-Blocker Use in Resistant
Hypertension: A Cohort Study
Gil F. Salles1, Claudia R.L. Cardoso1, Luciane L. Fonseca1, Roberto Fiszman1, Elizabeth S. Muxfeldt1
Methods
In a prospective study, 528 patients with resistant hypertension had
HR measured on clinical examination, electrocardiography (ECG), and
during ambulatory blood pressure monitoring. Primary endpoints
were a composite of fatal and nonfatal cardiovascular events, all-cause
and cardiovascular mortality. Multivariable Cox regression was used to
assess the associations between slow HR (< 60 bpm or < 55 bpm for
nighttime HR) and fast HR (> 75 bpm or > 70 bpm for nighttime HR)
and the occurrence of endpoints in relation to the reference middle HR
(60–75 bpm) subgroup.
Results
After a median follow-up of 4.8 years, 62 patients died, 44 from cardiovascular causes; and 94 cardiovascular events occurred. Fast and slow
HRs were mainly predictors of mortality, and ambulatory HRs were
more significant risk markers than clinic or ECG HR. A slow 24-hour HR
was a predictor of the composite endpoint (hazard ratio, 2.0; 95% confidence interval [CI], 1.2–3.4), whereas both slow and fast ambulatory
HRs were predictors of cardiovascular mortality (hazard ratio, 2.3; 95%
CI, 1.1–5.1). Four hundred and seventeen patients (79%) were using
beta-blockers and this affected the HR prognostic value. A fast HR was
a more significant risk marker in patients using beta-blockers, whereas
a slow HR was a predictor mainly in those not using beta-blockers.
Conclusions
There is an overall U-shaped relationship between HRs, particularly
when measured during ambulatory monitoring, and prognosis in
resistant hypertension. A fast HR is a significant predictor in patients
using beta-blockers, while a slow heart rate is a more important predictor in those not using beta-blockers.
Keywords: beta-blocker; cardiovascular risk; heart rate; mortality;
resistant hypertension.
doi:10.1093/ajh/hps004
A fast baseline resting heart rate (HR) has been consistently demonstrated to represent a risk marker for all-cause
and cardiovascular mortality in population-based samples:1–3 and in several cardiovascular diseases, such as acute
or chronic coronary heart disease,4,5 and in heart failure or
left ventricular dysfunction.6 Moreover, in these 2 conditions
treatment-induced HR reductions have been demonstrated
to contribute to a better prognosis.7,8
Nevertheless, the prognostic value of fast HRs in hypertensive individuals is more controversial. This value has
been demonstrated in some studies,9,10 but not in others,11,12
whereas one study13 found a J-shaped relationship between
HR and adverse outcomes. More recently, this relationship
was further influenced by the so-called the “beta-blocker
paradox,”14 in which the protective cardiovascular effect of
beta-blocker treatment was lower than for other antihypertensive treatments, particularly for stroke occurrence, in
spite of similar blood pressure (BP) reductions. This unexpected, unfavorable effect may be attributed to beta-blockers’
adverse metabolic effects, such as worsening of the lipid profile and decreased insulin sensitivity, and also to higher central BPs.15 Furthermore, it was shown that slower HRs while
on beta-blocker treatment were associated with increased
risk of cardiovascular events and mortality in hypertensive
patients.16
Resistant hypertension, defined as the lack of clinic BP
control despite an optimal antihypertensive treatment with
at least 3 drugs including a diuretic,17 is a rather common but
understudied condition that has a very high cardiovascular
risk profile.18 As far as we know, HR has never been evaluated
Correspondence: Gil F. Salles ([email protected]).
1Department of Internal Medicine, University Hospital Clementino
Introduction
Initially submitted April 17, 2012; date of first revision July 10, 2012;
accepted for publication July 28, 2012.
218 American Journal of Hypertension 26(2) February 2013
Fraga Filho, Faculdade de Medicina, Universidade Federal do Rio de
Janeiro, Brazil.
© American Journal of Hypertension, Ltd 2012. All rights reserved.
For Permissions, please email: [email protected]
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
Background
The prognostic significance of heart rate (HR) and its relationship with
beta-blocker use are controversial and have never been evaluated in
resistant hypertension.
Heart Rate and Prognosis in Resistant Hypertension
as a prognostic marker in individuals with resistant hypertension. Because these patients were already stable on 3 or
more antihypertensive drug treatments at baseline, we were
able to investigate not only the prognostic value of HR but
also the influence of any HR-limiting therapy, particularly
beta-blocker treatment. Therefore, we analyzed the prognostic significance of both slow and fast HRs, measured in different contexts (clinic, electrocardiography (ECG), and during
ambulatory BP monitoring), for cardiovascular morbidity
and mortality, as well as the influence of beta-blocker therapy
on these prognostic relationships, in a large cohort of patients
with resistant hypertension with up to 9 years of follow-up.
