| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Hypertension. 2007;49:1242.)
© 2007 American Heart Association, Inc.
Original Articles |
From the Baker Heart Research Institute (A.M.D., Y-L.L., K.L.B., G.L.R.J., C.I.J., B.A.K.), Melbourne, Victoria, Australia; the University of Melbourne (C.D.G.), Melbourne, Victoria, Australia; the Second Australian National Blood Pressure Study Group (K.W., L.M.H.W., CM.R., P.R., L.J.B., G.L.R.J., C.I.J., J.J.M., G.J.M., T.O.M., M.J.W.), Melbourne, Victoria, Australia; and LaTrobe University (J.D.C.), Melbourne, Victoria, Australia.
Correspondence to Anthony M. Dart, Baker Heart Research Institute, PO Box 6492, St Kilda Rd Central, Melbourne, Victoria 8008, Australia. E-mail a.dart{at}alfred.org.au
| Abstract |
|---|
|
|
|---|
Key Words: central blood pressure brachial blood pressure ACE inhibition diuretic clinical science antihypertensive drugs elderly
| Introduction |
|---|
|
|
|---|
2 treatment regimens.13 In almost all instances, the blood pressure response was determined from peripheral arterial sites, generally using brachial artery sphygmomanometry (eg, References 4 and 5). The pathophysiological consequences of elevated arterial pressure include left ventricular hypertrophy, coronary artery disease and its complications, cerebrovascular disease, renal disease, and aortic and peripheral artery disease. In several, though not all, of these sites of end organ damage, the prevailing pressure will more closely follow proximal aortic or carotid pressures rather than brachial pressures. However, the acquisition of such pressures has not generally been available without invasive cannulation. In recent years, techniques have been developed to allow estimates of central waveforms and pressures to be obtained without the need for arterial cannulation. Furthermore, there is emerging evidence that central aortic pressures may be independent predictors of end organ damage and clinical outcomes.69 Such techniques have used noninvasive applanation tonometry to record arterial waveforms, which are then calibrated from brachial pressures measured indirectly by conventional noninvasive means. In some instances, applanation of the radial artery is performed and the waveform manipulated by a transfer function to derive a central pressure10; in others, applanation of the carotid artery is performed directly8,9,11; in both methods, equivalence of mean and diastolic pressures at carotid and brachial sites is assumed.
Such noninvasive estimates of central pressures have been performed in subsets of 2 large outcome trials in hypertensive subjects, the Conduit Artery Function Evaluation (CAFE) substudy of the Anglo-Scandinavian Cardiovascular Outcome Trial (ASCOT)10 and the Second Australian National Blood Pressure study (ANBP2).11 The main finding of ANBP2 was that initiation of antihypertensive treatment involving angiotensin-converting enzyme inhibitors (ACE-Is) in older subjects, particularly men, appears to lead to better treatment outcomes than treatment with diuretic agents, despite similar reductions in brachial blood pressure.12 We report here on ANBP211,12 and, in particular, whether the treatment regimens had different effects on central as opposed to brachial pressures. In addition, we examined whether such differential central and peripheral effects could underlie the better outcomes with ACE-Ibased treatment in men compared with women observed in ANBP2 despite similar reductions in brachial blood pressure in both sexes.12
| Methods |
|---|
|
|
|---|
Arterial waveforms were obtained with participants recumbent in a quiet air-conditioned room (Figure). After
10 minutes of rest, a carotid pressure waveform was obtained by applanation tonometry of the proximal right carotid artery using a pencil type transducer (Micro-tip SPT-301, Millar Instruments; Figure). Brachial blood pressure was measured with a Dinamap 1846 SXP (Critikon) in triplicate, and the mean of these values was used for subsequent calculations.13 From the waveform recordings,
3 and a maximum of 10 representative cardiac cycles were selected by a single operator (K.L.B.), and the parameters obtained from each waveform were averaged. Waveforms were linearly detrended assuming equality of the start and end of each cardiac cycle and scaled to brachial diastolic and mean blood pressure. Other than selection of representative waveforms, all of the analysis was fully automated. Measurements were made before random assignment and after an average of 4 years of treatment.
|
Power calculations based on experience with the same methodologies indicated group sizes of
230 to 250 would be required to detect a 5-mm Hg difference in systolic blood pressure with 1-ß of 90% (which would also detect a 4-mm Hg difference in pulse pressure). Repeatability in 40 subjects with identical methods over a 3-week period showed a mean difference in carotid systolic pressure of <1 mm Hg with a 95% CI of 5 to +4 mm Hg (coefficient of variation: 6.6%).14 Subjects in the present study were recruited from the 853 subjects in whom technically satisfactory carotid artery waveforms (and aortic ultrasound) were obtained at their baseline visit.15 These 853 subjects with complete and satisfactory recordings composed 77% of the 1153 subjects in which arterial waveform and aortic ultrasound measurements were attempted. Mortality rates in this group were similar to the overall study,12 in which they did not differ between treatment groups. Based on the power calculations above, participants in 2 recruiting centers were invited to participate in the follow-up study. Of those invited to reattend for the follow-up measurements, only 3% were unable to accept because of personal ill health, and satisfactory carotid waveforms were obtained in all of those who returned for evaluation. Data were available from 479 subjects for the follow-up analysis.
