Is cardiovascular remodeling in patients with essential hypertension related to more than high blood pressure? A LIFE substudy

Is cardiovascular remodeling in patients with essential hypertension related to more than high blood pressure? A LIFE substudy Michael Hecht Olsen, MD,a,b Kristian Wachtell, MD,b Kirstine L. Hermann, MD,c Erik Frandsen, MSc,a Harriet Dige-Petersen, MD,a Jens Rokkedal, MD,b Richard B. Devereux, MD,d and Hans Ibsen, MDb Copenhagen, Denmark, and New York, NY

Background Blocking the renin–aldosterone-angiotensin II system has been hypothesized to induce blood pressuredependent as well as blood pressure-independent regression of cardiovascular hypertrophy. However, the relative influence of elevated blood pressure (BP) and various neurohormonal factors on cardiovascular remodeling in hypertension is unclear.

Methods In 43 untreated patients with hypertension with electrocardiographic left ventricular hypertrophy, we measured relative wall thickness and left ventricular mass index by echocardiography and by magnetic resonance imaging (n ⫽ 32), intima-media cross-sectional area, and distensibility of the common carotid arteries by ultrasound, media/lumen ratio of isolated subcutaneous resistance arteries by myography, and median 24-hour systolic BP (n ⫽ 40), serum insulin, and plasma levels of epinephrine, norepinephrine, renin, angiotensin II, aldosterone, and endothelin.

Results In multiple regression analyses, left ventricular mass index by echocardiography (R2 ⫽ 0.14, P ⬍ .05) and by magnetic resonance imaging (R2 ⫽ 0.32, P ⫽ .001) were associated with 24-hour systolic BP, whereas relative wall thickness was associated with plasma epinephrine (R2 ⫽ 0.12, P ⬍ .05) and aldosterone (R2 ⫽ 0.10, P ⬍ .05). Intimamedia cross-sectional area/height was associated with 24-hour systolic BP (␤ ⫽ 0.40) and plasma epinephrine (␤ ⫽ 0.43) (adjusted R2 ⫽ 0.32, P ⬍ .001), whereas carotid distensibility was associated with 24-hour systolic BP (␤ ⫽ 0.40) and plasma angiotensin II (␤ ⫽ ⫺0.41) (adjusted R2 ⫽ 0.30, P ⬍ .001). Media/lumen ratio in subcutaneous resistance arteries was associated with plasma epinephrine (R2 ⫽ 0.22, P ⬍ .01).

Conclusion Apart from being associated with a high BP burden, cardiovascular remodeling was associated with high levels of circulating epinephrine, aldosterone, as well as angiotensin II, suggesting a beneficial effect above and beyond the effect of BP reduction when using antihypertensive agents blocking the receptors of these neurohormonal factors. (Am Heart J 2002;144:530-7.)

Patients with essential hypertension often have cardiovascular changes, the severity of which do not always relate to the blood pressure (BP), suggesting that neurohormonal factors may contribute to the changes.1 Cardiac myocyte growth is believed to be primar-

From the aDepartments of Clinical Physiology and Nuclear Medicine, and bInternal Medicine, Glostrup Hospital, University of Copenhagen, Glostrup, the cDepartment of Magnetic Resonance Images, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark, and the dDepartment of Medicine, Department of Echocardiography, Weill Medical College of Cornell University, New York, NY. Supported in part by grants from The Danish Medical Association Research Fund, Copenhagen, The Becket-Foundation, Copenhagen, Denmark, and Merck and Co, Inc, West Point, Pa, as part of the LIFE-ICARUS substudy. Submitted September 5, 2001; accepted March 12, 2002. Reprint requests: Michael Hecht Olsen, MD, Department of Clinical Physiology and Nuclear Medicine, Glostrup Hospital, University of Copenhagen, DK-2600 Glostrup, Denmark. E-mail: [email protected] © 2002, Mosby, Inc. All rights reserved. 0002-8703/2002/$35.00 ⫹ 0 4/1/124863 doi:10.1067/mhj.2002.124863

ily governed by loading conditions, whereas smooth muscle cell and fibroblast proliferation is primarily regulated by neurohumoral mechanisms.2 The importance of the renin-angiotensin system and the sympathetic nervous system in the development of vascular hypertrophy and remodeling has been discussed for many years.3 In addition, associations between left ventricular (LV) hypertrophy and peripheral vascular remodeling,4 as well as conduit artery hypertrophy5 and stiffness6 have been documented. Some data indicate that reversal of vascular changes can add to regression of LV hypertrophy in conjunction with BP reduction,7 but the possibility of any causality in these associations is still unclear. Treatment with low-dose angiotensinconverting enzyme inhibitors has, in rats, been demonstrated to induce regression of vascular remodeling without reduction in blood pressure,8 supporting the hypothesis that blocking the renin–aldosterone-angiotensin II system in humans may have beneficial effects

American Heart Journal Volume 144, Number 3

Olsen et al 531

Table I. Relationship between blood pressure, circulating hormones, and cardiovascular remodeling

LVMIecho RWT LVMIMRI MFVR Common carotid arteries IMT IMA/height IMA/BSA Distensibility Resistance arteries in vitro Media/lumen ratio Distensibility

24-hour SBP

Epinephrine

Norepinephrine

Angiotensin II

0.37* 0.14 0.57† 0.40*

0.02 0.35* 0.07 –0.23

0.08 –0.01 0.01 –0.14

–0.31* 0.16 –0.17 –0.27

0.37* 0.40* 0.43† –0.42†

0.33* 0.41† 0.43† –0.05

0.18 0.38* 0.39* –0.12*

0.04 –0.02 –0.00 –0.42†

0.20 –0.06

0.48† 0.06

0.12 –0.10

–0.05 0.33*

SBP, Systolic blood pressure; LVMIecho, left ventricular mass index assessed by echocardiography; LVMIMRI, left ventricular mass index assessed by magnetic resonance imaging; RWTecho, relative wall thickness assessed by echocardiography; IMT, intima media thickness of the common carotid arteries; IMA, intima media cross-sectional area of the common carotid arteries indexed by height; IMA/BSA, intima media cross-sectional area of the common carotid arteries indexed by body surface area; MFVR, minimal forearm vascular resistance. *P ⬍ .05. † P ⬍ .01.

on regression of cardiovascular remodeling above and beyond the effects of blood pressure reduction. This study was undertaken to investigate the association of increased BP and various neurohormonal factors to cardiovascular hypertrophy and remodeling in patients with hypertension and electrocardiographic LV hypertrophy to sort out the relative importance of hemodynamic burden versus neurohormonal factors on cardiovascular hypertrophy and remodeling. Furthermore, we wanted to describe the relationship between LV hypertrophy and vascular hypertrophy, remodeling, and stiffness. LV mass was assessed by echocardiography, as well as by magnetic resonance imaging— combining the most clinically used method with the more accurate method.9

Methods Subjects After 2 to 3 weeks of placebo treatment, we studied 43 patients (8 females, 35 males), mean age 65 years, ranging from 56 to 77, with essential hypertension, LV hypertrophy as determined by screening electrocardiogram and LV ejection fraction ⬎40% on echocardiography before study entry. All patients were participants in the Insulin CARotids United States Scandinavia (ICARUS) substudy10 of the Losartan Intervention For Endpoint-Reduction in Hypertension (LIFE) trial, and met the study’s inclusion and exclusion criteria.11 Body mass index was 28.4 (27.6-29.3) kg/m2. BP was 174 (170179) mm Hg/95 (92-98) mm Hg. LV mass assessed by echocardiography (LV massecho) was 243 (224-261) g and LV mass index assessed by echocardiography (LVMIecho) was 123 (115-132) g/m2. There were 11 smokers, 17 exsmokers and 15 nonsmokers. Ten patients had newly diagnosed hypertension, of whom 6 had never received antihypertensive treatment. The mean duration of known hypertension was 6

years and 11 months, ranging from 0 to 35 years. Before the wash-out placebo period, 26 patients were receiving monotherapy with either calcium antagonists (13 patients), angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers (6 patients), ␤-receptor blockers (4 patients), or diuretics (3 patients). The remaining 11 had received combination therapy with either calcium antagonists and angiotensinconverting enzyme inhibitors/angiotensin II receptor blockers (5 patients), diuretics/␤-receptor antagonists and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers (4 patients), or diuretics and calcium antagonists (2 patients). Glucose intolerance was found in 31 patients, 15 of whom had overt diabetes according to World Health Organization criteria.12

Protocol and methods Echocardiography. An echocardiogram was performed after a previously used protocol.13 Standardized examinations included 2-dimensional-guided M-mode echocardiograms and selected 2-dimensional recordings. Studies were performed using a high-quality commercially available echocardiograph (VingMed CFM-800, VingMed, Sound, Norway) equipped with 3.0 to 3.5 MHZ and 2.0 to 2.5 MHZ probes and superVHS video recorder. Measurements were made blindly at the Echocardiography Reading Center at the New York Presbyterian Hospital-Weill Medical College of Cornell University in New York City by experienced physician readers using computerized review stations (Digisonics, Inc, Houston, Tex) equipped with digitizing tablet and monitor screen overlay for calibration and performance of each needed measurement. LV dimensions were measured according to recommendations of the American Society of Echocardiography. When optimal orientation of the LV or atrial imaging views could not be obtained, correctly oriented 2-dimensional linear dimension measurements were made by the leading-edge convention. LV end-diastolic diameter and wall thickness

American Heart Journal September 2002

532 Olsen et al

Table II. Relationship between cardiac and vascular hypertrophy and remodeling LVMIMRI No. MFVR (mm Hg ⫻ min/mL%) Common carotid arteries IMT (mm) IMA/height (mm2/m) IMA/BSA (mm2/m2) Distensibility (% mm Hg) Resistance arteries in vitro Media/lumen ratio (%) Distensibility (%/kPa)

LVMIecho r

No.

r

25

0.54†

30

0.50†

32 32 32 31

0.21 0.47† 0.45* ⫺0.24

42 42 42 42

0.36* 0.48† 0.41† ⫺0.05

31 30

0.08 ⫺0.42*

40 38

0.01 ⫺0.26

*P ⬍ .05. † P ⬍ .01.

were used to calculate relative wall thickness and LV mass by a formula that gives values closely related to autopsy LV weight (r ⫽ 0.90) and that showed excellent reproducibility in this group of experienced physician readers.13 Echocardiography was used to assess the hypertrophy and geometric pattern of the left ventricle. Magnetic resonance imaging. Of the 43 patients, 32 underwent magnetic resonance imaging; 11 did not because of technical breakdown of the machine, a strike among nurses, and claustrophobia in 2 patients. The 32 patients were comparable with the remaining 11 in regard to basic characteristics and all cardiovascular parameters measured. Two magnetic resonance series were performed, triggered by electrocardiography in systole as well as diastole using a Siemens Magnetom Vision 1.5 (Siemens, Erlangen, Germany) operating at 1.5 Tesla. The imaging techniques were CP BodyArray, single-phase high-resolution segmented FLASH, slice thickness 8 mm with 2-mm interslice gap, FOV 300⫻300, TE 7.3, NA ⫽ 5 and no breath holding. Guided from a transverse scout view through the center of the left ventricle, a triggered scout view parallel to the interventricular septum and through the middle of the left ventricle was made. From this scout view, the slice direction was chosen parallel to the most posterior part of the upper and lower part of the base of the left ventricle, and the left ventricle was sliced from the basis to the apex in 8 to 12 short axis slices. In each slice, LV chamber and myocardial areas were calculated and converted into volumes by multiplying by slice thickness. LV mass assessed by magnetic resonance imaging (LV massMRI) and LV mass index assessed by magnetic resonance imaging (LVMIMRI) were used to assess LV mass quantitatively because magnetic resonance imaging has been demonstrated to be more accurate,9 which is of great statistical importance in our rather small group of patients. Carotid artery ultrasound. Intima-media thickness and the lumen diameter of the common carotid arteries were measured by ultrasound using the ACUSON 128XP/10c (Acuson Corporation, Mountain View, Calif) and a linear 7 MHZ transducer. Intima-media thickness of the right and left common carotid artery was measured in the 1-cm segment proximal to the dilation of the carotid bulb as described by Howard et al14 using a previously described protocol.10 Based on end-

diastolic intima-media thickness and lumen diameter, the cross-sectional area of the intima-media complex (IMA) was calculated5 and indexed by height because lumen diameter and, thus, the cross-sectional area were correlated closer to height than to body surface area. The end-diastolic and endsystolic lumen diameter of the right as well as the left common carotid artery was measured using M-mode within the 1-cm segment proximal to the dilation of the carotid bulb, avoiding measuring at plaques. For both arteries, the relative change in lumen diameter was divided by the pulse pressure at the time of the investigation, and the mean value was used as a measure of carotid distensibility, the so-called pressure strain modulus. All readings were done on digitized pictures (Optimas 6.11, Optimas Corporation, Bothell) at The University of Michigan Medical Center in Ann Arbor. Intima-media thickness and IMA/height are used to assess vascular hypertrophy, whereas the pressure strain modulus is used to assess vascular stiffness. Forearm plethysmography. The maximal forearm blood flow was measured after 10 minutes of ischemia15 using strain gauge plethysmography (Model EC5R, Hokanson, Inc, Bellevue, Wash). A wrist-occluding cuff was used. Forearm blood flow was measured for 3 seconds every 6 seconds during the first minute of hyperemia and was expressed as milliliters flow per 100 mL tissue per minute. At the same time BP was measured 3 times with a semiautomatic sphygmanometer (Dinamap Model 1846 SX, Critikon, Johnson & Johnson, Medical Inc, Bellevue). Forearm vascular resistance was calculated as mean arterial BP divided by the corresponding forearm blood flow, and the minimal forearm vascular resistance was found. We chose to analyze minimal forearm vascular resistance in women and men separately, as minimal forearm vascular resistance values and standard variations were much higher in the more obese women (3.1 ⫾ 1.42 vs 2.3 ⫾ 0.48 mm Hg⫻min), probably because of a lower percentage of muscle tissue in their forearms.16 Minimal forearm vascular resistance is thought to represent an integrated measure of vascular remodeling and capillary rarefaction in the forearm and is used to assess peripheral vascular remodeling. Myograph investigations. Under local anesthesia, a subcutaneous biopsy measuring about one cm3 was taken from the gluteal region. The biopsy was placed in a cold physiological

American Heart Journal Volume 144, Number 3

saline solution. The small resistance arteries (100-500 ␮m) were isolated under microscope and mounted on a myograph (Automated Dual Wire Myograph System—Model 500A, J. P. Trading I/S, Århus, Denmark). The media thickness and lumen were measured by microscope, and the vessels were exposed to increasing radial stretch calculating internal circumference at 100 mm Hg distension pressure17 for calculation of the normalized media/lumen ratio. Based on data from the normalization, we calculated the distensibility as the ratio between the incremental relative deformation and incremental change in pressure at 100 mm Hg.18 Media/ lumen ratio is used as a measure of resistance artery remodeling, whereas resistance artery distensibility is a measure of passive stiffness. Ambulatory BP. After resting 1 hour, an automatic BP device that measured ambulatory BP using the cuff-oscillometric method (Takeda TM-2421, A&D Co Ltd, Tokyo, Japan)19 was applied to the left arm and worn for 24 hours. Oscillometric BP was measured every half hour during the daytime and every hour at night (11:00 pm to 6:00 am). There were ⬎20 readable measurements during 24 hours (in 33 men and 7 women) considered sufficient20 to calculate the median value in 24 hours.

Assays All patients met fasting and had a polyethylene cannula inserted into an antecubital vein for collection of blood samples. After lying supine in a quiet room with a constant temperature of 24° to 27°C for 1 hour, the blood samples were drawn and collected in specially prepared containers. Plasma glucose concentrations were measured using a Beckmann glucose analyzer 2 (Beckman Instruments Inc, Fullerton, Calif) and a glucose oxidase method. Serum insulin concentrations were determined by enzyme immunoassay, as described by Andersen et al.21 Plasma concentrations of epinephrine and norepinephrine were determined by radioenzymatic labeling and high pressure liquid chromatography.22 Plasma endothelin concentrations were measured by radioimmunoassay using rabbit antiserum (RAS 6901, Peninsula Laboratories, Belmont, Calif), 125I-endothelin (Amersham Life Science Ltd, Amersham, United Kingdom) and standards (Peptide Institute, Osaka, Japan). Blood was collected in EDTA aprotenin tubes, and plasma was extracted on Sep-pak C18 before analysis. Active plasma renin concentrations were determined using the principle of antibody trapping23 as modified by Millar et al.24 Plasma angiotensin II concentrations were measured according to the method of Kappelgaard et al25 with the modification that Sep-Pac C18 (Millipore Waters) was used for plasma extraction. Plasma aldosterone concentrations were measured using a commercial kit, DSL-8600, obtained from Diagnostic Systems Laboratories Inc, Webster, Tenn.

Statistical analysis For data management and statistical analyses, Statistica 5.1 (StatSoft, Inc, Tulsa, Okla) was used. Parametric statistics were used, calculating mean values and 95% CIs. We performed simple linear regression analyses to calculate the regression quotient (r). When performing stepwise, backward multiple linear regression analyses calculating the standard-

Olsen et al 533

ized regression quotient for each parameter (␤) and the common adjusted coefficient of determination for the model (adj.R2), all measured parameters and basic characteristics entered the model. Two-tailed P values ⬍.05 were considered statistically significant.

Results Influence of BP and neurohormonal factors In simple regression analyses, LVMIMRI (Figure 1, A), LVMIecho, intima-media thickness, IMA/height and distensibility of the common carotid arteries, as well as minimal forearm vascular resistance in men were positively correlated to the median 24-hour ambulatory systolic BP (Table I). Neither relative wall thicknessecho, media/lumen ratio, nor distensibility of the subcutaneous resistance arteries at 100 mm Hg correlated significantly to median 24-hour BPs (Table I). LV massMRI correlated positively to plasma endothelin (r ⫽ 0.35, P ⬍ .05) and negatively to serum insulin (r ⫽ ⫺0.37, P ⬍ .05), but did not correlate with serum insulin when taking diabetes into account. Neither LVMIMRI nor LVMIecho correlated with serum insulin or plasma concentrations of angiotensin II, renin, aldosterone, endothelin, epinephrine, or norepinephrine (Table I). However, relative wall thicknessecho was positively correlated to plasma epinephrine (Table I), as well as plasma aldosterone (r ⫽ 0.31, P ⬍ .05). Carotid artery hypertrophy and resistance artery remodeling were related to high plasma epinephrine. Low distensibility of the common carotid artery was related to high plasma angiotensin II (Table I). Minimal forearm vascular resistance did not relate to any of the growth hormones measured. Vascular structure and distensibility of the common carotid arteries and the subcutaneous resistance arteries were unrelated to plasma renin, plasma aldosterone, and plasma endothelin. However, media/lumen ratio of subcutaneous resistance arteries correlated positively to plasma epinephrine (Table I). In multiple regression analyses, plasma epinephrine (␤ ⫽ 0.41) was positively correlated to intima/media thickness independent of median 24-hour systolic BP (␤ ⫽ 0.36) (adj.R2 ⫽ 0.27, P ⫽ .001). Plasma epinephrine (␤ ⫽ 0.43) was also positively correlated to IMA/height independent of median 24-hour systolic BP (␤ ⫽ 0.40) (adj.R2 ⫽ 0.32, P ⬍ .001). Plasma angiotensin II (␤ ⫽ ⫺0.41) was negatively correlated to common carotid artery distensibility independent of median 24-hour systolic BP (␤ ⫽ ⫺0.40) (adj.R2 ⫽ 0.30, P ⬍ .001). Prior antihypertensive treatment and known duration of hypertension was not significantly related to the measured cardiovascular parameters and did not add to any of the models presented.

American Heart Journal September 2002

534 Olsen et al

Figure 1

of the common carotid arteries nor media/lumen ratio of resistance arteries (Table II). However, LVMIMRI correlated negatively to distensibility of the subcutaneous resistance arteries (Figure 1, C) (Table II). Relative wall thicknessecho did not correlate significantly to any of the vascular parameters. Intima-media thickness of the common carotid artery correlated positively to media/ lumen ratio of subcutaneous resistance arteries (r ⫽ 0.34, P ⬍ .05), and IMA/height showed a tendency toward a positive correlation to minimal forearm vascular resistance in men (r ⫽ 0.34, P ⫽ .06). In multiple regression analyses, LVMIMRI was correlated to IMA/height (␤ ⫽ 0.36) independently of median 24-hour systolic BP (␤ ⫽ 0.46) (adj.R2 ⫽ 0.40, P ⬍ .001), to minimal forearm vascular resistance (␤ ⫽ 0.35) independently of median 24-hour systolic BP (␤ ⫽ 0.49) (adj.R2 ⫽ 0.46, P ⬍ .001), and to resistance artery distensibility (␤ ⫽ ⫺0.34) independently of median 24-hour systolic BP (␤ ⫽ 0.48) (adj.R2 ⫽ 0.39, P ⬍ .001). The 24-hour systolic BP could, together with either IMA/height, minimal forearm vascular resistance, or resistance artery distensibility, explain about 40% of the variation in LVMIMRI. However, the correlates to LVMIecho were too weak to remain significant in multiple regression analyses.

Discussion

Relationship between LV mass index measured by magnetic resonance imaging and 24-systolic blood pressure (n: 30) (A), intimamedia cross-sectional area indexed by height (n: 32) (B), and distensibility of isolated subcutaneous resistance arteries at 100 mm Hg (n: 30) (C).

Relationship between cardiac and vascular hypertrophy and remodeling LV massMRI and LV massecho did correlate (r ⫽ 0.55, P ⫽ .001), and were positively correlated to IMA/ height (Figure 1, B) as well as to minimal forearm vascular resistance, but did not correlate to distensibility

First, we found that the relative association of BP and neurohormonal factors to remodeling of the myocardium, the conduit, and the resistance arteries, respectively, differed from 1 part of the cardiovascular system to the other. Higher LV mass index was associated with higher median 24-hour systolic BP, whereas higher relative wall thicknessecho, indicating concentric LV geometry, was associated with higher levels of circulating epinephrine and aldosterone. Higher intimamedia thickness of the common carotid arteries was associated with higher median 24-hour systolic BP as well as higher levels of circulating epinephrine, whereas lower distensibility was associated with higher median 24-hour systolic BP as well as higher levels of circulating angiotensin II. Remodeling of the subcutaneous resistance arteries, examined in vitro, was associated with higher levels of circulating epinephrine. Second, we found that the degree of cardiac and vascular hypertrophy was interrelated. LV hypertrophy was related to hypertrophy of the common carotid arteries, vascular remodeling in the forearm, and low distensibility of the subcutaneous resistance arteries partly independently of the BP.

Influence of BP and growth hormones Our study supports earlier findings that cardiac myocyte growth is primarily related to load, whereas

American Heart Journal Volume 144, Number 3

smooth muscle cell growth is related to neurohormonal stimuli,2 but indicate also an influence of circulating neurohormonal factors on LV geometry through LV remodeling.26 The lack of correlation between media/lumen ratio and BP in this study and the weak correlation found in other studies27 may, however, suggest a significant influence of other factors on vascular remodeling in subcutaneous resistance arteries.28 The weak relationship between BP and resistance artery remodeling is supported by Lever,29 who showed that hypophysectomy in rats prevented the structural vascular response to increased arterial pressure, and by Schiffrin et al,30 who showed that media/lumen ratio decreased in hypertensive patients treated with cilazapril, whereas it was unchanged in patients treated with atenolol despite a better BP reduction during treatment with atenolol. As angiotensin II can cause vascular hypertrophy in rats by a pressure-independent mechanism,29 we expected the plasma level of angiotensin II to correlate with cardiovascular hypertrophy in our patients. The lack of correlation might be explained by the relatively low levels of circulating angiotensin II, previous antihypertensive treatment, the long duration of hypertension, and the suggestion that angiotensin II primarily acts as an autocrine-paracrine growth factor.31 However, high plasma levels of circulating catecholamines, especially epinephrine, were associated with carotid artery hypertrophy independently of BP and resistance artery remodeling, suggesting a trophic influence of catecholamines on the vasculature, as previously demonstrated by others.32,33 Venous plasma levels of epinephrine resemble, to some extent, arterial plasma levels, whereas the venous plasma level of norepinephrine does not, because it consists primarily of spillover from peripheral nerve endings. This might explain why circulating epinephrine, but not norepinephrine, correlates to remodeling of the resistance arteries. We previously demonstrated an association between hyperinsulinemia and low carotid distensibility in a subgroup of never-treated hypertensive patients from the LIFE-ICARUS substudy10; therefore, the lack of association between hyperinsulinemia and cardiovascular hypertrophy in this study and in a previous publication34 is probably the result of the selection criteria for the LIFE study and the long duration of treated hypertension. Low distensibility of the common carotid arteries was related to high 24-hour ambulatory BP, which supported earlier findings by Cunha et al,35 as well as to high levels of plasma angiotensin II. However, low distensibility of the subcutaneous resistance arteries in vitro at a stretch resembling a pressure of 100 mm Hg was not related to high 24-hour ambulatory BP. This lack of relationship between BP and resistance artery distensibility has previously been found by Thybo et al18 and Intengan et al,36 both demonstrating normal

Olsen et al 535

distensibility in isolated subcutaneous resistance arteries in patients with essential hypertension. Our finding that higher plasma levels of angiotensin II are associated with lower distensibility of the common carotid artery is supported by experimental data by Himeno et al,37 who found that angiotensin II infusion induced an increase in aortic fibronectin mRNA in rats even without an increase in BP, and Albaladejo et al,38 who found that an angiotensin-converting enzyme inhibitor prevented the increase in aortic collagen in spontaneously hypertensive rats. Based on our data, we hypothesize that blocking the renin–aldosterone-angiotensin II system may have beneficial effects on LV geometry and conduit artery compliance independently of the BP reduction. Our data may indicate that blocking the ␤-receptors may have equally beneficial effects on LV geometry and vascular hypertrophy; however, several studies30 have demonstrated that ␤-receptor blockage does not induce regression of resistance artery remodeling.

Relationship between cardiac and vascular remodeling The moderate correlation between LV massecho LV massMRI can probably be explained by the fact that LV massecho is calculated by an empirically deduced formula, whereas LV massMRI is based on a more complete 3-dimensional visualization of LV myocardial volume derived from measurements in about 10 layers through the LV. We found that LV mass, but not relative wall thicknessecho, was associated with carotid artery hypertrophy, vascular remodeling in the forearm, and low distensibility of the subcutaneous resistance arteries partly independent of the BP. Several other investigators have found a relationship between LV hypertrophy and carotid artery hypertrophy,5 as well as peripheral vascular remodeling.4 Probably, we were able to demonstrate this in a small group of patients because we reduced the methodologic variation by use of ambulatory BP measurements and LVMIMRI.9 Our results extend previous observations39 documenting an association between cardiac and carotid artery structure independent of ambulatory BP. LV hypertrophy and geometry have previously been related to low distensibility of conduit arteries,6,40 but not to low distensibility of the resistance arteries. We found no significant correlation between LV mass index and media/ lumen ratio of subcutaneous resistance arteries, supporting the idea that LV hypertrophy and peripheral vascular remodeling evolve differently during longstanding hypertension. Changes in vascular structure, and especially increased stiffness of the conduit arteries, may increase the pulse wave velocity, and the reflected waves may then reach the central circulation in late systole, rather than in diastole, augmenting systolic BP and increasing cardiac afterload.41,42 Cardiac load

American Heart Journal September 2002

536 Olsen et al

is, therefore, dependent on brachial BP as well as vascular structure, supporting our finding of a relationship between cardiac mass and vascular structure and stiffness not explained by their common relationship to the BP.

Conclusion We analyzed the association between parameters describing cardiovascular remodeling on the one side, and on the other side, the BP burden and circulating neurohormonal factors. Different patterns appeared for each part of the cardiovascular system. LV hypertrophy was associated with higher BP, whereas concentric LV geometry was associated with higher plasma levels of epinephrine and aldosterone. Hypertrophy and low distensibility of the common carotid arteries were associated with higher BP as well as higher plasma levels of epinephrine and angiotensin II, respectively. Vascular remodeling of subcutaneous resistance arteries was associated with elevated levels of plasma epinephrine. Therefore, our data, to some extent, support the hypothesis of a potential beneficial effect on cardiovascular remodeling above and beyond the effect of blood pressure reduction when blocking the renin–aldosterone-angiotensin II system. However, our data also suggest a beneficial effect of ␤-receptor blockage. LV hypertrophy was related to hypertrophy of the common carotid arteries, forearm vascular remodeling, and low distensibility of the subcutaneous resistance arteries partly independent of ambulatory BP, indicating that LV hypertrophy is dependent not only on the BP burden, but also on vascular structure.

8.

9. 10.

11.

12. 13.

14.

15. 16.

17.

18.

19.

We thank Lisa Krause in Ann Arbor, Mich, for technical assistance in measuring intima-media thickness of the common carotid artery.

References 1. Glasser SP. Hypertension, hypertrophy, hormones, and the heart. Am Heart J 1998;135:S16-20. 2. Weber KT, Brilla CG, Cleland JG, et al. Cardioreparation and the concept of modulating cardiovascular structure and function. Blood Press 1993;2:6-21. 3. Rosendorff C. The renin-angiotensin system and vascular hypertrophy. J Am Coll Cardiol 1996;28:803-12. 4. Pierdomenico SD, Lapenna D, Guglielmi MD, et al. Vascular changes in hypertensive patients with different left ventricular geometry. J Hypertens 1995;13:1701-6. 5. Roman MJ, Saba PS, Pini R, et al. Parallel cardiac and vascular adaptation in hypertension. Circulation 1992;86:1909-18. 6. Saba PS, Roman MJ, Pini R, et al. Relation of arterial pressure waveform to left ventricular and carotid anatomy in normotensive subjects. J Am Coll Cardiol 1993;22:1873-80. 7. Shimamoto H, Shimamoto Y. Lisinopril reverses left ventricular hy-

20. 21.

22.

23.

24.

25.

26.

pertrophy through improved aortic compliance. Hypertension 1996;28:457-63. Rizzoni D, Castellano M, Porteri E, et al. Effects of low and high doses of fosinopril on the structure and function of resistance arteries. Hypertension 1995;26:118-23. Korner PI, Jennings GL. Assessment of prevalence of left ventricular hypertrophy in hypertension. J Hypertens 1998;16:715-23. Olsen MH, Fossum E, Hjerkinn E, et al. Relative influence of insulin resistance versus blood pressure on vascular changes in longstanding hypertension. ICARUS, a LIFE substudy. J Hypertens 2000;18:75-81. Dahlof B, Devereux RB, Julius S, et al. Characteristics of 9194 patients with left ventricular hypertrophy: the LIFE study. Hypertension 1998;32:989-97. WHO Technical Reports Series 727. 1985; p. 9. Palmieri V, Dahlof B, DeQuattro V, et al. Reliability of echocardiographic assessment of left ventricular structure and function: the PRESERVE study. Prospective Randomized Study Evaluating Regression of Ventricular Enlargement. J Am Coll Cardiol 1999;34: 1625-32. Howard G, Sharrett AR, Heiss G, et al. Carotid artery intimal-medial thickness distribution in general populations as evaluated by B-mode ultrasound. ARIC Investigators. Stroke 1993;24:1297-304. Folkow B. Physiological aspects of primary hypertension. Physiol Rev 1982;62:347-504. Utriainen T, Malmstrom R, Makimattila S, et al. Methodological aspects, dose-response characteristics and causes of interindividual variation in insulin stimulation of limb blood flow in normal subjects. Diabetologia 1995;38:555-64. Mulvany MJ, Halpern W. Contractile properties of small arterial resistance vessels in spontaneously hypertensive and normotensive rats. Circ Res 1977;41:19-26. Thybo NK, Mulvany MJ, Jastrup B, et al. Some pharmacological and elastic characteristics of isolated subcutaneous small arteries from patients with essential hypertension. J Hypertens 1996;14: 993-8. Imai Y, Sasaki S, Minami N, et al. The accuracy and performance of the A&D TM 2421, a new ambulatory blood pressure monitoring device based on the cuff-oscillometric method and the Korotkoff sound technique. Am J Hypertens 1992;5:719-26. Conway J, Coats A. Value of ambulatory blood pressure monitoring in clinical pharmacology. J Hypertens 1989;7(Suppl):S29-32. Andersen L, Dinesen B, Jorgensen PN, et al. Enzyme immunoassay for intact human insulin in serum or plasma. Clin Chem 1993;39: 578-82. Appel E, Bayer P, Hajdu P, et al. Determination of plasma catecholamines by means of radioenzymatic labelling and high pressure liquid chromatographic separation. Naunyn Schmiedebergs Arch Pharmacol 1981;315:233-9. Poulsen K, Jorgensen J. An easy radioimmunological microassay of renin activity, concentration and substrate in human and animal plasma and tissues based on angiotensin I trapping by antibody. J Clin Endocrinol Metab 1974;39:816-25. Millar JA, Leckie BJ, Morton JJ, et al. A microassay for active and total renin concentration in human plasma based on antibody trapping. Clin Chim Acta 1980;101:5-15. Kappelgaard AM, Nielsen MD, Giese J. Measurement of angiotensin II in human plasma: technical modifications and practical experience. Clin Chim Acta 1976;67:299-306. Muscholl MW, Schunkert H, Muders F, et al. Neurohormonal ac-

American Heart Journal Volume 144, Number 3

27.

28.

29. 30.

31.

32. 33.

34.

tivity and left ventricular geometry in patients with essential arterial hypertension. Am Heart J 1998;135:58-66. Cooper A, Heagerty AM. Blood pressure parameters as determinants of small artery structure in human essential hypertension. Clin Sci Colch 1997;92:551-7. Dzau VJ, Gibbons GH. Endothelium and growth factors in vascular remodeling of hypertension. Hypertension 1991;18(Suppl): III115-21. Lever AF. Slow developing pressor effect of angiotensin II and vascular structure. J Hypertens 1993;11(Suppl):S27-8. Schiffrin EL, Deng LY, Larochelle P. Effects of a beta-blocker or a converting enzyme inhibitor on resistance arteries in essential hypertension. Hypertension 1994;23:83-91. Gibbons GH. Mechanisms of vascular remodeling in hypertension: role of autocrine-paracrine vasoactive factors. Curr Opin Nephrol Hypertens 1995;4:189-96. Head RJ. Hypernoradrenergic innervation and vascular smooth muscle hyperplastic change. Blood Vessels 1991;28:173-8. Baumbach GL, Heistad DD, Siems JE. Effect of sympathetic nerves on composition and distensibility of cerebral arterioles in rats. J Physiol 1989;416:123-40. Olsen MH, Andersen UB, Wachtell K, et al. A possible link between endothelial dysfunction and insulin resistance in hypertension. A LIFE substudy. Blood Press 2000;9:132-9.

Olsen et al 537

35. Cunha RS, Benetos A, Laurent S, et al. Distension capacity of the carotid artery and ambulatory blood pressure monitoring. Effects of age and hypertension. Am J Hypertens 1995;8:34352. 36. Intengan HD, Deng LY, Li JS, et al. Mechanics and composition of human subcutaneous resistance arteries in essential hypertension. Hypertension 1999;33:569-74. 37. Himeno H, Crawford DC, Hosoi M, et al. Angiotensin II alters aortic fibronectin independently of hypertension. Hypertension 1994; 23:823-6. 38. Albaladejo P, Bouaziz H, Duriez M, et al. Angiotensin converting enzyme inhibition prevents the increase in aortic collagen in rats. Hypertension 1994;23:74-82. 39. Jones EC, Devereux RB, O’Grady MJ, et al. Relation of hemodynamic volume load to arterial and cardiac size. J Am Coll Cardiol 1997;29:1303-10. 40. Roman MJ, Ganau A, Saba PS, et al. Impact of arterial stiffening on left ventricular structure. Hypertension 2000;36:489-94. 41. Westerhof N, Sipkema P, Bos GC, et al. Forward and backward waves in the arterial system. Cardiovasc Res 1972;6:648-56. 42. O’Rourke MF, Yaginuma T, Avolio AP. Physiological and pathophysiological implications of ventricular/vascular coupling. Ann Biomed Eng 1984;12:119-34.