Patients and baseline procedures
This was a retrospective analysis of prospectively collected
data and included 528 patients with resistant hypertension
enrolled between January 1999 and December 2004 in the
hypertension outpatient clinic of the University Hospital
Clementino Fraga Filho. Study protocols were approved by
the Research Ethics Committee of Faculty of Medicine and
University Hospital, and informed consent was obtained
from all participants. The enrollment criteria, baseline protocol, and diagnostic definitions have been detailed previously.18–20 In brief, all patients referred to our clinic who
fulfilled the criteria for resistant hypertension (office BP
≥ 140/90 mm Hg, using ≥ 3 antihypertensive drugs in full
dosages, always including a diuretic, and considered at least
moderately adherent by a validated questionnaire21) were
submitted to a standard protocol that included a thorough
clinical examination (with particular attention to the presence of cardiovascular risk factors and target-organ damage;
diagnostic definitions have been previously detailed19,20), a
laboratory evaluation, a 12-lead ECG, and 24-hour ambulatory BP monitoring (ABPM). Patients with secondary
hypertension were excluded from the cohort, except those
with sleep apnea syndromes, which was not routinely investigated, and those with chronic parenchymal kidney diseases (only 17 patients had an estimated creatinine clearance
≤ 30 ml/min/1.73 m2 at entry). For this analysis, 20 patients
with nonsinusal rhythms on ECG were also excluded. With
patients in the sitting position, office BP was measured
twice by a trained physician, using a calibrated mercury
sphygmomanometer and suitably sized cuffs; the BP was
the mean between the 2 readings. HR was measured twice
immediately before each BP measurement by radial artery
palpation during 30s; the clinic HR was the mean between
the 2 measurements. Electrocardiographic HR was automatically measured from standard resting 12-lead ECGs
(CardioFax V electrocardiograph, Nihon-Kohden). ABPM
was recorded using Mobil O Graph (version 12) equipment
(Dynamapa, Cardios, São Paulo, Brazil), approved by the
British Society of Hypertension. A reading was taken every
15 minutes throughout the day and every 30 minutes at
night. The nighttime period was ascertained for each patient
from registered diaries. Parameters evaluated were mean
24-hour, daytime, and nighttime systolic and diastolic BPs,
and HRs. Nocturnal BP and HR dipping was defined as a
Follow-up and endpoints
Patients were followed up regularly until December 2007.
All patients on beta-blocker treatment at baseline remained
on treatment throughout the follow-up. The observation
period was considered as the number of months from the
first clinical evaluation to the date of the last clinical visit
or the first endpoint. Thirty-five subjects (6.6%) were lost to
follow-up and were considered censored observations at the
date of their last hospital visit. The primary endpoints were
a composite of all fatal or nonfatal cardiovascular events,
all-cause and cardiovascular mortalities. Definitions of endpoints have been recently detailed.18 In brief, cardiovascular
events were as follows: fatal and nonfatal acute myocardial
infarctions (AMI), sudden cardiac deaths, new-onset heart
failure, death from progressive heart failure, any myocardial
revascularization procedure, fatal and nonfatal strokes, any
aortic or lower limb revascularization procedure, amputation above the ankle, and deaths from aortic or peripheral
arterial disease. Endpoints were ascertained from medical
records, death certificates, and interviews with attending
physicians and patient families, using a standard questionnaire reviewed by an independent observer.
Statistical analysis
Continuous data were expressed as means and SD if normally distributed or as medians and interquartile range (IQR)
if asymmetrically distributed. Patients were divided into
3 subgroups according to HRs measured at the 5 contexts
(clinic, ECG, 24-hour, daytime, and nighttime) as normal
(between 60 and 75 bpm), slow (< 60 bpm), and fast (> 75
bpm) HRs, except for the nighttime HR, which had a 5-bpm
lower cut-off value. The upper cut-off value (75 bpm) was
chosen because most previous studies that demonstrated
any prognostic value of fast HRs reported threshold cut-offs
between 70 bpm and 90 bpm7-10 and also because it divided
our patients into subgroups with sample sizes large enough to
perform comprehensive statistical analyses. The lower cut-off
value (60 bpm) is the traditional one used to define bradycardia. Baseline characteristics between patients with normal,
slow, and fast HRs were compared by one-way analysis of
variance, Kruskal–Wallis test, or χ2 test. The assessment of the
fast and slow HRs as independent predictors of the endpoints
was performed by Kaplan–Meier estimation of survival
curves, compared by log-rank test, and by multivariable Cox
proportional hazards analysis, with the normal HR subgroup
as the reference. Cox models were first adjusted for age and
sex and then, for the composite endpoint, further adjusted for
all potential risk factors: age, sex, smoking, physical inactivity, diabetes, dyslipidemia, previous cardiovascular diseases,
serum creatinine, number of antihypertensive drugs in use,
ambulatory 24-hour systolic BP, and dipping pattern. Because
American Journal of Hypertension 26(2) February 2013 219
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
Methods
≥ 10% reduction in the average values at night compared
with the mean daytime values. BP variability was expressed
as the SD of all valid 24-hour BP measurements. Patients
were also classified as uncontrolled hypertension (24-hour
BP ≥ 130/80 mm Hg) or controlled resistant hypertension
(24-hour BP < 130/80 mm Hg).19
Salles et al.
Results
Baseline characteristics and follow-up endpoints
Distributions of clinic, ECG, daytime, and 24-hour HR
were similar: mean (SD) and median (IQR) were respectively
72 (12), 68 (64–80); 70 (14), 68 (61–78); 71 (12), 69 (63–78);
and 69 (11), 67 (61–76) bpm; whereas nighttime HR was
slower: 63 (10), 61 (55–69) bpm. Table 1 outlines the baseline
characteristics of patients with normal (60–75 bpm), slow (<
60 bpm), and fast (> 75 bpm) mean 24-hour ambulatory HR.
Patients with slow HRs were older and leaner, had a lower
prevalence of diabetes and lower fasting glycemia levels, and
had lower diastolic BPs than patients with normal HRs. On
the other hand, patients with elevated HRs were younger, were
more frequently females and current smokers, had higher
prevalence of diabetes and higher fasting glycemia levels,
used less antihypertensive medications (mainly beta-blockers) but used more calcium channel blockers (particularly of
the non-dihydropyridine class) had higher ambulatory BPs,
and had higher prevalence of true (uncontrolled) resistant
hypertension than patients with normal HRs. Otherwise,
short-term BP variability parameters did not differ among
HR subgroups. Categorization into the same subgroups by
clinic or ECG HR followed similar patterns of characteristics.
Overall, 417 patients (79%) were using beta-blockers at baseline, primarily (96% of them) propranolol or atenolol. None
were using the new vasodilating beta-blockers.
