1. No category

Risk/Benefit Assessment of β-Blockers and Diuretics Precludes Their Use for First-Line
Therapy in Hypertension
Franz H. Messerli, Sripal Bangalore and Stevo Julius
Circulation. 2008;117:2706-2715
doi: 10.1161/CIRCULATIONAHA.107.695007
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2008 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/117/20/2706
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/
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
CONTROVERSIES IN
CARDIOVASCULAR MEDICINE
Should ␤-blockers and diuretics remain as
first line therapy for hypertension?
Thiazide-Type Diuretics and ␤-Adrenergic Blockers as
First-Line Drug Treatments for Hypertension
Jeffrey A. Cutler, MD, MPH; Barry R. Davis, MD, PhD
S
ystemic arterial hypertension affects 72 million US adults
and an additional hundreds of millions of persons worldwide.1,2 Most of these are candidates for pharmacological
treatment to reduce risk of cardiovascular disease (CVD)
events, based primarily on a very large body of epidemiological and intervention research in humans. Because of this high
prevalence and the cardiovascular consequences of untreated
or inadequately treated hypertension, the selection of drugs
for initial and continuing, long-term treatment has large
public health and economic implications. Fortunately, such
decisions and the expert recommendations that seek to guide
them can call on evidence from 4 decades of randomized
multicenter clinical trials evaluating effects of treatment on
clinical CVD. We summarize that evidence in this article, in
approximate chronological order, and we comment on the
related treatment guidelines. We close with some of the major
clinical questions yet to be resolved.
Response by Messerli et al p 2705
Trials of Blood Pressure Reduction
Before the current era beginning in the early 1990s of
emphasis on positive-control trials, which directly compare
different drug regimens, there were 3 decades of trials that
compared an active regimen with placebo or, in a few cases,
“usual care.” For most of this period, the mainstay of
treatment was generally a thiazide-type diuretic (hereinafter
called thiazides) or, to a lesser extent, a ␤-adrenergic blocker
(termed ␤-blockers). With few exceptions, these trials, especially those with high statistical power and thiazide-based
regimens, showed benefit for CVD outcomes.3– 8 This evidence, which provided a basis for recommending thiazides or
␤-blockers as first-step drugs in most editions of US guidelines through 1997,9 needs to continue to be given due weight
in practice and practice guidelines.
The largest and most consistent benefits from the earlier
trials were for stroke and (where reported) heart failure.
Benefits for coronary heart disease (CHD) were less clear,
and commentators frequently speculated that potentially adverse metabolic effects— on potassium (for diuretics), lipids,
and glucose— of thiazides and ␤-blockers were related to
this shortfall in reducing CHD outcomes. Two reports in
the early 1990s reduced concern about the CHD shortfall
considerably. A meta-analysis of essentially all randomized antihypertensive treatment trials with clinical events
outcomes, placed in an epidemiological context to address
the question of what effects would be expected on the basis
of risks of various events at different blood pressure (BP)
Dr Cutler is a consultant to the National Heart, Lung, and Blood Institute. The opinions expressed in this article are his own and those of Dr Davis
and are not necessarily those of the National Heart, Lung, and Blood Institute or the US Department of Health and Human Services, the editors, or the
American Heart Association.
From the National Heart, Lung, and Blood Institute, Bethesda, Md (J.A.C.), and the University of Texas Health Science Center at Houston School of
Public Health, Houston, Tex (B.R.D.).
This article is Part I of a 2-part article. Part II appears on page 2706.
Correspondence to Barry R. Davis, MD, PhD, University of Texas Health Science Center at Houston, School of Public Health, 1200 Herman Pressler
St, Suite E801, Houston, TX 77030. E-mail [email protected]
(Circulation. 2008;117:2691-2705.)
© 2008 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org
DOI: 10.1161/CIRCULATIONAHA.107.709931
2691
Downloaded from http://circ.ahajournals.org/
by guest on September 9, 2014
2692
Circulation
May 20, 2008
Figure 1. Percentage of drug mentions by class of
antihypertensive agent, 1982–1993.21 Data from
IMS America.59 Reprinted from Manolio et al,21
with permission. Copyright 1995, American Medical Association.
levels, showed that treatment significantly reduced nonfatal
myocardial infarction (MI) or CHD death (major CHD) by
14%, which represented approximately two thirds of the
epidemiological expectation.10,11 The authors attributed the
shortfall to the short duration of treatment in the trials,
averaging ⬇2 to 3 years to CHD events, which was a
plausible conclusion. Shortly thereafter, the Systolic Hypertension in the Elderly Program (SHEP) reported that a
thiazide-based regimen (using low-dose chlorthalidone) not
only reduced fatal and nonfatal stroke by 36% but also
lowered major CHD by 27%.8
The results of these trials have also provided a basis for
guidelines on drug choice through indirect comparisons
among trials and groups of trials. Subsequent to SHEP, the
similarly designed Systolic Hypertension in Europe (SystEur) trial reported that treatment with a dihydropyridine
calcium channel blocker (CCB) as the main drug reduced BP,
stroke, and all cardiac events to an extent similar to that in
SHEP.12 However, neither the effects on CHD or heart failure
were separately statistically significant. For angiotensinconverting enzyme (ACE) inhibitors, there was no large trial
focused on hypertension, but several trials reported that these
drugs reduced mortality and morbidity in patients with heart
failure and/or CHD13–16; the Studies of Left Ventricular
Dysfunction (SOLVD) prevention trial showed that enalapril
lowered the risk of MI and overt heart failure in patients with
reduced ejection fraction.17 The size of the heart failure effect
was, however, only 20%, in contrast with the 49% benefit in
SHEP.18
These newer trials provided important added justification
for earlier recommendations concerning use of CCBs and
ACE inhibitors as first-line drugs in treating hypertension in
the 1988 Report of the Joint National Committee (JNC) on
Prevention, Detection, Evaluation, and Treatment of High
Blood Pressure.19 However, both JNCV and JNCVI recommended thiazides and ␤-blockers as preferred initial drugs for
antihypertensive treatment.9,20 As in the case of ACE inhib-
itors, the trials of secondary prevention of heart disease
continued to play a large role in the thinking about benefits of
␤-blockers in hypertension.
Despite all of these factors, that is, the weight of the
evidence, the consensus incorporated in JNC recommendations, and the low acquisition costs of several generic agents,
the use of thiazides and ␤-blockers to treat hypertension
declined dramatically after 1981–1983 through the early
1990s (Figure 1).21 These trends can be attributed to the
following: (1) concerns raised regarding metabolic effects;
(2) results of trials that used doses of thiazides much higher
than in general use today22; and (3) effective marketing of
newer patented drugs, in part based on their effects on
intermediate markers of various disease processes. Thiazide
dosing issues were eventually addressed on the basis of the
dose-response curves for BP effects versus potassium depletion, the SHEP trial results, and, most persuasively, by a
meta-analysis that separated use of “low-dose” (actually,
low-to-moderate dose) from high-dose thiazides among cardiovascular events trials.23 Nevertheless, it was increasingly
recognized that large, direct comparison trials would be
needed to provide a solid scientific basis for drug selection.
One of the first such trials, and the largest in terms of total
study population as well as patients per treatment arm, was
the Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).
The Continuing Case for ThiazideType Diuretics
Direct Comparison Trials
ALLHAT: Design and Prespecified Outcomes
ALLHAT was designed to address the issue of which class of
drugs should be used for initial therapy for hypertension. It
was planned as a practice-based trial to mirror community
treatment of hypertension, obtain sufficient patients, and
capture the diversity of hypertensive patients (by age, sex,
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Thiazide-Type Diuretics and ␤-Blockers
Cutler and Davis
Table 1.
2693
Clinical Outcomes by Antihypertensive Treatment Group25
Chlorthalidone
Primary end point
CHD*
Secondary end points
All-cause mortality
Combined CHD†
Stroke
Combined CVD†
End-stage renal
disease
Components of
secondary outcomes
Heart failure
Hospitalized/fatal
heart failure
Angina (hospitalized
or treated)
Angina (hospitalized)
Coronary
revascularizations
Peripheral arterial
disease (hospitalized
or treated)
Amlodipine
Lisinopril
No. of
Total
Events
6-Year Rate
per 100
Persons
(SE)
No. of
Total
Events
6-Year Rate
per 100
Persons
(SE)
No. of
Total
Events
6-Year Rate
per 100
Persons
(SE)
RR (95% CI)
Z Score
P
RR (95% CI)
Z Score
P
1362
11.5 (0.3)
798
11.3 (0.4)
796
11.4 (0.4)
0.98 (0.90–1.07)
⫺0.46
0.65
0.99 (0.91–1.08)
⫺0.24
0.81
2203
2451
675
3941
193
17.3 (0.4)
19.9 (0.4)
5.6 (0.2)
30.9 (0.5)
1.8 (0.1)
1256
1466
377
2432
129
16.8 (0.5)
19.9 (0.5)
5.4 (0.3)
32.0 (0.6)
2.1 (0.2)
1314
1505
457
2514
126
17.2 (0.5)
20.8 (0.5)
6.3 (0.3)
33.3 (0.6)
2.0 (0.2)
0.96 (0.89–1.02)
1.00 (0.94–1.07)
0.93 (0.82–1.06)
1.04 (0.99–1.09)
1.12 (0.89–1.40)
⫺1.27
0.04
⫺1.09
1.55
0.98
0.20
0.97
0.28
0.12
0.33
1.00 (0.94–1.08)
1.05 (0.98–1.11)
1.15 (1.02–1.30)
1.10 (1.05–1.16)
1.11 (0.88–1.38)
0.12
1.35
2.31
3.78
0.87
0.90
0.18
0.02
⬍0.001
0.38
870
724
7.7 (0.3)
6.5 (0.3)
706
578
10.2 (0.4)
8.4 (0.4)
612
471
8.7 (0.4)
6.9 (0.4)
1.38 (1.25–1.52)
1.35 (1.21–1.50)
6.29
5.37
⬍0.001
⬍0.001
1.19 (1.07–1.31)‡
1.10 (0.98–1.23)‡
3.33
1.59
⬍0.001
0.11
1567
12.1 (0.3)
950
12.6 (0.4)
1019
13.6 (0.4)
1.02 (0.94–1.10)
0.42
0.67
1.11 (1.03–1.20)
2.59
0.01
1078
1113
8.6 (0.3)
9.2 (0.3)
630
725
8.4 (0.4)
10.0 (0.4)
693
718
9.6 (0.4)
10.2 (0.4)
0.98 (0.89–1.08)
1.09 (1.00–1.20)
⫺0.41
1.88
0.68
0.06
1.09 (0.99–1.20)
1.10 (1.00–1.21)
1.85
1.95
0.06
0.05
510
4.1 (0.2)
265
3.7 (0.2)
311
4.7 (0.4)
0.87 (0.75–1.01)
⫺1.86
0.06
1.04 (0.90–1.19)
0.48
0.63
Amlodipine/Chlorthalidone
Lisinopril/Chlorthalidone
CHD includes nonfatal MI and fatal CHD; end-stage renal disease: kidney disease death, kidney transplant, or start of chronic renal dialysis; and heart failure: fatal,
nonfatal hospitalized, or treated. Adapted from Reference 25, with permission. Copyright 2002, American Medical Association.
*Nonfatal MIs comprise 64% to 66% of the primary outcome.
†Combined CHD indicates CHD death, nonfatal MI, coronary revascularization procedures, and hospitalized angina. Combined CVD indicates CHD death, nonfatal
MI, stroke, coronary revascularization procedures, hospitalized or treated angina, treated or hospitalized heart failure, and peripheral arterial disease (hospitalized or
outpatient revascularization).
‡Proportional hazards assumption violated; data are RRs from a 2⫻2 table.
ethnicity, and diabetic status). More specifically, the study
was a randomized, double-blind, multicenter clinical trial,
designed to determine whether the incidence of CHD is
reduced in high-risk patients with hypertension by a CCB
(represented by amlodipine), an ACE inhibitor (represented
by lisinopril), or an ␣-blocker (represented by doxazosin),
each compared with diuretic treatment (represented by
chlorthalidone).24 The overall findings of the trial showed that
CHD risk was not improved for any of the 3 newer agents
compared with chlorthalidone25,26 and that total mortality was
also similar for the 4 groups (Tables 1 and 2).
However, diuretic-based therapy was superior to ␣-blocker– based, ACE inhibitor– based, and CCB-based therapy in
preventing 1 or more major forms of CVD, including heart
failure and (in some comparisons) stroke.25,26 Results were
consistent for all outcomes by age, sex, diabetic status, and
ethnicity, except for stroke and combined CVD. For these end
points, significant heterogeneity was seen in the lisinoprilchlorthalidone comparison by ethnicity: Black persons assigned to chlorthalidone had a greater reduction in risk for
stroke and CVD, in agreement with larger BP differences.
On the basis of these findings, the ALLHAT investigators
recommended that diuretics should be the drug of choice for
initial hypertension therapy and, because most hypertensive
patients require ⬎1 drug, diuretics should generally be part of
any antihypertensive regimen. With regard to applicability of
these conclusions, although ALLHAT was conducted in
high-risk patients to ensure that enough outcome events
would occur during the study to detect important treatment
differences, its findings (just as those from most trials) can
and should be reasonably extrapolated beyond the exact
sample in which it was conducted.
BP Differences in ALLHAT
One of the main criticisms of ALLHAT was that its outcome
findings (especially the subgroup findings for stroke) could
be explained by observed BP differences among treatment
groups. The chlorthalidone-based regimen happened to be the
most effective in reducing clinical outcomes and, to a small
degree, in BP lowering (Figure 2).25
Meta-regressions of effects of BP differences on trial
results27 suggest that they offer a partial explanation, except
for heart failure. In ALLHAT, the small differences in
achieved mean systolic BP and diastolic BP (⬍1 mm Hg)
between the amlodipine and chlorthalidone groups overall
and the lisinopril and chlorthalidone groups in nonblack
persons should have had a negligible effect on cardiovascular
event rates. In these respective comparisons, the observed
rates of heart failure were higher with amlodipine (relative
risk [RR]⫽1.38; 95% CI, 1.25 to 1.52) and lisinopril
(RR⫽1.15; 95% CI, 1.01 to 1.30) than with chlorthalidone.25
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2694
Circulation
May 20, 2008
Table 2.
Outcomes in the BP Component of ALLHAT by Treatment Group as of February 15, 200026
Total No. of Patients With Outcomes
Outcomes
CHD (nonfatal MI⫹fatal CHD)
Chlorthalidone Group
Doxazosin Group
4-Year Rate per 100 (SE)
Chlorthalidone Group
(n⫽15 255)
Doxazosin Group
(n⫽9061)
RR (D/C), 95% CI
Z Score
P*
818
499
7.76 (0.30)
7.91 (0.39)
1.03 (0.92–1.15)
0.49
0.62
All-cause mortality
1258
769
10.51 (0.32)
11.04 (0.43)
1.03 (0.94–1.13)
0.68
0.50
Cardiovascular
551
377
4.74 (0.22)
5.60 (0.32)
1.15 (1.01–1.32)
2.15
0.03
184
105
1.65 (0.13)
1.76 (0.19)
0.96 (0.76–1.22)
⫺0.32
0.75
Definite CHD
57
39
0.54 (0.08)
0.54 (0.10)
1.16 (0.77–1.74)
0.70
0.49
Possible CHD
62
43
0.50 (0.08)
0.63 (0.11)
1.17 (0.79–1.73)
0.79
0.43
Stroke
92
76
0.79 (0.10)
1.25 (0.16)
1.39 (1.03–1.89)
2.14
0.03
Heart failure
59
42
0.60 (0.09)
0.65 (0.11)
1.20 (0.81–1.78)
0.91
0.36
Other cardiovascular
97
72
0.88 (0.10)
1.12 (0.15)
1.25 (0.92–1.70)
1.44
0.15
561
317
4.82 (0.23)
4.72 (0.30)
0.95 (0.83–1.09)
⫺0.67
0.50
314
162
2.78 (0.18)
2.43 (0.21)
0.87 (0.72–1.05)
⫺1.43
0.15
Kidney disease
12
12
0.11 (0.04)
0.24 (0.09)
1.69 (0.76–3.77)
1.29
0.20
Accident/suicide/homicide
39
28
0.33 (0.06)
0.40 (0.09)
1.21 (0.75–1.97)
0.78
0.44
196
115
1.75 (0.14)
1.84 (0.19)
0.99 (0.79–1.25)
⫺0.08
0.93
146
75
1.37 (0.13)
1.18 (0.16)
0.87 (0.66–1.15)
⫺1.00
0.32
1642
1040
14.87 (0.39)
16.00 (0.53)
1.07 (0.99–1.16)
1.82
0.07
434
325
4.08 (0.22)
5.49 (0.35)
1.26 (1.10–1.46)
3.20
0.001
MI
Noncardiovascular
Cancer
Other noncardiovascular
Unknown
Combined CHD†
Stroke
2829
1947
25.09 (0.48)
28.56 (0.64)
1.20 (1.13–1.27)
6.13
⬍0.001
Heart failure (fatal,
hospitalized, treated)
546
584
5.35 (0.26)
8.89 (0.42)
1.80 (1.61–2.02)§
10.27
⬍0.001
Heart failure (fatal,
hospitalized)
440
434
4.41 (0.24)
6.63 (0.37)
1.66 (1.46–1.89)§
7.72
⬍0.001
Combined CVD‡
Coronary revascularization
Hospitalized or treated
angina
Lower-extremity peripheral
arterial disease
End-stage renal disease
770
508
7.08 (0.28)
8.02 (0.40)
1.12 (1.00–1.25)
1.97
0.05
1227
811
10.81 (0.33)
11.82 (0.45)
1.13 (1.03–1.23)
2.65
0.01
376
217
3.68 (0.21)
3.49 (0.27)
0.97 (0.82–1.15)
⫺0.31
0.76
104
64
1.10 (0.13)
1.08 (0.17)
1.04 (0.76–1.42)
0.26
0.80
Adapted from Reference 26, with permission from Lippincott Williams & Wilkins. Copyright 2003, American Heart Association.
*To adjust for multiple comparisons, compare the P value with 0.018 rather than 0.05.
†Combined CHD includes CHD death, nonfatal MI, coronary revascularization procedures, and hospitalized angina.
‡Combined CVD includes CHD death, nonfatal MI, stroke, coronary revascularization procedures, hospitalized or treated angina, treated or hospitalized CHF, and
peripheral arterial disease (hospitalized, or outpatient revascularization).
§Proportional hazards assumption violated; numbers given are RRs.
Extrapolating Findings to Drugs Within a Class
Could the ALLHAT results be extrapolated to other drugs of
the same class? For ␣-blockers, ACE inhibitors, and dihydropyridine CCBs, such extrapolation seems reasonable.28 Data
from studies using various thiazide-type diuretics (chlorthalidone, hydrochlorothiazide, indapamide, and bendrofluazide)
suggest similar benefits among equivalent doses of all
thiazide-type diuretics tested in CVD prevention trials against
placebo, usual care, or another drug class.5,6,8,29 –34 However,
a few studies suggest that the longer duration of action of
chlorthalidone may provide some advantage in CVD prevention over hydrochorothiazide.22,35
Results From Other Trials
After the ALLHAT results appeared, several other active comparator trials were reported. The Second Australian National
Blood Pressure Study (ANBP2), a practice-based open-label
trial, was the only other large trial besides ALLHAT to
compare diuretic-based (hydrochlorothiazide recommended)
with ACE inhibitor– based (enalapril recommended) antihypertensive treatment.34 A total of 6083 participants, aged 65
to 84 years, were treated and followed up for a mean of 4.1
years. The primary end point was a composite of all cardiovascular events (including recurrent events, an unusual design) plus all-cause mortality. Cardiovascular events included
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
Thiazide-Type Diuretics and ␤-Blockers
2695
Figure 2. Mean systolic and diastolic BP by year during follow-up.25 Reprinted from Reference 25, with permission. Copyright 2002,
American Medical Association.
major coronary events, stroke and transient ischemic attacks,
heart failure (not otherwise defined), acute occlusion of any
other major artery, and dissecting or ruptured aortic aneurysm. The results for the primary end point favored the ACE
inhibitor group with marginal significance (RR, 0.89; 95%
CI, 0.79 to 1.00; P⫽0.05). The corresponding results with the
use of first cardiovascular event had essentially the same RR
(P⫽0.07). The RR for heart failure was 0.85 (95% CI, 0.62 to
1.18; P⫽0.33). Frohlich36 weighed the supposedly conflicting results of ANBP2 and ALLHAT, suggesting possible
explanations such as the patients studied (many more black
patients in ALLHAT) and the specific drugs used. Additionally, there were almost 8 times as many cardiovascular events
in the 2 comparable arms in ALLHAT as in ANBP2, and only
ALLHAT was double blind. ANBP2 used a prospective,
randomized, open-label, blinded end point (PROBE) design,
increasing the potential for bias in the reporting of events (the
rates of some outcomes might have been “expected” to be
lower with the ACE inhibitor) even though ANBP2 relied on
end point committee–adjudicated outcomes. Doses of agents
in ANBP2 were left up to the local investigator and were not
reported; thus, it is not possible to assess whether appropriate
doses of hydrochlorothiazide were used. Even given the
possible biases, however, the results of ANBP2 are consistent
with those of ALLHAT if the upper confidence limit for the
RR in ANBP2 is compared with the estimates of RR in
ALLHAT.37
The International Nifedipine GITS study [Intervention as a
Goal in Hypertension Treatment (INSIGHT)] was the other
large trial besides ALLHAT to compare diuretic-based (coamilozide) with CCB-based (nifedipine) antihypertensive
treatment on cardiovascular mortality and morbidity in highrisk patients with hypertension.33 It was a randomized,
double-blind trial in 6321 patients aged 55 to 80 years with
hypertension. Patients had at least 1 additional cardiovascular
risk factor and were randomly assigned to nifedipine (30 to
60 mg in a long-acting gastrointestinal-transport system
[GITS] formulation) or co-amilozide (hydrochlorothiazide 25
to 50 mg plus amiloride). The primary outcome was cardiovascular death, MI, heart failure, or stroke. Primary outcomes
occurred in 200 patients (6.3%) in the nifedipine group and in
182 (5.8%) in the co-amilozide group (RR⫽1.10; 95% CI,
0.91 to 1.34; P⫽0.35). The CCB was not superior to the
diuretic in preventing cardiovascular morbidity and mortality.
Nonfatal heart failure was more common in the CCB arm
(RR⫽2.20; 95% CI, 1.07 to 4.49; P⫽0.028). There were only
3 fatal heart failure events: 2 in the CCB arm and 1 in the
diuretic arm.
Further Details on Heart Failure in ALLHAT
Another major ALLHAT criticism concerned the heart failure
findings. Specifically, were the findings real, and could they
be explained by withdrawal from antihypertensive medications, such as diuretics and ACE inhibitors, on entry into
ALLHAT?
Several articles have addressed these 2 questions in detail.38 – 41 The reliability of the heart failure diagnosis during
the trial was examined in depth via the ALLHAT Heart
Failure Validation Study (HFVS).40 This study was designed
to validate and elucidate the significance of heart failure
events in ALLHAT. This study involved all hospitalized heart
failure events and relevant hospital records related to these
events. Cardiology fellows, external to ALLHAT and blinded
to treatment assignment, centrally abstracted the documentation for each heart failure hospitalization (2778 in 1935
patients; 2 independent reviews per case). ALLHAT and
Framingham criteria were assigned by a computer algorithm;
the reviewers also rendered a global clinical judgment.
Percent agreements with site physician diagnoses were 71%,
80%, and 84% for ALLHAT, Framingham, and reviewers’
judgment, respectively. On the basis of these 3 criteria, RRs
(95% CI) for new-onset hospitalized heart failure compared
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2696
Circulation
May 20, 2008
Figure 3. Incident hospitalized (Hosp) heart failure (HF) outcomes by antihypertensive treatment group (amlodipine/lisinopril vs chlorthalidone).40 *Prespecified end point of treated,
hospitalized, or fatal heart failure. Reprinted
from Einhorn et al,40 with permission. Copyright
2007, Elsevier.
with chlorthalidone were, respectively, 1.46 (1.27 to 1.68),
1.42 (1.25 to 1.62), and 1.45 (1.28 to 1.64) for amlodipine;
1.18 (1.02 to 1.28), 1.13 (0.99 to 1.30), and 1.15 (1.01 to
1.32) for lisinopril (Figure 3); and 1.79 (1.51 to 2.11), 1.71
(1.46 to 2.00), and 1.80 (1.55 to 2.10) for doxazosin.40
Although there was early divergence of the heart failure
incidence curves in ALLHAT, it continued after the first year
for doxazosin and amlodipine versus chlorthalidone. For
lisinopril versus chlorthalidone, the curves also separated
early but appeared to converge between years 6 and 7.25
Diagnostic analysis revealed that the proportional hazards
assumption of constant relative risk over time was not valid.39
A more appropriate model showed that RRs of amlodipine or
lisinopril versus chlorthalidone during year 1 were 2.22 (1.69
to 2.91; P⬍0.001) and 2.08 (1.58 to 2.74; P⬍0.001); after
year 1, they were 1.22 (1.08 to 1.38; P⬍0.001) and 0.96 (0.85
to 1.10; P⫽0.58) (Figure 4).39
In addition, information about previous medication use was
collected on the hospitalized and fatal cases of heart failure as
a follow-up to the HFVS.41 When case-only design theory
was used to assess interactions,42 the analyses did not support
an effect of preentry diuretic use on the observed heart failure
differences. However, the addition of second- and third-line
drugs (⬇30% at year 1) probably contributed to the lessening
of the divergence starting at 6 to 12 months after randomization. Given these additional examinations, the original conclusions remained the same. Thiazide-type diuretics should
be the preferred first-step therapy for prevention of heart
failure in high-risk patients with hypertension.
Diabetes in ALLHAT
As noted above, the initial ALLHAT reports showed consistent findings for those in the prespecified subgroups with and
without a baseline history of diabetes.25,26 A subsequent
report on comparisons among the chlorthalidone, amlodipine,
and lisinopril arms utilized baseline fasting glucose levels in
addition to history to classify participants into those with
diabetes, impaired fasting glucose (110 to 125 mg/dL), and
fasting normoglycemia. Results were similar in all 3 subgroups, showing, in particular, the superiority of chlorthalidone for heart failure and the absence of any outcome
(including end-stage renal disease) for which another arm
was superior.43
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
Thiazide-Type Diuretics and ␤-Blockers
2697
Figure 4. Cumulative event rates for hospitalized (Hosp)/fatal heart failure (HF) by treatment group.39
Among those without diabetes mellitus (DM) at baseline,
the mean fasting glucose was ⬇93.5 mg/dL. Changes in
fasting glucose and percentage of incident diabetes mellitus
(IDM) at 4 years, although not prespecified outcomes, were
⫹10.8 mg/dL and 11.0% in the chlorthalidone group; ⫹9.3
mg/dL and 9.3% in the amlodipine group, and ⫹6.8 mg/dL
and 7.8% in the lisinopril group, respectively.25 It was
observed that these and other metabolic differences did not
translate into any overall disadvantage for the diuretic arm
during the mean follow-up of 4.9 years (range, 4 to 8 years).
Nevertheless, further epidemiological-type analyses were
conducted to examine the association of glucose changes with
CVD and renal outcomes.44 There was no significant association of 2-year fasting glucose change with subsequent
events, overall or in the chlorthalidone arm separately. In
addition, among those who developed IDM by 2 years
compared with those who did not, there was no significant
increase in subsequent risk for any major disease outcomes
except CHD, eg, the RRs for all CVD combined were 1.04
(95% CI, 0.80 to 1.35) for all arms and 0.96 (0.66 to 1.37) for
chlorthalidone; for total mortality, they were 1.31 (0.96 to
1.81) and 1.05 (0.66 to 1.67), respectively. For CHD, the risk
associated with IDM was 1.64 (1.15 to 2.33) overall but only
1.46 (0.88 to 2.42) in the chlorthalidone arm (Figure 5).
Is this tendency for IDM occurring during low/moderatedose chlorthalidone treatment to impart less risk for adverse
clinical events than during other regimens real? If so, what
could be the explanation? First, in the only other long-term
follow-up data in treated hypertensive patients that are
relatively uncontaminated with concomitant drugs, the findings in the SHEP 14-year extended follow-up study showed a
similar phenomenon: a contrast between an increased CVD
risk associated with IDM in the placebo arm but not in the
chlorthalidone arm.45 Second, it must be recognized that in a
typical hypertensive population, the great majority of IDM
that occurs while taking a thiazide is not drug induced but is
due to typical causes of type 2 diabetes and therefore would
be expected to carry the same risk as any such occurrence.
From ALLHAT data on IDM at 4 years, such a fraction can
be estimated as 83%, if it is assumed that the CCB is
metabolically neutral.46 If only 1 of 5 cases of IDM in patients
prescribed a thiazide is due to the drug but these cases cannot
be separated from the majority, the lower risk in such patients
would serve to somewhat lower but not eliminate the overall
DM-associated risk.
For the reason why thiazide-induced IDM could carry less
risk than “naturally occurring” DM, one needs to consider the
abundant, but mostly older and often forgotten, literature on
potassium depletion and glucose disorders.46,47 To the extent
that glucose disorders are due to this mechanism, it is
plausible that its natural history is quite different. Furthermore, as pointed out in Reference 46, “In the diuretic-treated
patient hypokalemia is likely to be intermittent, due to dietary
and drug adherence variation, and potassium-sparing therapeutic intervention [which] may also translate to dysglycemia
that is intermittent. . .and thus confer little risk of diabetic
complications.”
Network Meta-Analysis
The role of the efficacy of various antihypertensive therapies
used as first-line agents in preventing major cardiovascular
disease outcomes has been assessed with network meta-analysis.48 This type of analysis combines direct within-trial,
between-drug comparisons with indirect evidence from the
other trials. The indirect comparisons, which preserve the
within-trial randomized findings, were constructed from trials
that had 1 treatment in common.
Data were combined from 42 clinical trials that included
192 478 patients randomized to 7 major treatment strategies,
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2698
Circulation
May 20, 2008
Figure 5. Cox regression models showing the hazard ratios (95% CIs) associated with IDM during the first 2 years of follow-up on subsequent CVD and renal end points in those without DM at baseline.44 All hazard ratios controlled for treatment group (total cohort),
2-year blood pressure, age, race, sex, smoking status, baseline fasting glucose level, baseline body mass index, 2-year serum potassium level, and atenolol and statin administration at 2 years. NA indicates not applicable (too few events for analysis). Reprinted from
Barzilay et al,44 with permission. Copyright 2006, American Medical Association.
including placebo. For all outcomes, low-dose diuretics were
superior to placebo (Figure 6): CHD (RR, 0.79; 95% CI, 0.69
to 0.92); congestive heart failure (RR, 0.51; 95% CI, 0.42 to
0.62); stroke (RR, 0.71; 0.63 to 0.81); CVD events (RR, 0.76;
95% CI, 0.69 to 0.83); CVD mortality (RR, 0.81; 95% CI,
0.73 to 0.92); and total mortality (RR, 0.90; 95% CI, 0.84 to
0.96). None of the other first-line treatment strategies—␤blockers, ACE inhibitors, CCBs, ␣-blockers, and angiotensin
receptor blockers—was significantly better than low-dose
diuretics for any outcome. BP changes were similar between
comparison treatments. On the basis of this network metaanalysis, the authors concluded that low-dose diuretics were
the most effective first-line treatment for preventing the
occurrence of CVD morbidity and mortality.
JNC7 Recommendations
In 2003, the Seventh Report of the Joint National Committee
on Prevention, Detection, Evaluation, and Treatment of High
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
Thiazide-Type Diuretics and ␤-Blockers
2699
Figure 6. Network meta-analysis of first-line treatment strategies in randomized controlled clinical trials in hypertension.48 Asterisks,
placed after the closed parentheses of the 95% CI, indicate that ␤-blockers (P⬍0.05), angiotensin-converting enzyme inhibitors
(P⬍0.05), calcium channel blockers (P⬍0.05), and angiotensin receptor blockers (P⬍0.05) were significantly better than placebo for
outcome. ␣-Blockers were not significantly better than placebo for any outcome (P⬎0.05). Reprinted from Psaty et al,48 with permission. Copyright 2003, American Medical Association. All rights reserved.
Blood Pressure (JNC7) issued its recommendations on the
basis of the ALLHAT results and other trial evidence that was
available at that time.49 Its major conclusions and recommendations were as follows.
In trials comparing thiazide-type diuretics with other
classes of antihypertensive agents, they are (1) well tolerated;
(2) effective and relatively safe for the management of
hypertension despite potential adverse metabolic effects; and
(3) unsurpassed in preventing the cardiovascular complications of hypertension. They are also less expensive and
underutilized. The doses of thiazide-type diuretics used in
successful morbidity trials of low/moderate-dose diuretics
should be used (generally the equivalent of 25 to 50 mg of
hydrochlorothiazide or 12.5 to 25 mg of chlorthalidone),
although therapy may be initiated at lower doses and titrated
to these doses if tolerated.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2700
Circulation
May 20, 2008
Figure 7. JNC7 algorithm for treatment of
hypertension.49 SBP indicates systolic BP;
DBP, diastolic BP; ACEI, ACE inhibitors;
ARB, angiotensin receptor blockers; and
BB, ␤-blockers. Reprinted from Chobanian
et al,49 with permission from Lippincott Williams & Wilkins. Copyright 2003, American
Heart Association.
The algorithm for the treatment of hypertensive patients is
to begin with lifestyle modification, and if the BP goal is not
achieved, thiazide-type diuretics should be used as initial
therapy for most patients, either alone or in combination with
1 of the other classes (ACE inhibitors, angiotensin receptor
blockers, ␤-blockers, CCBs) that have also been shown to
reduce 1 or more hypertensive complications in randomized
controlled outcome trials (Figure 7). Selection of 1 of these
other agents as initial therapy is recommended when a
diuretic cannot be used or when a compelling indication is
present that requires the use of a specific drug. If the initial
drug selected is not tolerated or is contraindicated, then a drug
from 1 of the other classes proven to reduce cardiovascular
events should be substituted. Because most hypertensive
patients will require ⱖ2 antihypertensive medications to
achieve their BP goals, addition of a second drug from a
different class should be initiated when use of a single agent
in adequate doses fails to achieve the goal. It is further
recommended that patients with stage 2 hypertension be
started on 2 drugs initially, 1 of which should be a thiazide.
The Evidence for and Against ␤-Blockers
The British Medical Research Council (BMRC) trial of
treatment of mild hypertension was the first clinical events
trial with a ␤-blocker– based arm (propanolol) in addition to
a thiazide arm (bendrofluazide).6 Compared with placebo,
only the thiazide significantly reduced stroke, likely as a
result of the greater BP reduction than with the ␤-blocker.
Both active treatment arms showed reductions in major
cardiovascular events. Along with trials that found improved
survival in post-MI patients with ␤-blockers, this experience
was sufficient for guidelines to begin recommending thiazides and ␤-blockers as relatively equivalent alternatives for
initiating treatment. This remained the situation through the
1990s, despite publication of results from the BMRC trial of
treatment of hypertension in older adults, the ␤-blocker arm
of which showed no significant reduction in either stroke or
CHD events compared with placebo.7 In keeping with practice trends, a cardioselective ␤-blocker, atenolol, was selected
in this trial; in this arm, systolic BP was less well controlled
for the first year than in the hydrochlorothiazide arm.
Although 1 observer began questioning the role of the class
for first-line treatment,50 during the era of direct comparison
trials, the arm representing traditional classes of drugs most
often offered participating clinicians a choice of thiazides or
␤-blockers or selected a ␤-blocker (most commonly, atenolol) as the comparator. Generally, before 2005 there were few
differences for major cardiovascular event rates in individual
trials between such regimens and those based on newer
classes; an exception was the Losartan Intervention For End
point reduction in hypertension (LIFE) trial, in which a
composite of stroke, MI, and cardiovascular death was 13%
(P⫽0.02) less frequent in the losartan-based than in the
atenolol-based group, owing mostly to an advantage for
stroke.51 This small disadvantage for the ␤-blocker in a
special study population selected for left ventricular hypertrophy, in the light of numerous trials not showing such a
disadvantage, was not sufficient for the JNC7 group to
downgrade the role of ␤-blockers as alternative first-line
drugs.49
The balance of evidence changed substantially with the
reporting of results of the Anglo-Scandinavian Cardiac Outcomes Trial—Blood Pressure Lowering Treatment Arm
(ASCOT-BPLA).52 This large open-label trial (PROBE design) compared atenolol-based treatment (bendroflumethiazide as an add-on) with amlodipine-based treatment (perindopril as add-on); it was stopped early because of 11% lower
all-cause mortality in the amlodipine group (P⫽0.02). When
the study was terminated early, there was no advantage in this
arm for the primary CHD end point (hazard ratio⫽0.90;
P⫽0.11), and stroke was also significantly reduced, by 23%.
This may have been related to the lower mean BP: 2.7/
1.9 mm Hg on average, with a greater difference during the
first year. Both effects may have in turn been related to the
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
Thiazide-Type Diuretics and ␤-Blockers
2701
Figure 8. Outcome data for all ␤-blockers vs other antihypertensive treatment.53 CONVINCE indicates Controlled Onset Verapamil
Investigation of Cardiovascular End Points; ELSA, European Lacidipine Study on Atherosclerosis; HAPPHY, Heart Attack Primary Prevention in Hypertension; INVEST, International Verapamil-Trandolapril Study; MRC, Medical Research Council; NORDIL, Nordic Diltiazem Study; STOP-2, Swedish Trial in Old Patients With Hypertension; and UKPDS, UK Prospective Diabetes Study. Reprinted from
Lindholm et al,53 with permission. Copyright 2005, Elsevier.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2702
Circulation
May 20, 2008
once daily use of atenolol and/or the relatively low dose of
the thiazide added to atenolol when needed for BP control.
After the publication of ASCOT, several meta-analyses
were produced that led to recommendations unfavorable to
␤-blockers (while continuing to acknowledge that evidence
on non-atenolol ␤-blockers in a primary prevention setting is
scanty).53,54 Although the trials included differed somewhat
between the meta-analyses and the Lindholm article did not
address heart failure, they both concluded that ␤-blockers
were less effective than other major antihypertensive classes
in preventing stroke (Figure 8). Therefore, the recommendations were to delete ␤-blockers as a first-line antihypertensive
treatment, except in cases in which there are compelling
indications, mainly for patients with CHD.
Conclusions
We believe that the JNC7 report came to the proper conclusions about thiazides, on the basis of (1) results of the
abundant trials aimed at the effect of BP reduction on clinical
events comparing thiazide-based regimens with placebo or
lesser treated controls; (2) ALLHAT results in the context of
the few other trials evaluating thiazide-based treatment versus
another class (trials testing a ␤-blocker and/or thiazide
protocol versus another class are not very informative); and
(3) the network meta-analysis by Psaty et al48 that summarized all of this evidence. As noted in JNC7, thiazides (1) are
well tolerated, better than other classes in the double-blind
setting of ALLHAT, in which a heterogeneous patient population was treated in widely diverse clinical settings; (2) are
at least as effective as other classes for BP control in most
patients, with the possible exception of younger white men55;
(3) are unsurpassed in preventing cardiovascular events and
improving survival, including an advantage in prevention of
heart failure (in the short term versus ACE inhibitors and in
the long term versus CCBs); and (4) have very low acquisition costs. Although frequency of their use in antihypertensive regimens has increased substantially since the ALLHAT
and JNC7 publications,56 they are still underutilized, and
organized efforts to improve performance in this regard are
continuing, most impressively in the Department of Veterans
Affairs medical system (William Cushman, MD, personal
communication, September 20, 2006).
Some commentators have focused on the small increase in
IDM associated with thiazide use while virtually ignoring the
advantage for heart failure prevention. This seems backward.
Regarding diabetes, naturally occurring disease carries appreciable long-term cardiovascular risk, but it is not clear that
thiazide-induced cases do, and they are largely preventable or
reversible by management of potassium balance. On the other
hand, heart failure imparts high functional impact and mortality risk in the short to medium term. It seems that this
benefit should be given considerable weight, except in
patients known to be at very low risk of heart failure. Thus,
the British recommendation to the effect that thiazides and
CCBs are equally desirable choices as first-line drugs in older
patients seems ill considered.54
As noted above, the role of other major drug classes (ACE
inhibitors, angiotensin receptor blockers, ␤-blockers, and
CCBs) in JNC7 is in cases in which a thiazide is not tolerated
or contraindicated, which is an infrequent situation; for listed
compelling indications; or in combination treatment, which is
likely to be required to control BP in most patients. There is
no update of JNC guidelines currently available that might
have changed the recommendations regarding ␤-blockers, but
British guidelines have placed them as a lesser choice than
the other classes except for compelling indications, ie, situations in which there is another indication for their use. This
conclusion seems reasonable to us, even though the data are
scant for non-atenolol ␤-blockers and the evidence for benefit in secondary prevention is much stronger for other
␤-blockers.57 Interestingly, the parallel situation for ACE
inhibitors in younger patients (lack of cardiovascular outcome data) did not deter the British report from recommending them as first choice in this demographic group. Once
again, the recommendation seems to turn on effects on
dysglycemia without really knowing their importance. However, to our knowledge, there is no reassuring evidence
regarding associated risks and mechanisms for ␤-blockers as
exists for thiazides.
Additional Research
Although a strong evidence base exists for a consensus on
preferred and other acceptable drug classes for initiating
antihypertensive pharmacotherapy, there are several other
major research questions for ongoing and needed trials,
several of which were highlighted in a report from a 2003
National Heart, Lung, and Blood Institute workshop.58 These
include the following: (1) What is the optimal second drug
class to add to a thiazide? (2) What goal BP, especially
systolic, should be sought for minimizing CVD risk? From
the evidence on heart failure in ALLHAT and the importance
of the condition generally, we believe that heart failure should
be included in any composite primary CVD end point for
future hypertension treatment trials.
Acknowledgment
The authors acknowledge the invaluable technical assistance of Sara
L. Pressel, MS.
Disclosures
Dr Davis has served as a consultant for Biomarin, GlaxoSmithKline,
Proctor & Gamble, and Takeda Pharmaceutical Company Limited.
Dr Cutler has no conflicts of interest.
References
1. Rosamond W, Flegal K, Friday G, Furie K, Go A, Greenlund K, Haase N,
Ho M, Howard V, Kissela B, Kittner S, Lloyd-Jones D, McDermott M,
Meigs J, Moy C, Nichol G, O’Donnell CJ, Roger V, Rumsfeld J, Sorlie
P, Steinberger J, Thom T, Wasserthiel-Smoller S, Hong Y; American
Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics—2007 update: a report from the
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
American Heart Association Statistics Committee and Stroke Statistics
Subcommittee. Circulation. 2007;115:e69 – e171.
WHO. The World Health Report 2002: reducing risks, promoting healthy
life. Available at: http://www.who.int/whr/2002/en/. Accessed June 25,
2007.
Veterans Administration Cooperative Study Group on Antihypertensive
Agents. Effects of treatment on morbidity in hypertension: results in
patients with diastolic blood pressures averaging 115 through
129 mm Hg. JAMA. 1967;202:1028 –1034.
Veterans Administration Cooperative Study Group on Antihypertensive
Agents. Effects of treatment on morbidity in hypertension, II: results in
patients with diastolic blood pressure averaging 90 through 114 mm Hg.
JAMA. 1970;213:1143–1152.
Hypertension Detection and Follow-up Program Cooperative Group.
Five-year findings of the Hypertension Detection and Follow-up
Program, I: reduction in mortality of persons with high blood pressure,
including mild hypertension. JAMA. 1979;242:2562–2571.
Medical Research Council Working Party. MRC trial of treatment of mild
hypertension: principal results. BMJ (Clin Res Ed). 1985;291:97–104.
MRC Working Party. Medical Research Council trial of treatment of
hypertension in older adults: principal results. BMJ. 1992;304:405– 412.
SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program
(SHEP). JAMA. 1991;265:3255–3264.
The sixth report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure. Arch
Intern Med. 1997;157:2413–2446.
MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R,
Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart
disease, part 1: prolonged differences in blood pressure: prospective
observational studies corrected for the regression dilution bias. Lancet.
1990;335:765–774.
Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA,
Godwin J, Qizilbash N, Taylor JO, Hennekens CH. Blood pressure,
stroke, and coronary heart disease, part 2: short-term reductions in blood
pressure: overview of randomised drug trials in their epidemiological
context. Lancet. 1990;335:827– 838.
Staessen JA, Fagard R, Thijs L, Celis H, Arabidze GG, Birkenhager WH,
Bulpitt CJ, de Leeuw PW, Dollery CT, Fletcher AE, Forette F, Leonetti
G, Nachev C, O’Brien ET, Rosenfeld J, Rodicio JL, Tuomilehto J,
Zanchetti A; The Systolic Hypertension in Europe (SYST-EUR) Trial
Investigators. Randomised double-blind comparison of placebo and
active treatment for older patients with isolated systolic hypertension.
Lancet. 1997;350:757–764.
The SOLVD Investigators. Effect of enalapril on survival in patients with
reduced left ventricular ejection fractions and congestive heart failure.
N Engl J Med. 1991;325:293–302.
Rutherford JD, Pfeffer MA, Moye LA, Davis BR, Flaker GC, Kowey PR,
Lamas GA, Miller HS, Packer M, Rouleau JL; SAVE Investigators.
Effects of captopril on ischemic events after myocardial infarction: results
of the Survival and Ventricular Enlargement trial. Circulation. 1994;90:
1731–1738.
Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G; The Heart
Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events
in high-risk patients. N Engl J Med. 2000;342:145–153.
Fox KM; EURopean trial On reduction of cardiac events with Perindopril
in stable coronary Artery disease Investigators. Efficacy of perindopril in
reduction of cardiovascular events among patients with stable coronary
artery disease: randomised, double-blind, placebo-controlled, multicentre
trial (the EUROPA study). Lancet. 2003;362:782–788.
The SOLVD Investigators. Effect of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. N Engl J Med. 1992;327:685– 691.
Kostis JB, Davis BR, Cutler J, Grimm RH Jr, Berge KG, Cohen JD, Lacy
CR, Perry HM Jr, Blaufox MD, Wassertheil-Smoller S, Black HR, Schron
E, Berkson DM, Curb JD, Smith WM, McDonald R, Applegate WB;
SHEP Cooperative Research Group. Prevention of heart failure by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA. 1997;278:212–216.
Thiazide-Type Diuretics and ␤-Blockers
2703
19. The 1988 report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med. 1988;
148:1023–1038.
20. The fifth report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC V). Arch Intern Med.
1993;153:154 –183.
21. Manolio TA, Cutler JA, Furberg CD, Psaty BM, Whelton PK, Applegate
WB. Trends in pharmacologic management of hypertension in the United
States. Arch Intern Med. 1995;155:829 – 837.
22. Multiple Risk Factor Intervention Trial Research Group. Mortality after
10 1/2 years for hypertensive participants in the Multiple Risk Factor
Intervention Trial. Circulation. 1990;82:1616 –1628.
23. Psaty BM, Smith NL, Siscovick DS, Koepsell TD, Weiss NS, Heckbert
SR, Lemaitre RN, Wagner EH, Furberg CD. Health outcomes associated
with antihypertensive therapies used as first-line agents: a systematic
review and meta-analysis. JAMA. 1997;277:739 –745.
24. Davis BR, Cutler JA, Gordon DJ, Furberg CD, Wright JT Jr, Cushman
WC, Grimm RH, LaRosa J, Whelton PK, Perry HM, Alderman MH, Ford
CE, Oparil S, Francis C, Proschan M, Pressel S, Black HR, Hawkins CM;
ALLHAT Research Group. Rationale and design for the Antihypertensive
and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).
Am J Hypertens. 1996;9:342–360.
25. ALLHAT Officers and Coordinators for the ALLHAT Collaborative
Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel
blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment
to Prevent Heart Attack Trial (ALLHAT). JAMA. 2002;288:2981–2997.
26. Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack
Trial Collaborative Research Group. Diuretic versus alpha-blocker as
first-step antihypertensive therapy: final results from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial
(ALLHAT). Hypertension. 2003;42:239 –246.
27. Turnbull F; Blood Pressure Lowering Treatment Trialists’ Collaboration.
Effects of different blood-pressure-lowering regimens on major cardiovascular events: results of prospectively-designed overviews of randomised trials. Lancet. 2003;362:1527–1535.
28. Chalmers J. All hats off to ALLHAT: a massive study with clear
messages. J Hypertens. 2003;21:225–228.
29. PROGRESS Collaborative Group. Randomised trial of a perindopril-based blood-pressure-lowering regimen among 6,105 individuals with
previous stroke or transient ischaemic attack. Lancet. 2001;358:
1033–1041.
30. Amery A, Birkenhager W, Brixko P, Bulpitt C, Clement D, Deruyttere M,
De Schaepdryver A, Dollery C, Fagard R, Forette F. Mortality and
morbidity results from the European Working Party on High Blood
Pressure in the Elderly trial. Lancet. 1985;1:1349 –1354.
31. Hansson L, Lindholm LH, Ekbom T, Dahlof B, Lanke J, Schersten B,
Waster PO, Hedner T, de Faire U. Randomised trial of old and new
antihypertensive drugs in elderly patients: cardiovascular mortality and
morbidity the Swedish Trial in Old Patients with Hypertension-2 study.
Lancet. 1999;354:1751–1756.
32. Black HR, Elliott WJ, Grandits G, Grambsch P, Lucente T, White WB,
Neaton JD, Grimm RH Jr, Hansson L, Lacourciere Y, Muller J, Sleight P,
Weber MA, Williams G, Wittes J, Zanchetti A, Anders RJ; CONVINCE
Research Group. Principal results of the Controlled Onset Verapamil
Investigation of Cardiovascular End Points (CONVINCE) trial. JAMA.
2003;289:2073–2082.
33. Brown MJ, Palmer CR, Castaigne A, de Leeuw PW, Mancia G, Rosenthal
T, Ruilope LM. Morbidity and mortality in patients randomised to
double-blind treatment with a long-acting calcium-channel blocker or
diuretic in the International Nifedipine GITS study: Intervention as a Goal
in Hypertension Treatment (INSIGHT). Lancet. 2000;356:366 –372.
34. Wing LM, Reid CM, Ryan P, Beilin LJ, Brown MA, Jennings GL,
Johnston CI, McNeil JJ, Macdonald GJ, Marley JE, Morgan TO, West
MJ; Second Australian National Blood Pressure Study Group. A comparison of outcomes with angiotensin-converting-enzyme inhibitors and
diuretics for hypertension in the elderly. N Engl J Med. 2003;348:
583–592.
35. Ernst ME, Carter BL, Goerdt CJ, Steffensmeier JJG, Phillips BB, Zimmerman MB, Bergus GR. Comparative antihypertensive effects of hydro-
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2704
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
Circulation
May 20, 2008
chlorothiazide and chlorthalidone on ambulatory and office blood
pressure. Hypertension. 2006;47:352–358.
Frohlich ED. Treating hypertension: what are we to believe? N Engl
J Med. 2003;348:639 – 641.
Cutler JA. The ANBP2 and ALLHAT: conflicting or consistent? J Clin
Hypertens. 2003;5:192–195.
Piller LB, Davis BR, Cutler JA, Cushman WC, Wright JT Jr, Williamson
JD, Leenen FH, Einhorn PT, Randall OS, Golden JS, Haywood LJ; The
ALLHAT Collaborative Research Group. Validation of heart failure
events in the Antihypertensive and Lipid Lowering Treatment to Prevent
Heart Attack Trial (ALLHAT) participants assigned to doxazosin and
chlorthalidone. Curr Control Trials Cardiovasc Med. 2002;3:10.
Davis BR, Piller LB, Cutler JA, Furberg C, Dunn K, Franklin S, Goff D,
Leenen F, Mohiuddin S, Papademetriou V, Proschan M, Ellsworth A,
Golden J, Colon P, Crow R; Antihypertensive and Lipid-Lowering
Treatment to Prevent Heart Attack Trial Collaborative Research Group. Role
of diuretics in the prevention of heart failure: the Antihypertensive and
Lipid-Lowering Treatment to Prevent Heart Attack Trial. Circulation. 2006;
113:2201–2210.
Einhorn PT, Davis BR, Massie BM, Cushman WC, Piller LB, Simpson
LM, Levy D, Nwachuku CE, Black HR; ALLHAT Collaborative
Research Group. The Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial (ALLHAT) heart failure validation study:
diagnosis and prognosis. Am Heart J. 2007;153:42–53.
Grimm RF, Davis BR, Cutler JA, Piller LB, Margolis K, Barzilay J, Dart
R, Graumlich JF, Murden R, Randall O, for the ALLHAT Collaborative
Research Group. Did blood pressure medication withdrawal prior to
randomization influence subsequent rates of heart failure in the Antihypertensive and Lipid-Lowering Treatment of Prevent Heart Attack Trial?
Presented at the 2007 American College of Cardiology meeting; March
24 –27, 2007; New Orleans, La.
Piegorsch WW, Weinberg CR, Taylor JA. Non-hierarchical logistic
models and case-only designs for assessing susceptibility in
population-based case-control studies. Stat Med. 1994;13:153–162.
Whelton PK, Barzilay J, Cushman WC, Davis BR, Ilamathi E, Kostis JB,
Leenen FH, Louis GT, Margolis KL, Mathis DE, Moloo J, Nwachuku C,
Panebianco D, Parish DC, Pressel S, Simmons DL, Thadani U; ALLHAT
Collaborative Research Group. Clinical outcomes in antihypertensive
treatment of type 2 diabetes, impaired fasting glucose concentration, and
normoglycemia: Antihypertensive and Lipid-Lowering Treatment to Prevent
Heart Attack Trial (ALLHAT). Arch Intern Med. 2005;165:1401–1409.
Barzilay JI, Davis BR, Cutler JA, Pressel SL, Whelton PK, Basile J,
Margolis KL, Ong ST, Sadler LS, Summerson J; ALLHAT Collaborative
Research Group. Fasting glucose levels and incident diabetes mellitus in
older nondiabetic adults randomized to receive 3 different classes of
antihypertensive treatment: a report from the Antihypertensive and LipidLowering Treatment to Prevent Heart Attack Trial (ALLHAT). Arch
Intern Med. 2006;166:2191–2201.
Kostis JB, Wilson AC, Freudenberger RS, Cosgrove NM, Pressel SL,
Davis BR, SHEP Collaborative Research Group. Long-term effect of
diuretic-based therapy on fatal outcomes in subjects with isolated systolic
hypertension with and without diabetes. Am J Cardiol. 2005;95:29 –35.
Cutler JA. Thiazide-associated glucose abnormalities: prognosis,
etiology, and prevention: is potassium balance the key? Hypertension.
2006;48:198 –200.
47. Zillich AJ, Garg J, Basu S, Bakris GL, Carter BL. Thiazide diuretics,
potassium, and the development of diabetes: a quantitative review.
Hypertension. 2006;48:219 –224.
48. Psaty BM, Lumley T, Furberg CD, Schellenbaum G, Pahor M, Alderman
MH, Weiss NS. Health outcomes associated with various antihypertensive therapies used as first-line agents: a network meta-analysis.
JAMA. 2003;289:2534 –2544.
49. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL
Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ; Joint
National Committee on Prevention, Detection, Evaluation, and Treatment
of High Blood Pressure National Heart, Lung, and Blood Institute
National High Blood Pressure Education Program Coordinating Committee.
Seventh report of the Joint National Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood Pressure. Hypertension. 2003;42:
1206–1252.
50. Messerli FH, Grossman E, Goldbourt U. Are beta-blockers efficacious as
first-line therapy for hypertension in the elderly? A systematic review.
JAMA. 1998;279:1903–1907.
51. Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U,
Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm
LH, Nieminen MS, Omvik P, Oparil S, Wedel H; LIFE Study Group.
Cardiovascular morbidity and mortality in the Losartan Intervention for
Endpoint Reduction in Hypertension study (LIFE): a randomised trial
against atenolol. Lancet. 2002;359:995–1003.
52. Dahlof B, Sever PS, Poulter NR, Wedel H, Beevers DG, Caulfield M,
Collins R, Kjeldsen SE, Kristinsson A, McInnes GT, Mehlsen J,
Nieminen M, O’Brien E, Ostergren J; ASCOT Investigators. Prevention
of cardiovascular events with an antihypertensive regimen of amlodipine
adding perindopril as required versus atenolol adding bendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes
Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomised controlled trial. Lancet. 2005;366:895–906.
53. Lindholm LH, Carlberg B, Samuelsson O. Should beta blockers remain
first choice in the treatment of primary hypertension? A meta-analysis.
Lancet. 2005;366:1545–1553.
54. National Collaborating Centre for Chronic Conditions. Hypertension:
Management of Hypertension in Adults in Primary Care: Partial Update.
London, UK: Royal College of Physicians; 2006.
55. Materson BJ, Reda DJ, Cushman WC, Massie BM, Freis ED, Kochar MS,
Hamburger RJ, Fye C, Lakshman R, Gottdiener J; The Department of
Veterans Affairs Cooperative Study Group on Antihypertensive Agents.
Single-drug therapy for hypertension in men: a comparison of six antihypertensive agents with placebo. N Engl J Med. 1993;328:914 –921.
56. Stafford RS, Monti V, Furberg CD, Ma J. Long-term and short-term
changes in antihypertensive prescribing by office-based physicians in the
United States. Hypertension. 2006;48:213–218.
57. Freemantle N, Cleland J, Young P, Mason J, Harrison J. ␤ Blockade after
myocardial infarction: systematic review and meta regression analysis. Br
Med J. 1999;318:1730 –1737.
58. The National Heart, Lung, and Blood Institute Working Group on Future
Directions in Hypertension Treatment Trials. Major clinical trials: what
should be done? Hypertension. 2005;46:1– 6.
59. IMS America. National Disease and Therapeutic Index. Ambler, Pa: IMS
America; 1993.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Cutler and Davis
Thiazide-Type Diuretics and ␤-Blockers
2705
Response to Cutler and Davis
Franz H. Messerli, MD; Sripal Bangalore, MD, MHA; and Stevo Julius, MD
We are delighted that Drs Cutler and Davis find the conclusion of the British guidelines to remove ␤-blockers from their
first-line status in hypertension “reasonable”—a position we have taken a decade ago. Indeed, one may wonder why
␤-blockers were ever elevated to that status; ever since the MRC trial, every single study comparing ␤-blockers with either
placebo or other therapy has been negative. Never has a drug class been prescribed so much on the basis of so little
evidence. In the United States alone, more than 120 million prescriptions for ␤-blockers were written in 2007. This means
that numerous patients still are exposed to the adverse effects, cost, and inconvenience of a drug class without ever having
any benefit. We also agree with our colleagues that, on the basis of the ALLHAT study, “thiazide-type diuretics should be
the preferred first-step therapy for prevention of heart failure in high-risk patients with hypertension,” although this is true
for chlorthalidone only. However, we strongly object to the authors’ suggestion that evidence gained in high-risk patients
such as ALLHAT “should be reasonably extrapolated beyond the exact sample in which it was conducted.” That this is pure
conjecture and that there is no scientific basis for extrapolation has now been documented by ACCOMPLISH, which now
relegates the thiazides to third-line therapy. As to the cost of thiazides, the authors’ own recent ALLHAT analysis showed
that amlodipine tended to be preferred on the basis of the fact that it provided a (nonsignificantly) greater survival benefit
than the other drugs. This analysis is misleading in that it did not even consider the now generic price of these drugs, nor
did it assess cost of new-onset diabetes associated with them. On the basis of our meta-analysis, we estimate that ␤-blockers
and diuretics could account for more than 100 000 cases of new-onset diabetes in the United States every year. Recent data
on outcome, cost, and adverse events allow us now to assign antihypertensive drugs to their proper place in the therapeutic
arsenal; keeping thiazides or ␤-blockers as first-line therapy seems to us, as Ralph Waldo Emerson said, “a foolish
consistency. . .the hobgoblin of little minds.”
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
CONTROVERSIES IN
CARDIOVASCULAR MEDICINE
Should ␤-blockers and diuretics remain as
first-line therapy for hypertension?
Risk/Benefit Assessment of ␤-Blockers and Diuretics
Precludes Their Use for First-Line Therapy in Hypertension
Franz H. Messerli, MD; Sripal Bangalore, MD, MHA; Stevo Julius, MD
I
n 1977, the first report of the Joint National Committee
(JNC) on Detection, Evaluation, and Treatment of High
Blood Pressure was published.1 On the basis of the available
information, the committee suggested an algorithm in which
the “first step drugs should usually be a thiazide diuretic.”
The algorithm suggested either reserpine, methyldopa, or
propranolol as a second step. In 1984, JNC III recommended
␤-blockers as a class for first-line therapy on an equal basis
with the thiazide diuretics.2 From then on, in all the subsequent JNC reports until JNC 7 in 2003, ␤-blockers and
diuretics remained first-line antihypertensive drugs.3 JNC 7
suggested that thiazide diuretics should be used as initial
therapy in most patients. If these were not tolerated or
contraindicated, then a ␤-blocker, on an equal basis with
angiotensin-converting enzyme (ACE) inhibitors, angiotensin
receptor inhibitors, and calcium antagonists, could be substituted. Thus, according to JNC 7, both thiazides and
␤-blockers are meant to be given to hypertensive patients
regardless of their risk factor status and regardless of their
concomitant metabolic abnormalities.
Response by Cutler and Davis p 2715
Diabetes and Blood Pressure
Within the same time frame, from 1980 to 2004, the prevalence of diabetes more than doubled in the United States, and
almost 10% of people aged ⬎20 years are currently suffering
from this disease.4 Patients with hypertension are known to be
at a higher risk of developing new-onset diabetes than are
normotensive subjects.5,6 In the Atherosclerosis Risk in Communities (ARIC) study, type 2 diabetes was almost twice as
likely to develop in hypertensive than in normotensive
subjects.5 In the Women’s Health Study, in a prospective
cohort of 38 172 women free of diabetes and cardiovascular
disease, baseline blood pressure and blood pressure progression were strong and independent predictors of incident type
2 diabetes.6 Women with baseline hypertension had a hazard
ratio for incident diabetes in the 4-year follow-up of 2.39
(95% CI, 1.95 to 2.93).6
Conversely, hypertension is exceedingly common in patients with type 2 diabetes, even in children and adolescents.
Upchurch et al7 reported that 55% of young people had
systolic blood pressures ⬎90 percentile at the time of
diagnosis of type 2 diabetes. In fact, hypertension at diagnosis
was as much as 8 times more common in adolescents with
type 2 diabetes compared with those with type 1 diabetes.
Thus, hypertension begets type 2 diabetes, and, conversely,
diabetes begets hypertension.
New-Onset Diabetes With Thiazide Diuretics
In the Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial (ALLHAT), ⬇10% of all patients
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.
From the Department of Medicine, Division of Cardiology, St Luke’s–Roosevelt Hospital and Columbia University College of Physicians and
Surgeons, New York, NY (F.H.M., S.B.), and the Department of Internal Medicine, University of Michigan, Ann Arbor (S.J.).
This article is Part II of a 2-part article. Part I appears on page 2691.
Reprint requests to Franz H. Messerli, MD, Division of Cardiology, St Luke’s–Roosevelt Hospital, Columbia University College of Physicians and
Surgeons, 1000 10th Ave, Suite 3B-30, New York, NY 10019. E-mail [email protected]
(Circulation. 2008;117:2706-2715.)
© 2008 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org
DOI: 10.1161/CIRCULATIONAHA.107.695007
2706
Downloaded from http://circ.ahajournals.org/
by guest on September 9, 2014
Messerli et al
Thiazide-Type Diuretics and ␤-Blockers
2707
Figure 1. Relative risk of new-onset diabetes mellitus in studies using diuretics against other agents (including placebo). We performed
a meta-analysis of all randomized controlled trials of patients on diuretics for hypertension with follow-up for at least 1 year and that
reported the incidence of new-onset diabetes. The sizes of the data markers relate to study sample size and the inverse of the SE of
each study. EWPHE indicates European Working Party for Hypertension in the Elderly; ACEi, ACE inhibitors; CCB, calcium channel
blockers; ANBP-2, Second Australian National Blood Pressure Study; and INSIGHT, Intervention as a Goal in Hypertension Treatment.
developed new-onset diabetes throughout the 4- to 5-year
duration of the study.8 However, this percentage was between
18% and 40% higher in patients in the chlorthalidone arm
than in those in the amlodipine and lisinopril arms, respectively.9 In the Antihypertensive Treatment and Lipid Profile
in a North of Sweden Efficacy Evaluation (ALPINE) study,10
hypertensive patients who had never been treated before were
randomized to either candesartan-based therapy (plus amlodipine if needed) or thiazide-based therapy (plus atenolol if
needed). After only 1 year, 18 patients in the diuretic group
fulfilled criteria of the metabolic syndrome, and 9 had
developed new-onset diabetes. The corresponding numbers in
the candesartan arm were 5 and 1, respectively.
Meta-Analysis of Diuretic-Based Studies
Figure 1 shows the results of a meta-analysis we prepared for
this review. We conducted a MEDLINE/PUBMED/EMBASE
search of studies using the terms diuretics and hypertension.
We limited our search to studies in human subjects and
English language in peer-reviewed journals from 1966 to
October 2007. All randomized controlled trials of patients
with hypertension treated with diuretic therapy with
follow-up for at least 1 year and that reported the incidence of
new-onset diabetes were included. Statistical analysis was
done with the use of standard software (Stata 9.0, Stata
Corporation) with the METAN program.11 The pooled effect
for each grouping of trials was derived from the point
estimate for each separate trial weighted by the inverse of the
variance (1/SE2). Heterogeneity was assessed visually with
funnel plots, Q (␹2) statistics, and/or I2 statistics.12 If trials
were homogeneous (P⬎0.05), a fixed-effect model was used
to calculate pooled effect sizes. Otherwise, a random-effect
model of DerSimonian and Laird13 was applied to calculate
overall differences. Publication bias was estimated with the
weighted regression test of Egger.
In the analysis of 6 trials enrolling 30 842 patients with
hypertension, diuretics resulted in a 32% increased risk of
new-onset diabetes compared with placebo or non–␤-blocker
antihypertensive agents. Compared with placebo, diuretics
resulted in a strong trend toward a 22% increased risk of
new-onset diabetes, suggesting that the risk is due to the
medication itself. When compared with antihypertensive
agents other than ␤-blockers, diuretics conferred a 35%
increased risk of new-onset diabetes. We also performed a
meta-regression analysis to evaluate the relationship between
follow-up duration of therapy and the risk of new-onset
diabetes. The risk of new-onset diabetes increased with
increasing duration of diabetic therapy with these agents
(Figure 2).
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2708
Circulation
May 20, 2008
diabetes a predefined end point. This requires a cautious
interpretation of the data.
Differences Among Antihypertensive Drugs
Figure 2. Relative risk of new-onset diabetes mellitus in the
studies using diuretics as a function of follow-up duration.
Meta-regression analysis was performed to evaluate the relationship between risk of new-onset diabetes mellitus and the
length of follow-up of patients on diuretic therapy. The diameter
of the circles represents the variance of the relative risk of each
individual trial. The line represents the regression fit with 95% CI
for the effect sizes. Abbreviations are as defined in Figure 1
legend.
New-Onset Diabetes With ␤-Blockers
The prodiabetic effect of the ␤-blockers has been less well
documented. In the ARIC study, hypertensive patients taking
␤-blockers had a 28% increased risk of developing diabetes
compared with those hypertensive subjects who did not take
any medication.5
Meta-Analysis of ␤-Blocker–Based Studies
Figures 3 and 4 shows the results of our meta-analysis of all
randomized controlled trials enrolling patients with hypertension treated with ␤-blocker therapy with a follow-up for at
least 1 year and that reported the incidence of new-onset
diabetes. We conducted a MEDLINE/PUBMED/EMBASE
search of studies using the terms adrenergic beta antagonists,
beta blockers, and hypertension. We limited our search to
studies in human subjects and English language in peerreviewed journals from 1966 to October 2007. In the analysis
of 6 trials enrolling 55 675 patients with hypertension,
␤-blockers conferred a 32% increased risk of new-onset
diabetes compared with placebo or nondiuretic antihypertensive agents. When compared with nondiuretic antihypertensive agents, ␤-blockers resulted in a 31% increased risk of
new-onset diabetes (Figure 3). The risk of new-onset diabetes
with ␤-blockers increased with duration of therapy (Figure 4).
However, in the mixed ␤-blocker/diuretic trials, in which
patients could be randomized to a ␤-blocker, a diuretic, or
their combination, ␤-blockers/diuretics resulted in an 11%
increased risk of new-onset diabetes compared with other antihypertensive agents (Figure 5). Of note, neither in ALLHAT nor
in any of the aforementioned analyzed studies was new-onset
The Glycemic Effects in Diabetes Mellitus: CarvedilolMetoprolol Comparison in Hypertensives (GEMINI) trial14
reported that with regard to metabolic adverse effects, not all
␤-blockers seem to be created equal. In patients with hypertension and type 2 diabetes who were treated with a blocker
of the renin-angiotensin system, the addition of carvedilol,
compared with metoprolol, had a favorable effect on glycemic control, insulin resistance, microalbuminuria, and body
weight. In a recent thorough and comprehensive network
meta-analysis, Elliott and Meyer15 documented the odds ratio
of new-onset diabetes to be 0.85, 0.90, 1.05, 1.25, and 1.34
with angiotensin receptor blockers, ACE inhibitors, calcium
antagonists, ␤-blockers, and diuretics, respectively, with placebo having a reference value of 1.0 (Figure 6).
Does Drug-Associated Diabetes Mellitus
Increase Morbidity and Mortality?
In regard to the clinical significance of drug-associated
new-onset diabetes, the ALLHAT investigators attempted to
reassure us by stating that “overall these metabolic differences did not translate into more cardiovascular events or into
higher all-cause mortality in the chlorthalidone group.”8 This
is a correct but perhaps somewhat myopic statement. We
should consider that antihypertensive therapy is lifelong and
that a follow-up period of 4 to 6 years, as was the case in
ALLHAT, is unlikely to give us any clues regarding the
long-term sequelae of the diabetes risk associated with either
␤-blockers or diuretics.9
Several studies have scrutinized this critical issue, as
follows.
1. Verdecchia et al16 reported a 16-year follow-up of
almost 800 initially untreated hypertensive patients,
6.5% of whom had diabetes at the onset and 5.8% of
whom developed new-onset diabetes throughout the
study. Fasting blood sugar at entry and diuretic treatment on follow-up were independent predictors of
new-onset diabetes (P⬍0.0001 and P⬍0.004, respectively). Compared with subjects who never developed
diabetes, the risk for cardiovascular disease during the
follow-up was very similar in patients who developed
diabetes (odds ratio, 2.92; 95% CI, 1.33 to 6.41;
P⫽0.007) and in the group that had preexisting diabetes
(odds ratio, 3.57; 95% CI, 1.65 to 7.73; P⫽0.001).
Patients with new-onset diabetes and those with a
previous diagnosis of diabetes were almost 3-fold more
likely to develop subsequent cardiovascular disease
than those who remained free of diabetes.
2. In a study involving almost 7000 patients, Alderman et
al17 showed that cardiovascular disease increased in
hypertensive diuretic users who developed hyperglycemia even when blood pressure was well controlled. The
authors stated, “Cardiovascular disease incidence has a
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Messerli et al
Thiazide-Type Diuretics and ␤-Blockers
2709
Figure 3. Relative risk of new-onset diabetes mellitus in studies using ␤-blockers (BB) against other agents (including placebo). We
performed a meta-analysis of all randomized controlled trials of patients on ␤-blockers for hypertension with follow-up for at least 1
year and that reported the incidence of new-onset diabetes. The sizes of the data markers relate to study sample size and the inverse
of the SE of each study. MRC indicates Medical Research Council; MRC-O, Medical Research Council trial of treatment of hypertension in older adults; AASK, African American Study of Kidney Disease and Hypertension; ASCOT, Anglo-Scandinavian Cardiac Outcomes Trial; and LIFE, Losartan Intervention For End point reduction in hypertension trial. Other abbreviations are as defined in Figure
1 legend.
direct dose response relation with diuretic used with
frequent users having the highest rate.”
3. In the 11 645-patient Multiple Risk Factor Intervention
Trial (MRFIT),18 the occurrence of new-onset diabetes
was increased by ⬎30% in diuretic-treated patients and
showed an excess mortality risk in the 18-year posttrial
follow-up.
4. Almgren et al19 followed a population sample of 7500
men in which 20.4% of treated hypertensive patients
developed diabetes. New-onset diabetes implied a significant increased risk for stroke, myocardial infarction,
and mortality (hazard ratio, 1.67, 1.66, and 1.42, respectively). The authors concluded that diabetes in treated
hypertensive patients was alarmingly common and carried a high risk for cardiovascular complications and
mortality.
5. Aksnes et al,20 in the Valsartan Antihypertensive Longterm Use Evaluation (VALUE) study, compared patients with diabetes at baseline and new-onset diabetes
with patients who did not develop diabetes (5250, 9995,
and 1298 patients, respectively). Patients with newonset diabetes had significantly higher cardiac morbidity, especially more congestive heart failure, than those
without diabetes (hazard ratio, 1.43). Patients were
followed up for an average of 4.2 years only.
6. In contrast to the aforementioned 5 studies, Kostis et
al,21 in a follow-up of the Systolic Hypertension in the
Elderly Program (SHEP), found no significant increase
in cardiovascular events in patients who had diabetes
associated with chlorthalidone therapy. In fact, these
patients had a better prognosis than those who had
preexisting diabetes, and diuretic treatment in subjects
who had diabetes was strongly associated with lower
long-term cardiovascular mortality and total mortality.
Thus, at present, the exact clinical significance of
diuretic-associated diabetes remains unknown, with 5
studies showing a detrimental outcome and 1 study
showing a neutral outcome.
In the present analysis using studies that reported the end
point of new-onset diabetes, the risks for other events—allcause mortality, cardiovascular mortality, myocardial infarction, stroke, and heart failure—are detailed in the Table.
Compared with non–␤-blocker antihypertensive agents, diuretics did not provide any incremental benefit for the end
points of all-cause mortality, cardiovascular mortality, myocardial infarction, and stroke but conferred a 33% reduction
in heart failure. The heart failure end point was driven mainly
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2710
Circulation
May 20, 2008
Figure 6. Risk of new-onset diabetes mellitus with antihypertensive treatment. ARB indicates angiotensin receptor blocker;
ACE-I, ACE inhibitor; and CCB, calcium channel blocker.
Reprinted from Elliott and Meyer,15 with permission. Copyright
2007, Lancet.
Figure 4. Relative risk of new-onset diabetes mellitus in the
studies using ␤-blockers as a function of follow-up duration.
Meta-regression analysis was performed to evaluate the relationship between risk of new-onset diabetes mellitus and the
length of follow-up of patients on ␤-blocker therapy. The diameter of the circles represents the variance of the relative risk of
each individual trial. The line represents the regression fit with
95% CI for the effect sizes. Abbreviations are as defined in Figure 1 and Figure 3 legends.
by the ALLHAT trial. Given the heterogeneity in the definition of heart failure used in various studies, the results should
be interpreted with caution. Similarly, compared with nondiuretic antihypertensive agents, ␤-blocker therapy resulted in
an 8% increased risk of all-cause mortality and 30% increased risk of stroke.
Conceivably, new-onset diabetes associated with diuretics
and/or ␤-blockers may be reversible on discontinuation of the
offending drug(s). Thus, treatment withdrawal could potentially separate the drug-induced new-onset diabetes from
spontaneously occurring new-onset diabetes. However, the
Figure 5. Relative risk of new-onset diabetes mellitus in studies using ␤-blockers (BB) or diuretics (mixed trials) against other agents
(including placebo). We performed a meta-analysis of all randomized controlled trials of patients on ␤-blockers/diuretics for hypertension with follow-up for at least 1 year and that reported the incidence of new-onset diabetes. The sizes of the data markers relate to
study sample size and the inverse of the SE of each study. CAPPP indicates Captopril Prevention Project; NORDIL, the Nordic Diltiazem study; and STOP-2, the second Swedish Trial in Old Patients with Hypertension. Other abbreviations are as defined in Figure 1
legend.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Messerli et al
Table.
Thiazide-Type Diuretics and ␤-Blockers
2711
Pooled Relative Risk of Cardiovascular Event Based on Our Meta-Analysis
All-Cause Deaths
Parameter
No.
RR (95% CI)
Cardiovascular Death
No.
Myocardial Infarction
Heart Failure
Stroke
RR (95% CI)
No.
RR (95% CI)
No.
RR (95% CI)
No.
RR (95% CI)
Diuretics
vs placebo
2
0.90 (0.79–1.02)
2
0.77 (0.63–0.94)
2
0.99 (0.43–2.30)
2
0.99 (0.70–1.41)
2
0.63 (0.50–0.80)
vs others
4
1.02 (0.98–1.06)
4
0.98 (0.91–1.04)
3
1.14 (0.81–1.61)
4
0.67 (0.57–0.79)
4
0.98 (0.86–1.11)
vs placebo
2
1.01 (0.88–1.16)
2
1.00 (0.83–1.20)
2
1.98 (0.53–7.37)
0.80 (0.64–1.01)
5
1.08 (1.02–1.14)
5
1.13 (0.98–1.31)
3
1.05 (0.94–1.18)
䡠䡠䡠
1.04 (0.89–1.21)
2
vs others
䡠䡠䡠
3
3
1.30 (1.18–1.44)
4
0.99 (0.92–1.07)
4
1.01 (0.89–1.13)
4
0.94 (0.83–1.06)
4
0.99 (0.84–1.15)
4
1.04 (0.87–1.26)
␤–Blockers
Mixed
vs others
RR indicates relative risk.
risk of new-onset diabetes continues to increase with duration
of therapy (Figure 4), and diabetes has been identified as a
coronary heart disease equivalent. Because antihypertensive
therapy is prescribed to prevent coronary heart disease, it
seems counterintuitive to select drug classes that indeed have
the exact opposite effect, ie, that can induce a coronary risk
equivalent.
Does Concomitant Renin-Angiotensin System
Blockade Reduce the Risk of New-Onset
Diabetes Associated With Diuretics
and ␤-Blockers?
Over the past few years, we have come to realize that even
patients with uncomplicated hypertension may need ⱖ2
drugs to bring their blood pressure under control. Because
most often diuretics are prescribed in combination with other
drugs such as ACE inhibitors or angiotensin receptor blockers, it has been argued that these drugs, by mitigating
hypokalemia associated with the thiazides, may abolish some
metabolic adverse effects. However, in the International
Verapamil-Trandolapril (INVEST) study, the addition of
hydrochlorothiazide in a dose-dependent way increased the
risk of new-onset diabetes in patients treated with either
atenolol or verapamil.22,23 The metabolic effects of the
thiazide diuretics in combination with renin-angiotensin system blockade were further investigated in the Study of
Trandolapril/Verapamil SR and Insulin Resistance (STAR)
study,24 in which patients with the metabolic syndrome were
randomized to either verapamil/trandolapril or losartan/hydrochlorothiazide. At the end of 1 year, losartan/hydrochlorothiazide increased plasma glucose significantly more than
verapamil/trandolapril after all oral glucose tolerance testing.
The differences were more pronounced with high-dose combinations than with low-dose combinations.
Similarly, in the GEMINI study,14 in which all patients
were treated with a blocker of the renin-angiotensin system,
we observed significant differences in glycosylated hemoglobin and insulin resistance between patients on metoprolol and
those on carvedilol. Thus, the presence of an ACE inhibitor or
an angiotensin receptor blocker did not abolish the effects of
␤-blockade on metabolic parameters.
Most recently, the Avoiding Cardiovascular Events in
Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) trial, which compared the fixed
combination of benazepril/amlodipine with benazepril/hydrochlorothiazide in more than 10 000 patients, was stopped
prematurely because of a 20% reduction in cardiovascular
mortality in the amlodipine/benazepril arm. Thus the
ACCOMPLISH data, in hypertension complicated by multiple risk factors, clearly establishes outcome superiority for a
CCB/ACE inhibitor combination over an ACE inhibitor/
diuretic combination, thereby relegating the thiazides to
third-line therapy.
Antihypertensive Therapy
and Dyslipoproteinemia
The effects of antihypertensive therapy on serum lipids were
analyzed by Kasiske et al25 in a meta-analysis of 474 clinical
trials in ⬎65 000 patients. Diuretics caused a relative increase
in cholesterol levels that was more pronounced with higher
doses and worse in black patients than in nonblacks.
Chlorthalidone caused a greater increase in low-density
lipoprotein cholesterol than did other diuretics. ␤-Blockers
caused an increase in triglyceride levels that was smaller for
agents with intrinsic sympathetic activity. A beneficial effect
on total cholesterol, low-density lipoprotein cholesterol, and
triglycerides was seen with the ␣-blockers, and a beneficial
effect on triglycerides and in diabetic patients on total
cholesterol was seen with ACE inhibitors. Calcium antagonists had no effects on lipids. Although, as with new-onset
diabetes, the long-term effects of these drug-associated
changes are not known, it seems unlikely that the increase in
total cholesterol and triglycerides with diuretics and
␤-blockers, respectively, will be beneficial. In GEMINI, there
were distinct differences in total cholesterol and triglycerides
between the metoprolol arm and the carvedilol arm, indicating again that blockade of the renin-angiotensin system did
not abolish the effect of ␤-blockers on lipoproteins.14 Signif-
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2712
Circulation
May 20, 2008
icantly more patients had to be started on a statin in the
metoprolol arm than in the carvedilol arm.14
Weight Change and Antihypertensive Therapy
Sharma et al,26 in a systematic analysis of prospective
randomized controlled trials, found that ␤-blockers increased
body weight by a median of 1.2 kg; weight gain occurred
primarily during the first few months. Of note, not all
␤-blockers seem to cause weight gain equally. We recently
showed no significant weight gain in the GEMINI study in
patients on carvedilol, whereas with metoprolol, after 6
months the average weight increased by 1.2 kg.27 Recently,
Scholze et al28 compared a weight loss regimen with sibutramine in a 16-week, double-blind, placebo-controlled, randomized, multicenter study of 171 obese hypertensive patients who were randomized to 3 different antihypertensive
regimens (felodipine/ramipril, verapamil/trandolapril, or
metoprolol/hydrochlorothiazide). Overall weight loss and
reduction of abdominal obesity were markedly attenuated in
the metoprolol/hydrochlorothiazide group compared with the
other 2 groups. Similarly, improvement in glucose tolerance
and hypertriglyceridemia with weight loss was abrogated in
the metoprolol/hydrochlorothiazide cohort compared with the
other 2 arms. This would indicate that antihypertensive
combination therapy with a renin-angiotensin system blocker
and a calcium blocker facilitates weight loss with a sibutramine regimen significantly more than does a ␤-blocker/diuretic– based regimen. This is further evidence against
␤-blocker/diuretic treatment for hypertension, especially in
the large majority of overweight and obese patients.
␤-Blockade in the Young Hypertensive Patient
Because numerous studies have shown that the young borderline hypertensive patient is characterized by an elevated
cardiac output and an “inappropriately” normal total peripheral resistance, it was thought that ␤-blockade would be the
most “physiological” antihypertensive therapy in this patient
population. However, a “normalization” of hemodynamics
does not necessarily prevent the transition from borderline to
more established hypertension. In addition, the ALPINE
study taught us that, if anything, young patients (aged ⬍55
years) were more susceptible to the development of newonset diabetes with diuretic-based therapy than patients aged
⬎55 years.29 Similarly, total cholesterol increased with diuretic therapy in the younger ALPINE study population,
whereas it fell in those aged ⬎55 years. Thus, ␤-blockers are
not more beneficial (or less detrimental) in the young than
they are in the older patient.
Other Adverse Effects of ␤-Blockers
and Diuretics
Neither ␤-blockers nor diuretics are well tolerated antihypertensive drugs. In most systematic analyses, they fare consistently worse than blockers of the renin-angiotensin system
and calcium antagonists. In the Cochrane meta-analysis, the
withdrawal rate of patients treated with ␤-blockers was twice
as high as the rate in patients treated with diuretics.30 Our
analysis allowed us to calculate that for every stroke or heart
attack prevented, 3 patients remained impotent and 8 experienced fatigue to the extent that they withdrew from such
therapy.31 Clearly, this is not an acceptable risk/benefit ratio
for a completely asymptomatic disease such as stage 1
hypertension.
The question of nephrotoxicity of long-term thiazide diuretic therapy continues to surface.32–35 Because most prospective studies last 6 years at most, solid documents attesting
to the safety of diuretic therapy (and of all antihypertensive
drugs) are lacking. Similarly, the issue of carcinogenicity
with long-term diuretic therapy has not been resolved. We
reported in a meta-analysis an association between diuretic
use and renal cell carcinoma with a pooled odds ratio of 1.54
in 10 independent case-control studies and 3 cohort studies.35
The association between renal cell carcinoma and diuretic
therapy remains a concern because the renal tubular cell, ie,
the cell that turns cancerous, is also the main target of the
diuretic pharmacological effect. Because diuretic-associated
carcinogenicity seemed to be cumulative in some studies, it
may be yet another reason not to expose young patients to
years and decades of thiazide therapy. Clearly, however, the
issue of carcinogenicity with hypertension and/or antihypertensive therapy is exceedingly complex, and hasty conclusions should be avoided.
Morbidity and Mortality Reduction With
␤-Blockers and Diuretics
The efficacy of ␤-blockers in reducing morbidity and mortality in hypertension has come into question recently.36 –38
Most “evidence” used in current guidelines for hypertension
stems from extrapolation from post–myocardial infarction
studies or cohorts of hypertensive patients specifically selected for high cardiovascular risk. For many years, extensive
marketing efforts by the pharmaceutical industry have
planted the seeds of ␤-blockers being “cardioprotective.”
However, in uncomplicated hypertension, particularly in the
elderly, outcome data with ␤-blockers showed either no
benefit or even harm.37–39
We are not debating that thiazide diuretics are powerful
drugs that repeatedly have been shown to reduce morbidity
and mortality in high-risk established essential hypertension
such as, for example, in the ALLHAT study. Clearly, in this
setting, the benefits of therapy distinctly outweigh the lowgrade negative metabolic adverse effects of the thiazides.
Thus, unlike with ␤-blockers, the risk/benefit ratio with
thiazide diuretics, specifically chlorthalidone, remains acceptable in such patients, and they should be considered
candidates for such therapy. We should also remember that
in the elderly, who are exposed to these drugs for a limited
time, the metabolic adverse effects may be more acceptable
than in the comparatively younger patient. However, in the
younger patients with stage 1 hypertension who will be
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Messerli et al
exposed to antihypertensive therapy for decades, the trade-off
of lowering blood pressure at the expense of increasing the
risk of new-onset diabetes by up to 10% yearly is not
acceptable. This is particularly true given that there are
efficacious and safe antihypertensive drugs other than diuretics and ␤-blockers that do not increase the risk of new-onset
diabetes or dyslipoproteinemia.40
Pseudoantihypertensive Effect and the
␤-Blocker Hypertension Paradox
In hypertension as well as coronary heart disease, heart rate
has been shown to be a blood pressure–independent risk
factor for morbidity and mortality. Conversely, Kjekshus41
has shown that the benefit of ␤-blockers in the post–
myocardial infarction population is related to the negative
chronotropic effect: the slower the heart rate, the greater is the
reduction in mortality. A similar observation was made in
congestive heart failure, in which ␤-blockers remain a cornerstone in the therapeutic arsenal. Thus, heart rate lowering
by ␤-blockade increased the survival rate in congestive heart
failure as well as in coronary artery disease. In contrast, we
recently showed that in hypertension, ␤-blockade–induced
bradycardia had a detrimental effect. In a meta-analysis of
almost 80 000 patients in 9 trials, heart rate correlated
negatively with cardiovascular mortality, nonfatal myocardial
infarction, stroke, and heart failure (all P⬍0.0001).42 The
reason for this hypertension paradox may have to do with the
reflected pulse wave. This wave should ideally return toward
the heart during diastole to augment diastolic filling. If it
returns earlier during the cardiac cycle (as is the case with
␤-blocker–induced bradycardia), it amplifies the outgoing
pressure wave and leads to an increase in central aortic
pressure. As a result, central aortic pressure will be lowered
less than brachial pressure, as was seen in the Conduit Artery
Function Evaluation (CAFÉ) study.43 Atenolol-based treatment was significantly less effective than amlodipine-based
treatment in lowering aortic systolic pressure and pulse
pressure despite identical brachial pressure in both treatment
arms. This pseudoantihypertensive effect could explain why
␤-blocker– based strategies are less effective than alternative
treatments at regressing end-organ damage and in preventing
stroke. This pseudoantihypertensive effect could also explain
why ␤-blocker–induced bradycardia may not necessarily be
beneficial in hypertensive patients.
Thiazide-Type Diuretics and ␤-Blockers
2713
placebo in hypertension. Obviously, the myth of universal
cardioprotection by ␤-blockade, a seed initially planted by the
pharmaceutical industry, is alive and doing well. This is also
documented by the fact that year after year, the ␤-blocker
market has continued to grow, and ⬎100 million prescriptions for ␤-blockers are written every year in the United
States alone, a number that has doubled over the past decade.
Cost-Effectiveness
The compliance of patients with medication depends on
multiple factors, including economic costs of the drugs.
Previously, the guideline committees endorsed thiazides and
␤-blockers as first-line agents given the actual cost of
medication. However, such an approach does not take into
consideration the efficacy and effectiveness of medication.
An ideal cost-effectiveness analysis should consider the
relative effectiveness of different drugs on clinical outcomes,
the direct and indirect costs associated with long-term sequelae with the medications (such as development of diabetes), and the actual cost of the drug itself. Such a formal
cost-effectiveness analysis was performed by the National
Institute for Health and Clinical Excellence. On the basis of
their health-economic model, they concluded that for 65year-old men and women with an annual cardiovascular
disease risk of 2%, heart failure risk of 1%, and diabetes risk
of 1.1%, the most cost-effective initial drug in this group was
a calcium antagonist.45 ␤-Blockers were the least costeffective option. Given the metabolic adverse effects of
␤-blockers, its direct and indirect costs, and the direct and
indirect costs of increased stroke risk in the elderly,
␤-blockers are not cost-effective for this indication.
Outcome Evidence Versus Surrogate
End Point
Glycemia, lipids, and blood pressure are well-known
surrogate end points that correlate reasonably well with
outcome, ie, morbidity and mortality. However, instances
of clear-cut surrogate end point failure have been documented for all 3 of these markers. Thus, the fact that
␤-blockers and diuretics have a negative impact on glycemia and lipids should be interpreted with caution as long as
we have no solid prospective data firmly connecting
surrogate end point with outcome.
Conclusions
Physicians’ Perception: The Myth
of Cardioprotection
In a recent survey, Veterans Affairs physicians were asked
the question, “Which of the following class of drugs has been
proven to reduce mortality in hypertensive patients?”
␤-Blockers, with 78%, received, by far, the highest vote,
followed by ACE inhibitors (65%), diuretics (53%), and
calcium channel blockers (17%).44 This is a sad state of the
art given that only diuretics and calcium channel blockers
have been shown to reduce morbidity and mortality against
Under the heading “Bottom Line,” there was an anonymous
statement in a widely read medical journal: “It can’t get
clearer. Diuretics—the least expensive and most effective
agents—should be the first line treatment for almost
everyone with hypertension, including patients with diabetes.”46 We beg to differ and think that the time has come
to differentiate patients with stage 1 hypertension from
those with a more advanced stage of the disease. Because
of their inefficacy and multiple adverse effects, ␤-blockers
should no longer be used for initial therapy in hypertension.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
2714
Circulation
May 20, 2008
In patients with advanced hypertensive cardiovascular disease, the benefits of lowering blood pressure outweigh the
small metabolic adverse events associated with diuretics and,
in some specific cases, even with ␤-blockers. However, in
uncomplicated hypertension, neither diuretics nor ␤-blockers
are acceptable for first-line treatment. Not quite unexpectedly, Thomas Sydenham’s dictum of primum nil nocere also
applies to first-line antihypertensive therapy.
Disclosures
Dr Messerli reports having served as an ad hoc consultant/speaker
for the following organizations: Abbott, GSK, Novartis, Pfizer,
AstraZeneca, Bayer, Boehringer Ingelheim, BMS, Forest, Sankyo,
Merck, Mars, and Sanofi. Dr Julius reports having served as a
consultant to Novartis and Servier and having received lecture fees
from Novartis and Merck and grant support from AstraZeneca and
Novartis. Dr Bangalore has no disclosures.
References
1. Report of the Joint National Committee on Detection, Evaluation, and
Treatment of High Blood Pressure: a cooperative study. JAMA. 1977;
237:255–261.
2. The 1984 Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med. 1984;
144:1045–1057.
3. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL
Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ. The
Seventh Report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7
report. JAMA. 2003;289:2560 –2572.
4. CDC’s Diabetes Program. National Diabetes Surveillance System: prevalence of diabetes. Available at: http://www.ede.gov/diabetes/statistics/
prev/national/source.htm. Accessed January 8, 2007.
5. Gress TW, Nieto FJ, Shahar E, Wofford MR, Brancati FL. Hypertension
and antihypertensive therapy as risk factors for type 2 diabetes mellitus:
Atherosclerosis Risk in Communities Study. N Engl J Med. 2000;342:
905–912.
6. Conen D, Ridker PM, Mora S, Buring JE, Glynn RJ. Blood pressure and
risk of developing type 2 diabetes mellitus: the Women’s Health Study.
Eur Heart J. 2007;28:2937–2943.
7. Upchurch SL, Brosnan CA, Meininger JC, Wright DE, Campbell JA,
McKay SV, Schreiner B. Characteristics of 98 children and adolescents
diagnosed with type 2 diabetes by their health care provider at initial
presentation. Diabetes Care. 2003;26:2209.
8. Furberg CD, Wright JT, Davis BR, Cutler JA, Alderman M, Black H,
Cushman W, Grimm R, Haywood LJ, Franz Leenen F, Oparil S, Perry
HM, Probstfield J, Whelton P, Payne G, Nwachuku C, Gordon D,
Proschan M, Frommer P, Einhorn P, Hawkins CM, Ford C, Pressel S,
Piller L, Lusk C, Bettencourt J, Kimmel B, Geraci T, Walsh S, Rahman
M, Juratovac A, Pospisil R, Brennan K, Carroll L, Sullivan S, Barone G,
Christian R, Feldman S, Lucente T, Lewis CE, Jenkins K, McDowell P,
Johnson J, Kingry C, Letterer R, Margolis K, Holland L, Jaeger-Fox B,
Williamson J, Louis G, Ragusa P, Williard A, Ferguson RS, Tanner J,
Eckfeldt J, Crow R, Pelos J; ALLHAT Collaborative Research Group.
Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart
Attack Trial (ALLHAT). JAMA. 283:1967–1975.
9. Messerli FH, Weber MA. Long-term cardiovascular consequences of
diuretics vs calcium channel blockers vs angiotensin-converting enzyme
inhibitors. JAMA. 2003;289:2067–2068.
10. Lindholm LH, Persson M, Alaupovic P, Carlberg B, Svensson A, Samuelsson O. Metabolic outcome during 1 year in newly detected hypertensives: results of the Antihypertensive Treatment and Lipid Profile in a
North of Sweden Efficacy Evaluation (ALPINE study). J Hypertens.
2003;21:1563–1574.
11. Bradburn MJ, Deeks JJ, Altman DG. Sbe24: METAN: an alternative
meta-analysis command. Stata Tech Bull Reprints. 1998;8:86 –100.
12. Galbraith RF. A note on graphical presentation of estimated odds ratios
from several clinical trials. StatMed. 1988;7:889 – 894.
13. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
Trials. 1986;7:177–188.
14. Bakris GL, Fonseca V, Katholi RE, McGill JB, Messerli FH, Phillips RA,
Raskin P, Wright JT Jr, Oakes R, Lukas MA, Anderson KM, Bell DS.
Metabolic effects of carvedilol vs metoprolol in patients with type 2
diabetes mellitus and hypertension: a randomized controlled trial. JAMA.
2004;292:2227–2236.
15. Elliott WJ, Meyer PM. Incident diabetes in clinical trials of antihypertensive drugs: a network meta-analysis. Lancet. 2007;369:201–207.
16. Verdecchia P, Reboldi G, Angeli F, Borgioni C, Gattobigio R, Filippucci
L, Norgiolini S, Bracco C, Porcellati C. Adverse prognostic significance
of new diabetes in treated hypertensive subjects. Hypertension. 2004;43:
963–969.
17. Alderman MH, Cohen H, Madhavan S. Diabetes and cardiovascular
events in hypertensive patients. Hypertension. 1999;33:1130 –1134.
18. Davey Smith G, Bracha Y, Svendsen KH, Neaton JD, Haffner SM, Kuller
LH. Incidence of type 2 diabetes in the randomized multiple risk factor
intervention trial. Ann Intern Med. 2005;142:313–322.
19. Almgren T, Wilhelmsen L, Samuelsson O, Himmelmann A, Rosengren
A, Andersson OK. Diabetes in treated hypertension is common and
carries a high cardiovascular risk: results from a 28-year follow-up.
J Hypertens. 2007;25:1311–1317.
20. Aksnes TA, Kjeldsen SE, Rostrup M, Omvik P, Hua TA, Julius S. Impact
of new-onset diabetes mellitus on cardiac outcomes in the Valsartan
Antihypertensive Long-term Use Evaluation (VALUE) trial population.
Hypertension. 2007;50:467– 473.
21. Kostis JB, Wilson AC, Freudenberger RS, Cosgrove NM, Pressel SL,
Davis BR. Long-term effect of diuretic-based therapy on fatal outcomes
in subjects with isolated systolic hypertension with and without diabetes.
Am J Cardiol. 2005;95:29 –35.
22. Pepine CJ, Handberg EM, Cooper-DeHoff RM, Marks RG, Kowey P,
Messerli FH, Mancia G, Cangiano JL, Garcia-Barreto D, Keltai M, Erdine
S, Bristol HA, Kolb HR, Bakris GL, Cohen JD, Parmley WW. A calcium
antagonist vs a non-calcium antagonist hypertension treatment strategy
for patients with coronary artery disease: the International VerapamilTrandolapril Study (INVEST): a randomized controlled trial. JAMA.
2003;290:2805–2816.
23. Cooper-Dehoff R, Cohen JD, Bakris GL, Messerli FH, Erdine S, Hewkin
AC, Kupfer S, Pepine CJ. Predictors of development of diabetes mellitus
in patients with coronary artery disease taking antihypertensive medications (findings from the INternational VErapamil SR-Trandolapril
STudy [INVEST]). Am J Cardiol. 2006;98:890 – 894.
24. Bakris G, Molitch M, Hewkin A, Kipnes M, Sarafidis P, Fakouhi K,
Bacher P, Sowers J. Differences in glucose tolerance between fixed-dose
antihypertensive drug combinations in people with metabolic syndrome.
Diabetes Care. 2006;29:2592–2597.
25. Kasiske BL, Ma JZ, Kalil RS, Louis TA. Effects of antihypertensive
therapy on serum lipids. Ann Intern Med. 1995;122:133–141.
26. Sharma AM, Pischon T, Hardt S, Kunz I, Luft FC. Hypothesis: betaadrenergic receptor blockers and weight gain: a systematic analysis.
Hypertension. 2001;37:250 –254.
27. Messerli FH, Bell DS, Fonseca V, Katholi RE, McGill JB, Phillips RA,
Raskin P, Wright JT Jr, Bangalore S, Holdbrook FK, Lukas MA,
Anderson KM, Bakris GL; GEMINI Investigators. Body weight changes
with beta-blocker use: results from GEMINI. Am J Med. 2007;120:
610 – 615.
28. Scholze J, Grimm E, Herrmann D, Unger T, Kintscher U. Optimal
treatment of obesity-related hypertension: the Hypertension-ObesitySibutramine (HOS) study. Circulation. 2007;115:1991–1998.
29. Lindholm LH, Carlberg B, Samuelsson O. Beta blockers in primary
hypertension: do age and type of beta-blocker matter? J Hypertens.
2006;24:2143–2145.
30. Wiysonge CS, Bradley H, Mayosi BM, Maroney R, Mbewu A, Opie LH,
Volmink J. Beta-blockers for hypertension. Cochrane Database Syst Rev.
2007:CD002003.
31. Messerli FH, Grossman E. Beta-blocker therapy and depression. JAMA.
2002;288:1845–1846.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014
Thiazide-Type Diuretics and ␤-Blockers
Messerli et al
32. Hawkins RG, Houston MC. Is population-wide diuretic use directly
associated with the incidence of end-stage renal disease in the United
States? A hypothesis. Am J Hypertens. 2005;18:744 –749.
33. Reungjui S, Hu H, Mu W, Roncal CA, Croker BP, Patel JM, Nakagawa
T, Srinivas T, Byer K, Simoni J, Wesson D, Sitprija V, Johnson RJ.
Thiazide-induced subtle renal injury not observed in states of equivalent
hypokalemia. Kidney Int. 2007;72:1483–1492.
34. Rovin BH, Hebert LA. Thiazide diuretic monotherapy for hypertension:
diuretic’s dark side just got darker. Kidney Int. 2007;72:1423–1426.
35. Grossman E, Messerli FH, Goldbourt U. Does diuretic therapy increase
the risk of renal cell carcinoma? Am J Cardiol. 1999;83:1090 –1093.
36. Lindholm LH, Carlberg B, Samuelsson O. Should beta blockers remain
first choice in the treatment of primary hypertension? A meta-analysis.
Lancet. 2005;366:1545–1553.
37. Messerli FH, Grossman E, Goldbourt U. Are beta-blockers efficacious as
first-line therapy for hypertension in the elderly? A systematic review.
JAMA. 1998;279:1903–1907.
38. 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.
39. Dahlof B, Sever PS, Poulter NR, Wedel H, Beevers DG, Caulfield M,
Collins R, Kjeldsen SE, Kristinsson A, McInnes GT, Mehlsen J,
Nieminen M, O’Brien E, Ostergren J. Prevention of cardiovascular events
with an antihypertensive regimen of amlodipine adding perindopril as
40.
41.
42.
43.
44.
45.
46.
2715
required versus atenolol adding bendroflumethiazide as required, in the
Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering
Arm (ASCOT-BPLA): a multicentre randomised controlled trial. Lancet.
2005;366:895–906.
Messerli FH, Grossman E, Leonetti G. Antihypertensive therapy and new
onset diabetes. J Hypertens. 2004;22:1845–1847.
Kjekshus JK. Importance of heart rate in determining beta-blocker
efficacy in acute and long-term acute myocardial infarction intervention
trials. Am J Cardiol. 1986;57:43F– 49F.
Bangalore S, Sawhney S, Messerli FH. Heart rate and anti-hypertensive
therapy: the ␤-blocker paradox. J Am Coll Cardiol. In press.
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.
Kaboli PJ, Shivapour DM, Henderson MS, Barnett MJ, Ishani A, Carter
BL. Patient and provider perceptions of hypertension treatment: do they
agree? J Clin Hypertens. 2007;9:416 – 423.
National Collaborating Centre for Chronic Conditions. Hypertension:
Management of Hypertension in Adults in Primary Care: Partial Update.
London, UK: Royal College of Physicians; 2006.
POEM*: diuretics should be first line treatment for hypertension. BMJ.
2003;327:DOI:10.1136/bmj.327.7413.0-e Accessed May 12, 2008.
Response to Messerli et al
Jeffrey A. Cutler, MD, MPH; Barry R. Davis, MD, PhD
The article by Messerli and colleagues argues against first-line diuretic treatment for those with stage I hypertension on the
basis of selective evidence from many nonrandomized studies and analyses. The crux of their case gives undue weight to
small differences in new-onset diabetes among antihypertensive drug treatments. We disagree with this point for the
following reasons: (1) As the authors acknowledge (but fail to incorporate in their conclusions), type 2 diabetes mellitus
is a surrogate outcome and is important primarily because of its status as a risk factor, presumably causal, for
cardiovascular-renal outcomes. Thiazide-type diuretics have an advantage over other drugs as a first-step treatment for
prevention of such outcomes. In ALLHAT, these advantages persisted in subgroups of patients with diabetes or the
metabolic syndrome. (2) The cause of the diabetes epidemic is increasing overweight/obesity and physical inactivity. The
epidemic can be reversed in the population, and diabetes can be prevented (and treated in its early stage) in the individual
by attention to these factors. (3) Most new-onset diabetes in diuretic-treated patients is not diuretic induced. Cases that are
can likely be prevented by avoiding significant potassium depletion. There are no direct data on diuretic-induced diabetes
and cardiovascular risk. The authors purport to refute JNC7, a broadly based set of consensus, evidence-based guidelines.
The process for generating a JNC8 is getting under way, and we are confident that the best evidence will prevail. We have
no major disagreement with these authors regarding a secondary role for ␤-blockers, despite the deficiencies in evidence
regarding non-atenolol ␤-blockers.
Downloaded from http://circ.ahajournals.org/ by guest on September 9, 2014