All of the group data are presented as mean±SEM except where otherwise indicated after list-wise deletion of case subjects with missing data. Comparison of groups was made using paired and unpaired t tests as appropriate.
| Results |
|---|
|
|
|---|
|
|
Neither brachial nor central pressures differed between the groups before random assignment (Table 2). After approximately 4 years of treatment, brachial systolic and pulse pressures were equivalent in the groups (Table 3). Similarly carotid systolic and pulse pressures were not different between the groups. Heart rate after 4 years in the trial was also similar in both groups. Changes in systolic and pulse pressure with treatment were almost identical at brachial and central sites (Table 3). None of the differences between the diuretic and ACE-I groups were significant.
|
Blood pressure responses were also analyzed separately for men and women (Tables 2 and 4
) in view of the results from the main trial suggesting that men and women may have responded differently. For both men and women, central systolic and pulse pressures, measured after
4 years of treatment, were virtually identical between those randomly assigned to ACE-I or diuretic treatment (Table 4). In men, the changes in pressure from baseline were slightly greater peripherally than centrally, but differences were not significant (P>0.3 for both systolic and pulse pressure; Table 4). Females randomly assigned to ACE-I showed somewhat bigger absolute falls centrally than peripherally (P>0.05 and P<0.1 for both systolic and pulse pressure), whereas the converse was seen in those randomly assigned to diuretic treatment (P>0.3 for both systolic and pulse pressure). The differences between treatments for women was not significant for either systolic or pulse pressure (P=0.06 and P=0.08, respectively). When analyzed as percentage of change from baseline, carotid pulse pressure fell more on ACE-I than diuretic in both men and women, but differences between treatments were also not significant.
|
When patients receiving ß-blockers were excluded from the analysis, central systolic and pulse pressures at the end of the study were 144±2, 145±2, 71±2, and 72±2 mm Hg, respectively, for ACE-I and diuretic groups. None of these (nor their respective
and %
changes, data not shown) were significantly different between ACE-I and diuretic groups. The respective values for those not receiving ß-blockers and/or calcium channel blockers were 146±2, 144±2, 72±2, and 71±2 mm Hg.
| Discussion |
|---|
|
|
|---|
The slight benefit observed with angiotensin-converting enzyme inhibition over diuretic-based therapy in the main study may have been because of factors other than blood pressure lowering, such as vessel wallspecific actions,16,17 better efficacy in reducing left ventricular hypertrophy,1820 or relatively deleterious effects of diuretics on glucose intolerance and hyperuricemia.21 Another possibility that has been raised in relation to previous trials, such as the Heart Outcomes Prevention Evaluation Study,16 is that, although the brachial pressures on treatment were similar, there could have been a difference in central (ie, proximal aortic) pressures that was not apparent from the in-trial measures of blood pressure obtained at the brachial artery. The current study demonstrates that this is unlikely to be the explanation of the outcome differences observed in ANBP2, where an ACE-Ibased regimen was compared with a diuretic-based regimen.
Central pressure measurements in the current study were available before random assignment and after
4 years in trial treatment. It is, therefore, possible that the findings after 4 years were distorted by differential event rates in the 2 treatment arms over the 4-year period. However, this is unlikely to be a major factor, because there was no significant difference in mortality between treatment arms in ANBP2 in which differences between treatments related to combined fatal and nonfatal end points.12 Furthermore, the baseline characteristics of the substudy cohort were very similar to those of the main study except for a slight excess in the percentage of female participants. The study was designed to detect a difference of
5 mm Hg in systolic blood pressure (and 4 mm Hg in pulse pressure), similar to that in the recently published CAFE substudy of ASCOT.10 It is possible that a smaller difference between treatments could be clinically significant, however, there was no trend to suggest any such difference. Furthermore, when men and women were analyzed separately, the differences were the opposite of those that would account for the outcome differences in the parent study. Thus, absolute pressures fell more centrally than peripherally in women on ACE-I with the opposite occurring in men, whereas clinical end point differences between treatments were evident in men but not women.