After a median follow-up of 4.8 years (range, 1 month–
9 years), 94 fatal or nonfatal cardiovascular events occurred
(incidence rate: 3.9 per 100 patient-years of follow-up): 42
strokes, 21 AMIs, 14 myocardial revascularizations, 9 newonset heart failures, 4 sudden deaths, and 4 aortic or peripheral
artery events. There were 62 all-cause deaths (incidence: 2.4 per
100 patient-years), 44 from cardiovascular causes (incidence:
1.7 per 100 patient-years). Patients with either fast or slow HRs
220 American Journal of Hypertension 26(2) February 2013
had an increased incidence rate of endpoints in relation to the
reference group with HRs between 60 and 75 bpm (Table 1).
Prognostic value of fast and slow HRs and influence
of beta-blocker use
Table 2 presents the results of Cox survival analyses for
the association between the 5 measurements of HR and
the 3 primary endpoints. In general, ambulatory HRs were
more significant prognostic predictors than clinic or ECG
HRs and, more important, for all-cause and cardiovascular mortalities than for the composite endpoint. Clinic HR
was not a predictor of any outcome, and a slow ECG HR
was a predictor only for the composite endpoint. Otherwise,
fast ambulatory HRs, 24-hour, daytime, or nighttime, were
significant predictors of all-cause and cardiovascular mortalities with hazard ratios ranging from 1.9 to 2.3. Slow nocturnal HR was also a predictor of all-cause mortality, and
slow 24-hour HR was predictive of cardiovascular mortality
and of the composite endpoint. The worse prognosis associated with both fast and slow ambulatory HRs was also
demonstrated by Kaplan–Meier survival curves (Figure 1).
A blunted nocturnal HR dipping was not a predictor of any
of the endpoints.
Table 3 shows the prognostic value of fast and slow HRs in
patients stratified by beta-blocker use at baseline. The worse
prognosis associated with fast HRs was observed exclusively
in patients using beta-blockers, whereas the worse prognosis of slow HRs was mainly observed in patients not using
beta-blockers. The only exception was the predictive value
of slow 24-hour HR for the composite endpoint, in which
a significant 2-fold increased risk in patients using betablockers was found, although the risk was higher (3-fold)
in the subgroup not using beta-blockers. The use of betablockers alone had no independent predictive value, neither harm nor benefit, for any of the endpoints evaluated.
Short-term BP variability parameters also did not have any
prognostic value and did not affect the prognostic implications of HR.
Discussion
As a result of this prospectively collected data study,
we made three important findings. First, we observed a
U-shaped relationship between HR and outcome in patients
treated for resistant hypertension: both slow (< 60 bpm) and
fast (> 75 bpm) HRs were significant predictors of worse
prognosis, mainly for all-cause and cardiovascular mortality.
Second, ambulatory HRs were more significant predictors
than clinic or ECG HRs. Third, HR reduction by betablocker therapy did not appear to be harmful. However, this
cannot be completely ruled out because we found a significantly increased risk of slow 24-hour HR for the composite endpoint of fatal and nonfatal cardiovascular events in
patients using beta-blockers. With this exception, all other
increased risks associated with slow HRs were observed in
patients not using beta-blockers. Otherwise, the adverse
outcomes associated with fast HRs were always observed in
patients on beta-blocker treatment.
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
of the smaller number of events and to avoid overfitting, the
multivariable Cox analyses for all-cause and cardiovascular
mortalities were adjusted only for their significant predictors. Furthermore, all analyses were also adjusted for any
HR-limiting therapy (beta-blockers, non-dihydropyridine
calcium channel blockers, or clonidine), and an interaction
term between HR and beta-blocker use was tested in all multivariate models and kept in the models interaction product
terms with P < 0.10. Regardless of whether any interaction
between the prognostic value of HR and beta-blocker use
was detected, all survival analyses were repeated in patients
stratified by beta-blocker use. Because of the small number
of patients and consequently of endpoints in the subgroup
without beta-blockers at baseline, the survival analyses in this
subgroup was adjusted only for age and sex. Results were presented as hazard ratios with their 95% confidence intervals.
The proportional hazards assumption was tested and no violation was observed. Other interactions between HR subgroups
and age (< 65 and ≥ 65 years), sex, and presence of diabetes
and of cardiovascular disease at baseline were also examined,
but none had P values < 0.10. Statistics were performed with
SPSS version 19.0 package (SPSS Inc., Chicago, IL), and a
2-tailed probability value < 0.05 was considered significant.
Heart Rate and Prognosis in Resistant Hypertension
Table 1. Baseline characteristics and crude incidence rates of endpoints during follow-up in patients grouped according to 24-hour
ambulatory heart rate (normal, 60–75 bpm; slow, < 60 bpm; and fast, > 75 bpm)
Patients with 24-h
Patients with 24-h
Patients with 24-h
Characteristic
HR 60–75 bpm (n = 262)
HR < 60 bpm (n = 112)
HR > 75 bpm (n = 154)
P value
Sex (% male)
30.3
34.2
23.9
0.18
Age (years)
66.2 (11.2)
70.4 (9.7)
60.9 (11.3)
<0.001
Body mass index (kg/m2)
30.2 (6.2)
28.9 (5.4)
30.7 (5.3)
0.04
Physical inactivity (%)
74.8
76.6
73.9
0.90
Diabetes (%)
37.4
22.5
49.3
<0.001
7.5
8.1
16.9
0.009
90.2
86.5
84.5
0.23
Previous cardiovascular disease (%)
44.9
49.5
43.0
0.57
Echocardiographic left ventricular hypertrophy (%)
74.2
77.7
73.3
0.72
Fasting glycemia (mmol/L)
6.5 (2.3)
5.9 (1.4)
7.3 (3.4)
<0.001
Total cholesterol (mmol/L)
5.82 (1.35)
5.74 (1.32)
5.56 (1.21)
0.16
High-density lipoprotein cholesterol (mmol/L)
1.22 (0.34)
1.25 (0.33)
1.19 (0.30)
0.32
1.57 (1.15–2.25)
1.51 (1.05–2.26)
1.45 (1.07–2.11)
0.34
Serum creatinine (μmol/L)
80 (71–106)
80 (71–106)
80 (62–106)
0.56
Albuminuria (mg/24 h)
19 (10–53)
17 (9–35)
21 (11–50)
0.14
Number of antihypertensive drugs in use
4 (3–4)
4 (3–4)
3 (3–4)
0.03
Angiotensin-converting enzyme inhibitors/
angiotensin II receptor blockers (%)
90.9
89.2
88.0
0.64
βeta-blockers (%)
82.7
88.3
66.2
<0.001
Calcium channel blockers (%)
46.9
50.5
59.2
0.06
Dihydropyridines (%)
38.2
47.8
43.2
0.001
Triglycerides (mmol/L)
Antihypertensive treatment
Non-dihydropyridines (%)
8.7
2.7
16.0
Direct vasodilators (%)
34.6
33.3
31.0
0.76
Central α agonists (%)
Clinic SBP (mm Hg)
17.7
11.7
13.4
0.26
179 (28)
177 (28)
175 (25)
0.39
Clinic DBP (mm Hg)
99 (18)
93 (17)
102 (17)
<0.001
24-h SBP (mm Hg)
136 (19)
133 (19)
141 (19)
0.002
79 (12)
73 (11)
85 (13)
<0.001
24-h SBP variability (SD) (mm Hg)
24-h DBP (mm Hg)
17.2 (4.4)
17.0 (4.1)
17.2 (4.1)
0.91
24-h DBP variability (SD) (mm Hg)
12.0 (2.9)
11.7 (2.8)
12.4 (2.7)
0.12
62.7
70.3
60.6
0.25
Nondipping SBP (%)
Nondipping heart rate (%)
41.0
47.3
35.4
0.17
Uncontrolled ambulatory blood pressure (%)
66.5
57.7
76.1
0.008
Number of all-cause deaths (incidence rate,
per 100 patient-years of follow-up)
24 (2.04)
16 (3.56)
22 (3.26)
0.14
Number of cardiovascular deaths (incidence rate,
per 100 patient-years of follow-up)
15 (1.27)
13 (2.90)
16 (2.37)
0.04
Number of total cardiovascular events (incidence
rate, per 100 patient-years of follow-up)
38 (3.40)
25 (5.93)
31 (4.91)
0.05
Values are means (SDs) or proportions, except for triglycerides, serum creatinine, albuminuria, and number of antihypertensive drugs that
are medians (interquartile ranges) and endpoints occurrence that is absolute number (incidence rate).
Abbreviations: DBP, diastolic blood pressure; HR, heart rate; SBP, systolic blood pressure.
American Journal of Hypertension 26(2) February 2013 221
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
Current smoking (%)
Dyslipidemia (%)
Salles et al.
Table 2. Results of Cox survival analyses for the associations between different measurements of heart rate and primary endpoints
Endpoint
Age and sex adjusted
Multivariate adjusteda
Interaction with
hazard ratios (95% CI)
hazard ratios (95% CI)
beta-blocker use (P value)
0.73 (0.47–1.14)
0.77 (0.48–1.22)
Composite endpoint (n = 94)
Clinic heart rate
fast
slow
0.80 (0.37–1.77)
0.84 (0.38–1.87)
ECG heart rate
fast
0.63 (0.19–2.17)
0.46 (0.13–1.63)
slow
2.55 (0.92–7.06)
2.98 (1.06–8.37)*
24-h heart rate
Daytime heart rate
1.70 (1.01–2.88)*
1.62 (0.94–2.80)
1.69 (1.00–2.87)*
1.96 (1.15–3.36)*
fast
1.54 (0.93–2.56)
1.52 (0.90–2.56)
slow
1.39 (0.79–2.45)
1.56 (0.88–2.78)
fast
1.59 (0.93–2.70)
1.41 (0.81–2.43)
slow
Nondipping heart rate
0.007
0.49
0.22
0.38
1.38 (0.83–2.28)
1.58 (0.93–2.66)
1.24 (0.81–1.91)
1.05 (0.68–1.63)
0.52
0.09
Total mortality (n = 62)
Clinic heart rate
ECG heart rate
fast
0.76 (0.23–2.49)
0.47 (0.14–1.64)
slow
1.98 (0.38–10.26)
2.71 (0.52–4.23)
fast
1.71 (0.99–2.95)
1.42 (0.80–2.50)
slow
0.62 (0.29–1.32)
0.61 (0.28–1.31)
24-h heart rate
fast
2.08 (1.13–3.85)*
1.92 (1.02–3.62)*
slow
1.50 (0.79–2.85)
1.64 (0.86–3.13)
Daytime heart rate
fast
2.30 (1.27–4.16)*,*
2.09 (1.13–3.88)*
slow
1.29 (0.64–2.61)
1.39 (0.68–2.83)
fast
2.38 (1.29–4.40)*,*
2.22 (1.19–4.13)*
slow
1.60 (0.86–2.96)
1.87 (1.00–3.50)*
1.33 (0.80–2.20)
1.19 (0.71–2.00)
0.78
0.77 (0.40–1.48)
0.63 (0.30–1.28)
0.22
Nighttime heart rate
Nondipping heart rate
0.49
0.47
0.59
0.56
Cardiovascular mortality (n = 44)
Clinic heart rate
fast
slow
0.47 (0.11–1.97)
0.36 (0.08–1.57)
ECG heart rate
fast
1.46 (0.76–2.79)
1.31 (0.68–2.52)
slow
0.46 (0.17–1.21)
0.48 (0.18–1.29)
24-h heart rate
Daytime heart rate
Nighttime heart rate
fast
2.46 (1.14–5.30)*
2.34 (1.07–5.13)*
slow
2.14 (1.00–4.61)*
2.30 (1.06–4.98)*
fast
2.25 (1.09–4.64)*
2.22 (1.06–4.67)*
slow
1.50 (0.65–3.43)
1.63 (0.70–3.77)
fast
2.21 (1.04–4.72)*
1.84 (0.84–4.01)
slow
Nondipping heart rate
1.63 (0.79–3.38)
1.76 (0.84–3.69)
1.17 (0.64–2.17)
0.98 (0.52–1.85)
0.16
0.41
0.39
0.21
0.84
Values are hazard ratios (95% CIs). Fast and slow heart rates were analyzed in relation to the reference group, with heart rates between
60 bpm and 75 bpm (55 bpm and 70 bpm for the nighttime period) and nondipping heart rate (nocturnal reduction < 10%) in relation to normal
dipping (reduction ≥ 10%).
a For the composite endpoint, adjusted for age, sex, smoking, physical inactivity, diabetes, dyslipidemia, previous cardiovascular diseases,
serum creatinine, number of antihypertensive drugs in use, ambulatory 24-hour systolic blood pressure, dipping pattern, and any rate-limiting
therapy (beta-blocker, non-dihydropyridine calcium channel blocker, or clonidine).
For total mortality, adjusted for age, sex, smoking, diabetes, serum creatinine, number of antihypertensive drugs in use, dipping pattern, and
any rate-limiting therapy.
For cardiovascular mortality, adjusted for age, sex, smoking, number of antihypertensive drugs in use, dipping pattern, and any rate-limiting
therapy. * P < 0.05, *,* P < 0.01. CI, confidence interval; ECG, electrocardiographic.
222 American Journal of Hypertension 26(2) February 2013
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
Nighttime heart rate
fast
slow
0.53
Heart Rate and Prognosis in Resistant Hypertension
0.40
Total cardiovascular events
A
0.35
Heart Rate < 60 bpm
0.30
Heart Rate > 75 bpm
0.25
Heart Rate 60 – 75 bpm
0.20
0.15
0.10
Log rank test
P = 0.052
0.05
0
Number of patients at risk:
Heart Rate 60 – 75 bpm:
Heart Rate < 60 bpm:
Heart Rate > 75 bpm:
24
48
72
96
120
Follow-up (months)
262
112
154
231
96
127
157
49
93
62
12
35
12
2
8
0.30
B
All-cause mortality
0.25
Heart Rate > 75 bpm
0.20
Heart Rate < 60 bpm
Heart Rate – 75 bpm
0.15
0.10
Log rank test
P = 0.14
0.05
0.00
0
24
48
72
96
120
Follow-up (months)
0.20
C
Cardiovascular mortality
Heart Rate > 75 bpm
Heart Rate < 60 bpm
0.15
Heart Rate 60 – 75 bpm
0.10
0.05
Log rank test
P = 0.039
0.00
0
Number of patients at risk:
Heart Rate 60 – 75 bpm:
Heart Rate < 60 bpm:
Heart Rate >75 bpm:
24
48
72
96
120
Follow-up (months)
262
112
154
240
100
132
170
56
104
68
14
40
12
3
9
Figure 1. Kaplan–Meier estimation of cumulative incident total cardiovascular events (A), all-cause deaths (B), and cardiovascular deaths (C) in patients
divided according to 24-hour mean heart rate.
American Journal of Hypertension 26(2) February 2013 223
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
0.00
Salles et al.
Table 3. Results of Cox survival analyses for the associations between different measurements of heart rate and primary endpoints in
patients stratified by beta-blocker use at baseline
Fast heart rate (> 75 bpm)
Endpoint
Age and sex
Multivariate
adjusted hazard ratios
adjusteda hazard ratios
(95% CI)
(95% CI)
Beta-blocker strata
Slow heart rate (< 60 bpm)
Age and sex
Multivariate
adjusted hazard ratios adjusteda hazard ratios
(95% CI)
(95% CI)
Composite endpoint (with BB n = 74, without BB n = 20)
Clinic heart rate
ECG heart rate
Daytime heart rate
Nighttime heart rate
0.83 (0.50–1.39)
0.56 (0.22–1.41)
with BB
1.13 (0.64–2.01)
without BB
0.63 (0.18–2.16)
with BB
1.98 (1.10–3.57)*
without BB
1.35 (0.35–3.18)
with BB
1.91 (1.08–3.38)*
without BB
1.54 (0.92–2.56)
with BB
1.64 (0.89–3.00)
without BB
1.51 (0.48–4.72)
0.89 (0.53–1.49)
0.91 (0.39–2.14)
0.90 (0.38–2.15)
0.54 (0.06–3.87)
1.01 (0.56–1.83)
0.80 (0.46–1.38)
0.81 (0.46–1.41)
4.40 (1.36–14.21)*
1.91 (1.04–3.50)*
1.72 (0.97–3.06)
1.98 (1.10–3.55)*
3.04 (0.68–13.58)
1.93 (1.07–3.45)*
1.58 (0.85–2.86)
1.75 (0.95–3.23)
1.56 (0.88–2.78)
1.47 (0.79–2.74)
1.23 (0.70–2.15)
1.35 (0.75–2.42)
4.81 (1.26–18.34)*
Total mortality (with BB n = 49, without BB n = 13)
Clinic heart rate
with BB
without BB
0.71 (0.22–2.37)
ECG heart rate
with BB
2.00 (1.08–3.71)*
without BB
1.06 (0.34–3.37)
24-h heart rate
with BB
2.60 (1.31–5.13)*
without BB
0.90 (0.24–3.45)
Daytime heart rate
with BB
2.82 (1.46–5.45)*,*
Nighttime heart rate
0.76 (0.40–1.44)
without BB
1.06 (0.30–3.74)
with BB
2.62 (1.32–5.24)*,*
without BB
2.36 (0.59–9.52)
0.72 (0.37–1.41)
0.40 (0.10–1.66)
1.74 (0.91–3.32)
0.61 (0.27–1.38)
2.21 (1.08–4.52)*
1.49 (0.74–2.99)
0.32 (0.08–1.39)
2.41 (0.36–16.03)
0.59 (0.26–1.36)
0.88 (0.10–7.60)
1.55 (0.76–3.16)
3.36 (0.50–22.58)
2.55 (1.27–5.11)*,*
1.33 (0.63–2.80)
1.41 (0.66–3.03)
1.82 (0.20–17.85)
2.34 (1.16–4.73)*
1.41 (0.72–2.76)
1.51 (0.76–3.00)
5.74 (1.04–31.84)*
Cardiovascular mortality (with BB n = 35, without BB n = 9)
Clinic heart rate
ECG heart rate
24-h heart rate
Daytime heart rate
Nighttime heart rate
with BB
0.72 (0.34–1.54)
without BB
1.06 (0.25–4.49)
with BB
2.04 (1.00–4.18)*
without BB
0.53 (0.11–2.67)
with BB
3.18 (1.35–7.50)*
without BB
0.82 (0.16–4.08)
with BB
3.20 (1.41–7.23)*,*
without BB
0.58 (0.13–2.68)
with BB
2.49 (1.07–5.82)*
without BB
1.33 (0.26–6.51)
0.67 (0.30–1.48)
0.28 (0.04–2.07)
0.19 (0.03–1.49)
1.98 (0.20–19.23)
1.85 (0.90–3.82)
0.42 (0.14–1.25)
0.42 (0.14–1.28)
1.43 (0.15–14.14)
2.94 (1.22–7.10)*
2.04 (0.89–4.74)
2.13 (0.90–5.03)
3.13 (0.50–19.37)
3.08 (1.32–7.14)*,*
1.53 (0.62–3.78)
1.65 (0.66–4.13)
1.93 (0.21–17.83)
2.40 (1.01–5.69)*
1.33 (0.60–3.02)
1.35 (0.59–3.12)
6.31 (1.29–38.60)*
Values are hazard ratios (95% CIs).
a Multivariate Cox analyses in patients using beta-blockers were adjusted for the same covariates detailed in Table 2. Multivariate analyses
were not performed in patients not using beta-blockers because of the small number of events. * P < 0.05, *,* P < 0.01.
Abbreviations: BB, beta-blockers; CI, confidence interval; ECG, electrocardiographic.
Although the prognostic importance of fast HRs has
been consistently established in population-based studies
and in coronary heart disease and heart failure patients,1–8
until 2006 only 4 studies have reported on this prognostic importance in hypertensive patients. 9,10,22,23 A 2006
224 American Journal of Hypertension 26(2) February 2013
consensus document24 recognized that the evidence linking
HR and outcomes remained incomplete and inconsistent
and advocated further research. Since that time, other studies have reported on hypertensive individuals,11-13,25-27 but
this issue remains controversial. Some studies did not find
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
24-h heart rate
with BB
without BB
Heart Rate and Prognosis in Resistant Hypertension
number of readings of ambulatory monitoring in relation
to clinic or ECG HR, which may more closely approach
patients’ usual HR. In our study, nighttime HR predicted
only all-cause mortality, but daytime HR also predicted cardiovascular mortality. This is contrary to findings of the population-based Ohasama study,30 in which both daytime and
nighttime HRs were predictors only of noncardiovascular
mortality. Furthermore, contrary to other studies,25,31 we did
not observe any independent prognostic importance of the
nocturnal HR dipping, although we found a trend toward an
increased risk of mortality in the small group (34 patients,
7%) with a rising pattern of nighttime HR (data not shown).
The reasons for these differences are not clear but may be
due to different study populations.
This study has some limitations that warrant discussion. First, some potential confounders of the relationship
between HR and outcomes, such as cardiorespiratory fitness
and hemoglobin concentration, were not available and could
not be accounted for. Second, due to the small number of
patients and consequently small number of events in the
subgroup of patients not using beta-blockers, a more comprehensive statistical analysis was not possible, and these
results should be reviewed with caution and confirmed in
larger cohorts. Moreover, the interaction analysis with betablocker use may have been affected by selection bias, because
a beta-blocker was not randomly prescribed. So, it is possible that patients with fast HRs were more frequently prescribed beta-blockers than those who already had slow HRs.
Therefore, the results regarding the differential prognostic
significance of fast and slow HRs in relation to beta-blocker
use should be considered, at most, as suggestive but not as
conclusive findings and should be confirmed in other prospective studies, ideally randomized clinical trials. Third,
because this study included only resistant hypertensive
patients, our findings may not be generalized to less severe
hypertensive populations. Otherwise, resistant hypertension
is a rather common clinical condition, with an estimated
prevalence of 15% to 20% of general hypertensives.
In conclusion, this prospective cohort study in a large
group of resistant hypertensive patients who were followed
up for up to 9 years provides evidence of a U-shaped overall relationship between HR, particularly when measured
during ambulatory monitoring, and adverse outcomes.
However, beta-blocker therapy may influence this association: a fast HR (> 75 bpm) is a risk predictor exclusively in
patients using beta-blockers, while a slow HR (< 60 bpm) is
primarily a risk marker in patients not using beta-blockers.
Further studies are necessary to determine whether direct
HR-limiting therapy can per se reduce adverse outcomes in
hypertensives with baseline fast HRs or whether an achieved
“normal” heart rate during antihypertensive treatment
merely reflects an overall improved cardiovascular status,
independent of any specific rate-limiting therapy.
Acknowledgments
This study was supported by grants from Conselho
Brasileiro de Desenvolvimento Científico e Tecnológico
American Journal of Hypertension 26(2) February 2013 225
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
any prognostic value of baseline HR,11,12 although some did
find a greater prognostic importance of changes in HR during follow-up than in baseline HR,12,13,26 while others11,25
observed a prognostic influence of HR only for mortality but not for nonfatal cardiovascular events. Moreover,
few studies evaluated the prognostic significance of slow
HRs13 or the influence of HR-limiting therapy.11,13,27 In this
regard, our results support the findings of the INVEST
(International VErapamil-SR/trandolapril STudy) trial,13
which reported a J-shaped relationship between on-treatment HR and adverse outcomes, with an increased risk of
fast HRs beginning at 75 bpm and a slow HR increased risk
at <60 bpm in hypertensive patients with stable coronary
artery disease, without influence of the randomized treatment arm (atenolol or verapamil). It should be noted that
the “baseline” HRs in our study are indeed “on-treatment”
HRs because all patients were already on a stable antihypertensive treatment with at least 3 drugs and most (85%)
were on rate-limiting therapy, mainly beta-blockers.
Our study extended the present knowledge to patients
with resistant hypertension and advanced by showing that
the prognostic importance of fast HRs was predominant in
patients using beta-blockers, whereas the predictive value
of slow HRs seemed to be most important in patients without any rate-limiting therapy. The prognostic significance of
fast HRs in patients using beta-blockers may be explained
as reflecting an exaggerated sympathetic overactivity, with
resistance to beta-blocker inducing HR reduction. Beyond
sympathetic overactivity, other potential mechanisms linking fast HRs to cardiovascular morbidity and mortality
involve progression of atherosclerotic lesions and plaque disruption due to hemodynamic shear stress, unbalanced myocardial oxygen demand and supply, ventricular arrhythmias,
left ventricular dysfunction, and poor cardiorespiratory fitness.7,8,28 Furthermore, patients with fast HRs at baseline had
higher prevalences of cardiovascular risk factors, such as
smoking, obesity, and diabetes, and higher ambulatory BPs.
The use of beta-blockers in these patients may worsen the
metabolic profile and increase central aortic pressure15 and
BP variability,29 contributing to adverse prognosis. On the
other hand, the physiopathological mechanisms underlying
the prognostic impact of slow HRs, observed predominantly
in patients without beta-blocker therapy, are unclear. Patients
with baseline slow HRs were more frequently elderly males
with lower body mass index and lower BPs, particularly evident in diastolic BP. These may be unspecific markers of poor
health or of more advanced disease states, which may explain
the association with increased mortality. Unfortunately,
other markers of poor health, such as hemoglobin concentration or serum albumin, were not available in our study.
Few studies evaluated ambulatory HRs in relation to
clinic or electrocardiographic HRs. In the Syst-Eur (Systolic
Hypertension in Europe) trial,10 ambulatory HR was no
better than clinic HR for mortality risk stratification, while
in the IDACO (International Database on Ambulatory
BP Monitoring in Relation to Cardiovascular Outcomes)
report,25 clinic HR was similar to 24-hour HR regarding
all-cause mortality prediction. We found that ambulatory
HRs were more significant predictors of adverse prognosis
than clinic or ECG HRs. This may simply reflect the greater
Salles et al.
(CNPq) and Fundação Carlos Chagas Filho de Amparo à
Pesquisa do Estado do Rio de Janeiro (FAPERJ).
Disclosure
The authors declared no conflict of interest.
References
226 American Journal of Hypertension 26(2) February 2013
Downloaded from http://ajh.oxfordjournals.org/ at University of British Columbia on February 15, 2013
1. Tverdal A, Hjellvik V, Selmer R. Heart rate and mortality from cardiovascular causes: a 12 year follow-up study of 379,843 men and women
aged 40–45 years. Eur Heart J 2008; 29:2772–2781.
2.Jouven X, Empana JP, Escolano S, Buyck JF, Tafflet M, Desnos M,
Ducimetière P. Relation of heart rate at rest and long-term (>20 years)
death rate in initially healthy middle-aged men. Am J Cardiol 2009;
103:279–283.
3. Nauman J, Janszky I, Vatten LJ, Wisløff U. Temporal changes in resting heart rate and deaths from ischemic heart disease. JAMA 2011;
306:2579–2587.
4. Kovar D, Cannon CP, Bentley JH, Charlesworth A, Rogers WJ. Does
initial and delayed heart rate predict mortality in patients with acute
coronary syndromes? Clin Cardiol 2004; 27:80–86.
5. Diaz A, Bourassa MG, Guertin MC, Tardif JC. Long-term prognostic
value of resting heart rate in patients with suspected or proven coronary
artery disease. Eur Heart J 2005; 26:967–974.
6. Fox K, Ford I, Steg PG, Tendera M, Robertson M, Ferrari R. Heart rate
as a prognostic risk factor in patients with coronary artery disease and
left-ventricular systolic dysfunction (BEAUTIFUL): a subgroup analysis of a randomised controlled trial. Lancet 2008; 372:817–821.
7. Fox K, Borer JS, Camm AJ, Danchin N, Ferrari R, Lopez Sendon JL,
Steg PG, Tardif JC, Tavazzi L, Tendera M; Heart Rate Working Group.
Resting heart rate in cardiovascular disease. J Am Coll Cardiol 2007;
50:823–830.
8. Palatini P. Elevated heart rate in cardiovascular diseases: a target to
treatment? Prog Cardiovasc Dis 2009; 52:46–60.
9.Gillman MW, Kannel WB, Belanger A, D’Agostino RB. Influence
of heart rate on mortality among persons with hypertension: the
Framingham Study. Am Heart J 1993; 125:1148–1154.
10. Palatini P, Thijs L, Staessen JA, Fagard RH, Bulpitt CJ, Clement DL, de
Leeuw PW, Jaaskivi M, Leonetti G, Nachev C, O’Brien ET, Parati G,
Rodicio JL, Roman E, Sarti C, Tuomilehto J. Predictive value of clinic
and ambulatory heart rate for mortality in elderly subjects with systolic
hypertension. Arch Intern Med 2002; 162:2313–2321.
11. Poulter NR, Dobson JE, Sever PS, Dahlöf B, Wedel H, Campbell NR.
Baseline heart rate, antihypertensive treatment, and prevention of
cardiovascular outcomes in ASCOT (Anglo-Scandinavian Cardiac
Outcomes Trial). J Am Coll Cardiol 2009; 54:1154–1161.
12. Paul L, Hastie CE, Li WS, Harrow C, Muir S, Connell JM, Dominiczak
AF, McInnes GT, Padmanabhan S. Resting heart rate pattern during
follow-up and mortality in hypertensive patients. Hypertension 2010;
55:567–574.
13.Kolloch R, Legler UF, Champion A, Cooper-Dehoff RM, Handberg
E, Zhou Q, Pepine CJ. Impact of resting heart rate on outcomes in
hypertensive patients with coronary artery disease: findings from the
INternational VErapamil-SR/trandolapril STudy (INVEST). Eur Heart
J 2008; 29:1327–1334.
14. Bangalore S, Messerli FH, Kostis JB, Pepine CJ. Cardiovascular protection using beta-blockers. A critical review of the evidence. J Am Coll
Cardiol 2007; 50:563–572.
15. Williams B, Lacy PS, Thom SM, Cruickshank K, Stanton A, Collier D,
Hughes AD, Thurston H, O’Rourke M. Differential impact of blood
pressure-lowering drugs on central aortic pressure and clinical outcomes: principal results of the Conduit Artery Function Evaluation
(CAFE) study. Circulation 2006; 113:1213–1225.
16. Bangalore S, Sawhney S, Messerli FH. Relation of beta-blocker-induced
heart rate lowering and cardioprotection in hypertension. J Am Coll
Cardiol 2008; 52:1482–1489.
17. Calhoun DA, Jones D, Textor S, Goff DC, Murphy TP, Toto RD, White
A, Cushman WC, White W, Sica D, Ferdinand K, Giles TD, Falkner
B, Carey RM. Resistant hypertension: diagnosis, evaluation and treatment. A scientific statement from the American Heart Association
Professional Education Committee of the Council for High Blood
Pressure Research. Hypertension 2008; 51:1403–1419.
18. Salles G, Cardoso C, Muxfeldt E. Prognostic influence of office and
ambulatory blood pressures in resistant hypertension. Arch Intern Med
2008; 168:2340–2346.
19.Muxfeldt ES, Bloch KV, Nogueira AR, Salles GF. Twenty-four hour
ambulatory blood pressure monitoring pattern of resistant hypertension. Blood Press Monit 2003; 8:181–185.
20. Salles G, Leocadio S, Bloch K, Nogueira AR, Muxfeldt E. Combined
QT interval and voltage criteria improve left ventricular hypertrophy detection in resistant hypertension. Hypertension 2005;
46:1207–1212.
21.Bloch KV, Melo AN, Nogueira AR. Prevalence of anti-hypertensive
treatment adherence in patients with resistant hypertension and validation of three indirect methods for assessing treatment adherence. Cad
Saude Publica 2008; 24:2979–2984.
22. Benetos A, Rudnichi A, Thomas F, Safar M, Guize L. Influence of heart
rate on mortality in a French population: role of age, gender and blood
pressure. Hypertension 1999; 33:44–52.
23.Thomas F, Rudnichi A, Bacri AM, Bean K, Guize L, Benetos A.
Cardiovascular mortality in hypertensive men according to presence of
associated risk factors. Hypertension 2001; 37:1256–1261.
24.Palatini P, Benetos A, Grassi G, Julius S, Kjeldsen SE, Mancia G,
Narkiewicz K, Parati G, Pessina AC, Ruilope LM, Zanchetti A.
Identification and management of the hypertensive patient with
elevated heart rate: statement of a European Society of Hypertension
Consensus Meeting. J Hypertens 2006; 24:603–610.
25. Hansen TW, Thijs L, Boggia J, Li Y, Kikuya M, Björklund-Bodegård K,
Richart T, Ohkubo T, Jeppesen J, Torp-Pedersen C, Lind L, Sandoya E,
Imai Y, Wang J, Ibsen H, O’Brien E, Staessen JA. Prognostic value of
ambulatory heart rate revisited in 6928 subjects from 6 populations.
Hypertension 2008; 52:229–235.
26.Okin PM, Kjeldsen SE, Julius S, Hille DA, Dahlöf B, Edelman JM,
Devereux RB. All-cause and cardiovascular mortality in relation to
changing heart rate during treatment of hypertensive patients with
electrocardiographic left ventricular hypertrophy. Eur Heart J 2010;
31:2271–2279.
27. Julius S, Palatini P, Kjeldsen SE, Zanchetti A, Weber MA, McInnes GT,
Brunner HR, Mancia G, Schork MA, Hua TA, Holzhauer B, Zappe D,
Majahalme S, Jamerson K, Koylan N. Usefulness of heart rate to predict
cardiac events in treated patients with high-risk systemic hypertension.
Am J Cardiol 2012; 109:685–692.
28.Palatini P. Role of heart rate in the development of cardiovascular
disease in hypertension. Hypertension 2011; 58:745–750.
29. Webb AJ, Fischer U, Mehta Z, Rothwell PM. Effects of antihypertensive-drug class on interindividual variation in blood pressure and
risk of stroke: a systematic review and meta-analysis. Lancet 2010;
375:906–915.
30.Hozawa A, Inoue R, Ohkubo T, Kikuya M, Metoki H, Asayama K,
Hara A, Hirose T, Kanno A, Obara T, Hoshi H, Hashimoto J, Totsune
K, Satoh H, Imai Y. Predictive value of ambulatory heart rate in the
Japanese general population: the Ohasama study. J Hypertens 2008;
26:1571–1576.
31. Ben-Dov IZ, Kark JD, Ben-Ishay D, Mekler J, Ben-Arie L, Bursztyn M.
Blunted heart rate dip during sleep and all-cause mortality. Arch Intern
Med 2007; 167:2116–2121.