Beruflich Dokumente
Kultur Dokumente
net/publication/13650850
CITATIONS READS
500 286
3 authors, including:
Some of the authors of this publication are also working on these related projects:
Predictors of 3016 cancer incidence cases in 10,000 men over 48 years View project
All content following this page was uploaded by Uri Goldbourt on 17 May 2014.
JAMA, June 17, 1998—Vol 279, No. 23 b-Blockers for Hypertension in the Elderly—Messerli et al 1903
©1998 American Medical Association. All rights reserved.
Downloaded from www.jama.com at YOKOSUKA HOSPITAL UWAMACHI, on December 25, 2005
Table 1.—Studies Included in the Meta-analysis RESULTS
No. of No. of Description of the Trials
Study Age, y Patients Controls Type of Drug
Taken together, the 10 trials included
Studies of Diuretics in the Elderly
a total of 16 164 patients, of whom 8217
Veterans Administrative Cooperative on .60 38 43 Hydrochlorothiazide
4
Antihypertensive Agents, 1972 received active treatment, followed up
Kuramoto et al,5 1981 .60 44 47 Thiazide for an average of approximately 5 years.
National Heart Foundation of Australia,6 1981 .60 293 289 Thiazide The characteristics of the trials are pre-
European Working Party on High Blood .60 416 424 Hydrochlorothiazide sented in Table 1. Seven trials used a
Pressure in the Elderly,7 1985 and triamterene diuretic as a first drug of choice.4-8,10,11
Hypertension Detection and Follow-up 60-69 1204 1172 Chorthalidone The Medical Research Council (MRC)
Program,8 1985
trial was a 3-arm trial comparing hydro-
Systolic Hypertension in the Elderly .60 443 108 Chorthalidone
Program Pilot,10 1989
chlorothiazide and amiloride hydrochlo-
Systolic Hypertension in the Elderly .60 2365 2371 Chorthalidone
ride or b-blocker with placebo.13 Taken
Program,11 1991 together, the diuretic trials include 5884
Medical Research Council Working 65-74 1081 2213 Hydrochlorothiazide patients in the active-treatment arm.
Party,13 1992 and amiloride hydrochloride Another study9 used a b-blocker as a
Total 5884 6667 first drug of choice and was therefore
Studies of b-Blockers in the Elderly included in our analysis. Combining this
Coope et al,9 1986 60-79 419 465 Atenolol study with the b-blocker arm of the MRC
Medical Research Council Working 65-74 1102 2213 Atenolol trial gives a total of 1521 patients in the
Party,13 1992 active treatment arm of b-blocker trials.
Total 1521 2678 The Swedish Trial in Old Patients With
Other Hypertension (STOP) study used either
Swedish Trial in Old Patients,12 1991 70-84 812* 815 b-Blockers or b-blocker (67%) or diuretic (33%) as
hydrochlorothiazide a first drug of choice.12 Only the blood
and amiloride hydrochloride
pressure responses were reported sepa-
*Sixty-seven percent of patients received b-blockers and 33% received hydrochlorothiazide and amiloride rately for the different treatment regi-
hydrochloride. mens, and these results were included in
the analysis.17 However, the results for
Table 2.—Response Rate to Antihypertensive Treatment in Elderly Patients With Hypertension the morbidity and mortality were not
No. of Response broken down by different treatment
Study Patients First Drug Rate, % regimens, and therefore they could not
Diuretics be included in our analysis.
Kuramoto et al,5 1981 44 Thiazide 79 Seven trials included only patients
European Working Party on High Blood 416 Hydrochlorothiazide 65 with diastolic hypertension.4-9,12 Two tri-
Pressure in the Elderly,7 1985 and triamterene
als, the Systolic Hypertension in the El-
Systolic Hypertension in the Elderly 443 Chlorthalidone 88
Program Pilot,10 1989
derly Program (SHEP) pilot and the
Systolic Hypertension in the Elderly 2365 Chlorthalidone 46 subsequent larger trial, were limited to
Program,11 1991 subjects with isolated systolic hyperten-
Swedish Trial in Old Patients,17 1991 246 Hydrochlorothiazide and 60 sion.10,11 The MRC trial included either
amiloride hydrochloride patients with moderate-to-severe iso-
Medical Research Council Working 1081 Hydrochlorothiazide and 62 lated systolic hypertension or patients
Party,13 1992 amiloride hydrochloride
with combined systolic and diastolic hy-
b-Blockers pertension.13
Coope et al,9 1986 419 Atenolol 33
Swedish Trial in Old Patients,17 1991 219 Metoprolol 22 Response Rate in Patients Treated
Swedish Trial in Old Patients,17 1991 180 Atenolol 32 With b-Blockers and Diuretics
Swedish Trial in Old Patients,17 1991 120 Pindolol 28
In several studies the response rate to
Medical Research Council Working Party,13 1992 1102 Atenolol 48
the first drug of choice was reported
(Table 2).5,7,9-11,13,17 Among 4595 patients
tive heart failure. In some studies, part of the computation of the CI are given by who received a diuretic as a first drug,
of the information was not assessed or Mehta et al.16 We pooled the estimates of about 66% were well controlled on mono-
reported. the odds ratios (ORs) of cerebrovascular therapy, and the remaining third re-
Quantitative analyses of outcomes events, stroke mortality, CHD, cardio- quired an additional agent. In contrast,
were based on intention-to-treat results. vascular mortality, and all-cause mortal- among 2040 patients who received b-
The Programs for Epidemiologic Analy- ity over studies to provide a pooled OR blocker as a first drug, less than a third
sis program CASECONT14 was used to and a CI for each end point, and we tested were controlled on monotherapy and
combine measures of associations from for heterogeneity between individual about two thirds required a diuretic as a
the different studies. The procedure study estimates. The methods used were supplement.
used computes x2 for each table (study) described in detail by Fleiss.15 We have
and computes the Cochran-Mantel- opted to apply the DerSimonian-Laird Reduction of Risk in Studies
Haenszel pooled x2 statistic for test of procedure for random-effects model With b-Blockers and Diuretics
association, with and without Yates cor- when the test for heterogeneity of OR Both treatment regimens reduced the
rection for continuity, its associated 90%, between studies was significant at the incidence of cerebrovascular events
95%, and 99% confidence intervals (CIs), .10 level. However, this did not turn out (Figure 1). Diuretic treatment reduced
and a test for heterogeneity based on the to be the case for any of the end points the odds for cerebrovascular events by
Cornfield-Gart procedure.15 The details considered. 39% (OR, 0.61; 95% CI, 0.51-0.72), and
1904 JAMA, June 17, 1998—Vol 279, No. 23 b-Blockers for Hypertension in the Elderly—Messerli et al
JAMA, June 17, 1998—Vol 279, No. 23 b-Blockers for Hypertension in the Elderly—Messerli et al 1905
©1998 American Medical Association. All rights reserved.
Downloaded from www.jama.com at YOKOSUKA HOSPITAL UWAMACHI, on December 25, 2005
Table 3.—Possible Reasons for Diminished Efficacy of b-Blockade in the Treatment of Hypertension in Elderly Patients
the speculation that either the diuretic mittee on Prevention, Detection, Evalu- blockers produce a hemodynamic effect
confers a specific benefit irrespective of ation, and Treatment of High Blood exactly opposite to that desired in an el-
the decrease in arterial pressure or, Pressure, which has changed its previ- derly patient. By shifting the hemody-
more concerning, that the b-blocker con- ous recommendation for treatment of namic profile from a normal cardiac out-
fers an ill effect on the cardiovascular the elderly by now stating, “When com- put, high vascular resistance pattern to a
system in the elderly that overrides the pared to each other, diuretics are supe- low cardiac output, high vascular resis-
beneficial effect of a decrease in arterial rior to the b-blocker atenolol.”1 tance pattern, b-blockers accelerate or en-
pressure. Nevertheless, some indirect evidence hance hemodynamic changes patients with
In all other studies in the elderly popu- suggests that b-blockers may have some hypertension experience as they age.27,28
lation in which b-blockers used to treat benefits in treating hypertension in b-Blockers have been used for the
hypertension were implied to reduce middle-aged and younger patients. In all treatment of hypertension for more
morbidity and mortality, they were used 3 trials (MRC, International Prospective than 3 decades.19 Despite their well-
in combination with a diuretic. Thus, in Primary Prevention Study in Hyperten- documented potential for lowering mil-
the STOP trial17 more than 70% of the sion, and Heart Attack Primary Preven- limeters of mercury, no study has shown
patients assigned to b-blockers were re- tion in Hypertension),23-25 the rate of that b-blockers, either alone or when
ceiving diuretics, and no information was myocardial infarction, stroke, and car- added to diuretic therapy, indepen-
available regarding the effects of a b- diovascular death with a diuretic was dently diminish CHD morbidity or car-
blocker as a first-line therapy on mor- similar to that with a b-blocker regimen. diovascular mortality and all-cause mor-
bidity and mortality. The study of Coope A meta-analysis analyzing the 3 studies tality when used to treat hypertension
and Warrender9 demonstrated a signifi- showed a trend toward a decrease in to- in elderly patients. Quite to the con-
cant reduction in the rate of strokes and tal cardiovascular mortality in men by trary, the present analysis shows few, if
was included as a b-blocker study in our 14% and an increase in women by 16% in any, benefits of b-blocker therapy when
analysis. However, whereas 70% of pa- the b-blocker group when compared compared with diuretic therapy. In this
tients in the treatment group were re- with non–b-blocker treatment.26 context it must be remembered that blood
ceiving atenolol, 60% were receiving Several points may possibly account for pressure is a surrogate end point that of-
bendroflumethiazide; the outcome data the inefficacy of b-blockers in reducing ten, but not always, correlates with real
were never reported separately.9 In the morbidity and mortality in the elderly end points, such as heart attacks, strokes,
SHEP study,11 only 32% of patients were hypertensive patient (Table 3). The most and sudden death. The reason for the in-
receiving atenolol (or reserpine), almost important of these points is perhaps the efficacy of b-blockers may lie in their in-
all of these in combination with a di- hemodynamic mismatch caused by b- herent unfavorable effect on the sys-
uretic. A recent subanalysis of SHEP by blockade in the elderly. The hemody- temic hemodynamics of elderly patients
Kostis et al21 did not identify any ben- namic profile of hypertension in the el- and on pathophysiologic findings in the
efits attributable to atenolol (or reser- derly is characterized by a low cardiac arterial tree, the heart, the kidneys, and
pine) per se that were independent of, or output and a high peripheral resis- the brain and to a lesser extent on the me-
in addition to, the ones conferred by the tance.27,28 Most b-blockers (with the ex- tabolism of lipids and carbohydrates.
diuretic. None of these studies allows us ception of few vasodilating b-blockers) Thus, although they have been shown to
to conclude that either the b-blocker lower arterial pressure by further de- be beneficial in patients after myocar-
alone or the addition of the b-blocker to creasing cardiac output and increasing sys- dial infarction,3 b-blockers appear to
the diuretic antihypertensive regimen temic vascular resistance.29 A review of expose the elderly patient with uncom-
significantly and independently reduced 85 studies on 10 different b-blockers plicated hypertension to the adverse ef-
morbidity and mortality. Conceivably, showed an increase in peripheral resis- fects of b-blockade while conferring few,
all benefits observed could be due to di- tance and a decrease in cardiac output with if any, true benefits. This present study
uretic therapy alone. The fact that b- short-term treatment, whereas with long- reinforces the recommendation of the
blockers are less appropriate first-line term treatment, cardiac output re- Joint National Committee VI that, in con-
therapy than diuretics in the elderly was mained depressed, although total periph- trast to diuretics, b-blockers are not ap-
emphasized by the Working Party on eral resistance fell somewhat but remained propriate first-line therapy of uncompli-
Hypertension in the Elderly22 and by the distinctly above normal levels.30 Thus, cated essential hypertension in the
sixth report of the Joint National Com- while lowering arterial pressure, b- elderly.
1906 JAMA, June 17, 1998—Vol 279, No. 23 b-Blockers for Hypertension in the Elderly—Messerli et al
JAMA, June 17, 1998—Vol 279, No. 23 b-Blockers for Hypertension in the Elderly—Messerli et al 1907
©1998 American Medical Association. All rights reserved.
Downloaded from www.jama.com at YOKOSUKA HOSPITAL UWAMACHI, on December 25, 2005
View publication stats