The only published outcome trial to date in which central pressures were obtained was ASCOT, in which an amlodipine±perindopril regimen was compared with an atenolol±thiazide regimen.10 In the ASCOT substudy (CAFE), there was a treatment difference in central pulse pressure but not in brachial pulse pressure. This trial differed in a number of ways from the current study. Importantly, recruitment into CAFE commenced 1 year after random assignment into ASCOT. Because CAFE did not acquire any pretreatment values, the fall in pressure with treatment is not known. The methodology for central pressure determination used radial tonometry and a generalized transfer function to estimate a central aortic pressure. However, in the present study a generalized transfer function was not required, because the central pressure waveform was measured by tonometry at the carotid artery. Finally, the ANBP2 cohort were approximately 1 decade older than the CAFE group, which likely contributes to the equivalence at baseline of central and brachial pressures in all of the groups in ANBP2 (including in women randomly assigned to ACE-I in whom mean values were slightly higher centrally than peripherally). The similarity between peripheral and central arterial pressures persisted even on treatment, indicating that a "passive" increase in arterial distensibility (secondary to a fall in blood pressure) did not unmask centralperipheral pressure differences Peripheral pulse pressure amplification shows a strong negative relationship with age, even in a normotensive, disease-free, nonmedicated population.22
In addition to the CAFE Study, there have been a number of previous studies reporting lower central pressure and/or lower central pressure augmentation with angiotensin-converting enzyme inhibition versus predominantly ß-blocker based regimens.2326 In these studies, heart rate was lower (by 7 to 13 bpm) after the ß-blocker regimen than with angiotensin-converting enzyme inhibition. Interestingly, similar effects have also been demonstrated acutely with single-dose atenolol compared with ramipril.27 Such data highlight the importance of heart rate as a determinant of central wave reflection, augmentation index, and, therefore, central blood pressure.2831 Indeed, lower heart rate is associated with higher augmentation index and, therefore, central pressure.28,31 The greater effects of ACE-I compared with ß-blockerbased regimens on central pressure may, therefore, relate to the fact that ACE-Is do not lower heart rate. In support of this contention, when heart rate was considered as a covariate, the difference in augmentation index between an angiotensin receptor blocker and atenolol intervention was abolished.32 In CAFE there was a mean difference in heart rate of 11 bpm at study end between the treatment groups, attributable to the differential use of atenolol. In the current study, there was no significant difference in heart rate between treatment groups for either gender. Thus, there was a marked difference in heart rate responses to treatment between CAFE and ANBP2, which likely contributed to the different findings on central blood pressure. In conclusion, ACE-Is and diuretic-based regimens had equivalent effects on both central and brachial blood pressure in these elderly hypertensive subjects.
Perspectives
Current methodology permits an assessment of central arterial pressures noninvasively either by direct applanation of the carotid artery or by using a transfer function applied to a peripheral (usually radial) arterial waveform. Marked differences in systolic (and, therefore, also pulse) pressures may exist between central and peripheral sites, particularly in younger subjects. It is therefore possible that treatment effects estimated by recording peripheral (usually brachial) pressures may not be indicative of effects on central pressures. Furthermore, it is therefore possible that different blood pressurelowering treatments may result in similar peripheral (brachial) pressures but different central arterial pressures. The current study suggests that this is less likely in older hypertensive subjects in whom differences between central and peripheral pressures are much less than in young subjects and in whom this approximate equivalence of central and peripheral pressures persists on treatment. It will therefore probably be more useful to target somewhat younger patient groups in future studies to determine the importance of estimating central pressures in the management of subjects with hypertension. Such studies will also need to evaluate the effects of different drug classes, including ß-adrenoceptor antagonists, on central and peripheral pressures.
| Acknowledgments |
|---|
This study was supported by the Australian Commonwealth Department of Health and Aging, the National Health and Medical Research Council of Australia, and Merck Sharp and Dohme Pty Ltd, Australia.
Disclosures
None.
Received December 7, 2006; first decision December 29, 2006; accepted March 31, 2007.
| References |
|---|
|
|
|---|
Related Article:
This article has been cited by other articles:
![]() |
C. M. McEniery, Yasmin, B. McDonnell, M. Munnery, S. M. Wallace, C. V. Rowe, J. R. Cockcroft, I. B. Wilkinson, and on Behalf of the Anglo-Cardiff Collaborative Trial Central Pressure: Variability and Impact of Cardiovascular Risk Factors: The Anglo-Cardiff Collaborative Trial II Hypertension, June 1, 2008; 51(6): 1476 - 1482. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Shimizu and K. Kario Review: Role of the augmentation index in hypertension Therapeutic Advances in Cardiovascular Disease, February 1, 2008; 2(1): 25 - 35. [Abstract] [PDF] |
||||
![]() |
M. F. O'Rourke, W. W. Nichols, and M. E. Safar Effects of Angiotensin-Converting Enzyme Inhibitors on Central Blood Pressure Hypertension, September 1, 2007; 50(3): e63 - e63. [Full Text] [PDF] |
||||
![]() |
A. V. Chobanian Isolated Systolic Hypertension in the Elderly N. Engl. J. Med., August 23, 2007; 357(8): 789 - 796. [Full Text] [PDF] |
||||
![]() |
V. Palmieri, R. Pini, and M. Chiara Cavallini Pulsology Reloaded: Commentary on Similar Effects of Treatment on Central and Brachial Blood Pressure in Older Hypertensive Subjects Hypertension, June 1, 2007; 49(6): 1210 - 1212. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hypertension Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2007 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |