- Email: [email protected]
Prevalence of specific variant carotid geometric patterns and incidence of cardiovascular events in older persons
Journal of the American College of Cardiology © 2004 by the American College of Cardiology Foundation Published by Elsevier Inc.
Vol. 43, No. 2, 2004 ISSN 0735-1097/04/$30.00 doi:10.1016/j.jacc.2003.08.035
Prevalence of Specific Variant Carotid Geometric Patterns and Incidence of Cardiovascular Events in Older Persons The Cardiovascular Health Study (CHS E-131) Angelo Scuteri, MD, PHD,* Teri A. Manolio, MD, PHD,† Emily K. Marino, MS,‡ Alice M. Arnold, PHD,‡ Edward G. Lakatta, MD* Baltimore and Bethesda, Maryland; and Seattle, Washington We hypothesized that variant geometric patterns of the common carotid artery (CCA) predict the incidence of cardiovascular disease (CVD), after accounting for CCA intimamedial thickness (IMT). BACKGROUND Common carotid artery intima-media thickness has been associated with the incidence of cardiovascular disease. METHODS Noninvasive measurements of IMT were made with high-resolution ultrasonography in 5,640 subjects 65 years of age or older participating in the Cardiovascular Health Study. New coronary and/or cerebrovascular events served as outcome variables over a median 10.2-year follow-up. To characterize different carotid structural geometric patterns (CGP), vascular mass (VM) was combined with the wall-to-lumen ratio (W/L). Normal values for W/L and VM were defined as age-adjusted, gender-specific 75th percentiles of the 1,899 normotensive subjects free of CVD at baseline. Four CGPs were defined: CGP1 ⫽ normal W/L ratio and VM; CGP2 ⫽ arterial remodeling (i.e., increased W/L ratio with normal VM); CGP3 ⫽ arterial hypertrophy (i.e., increased W/L ratio with increased VM); and CGP4 ⫽ arterial hypertrophy with dilation (i.e., normal W/L ratio and increased VM). RESULTS Coronary or cerebrovascular events (adjusted for age, gender, traditional risk factors, and IMT) were associated with CGP in subjects free of CVD at baseline. Specifically, the hazard ratio (Cox proportional-hazards analyses) for CGP3 (arterial hypertrophy) was 1.25 (95% confidence interval [CI] 1.03 to 1.53), and for CGP4 (arterial hypertrophy with dilation) was 1.43 (95% CI 1.16 to 1.75) compared with CGP1 (normal). CONCLUSIONS Arterial hypertrophy defined by variant CGP patterns is associated with the development of new CVD, independent of age, traditional risk factors, and CCA IMT. (J Am Coll Cardiol 2004;43:187–93) © 2004 by the American College of Cardiology Foundation OBJECTIVES
The blood vessel wall is an active organ continuously undergoing remodeling (1). The result is a geometric pattern that cannot be described by only a unidimensional variable, such as arterial wall thickness or diameter, but can be characterized by combining a measure of vascular mass with the wall-to-lumen ratio (W/L) similar to small-artery and arteriolar morphologic changes previously distinguished in hypertension as “remodeling” or “hypertrophy” (2– 4) and whose differential functional correlates have been described (5,6). Although this concept of small-vessel geometric remod-
From the *Laboratory of Cardiovascular Sciences, National Institute on Aging, National Institutes of Health, Baltimore, Maryland; †Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Bethesda, Maryland; and the ‡Collaborative Health Studies Coordinating Center, Department of Biostatistics, University of Washington, Seattle, Washington. The research reported in this article was supported by contracts N01-HC-85079 through N01-HC-85086, N01-HC-35129, and N01 HC-15103 from the National Heart, Lung, and Blood Institute. Dr. Scuteri is currently affiliated with UO Geriatria, INRCA, Rome, Italy. The Cardiovascular Health Study (CHS) was conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with the CHS Investigators. This manuscript has been reviewed by CHS and by NHLBI for scientific content and consistency of data interpretation with previous CHS publications; significant comments have been incorporated prior to submission for publication. Manuscript received February 14, 2003; revised manuscript received August 18, 2003, accepted August 25, 2003.
eling may have substantial pathophysiologic and therapeutic relevance (1), the geometric patterns of large arteries have not been examined as risk factors for cardiovascular disease (CVD). Traditional risk factors of coronary and cerebrovascular disease (CeVD) cause alterations in vascular structure and function (7–9) that increase the likelihood of a subsequent clinical event. We recently observed that specific vascular geometric patterns of the common carotid artery (CCA) were associated with characteristic profiles of vascular functional measurements such as distensibility, circumferential stress, and strain in a cross-sectional population-based survey in 1,315 normotensive and untreated hypertensive Taiwanese men and women (10). Although these functional profiles differed quantitatively in normotensive compared with hypertensive subjects and in younger compared with older subjects, the characteristic functional profile of a given carotid geometric pattern (CGP) was preserved regardless of age or blood pressure (BP) status. We hypothesized that the geometric rearrangement of large arteries may play a role in the genesis of cardiovascular events, and may add predictive information to traditional cardiovascular risk factors and subclinical vascular disease defined by unidimensional vascular measures such as intimal-medial thickness (IMT) or lumen diameter (7,9,11).
188
Scuteri et al. Carotid Geometry and Cardiovascular Events
Abbreviations and Acronyms BP ⫽ blood pressure CCA ⫽ common carotid artery CeVD ⫽ cerebrovascular disease CGP ⫽ carotid geometric pattern CHD ⫽ coronary heart disease CHS ⫽ Cardiovascular Health Study CI ⫽ confidence interval CVD ⫽ cardiovascular disease HR ⫽ hazard ratio IMT⫽ intima-media thickness SD ⫽ standard deviation VM ⫽ vascular mass W/L ⫽ wall-to-lumen ratio
METHODS Subjects and study design. The study subjects were participants in the Cardiovascular Health Study (CHS), a prospective multicenter study of men and women 65 years of age or older sponsored by the National Heart, Lung, and Blood Institute (12). A detailed description of the recruitment methods has been published elsewhere (13). Of the members of the original cohort, 94% were white and most of the rest were black. A second cohort of 687 black participants was enrolled between June 1992 and June 1993, three years after the enrollment of the original CHS cohort. The baseline examination for both cohorts included a medical history, physical examination, laboratory testing, and assessment of CVD status. The study design, qualitycontrol procedures, laboratory methods, and BP measurement have been previously reported (12,14,15). The algorithms for classifying new myocardial infarctions and strokes have also been published (16,17). The carotid arteries were evaluated with high-resolution B-mode ultrasonography (15). One longitudinal image of the CCA was acquired. Measurements were made at a central reading center by readers blind to all clinical information. The same readers were used for all readings. The maximal IMT of the CCA was defined as the mean of the maximal IMTof the far wall on both the left and right sides. The CCA diameter measurements were made in the portion of the CCA where the wall interfaces are parallel. The CCA diameter was the mean of the minimal distance between the near-wall intima-lumen interface to the farwall lumen-intima interface of the right and left CCAs. As in prior studies of heart (18,19) and small artery (2– 4) geometry, vascular mass (VM) was combined with the W/L to identify different structural CGPs. Normal values for W/L (W/L ⫽ 2 · carotid IMT/carotid diameter) and VM (VM ⫽ IMT2 ; where is the arterial wall density, i.e., 1.06 g/cm3) (20) were defined as those within the ageadjusted and gender-specific 75th percentile of the 1,899 normotensive subjects (BP ⬍140/90 mm Hg) without prevalent CVD. Four common CGP were defined: CGP1 ⫽ normal, with normal W/L ratio and VM; CGP2 ⫽
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
arterial remodeling, with increased W/L ratio with normal VM; CGP3 ⫽ arterial hypertrophy, with increased W/L ratio and increased VM; and CGP4 ⫽ arterial hypertrophy with dilation, with normal W/L ratio and increased VM. Statistical analysis. Prevalent CVD at first visit was defined as history of myocardial infarction, angina, coronary artery bypass surgery, coronary artery angioplasty, congestive heart failure, stroke, transient ischemic attack, or carotid endarterectomy confirmed by review of medical records. Incident CVD was defined as coronary heart disease (CHD), including angina pectoris, myocardial infarction, angioplasty, or aortocoronary bypass surgery; congestive heart failure; and CeVD, including stroke or transient ischemic attack; as confirmed by review of medical records. New definite CHD and CeVD (combined cardiovascular end points) were considered the end points for the present analyses. Participants with a history of claudication were excluded from analysis and those with incident claudication were censored at the time of the claudication event. Subclinical disease for this analysis was defined as carotid stenosis ⱖ75% and ankle-arm index ⬍0.9. Hypertension was defined as BP ⱖ140/90 mm Hg or a reported history of high BP in combination with use of antihypertensive medications. Diabetes was defined by the American Diabetes Association criteria: glucose intolerance was defined by fasting glucose ⬎110 mg/dl and diabetes was defined by fasting glucose ⬎126 mg/dl or use of insulin or oral hypoglycemics. All analyses were performed with SPSS statistical software (SPSS Inc., Chicago, Illinois). Mean values of demographics and CVD risk factors were calculated across CGPs by means of one-way analysis of variance. Significance of pair-wise comparisons was adjusted for multiple comparisons using a Bonferroni correction. Categorical variables were compared by a chi-square test. We evaluated the associations of incident and recurrent CVD and the CGP groups using Cox proportional-hazards regression models, adjusting for traditional risk factors, CCA IMT and diameter, and subclinical disease. Atrial fibrillation and creatinine levels were also introduced into the multivariable models, given that previous analyses in the CHS population showed these to be risk factors for CHD or CeVD (21,22). Prediction of subsequent cardiovascular events was analyzed separately in subjects without prevalent CVD (incident events) and in subjects with prevalent CVD at baseline (recurrent events). Participants were censored at the time of incident claudication, death, or loss to follow-up. Unadjusted Cox proportional-hazards regression was used to plot curves for incident CVD or CeVD and compare survival rates among the CGP groups.
RESULTS The population sample consisted of 4,291 (76%) subjects without prevalent CVD at baseline and 1,349 (24%) subjects with prevalent CVD. The baseline characteristics of
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93 Table 1. Baseline Characteristics of the Study Population Age (yrs) Female Family history of MI Current smoking BMI (kg/m2) Fasting glucose (mg/dl) Diabetes Normal IFG New or known diabetes Total cholesterol (mg/dl) HDL cholesterol (mg/dl) History of CHD History of CeVD History of either CHD or CeVD (CVD) Hypertension Controlled hypertension Use of antihypertensive medications SBP (mm Hg) DBP (mm Hg) PP (mm Hg) Creatinine (mg/dl) Atrial fibrillation on ECG Maximum common carotid IMT (mm) Average common carotid IMT (mm) Average common carotid diameter (mm) Carotid stenosis ⬎75% Ankle-arm index ⬍0.9 Normal (CGP1) Arterial concentric remodeling (CGP2) Arterial hypertrophy (CGP3) Arterial hypertrophy with dilation (CGP4)
72.8 (5.6) 3,283 (58.2) 1,632 (28.9) 653 (11.6) 26.7 (4.7) 110.7 (36.0) 3,935 (69.8) 751 (13.3) 894 (15.9) 211.3 (39.2) 54.4 (15.7) 1174 (20.8) 310 (5.5) 1,349 (23.9) 3,280 (58.2) 1,075 (19.1) 2,629 (46.6) 136.41 (21.83) 70.83 (11.34) 65.53 (18.67) 1.06 (0.40) 143 (2.5) 1.06 (0.21) 1.04 (0.26) 9.23 (0.98) 50 (0.9) 656 (11.6) 3,429 (60.8) 248 (4.4) 1,420 (25.2) 543 (9.6)
Entries are mean (SD) or N (%). Total: N ⫽ 5,640. BMI ⫽ body mass index; CeVD ⫽ cerebrovascular disease; CGP ⫽ carotid geometric pattern; CHD ⫽ coronary heart disease; CVD ⫽ cardiovascular disease; DBP ⫽ diastolic blood pressure; ECG ⫽ electrocardiogram; HDL ⫽ high-density lipoprotein; IFG ⫽ impaired fasting glucose; IMT ⫽ intima-media thickness; MI ⫽ myocardial infarction; PP ⫽ pulse pressure; SBP ⫽ systolic blood pressure.
the 5,640 study participants are given in Table 1. The median follow-up was 10.2 years (maximum was 11.1 years). A total of 1,878 (33.3%) participants had coronary events and 835 (14.8%) participants had cerebrovascular events during follow-up. Of the subjects without prevalent CVD, 1,453 (33.9%) experienced a new event during the follow-up time (1,210 [28.2%] CHD event and 720 [16.8%] CeVD event), whereas 749 (56%) of those with prevalent CVD experienced a subsequent event (668 [49.5%] CHD event and 115 [8.5%] CeVD event). The combined CVD end point was used after initial separate analyses with CHD and CeVD showed similar results. Prevalence of CGP. The prevalence of the four measured CGPs in the entire population is listed in Table 2. Subjects with prevalent CVD had a higher frequency of variant CGPs compared with those without prevalent CVD (44.6% vs. 37.5%, respectively, p ⬍ 0.001). Prevalent CVD was associated with a lower prevalence of normal (CGP1) and concentric remodeling (CGP2) and with a higher prevalence of hypertrophy (CGP3) and hypertrophy with dilation (CGP4) compared with subjects without prevalent CVD (Table 3).
189
Table 2. Prevalence of CGP by Gender
CGP1 CGP2 CGP3 CGP4 Total variant CGP
Men
Women
1,413 (59.9) 106 (4.5) 603 (25.6) 235 (10.0) 944 (40.1)
2,016 (61.4) 142 (4.3) 817 (24.9) 308 (9.4) 1,267 (38.6)
Chi-square 1.224: p value 0.269 for comparison of total variant CGP prevalence between men and women. CGP ⫽ carotid geometric pattern.
The demographic and risk factor profiles of each CGP are summarized in Table 4. Mean carotid IMT was higher for each variant CGP compared with CGP1, with a progressive increase in IMT values from arterial concentric remodeling (CGP2) to arterial hypertrophy with dilation (CGP4) to arterial hypertrophy (CGP3). Although carotid IMT was increased in subjects with concentric remodeling (CGP2) compared with CGP1, blood pressure values, glucose, and cholesterol levels were similar in both these patterns, and hypertension prevalence in concentric remodeling (CGP2) was lower than in CGP1. New CVD by CGP. As shown in Figure 1, in subjects without prevalent CVD, the rate of new combined cardiovascular events was not significantly different in CGP1 and concentric remodelling (CGP2), but was higher in arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4) than in normal geometry (CGP1) or concentric remodeling (CGP2). In subjects with prevalent CVD, the rate of new combined cardiovascular end points was much Table 3. Prevalence of Variant CGP by Prevalent CHD, CeVD, and CVD (CHD or CeVD)
CGP1* CGP2* CGP3* CGP4* Total variant CGP*
CGP1† CGP2† CGP3† CGP4† Total variant CGP†
CGP1‡ CGP2‡ CGP3‡ CGP4‡ Total variant CGP‡
No Prevalent CHD (n ⴝ 4,466)
Prevalent CHD (n ⴝ 652)
2,777 (62.2) 210 (4.7) 1,078 (24.1) 401 (9.0) 1,689 (37.8)
652 (55.5) 38 (3.2) 342 (29.1) 142 (12.1) 522 (44.5)
No Prevalent CeVD (n ⴝ 5,330)
Prevalent CeVD (n ⴝ 310)
3,275 (61.4) 240 (4.5) 1,311 (24.6) 504 (9.5) 2,055 (38.6)
154 (49.7) 8 (2.6) 109 (35.2) 39 (12.6) 156 (50.3)
No Prevalent CVD (n ⴝ 4,291)
Prevalent CVD (n ⴝ 747)
2,682 (62.5) 207 (4.8) 1,020 (23.8) 382 (8.9) 1,609 (37.5)
747 (55.4) 41 (3.0) 400 (29.7) 161 (11.9) 602 (44.6)
*Chi-square 17.219, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. †Chi-square 17.020, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. ‡Chi-square 21.883, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. Abbreviations as in Table 1.
190
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Table 4. Demographic and CVD Risk Profile by CGP Status (Means ⫾ SE or %)
Age (yrs) Female (%) Family history of MI (%) Current smoking (%) BMI (kg/m2) Fasting glucose (mg/dl) Diabetes (%) Total cholesterol (mg/dl) HDL cholesterol (mg/dl) Pre-existing CHD (%) Prevalent CeVD (%) Prevalent CVD (%) Hypertension (%) Controlled hypertension (%) Use of anti-hypertensive medications SBP (mm Hg) DBP (mm Hg) PP (mm Hg) Creatinine (mg/dl) Atrial fibrillation on ECG (%) Carotid stenosis ⬎75% (%) Ankle-arm index ⬍0.9 (%) Maximum carotid IMT (mm) Average carotid IMT (mm) Carotid diameter (mm) W/L ratio Vascular mass (mm2)
CGP1 Normal (n ⴝ 3429) (60.8%)
CGP2 Remodelling (n ⴝ 248) (4.4%)
CGP3 Hypertrophy (n ⴝ 1,420) (25.2%)
CGP4 Hypertrophy With Dilation (n ⴝ 543) (9.6%)
72.8 ⫾ 0.1 58.8 31.2 10.4 26.4 ⫾ 0.1 107.9 ⫾ 0.5 13.1 209.8 ⫾ 0.7 55.3 ⫾ 0.3 19.0 4.5 21.8 54.0 19.0 44.2 134.2 ⫾ 0.4 70.6 ⫾ 0.2 63.6 ⫾ 0.3 1.05 ⫾ 0.33 2.3 0.5 9.7 0.96 ⫾ 0.002 0.89 ⫾ 0.002 9.15 ⫾ 0.02 0.198 ⫾ 0.0003 2.70 ⫾ 0.01
72.1 ⫾ 0.3 57.3 29.9 7.3 25.9 ⫾ 0.2 107.1 ⫾ 1.9 11.3 213.7 ⫾ 2.3 56.7 ⫾ 1.1 15.3 3.2 16.5 44.4 15.7 38.7 131.8 ⫾ 1.2 71.1 ⫾ 0.7 60.7 ⫾ 1.1 1.04 ⫾ 0.24 2.9 0 7.7 1.00 ⫾ 0.007 1.01 ⫾ 0.004 8.16 ⫾ 0.03 0.248 ⫾ 0.001 3.41 ⫾ 0.03
72.9 ⫾ 0.2 57.5 32.6 14.0 27.1 ⫾ 0.1 115.8 ⫾ 1.2 22.1 215.4 ⫾ 1.1 52.4 ⫾ 0.4 24.1 7.7 28.2 65.8 19.6 50.2 140.2 ⫾ 0.6 70.9 ⫾ 0.3 69.3 ⫾ 0.5 1.09 ⫾ 0.53 2.7 1.6 17.7 1.265 ⫾ 0.006 1.348 ⫾ 0.007 9.44 ⫾ 0.02 0.286 ⫾ 0.001 6.31 ⫾ 0.08
72.8 ⫾ 0.1 56.7 32.3 14.7 27.8 ⫾ 0.2 117.1 ⫾ 1.9 20.9 209.2 ⫾ 1.7 53.1 ⫾ 0.6 26.2 7.2 29.7 71.2 19.9 56.7 142.5 ⫾ 1.0 72.1 ⫾ 0.5 70.3 ⫾ 0.8 1.07 ⫾ 0.41 3.3 1.7 12.4 1.156 ⫾ 0.006 1.135 ⫾ 0.004 10.26 ⫾ 0.04 0.222 ⫾ 0.001 4.32 ⫾ 0.03
p < 0.05 n.s. n.s. n.s. 3,4 ⬎ 1,2 4 ⬎ 3 ⬎ 1,2 3,4 ⬍ 1,2 3,4 ⬍ 1,2 3 ⬎ 1,4 1,2 ⬎ 3,4 3,4 ⬎ 1,2 3 ⬎ 1,2 3,4 ⬎ 1,2 3,4 ⬎ 1 ⬎ 2 n.s. 3,4 ⬎ 1,2 3,4 ⬎ 1,2 4⬎1 3,4 ⬎ 1,2 3⬎1 n.s. 3⬎1 3 ⬎ 1,2,4 3⬎4⬎2⬎ 3⬎4⬎2⬎ 4⬎3⬎1⬎ 3⬎2⬎4⬎ 3⬎4⬎2⬎
1 1 2 1 1
All p values are adjusted for all six two-way comparisons using Bonferroni’s method. W/L ⫽ wall-to-lumen ratio; other abbreviations as in Tables 1 and 3.
lower in concentric remodeling (CGP2) as compared with CGP1, and not significantly higher in arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4) as compared with CGP1.
Unadjusted Cox proportional hazard-regression in subjects without prevalent CVD confirmed a similar rate of events for CGP2 and CGP1 and a higher rate of events for CGP3 and CGP4. Event rates started to diverge from
Figure 1. Cumulative rate for incident and/or recurrent events over the 10.2 years follow-up according to carotid geometric pattern (CGP) status. CVD ⫽ cardiovascular disease.
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Figure 2. Unadjusted incidence of coronary heart disease or cerebrovascular disease according to carotid geometric pattern (CGP) status and presence or absence of prevalent cardiovascular disease (CVD).
CGP1 after approximately one year (Fig. 2). In subjects with prevalent CVD, unadjusted Cox proportional hazardsregression showed a much lower rate of recurrent events for concentric remodeling (CGP2) and a slightly higher rate of recurrent events for arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4). Rates started to diverge very early for concentric remodeling (CGP2) and divergence increased with time (Fig. 2). Multivariable Cox proportional hazard models in only those subjects without prevalent CVD showed that whereas CGP2 did not differ from CGP1, the variant “hypertrophic” carotid geometric patterns (CGP3 and CGP4) persisted as independent predictors of new combined cardiovascular events, after taking into account traditional risk factors, carotid IMT and diameter, and subclinical disease (Table 5). Specifically, hypertrophy (CGP3) was associated with a 25% (hazard ratio [HR] ⫽ 1.25, 95% confidence interval [CI] ⫽ 1.03 to 1.53) higher risk of new CVD, and hypertrophy with dilation (CGP4) with a 43% (HR ⫽ 1.43, 95% CI ⫽ 1.16 to 1.75) higher risk (Table 5). In multivariable Cox proportional hazard analyses limited to subjects with prevalent CVD, the CGP groups were not associated with increased risk of recurrent events after adjustment for traditional risk factors, carotid IMT, and subclinical disease (Table 5). Secondary analyses for more specific end points showed that when angina pectoris was deleted from the definition of CHD the results were similar to those presented in Table 5. The CGP3 and CGP4 were associated with incident myocardial infarction and a combined outcome of either myocardial infarction or congestive heart failure. When CGP was removed from the model for those free of CVD at baseline, and IMT, lumen diameter, and the product (interaction) of the IMT and diameter were included, the interaction term was significant (p ⬍ 0.03). The HR of incident CVD associated with a 1 standard deviation (SD) increment in IMT (0.26 mm) at the mean value of lumen diameter was 1.16 (95% CI 1.10 to 1.23), and the HR of a 1 SD increment in lumen diameter (0.98 mm) at the mean value of IMT was 1.15 (95% CI 1.08 to 1.24).
Scuteri et al. Carotid Geometry and Cardiovascular Events
191
The coefficient of the interaction term was negative, suggesting an increased risk of CVD associated with a 1 SD increase in IMT when the diameter is below the mean and vice versa. When both are above their means, the negative interaction results in a somewhat less than additive effect of wall thickness and lumen diameter on the HR. When CGP and lumen diameter were removed from the model for those free of CVD, IMT was significantly associated with the risk of CVD (HR ⫽ 1.18, 95% CI ⫽ 1.12 to 1.25, p ⬍ 0.001 per IMT SD ⫽ 0.26). Adjusting only for age and gender and using outcomes and exclusions similar to a prior CHS report (with less follow-up and outcomes restricted to myocardial infarction or stroke) by O’Leary et al. (11), the HR associated with a 1 SD increment in maximum carotid IMT was 1.32 (95% CI 1.25 to 1.40), compared with 1.35 (95% CI 1.25 to 1.45) reported by O’Leary et al. (11). Additional adjustment as performed by O’Leary et al. (11) yielded a HR per 1 SD increment in maximum carotid IMT of 1.23 (95% CI 1.15 to 1.31), compared with 1.27 (95% CI 1.17 to 1.38) reported by O’Leary et al. (11).
DISCUSSION Previous CHS analyses reported a positive association between IMT, a single parameter of carotid structure, and the incidence of new CHD or CeVD in subjects 65 years of age and older who did not have clinical manifestations of CVD (11). In the present study, we have shown that specific carotid structural geometric patterns are associated with the development of new cardiovascular events, independently of age, gender, traditional risk factors, and carotid IMT. The present observation linking large vessel geometry to the risk of CVD, thus adding an additional perspective regarding the impact of IMT as an independent risk factor for CVD. That the association of specific variant CGPs with subsequent cardiovascular events differed in subjects by presence of clinical cardiovascular disease at baseline is noteworthy. The CGP2 did not differ from CGP1 in its association with future events and showed a trend towards lower risk of subsequent cardiovascular events in subjects with prevalent CVD. In contrast, the “hypertrophic patterns,” such as CGP3 and CGP4, were associated with increased risk of incident cardiovascular events in subjects without prevalent CVD. The nonsignificant association of IMT with CVD in the multivariable models reported in the present study does not contradict previous results in the same CHS population (11), because CGP3 and CGP4 represent participants in the upper quartile of the IMT distribution. That both CGP3 and CGP4 are significantly related to outcome supports the association of IMT with a more broadly defined CVD outcome. When CGP and carotid diameter were removed from the model for those free of CVD, the association of carotid IMT with risk of myocardial infarction and stroke was almost identical to that of O’Leary et al.
192
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Table 5. Predictors of Subsequent Cardiovascular Events A. Among Those Free of CVD at Baseline
Age (per 1 standard deviation ⫽ 5.58 yrs) Female gender (vs. male) Family history of MI (yes vs. no) BMI (per 1 standard deviation ⫽ 4.71 kg/m2) SBP (per 1 standard deviation ⫽ 21.83 mm Hg) Anti-hyptertensive medication use (yes vs. no) Ankle-arm index ⬍ 0.9 (yes vs. no) Creatinine (per 1 SD ⫽ 0.40 mg/dl) Atrial fibrillation on ECG (yes vs. no) Diabetes Glucose intolerant vs. normal New or known diabetes vs. normal Average carotid IMT (per 1 SD ⫽ 0.26 mm) Average carotid diameter (per 1 SD ⫽ 0.98 mm) Carotid structural geometric patterns CGP2 vs. CGP1 CGP3 vs. CGP1 CGP4 vs. CGP1
HR
95% CI
p
1.32 0.85 1.14 0.99 1.19 1.13 1.35 1.09 2.15
1.25–1.40 0.74–0.97 1.02–1.29 0.93–1.05 1.13–1.26 1.01–1.27 1.13–1.60 1.03–1.14 1.59–2.89
⬍ 0.001 0.013 0.027 0.692 ⬍ 0.001 0.039 0.001 0.001 ⬍ 0.001 ⬍ 0.001
1.24 1.56 1.06 1.11
1.06–1.46 1.34–1.81 0.98–1.17 1.03–1.12
1.12 1.25 1.43
0.83–1.50 1.03–1.53 1.16–1.75
0.158 0.006 0.005
B. Among Those With a History of CVD at Baseline
Age (per 1 standard deviation ⫽ 5.58 yrs) Female gender (vs. male) Family history of MI (yes vs. no) BMI (per 1 SD ⫽ 4.71 kg/m2) SBP (per 1 SD ⫽ 21.83 mm Hg) Anti-hyptertensive medication use (yes vs. no) Ankle-arm index ⬍ 0.9 (yes vs. no) Creatinine (per 1 SD ⫽ 0.40 mg/dl) Atrial fibrillation on ECG (yes vs. no) Diabetes Glucose intolerant vs. normal New or known diabetes vs. normal Average carotid IMT (per 1 SD ⫽ 0.26 mm) Average carotid diameter (per 1 SD ⫽ 0.98 mm) Carotid structural geometric patterns CGP2 vs. CGP1 CGP3 vs. CGP1 CGP4 vs. CGP1
HR
95% CI
p
1.15 0.78 1.05 1.02 0.96 1.37 1.32 1.07 1.45
1.07–1.25 0.65–0.94 0.90–1.23 0.94–1.10 0.88–1.04 1.14–1.65 1.08–1.62 1.01–1.12 1.05–2.03
⬍ 0.001 0.007 0.534 0.689 0.258 0.001 0.007 0.018 0.027 ⬍ 0.001
1.18 1.58 1.02 1.13
0.93–1.48 1.32–1.90 0.90–1.16 1.01–1.25
0.64 0.94 1.04
0.37–1.09 0.70–1.26 0.79–1.37
0.729 0.026 0.406
CI ⫽ confidence interval; HR ⫽ heart rate. Other abbreviations as in Table 1.
(11), despite the longer follow-up period of the present analysis. In addition, modeling of both IMT and lumen diameter showed that both are predictive of incident CVD, extending the results of the previous paper. Our findings can also raise the issue of whether the CGP classification provides additional information on the relationship between carotid IMT and diameter. Because IMT is used to calculate both vascular mass and W/L, interpretation of mathematical combinations of these terms is difficult. Qualitative CGPs may offer a significant advantage in this regard. CGP3 and CGP4 capture participants in the upper quartile of IMT; CGP3 includes subjects with high IMT and normal or narrowed vessels and CGP4 subjects with high IMT and larger lumens. It is thus not surprising that when the CGP groups are modeled, IMT is no longer significant. Diameter remains significant because its effect is
never directly modeled in the numerator of any of the defined CGPs. It is difficult to speculate about potential mechanisms by which specific CGPs may be associated with cardiovascular events. The altered risk of CVD associated with variant CGPs may relate, in part, to differences in functional characteristics of CGPs that have previously been defined, that is, to specific functional characteristics impacted by the interplay of IMT and diameter (10). Although all deviant CGPs presented a thicker IMT and higher circumferential wall stress than normals (CGP1), concentric remodeling (CGP2) was characterized by similar central and brachial BP levels, pulse wave velocity and augmentation index, and higher (protective) shear stress compared with CGP1. Conversely, beyond their specific profiles, both hypertrophy (CGP3) and hypertrophy with dilation (CGP4) were char-
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
acterized by a stiffer vasculature (higher pulse wave velocity and augmentation index, lower distensibility) and a lower shear stress than normals. Thus, although we cannot conclude at present that CGP classification can be a diagnostic tool, clinically superior to carotid IMT or diameter measurement, the results of the present study associate the relevance of vessel geometry with the risk of cardiovascular events in older persons. It has long been recognized that the geometry and functional properties of large vessels becomes altered by aging per se, that is, in the absence of clinically defined vascular disease such as hypertension or atherosclerosis. Studies in healthy volunteers of the Baltimore Longitudinal Study (BLSA) cohort, uniquely and rigorously screened to exclude hidden and clinical coronary artery disease, have defined age-associated arterial changes that occur in the absence of clinical disease. These changes include an increase in IMT and vascular lumen, and reduced compliance (increased stiffness) (23). Thus, interactions between the subclinical vascular aging process and pathophysiologic mechanisms that underlie disease processes, including atherosclerosis or hypertension, may determine the clinical threshold, severity, and prognosis of specific vascular diseases (24,25). In summary, variant carotid geometric patterns are associated with the development of new cardiovascular events, independent of age, traditional risk factors, and various measures of subclinical vascular disease, including carotid IMT. Specific carotid geometric patterns associated with a higher risk of developing new cardiovascular events, such as CGP3 or hypertrophy and CGP4 or hypertrophy with dilation in subjects free of CVD at baseline. Therefore, the underlying processes affecting rearrangement of the geometry of large arteries should be explored for potential insights they may provide in the diagnosis and prevention of CVD. Reprint requests and correspondence: Dr. Angelo Scuteri, Laboratory of Cardiovascular Science, National Institute on AgingNIH, 5600 Nathan Shock Drive, Baltimore, Maryland 21224. E-mail: [email protected].
REFERENCES 1. Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling. N Engl J Med 1994;330:1431–8. 2. Folkow B. Physiological aspects of primary hypertension. Physiol Rev 1982;62:347–504. 3. Mulvany MM, Aalkjaer C. Structure and function of small arteries. Physiol Rev 1990;70:921–61. 4. Schriffin EL. Reactivity of small blood vessel in hypertension: relationship with structural changes. Hypertension 1992;19 Suppl II: II1–9. 5. Baumbach GL, Heistad DD. Remodeling of cerebral arterioles in chronic hypertension. Hypertension 1989;13:968 –72.
Scuteri et al. Carotid Geometry and Cardiovascular Events
193
6. Mourad JJ, Girerd X, Boutouyrie P, et al. Opposite effects of remodeling and hypertrophy on arterial compliance in hypertension. Hypertension 1998;31:529 –33. 7. Burke GL, Evans GW, Riley WA, et al. Arterial wall thickness is associated with prevalent cardiovascular disease in middle-aged adult: the Atherosclerosis Risk In Communities (ARIC) Study. Stroke 1995;26:386 –91. 8. Poli A, Tremoli E, Colombo A, et al. Ultrasonographic measurements of the common carotid artery wall thickness in hypercholesterolemic patients: a new model for the quantitation and follow-up of preclinical atherosclerosis in living human subjects. Atherosclerosis 1988;70:253– 61. 9. O’Leary DH, Polak JF, Kronmal RA, et al. Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study. Stroke 1992;23:1752–60. 10. Scuteri A, Chen CH, Yin FCP, et al. Functional correlates of central arterial geometric phenotypes. Hypertension 2001;38:1471–5. 11. O’Leary DH, Polak JF, Kronmal RA, et al. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. N Engl J Med 1999;340:14 –22. 12. Tell GS, Fried LP, Hermanson BH, et al. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol 1993;3:358 –66. 13. Price TR, Psaty B, O’Leary D, et al. Assessment of cerebrovascular disease in the Cardiovascular Health Study. Ann Epidemiol 1993;3: 504 –7. 14. Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991;1:263–6. 15. Bouthier JD, Delucas N, Safar ME, et al. Cardiac hypertrophy and arterial distensibility in essential hypertension. Am Heart J 1985;109: 1345–52. 16. Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and ascertainment of cardiovascular events: the Cardiovascular Health Study. Ann Epidemiol 1995;5:278 –85. 17. O’Leary DH, Polak JF, Wolfson SK Jr., et al. Use of sonography to evaluate carotid atherosclerosis in the elderly: the Cardiovascular Health Study. Stroke 1991;22:1155–63. 18. Nichols WW, O’Rourke M. MacDonald’s Blood Flow in Arteries. London: Edward Arnold, 1990:283–329. 19. Girerd X, Mourad JJ, Copie X, et al. Noninvasive detection of an increased vascular mass in untreated hypertensive patients. Am J Hypertens 1994;7:1076 –84. 20. Manolio TA, Kronmal RA, Burke GL, et al. Short-term predictors of incident stroke in older adults—the Cardiovascular Health Study. Stroke 1996;27:1479 –86. 21. Psaty BM, Furberg CD, Kuller LH, et al. Traditional risk factors and subclinical disease measures as predictors of first myocardial infarction in older adults: the Cardiovascular Health Study. Arch Intern Med 1999;159:1339 –47. 22. Fleg JL, Gerstenblith G, Zonderman AB, et al. Prevalence and prognostic significance of exercise-induced silent myocardial ischemia detected by thallium scintigraphy and electrocardiography in asymptomatic volunteers. Circulation 1990;81:428 –36. 23. Lakatta EG, Levy D. Arterial and cardiac aging: major shareholders in cardiovascular disease enterprises. Part I: aging arteries: a “set up” for vascular disease. Circulation 2003;107:139 –46. 24. Nagai J, Metter EJ, Earley CJ, et al. Increased carotid artery intimalmedial thickness in asymptomatic older subjects with exercise-induced myocardial ischemia. Circulation 1998;98:1504 –9. 25. Brooks WW, Bing OHL, Conrad CN, et al. Captopril modifies gene expression in hypertrophied and failing hearts of aged spontaneously hypertensive rats. Hypertension 1997;30:1362–8.
APPENDIX For a full list of the participating CHS Investigators and Institutions, please see the January 21, 2004, issue of JACC at www.cardiosource.com/jacc.html.
Vol. 43, No. 2, 2004 ISSN 0735-1097/04/$30.00 doi:10.1016/j.jacc.2003.08.035
Prevalence of Specific Variant Carotid Geometric Patterns and Incidence of Cardiovascular Events in Older Persons The Cardiovascular Health Study (CHS E-131) Angelo Scuteri, MD, PHD,* Teri A. Manolio, MD, PHD,† Emily K. Marino, MS,‡ Alice M. Arnold, PHD,‡ Edward G. Lakatta, MD* Baltimore and Bethesda, Maryland; and Seattle, Washington We hypothesized that variant geometric patterns of the common carotid artery (CCA) predict the incidence of cardiovascular disease (CVD), after accounting for CCA intimamedial thickness (IMT). BACKGROUND Common carotid artery intima-media thickness has been associated with the incidence of cardiovascular disease. METHODS Noninvasive measurements of IMT were made with high-resolution ultrasonography in 5,640 subjects 65 years of age or older participating in the Cardiovascular Health Study. New coronary and/or cerebrovascular events served as outcome variables over a median 10.2-year follow-up. To characterize different carotid structural geometric patterns (CGP), vascular mass (VM) was combined with the wall-to-lumen ratio (W/L). Normal values for W/L and VM were defined as age-adjusted, gender-specific 75th percentiles of the 1,899 normotensive subjects free of CVD at baseline. Four CGPs were defined: CGP1 ⫽ normal W/L ratio and VM; CGP2 ⫽ arterial remodeling (i.e., increased W/L ratio with normal VM); CGP3 ⫽ arterial hypertrophy (i.e., increased W/L ratio with increased VM); and CGP4 ⫽ arterial hypertrophy with dilation (i.e., normal W/L ratio and increased VM). RESULTS Coronary or cerebrovascular events (adjusted for age, gender, traditional risk factors, and IMT) were associated with CGP in subjects free of CVD at baseline. Specifically, the hazard ratio (Cox proportional-hazards analyses) for CGP3 (arterial hypertrophy) was 1.25 (95% confidence interval [CI] 1.03 to 1.53), and for CGP4 (arterial hypertrophy with dilation) was 1.43 (95% CI 1.16 to 1.75) compared with CGP1 (normal). CONCLUSIONS Arterial hypertrophy defined by variant CGP patterns is associated with the development of new CVD, independent of age, traditional risk factors, and CCA IMT. (J Am Coll Cardiol 2004;43:187–93) © 2004 by the American College of Cardiology Foundation OBJECTIVES
The blood vessel wall is an active organ continuously undergoing remodeling (1). The result is a geometric pattern that cannot be described by only a unidimensional variable, such as arterial wall thickness or diameter, but can be characterized by combining a measure of vascular mass with the wall-to-lumen ratio (W/L) similar to small-artery and arteriolar morphologic changes previously distinguished in hypertension as “remodeling” or “hypertrophy” (2– 4) and whose differential functional correlates have been described (5,6). Although this concept of small-vessel geometric remod-
From the *Laboratory of Cardiovascular Sciences, National Institute on Aging, National Institutes of Health, Baltimore, Maryland; †Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Bethesda, Maryland; and the ‡Collaborative Health Studies Coordinating Center, Department of Biostatistics, University of Washington, Seattle, Washington. The research reported in this article was supported by contracts N01-HC-85079 through N01-HC-85086, N01-HC-35129, and N01 HC-15103 from the National Heart, Lung, and Blood Institute. Dr. Scuteri is currently affiliated with UO Geriatria, INRCA, Rome, Italy. The Cardiovascular Health Study (CHS) was conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with the CHS Investigators. This manuscript has been reviewed by CHS and by NHLBI for scientific content and consistency of data interpretation with previous CHS publications; significant comments have been incorporated prior to submission for publication. Manuscript received February 14, 2003; revised manuscript received August 18, 2003, accepted August 25, 2003.
eling may have substantial pathophysiologic and therapeutic relevance (1), the geometric patterns of large arteries have not been examined as risk factors for cardiovascular disease (CVD). Traditional risk factors of coronary and cerebrovascular disease (CeVD) cause alterations in vascular structure and function (7–9) that increase the likelihood of a subsequent clinical event. We recently observed that specific vascular geometric patterns of the common carotid artery (CCA) were associated with characteristic profiles of vascular functional measurements such as distensibility, circumferential stress, and strain in a cross-sectional population-based survey in 1,315 normotensive and untreated hypertensive Taiwanese men and women (10). Although these functional profiles differed quantitatively in normotensive compared with hypertensive subjects and in younger compared with older subjects, the characteristic functional profile of a given carotid geometric pattern (CGP) was preserved regardless of age or blood pressure (BP) status. We hypothesized that the geometric rearrangement of large arteries may play a role in the genesis of cardiovascular events, and may add predictive information to traditional cardiovascular risk factors and subclinical vascular disease defined by unidimensional vascular measures such as intimal-medial thickness (IMT) or lumen diameter (7,9,11).
188
Scuteri et al. Carotid Geometry and Cardiovascular Events
Abbreviations and Acronyms BP ⫽ blood pressure CCA ⫽ common carotid artery CeVD ⫽ cerebrovascular disease CGP ⫽ carotid geometric pattern CHD ⫽ coronary heart disease CHS ⫽ Cardiovascular Health Study CI ⫽ confidence interval CVD ⫽ cardiovascular disease HR ⫽ hazard ratio IMT⫽ intima-media thickness SD ⫽ standard deviation VM ⫽ vascular mass W/L ⫽ wall-to-lumen ratio
METHODS Subjects and study design. The study subjects were participants in the Cardiovascular Health Study (CHS), a prospective multicenter study of men and women 65 years of age or older sponsored by the National Heart, Lung, and Blood Institute (12). A detailed description of the recruitment methods has been published elsewhere (13). Of the members of the original cohort, 94% were white and most of the rest were black. A second cohort of 687 black participants was enrolled between June 1992 and June 1993, three years after the enrollment of the original CHS cohort. The baseline examination for both cohorts included a medical history, physical examination, laboratory testing, and assessment of CVD status. The study design, qualitycontrol procedures, laboratory methods, and BP measurement have been previously reported (12,14,15). The algorithms for classifying new myocardial infarctions and strokes have also been published (16,17). The carotid arteries were evaluated with high-resolution B-mode ultrasonography (15). One longitudinal image of the CCA was acquired. Measurements were made at a central reading center by readers blind to all clinical information. The same readers were used for all readings. The maximal IMT of the CCA was defined as the mean of the maximal IMTof the far wall on both the left and right sides. The CCA diameter measurements were made in the portion of the CCA where the wall interfaces are parallel. The CCA diameter was the mean of the minimal distance between the near-wall intima-lumen interface to the farwall lumen-intima interface of the right and left CCAs. As in prior studies of heart (18,19) and small artery (2– 4) geometry, vascular mass (VM) was combined with the W/L to identify different structural CGPs. Normal values for W/L (W/L ⫽ 2 · carotid IMT/carotid diameter) and VM (VM ⫽ IMT2 ; where is the arterial wall density, i.e., 1.06 g/cm3) (20) were defined as those within the ageadjusted and gender-specific 75th percentile of the 1,899 normotensive subjects (BP ⬍140/90 mm Hg) without prevalent CVD. Four common CGP were defined: CGP1 ⫽ normal, with normal W/L ratio and VM; CGP2 ⫽
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
arterial remodeling, with increased W/L ratio with normal VM; CGP3 ⫽ arterial hypertrophy, with increased W/L ratio and increased VM; and CGP4 ⫽ arterial hypertrophy with dilation, with normal W/L ratio and increased VM. Statistical analysis. Prevalent CVD at first visit was defined as history of myocardial infarction, angina, coronary artery bypass surgery, coronary artery angioplasty, congestive heart failure, stroke, transient ischemic attack, or carotid endarterectomy confirmed by review of medical records. Incident CVD was defined as coronary heart disease (CHD), including angina pectoris, myocardial infarction, angioplasty, or aortocoronary bypass surgery; congestive heart failure; and CeVD, including stroke or transient ischemic attack; as confirmed by review of medical records. New definite CHD and CeVD (combined cardiovascular end points) were considered the end points for the present analyses. Participants with a history of claudication were excluded from analysis and those with incident claudication were censored at the time of the claudication event. Subclinical disease for this analysis was defined as carotid stenosis ⱖ75% and ankle-arm index ⬍0.9. Hypertension was defined as BP ⱖ140/90 mm Hg or a reported history of high BP in combination with use of antihypertensive medications. Diabetes was defined by the American Diabetes Association criteria: glucose intolerance was defined by fasting glucose ⬎110 mg/dl and diabetes was defined by fasting glucose ⬎126 mg/dl or use of insulin or oral hypoglycemics. All analyses were performed with SPSS statistical software (SPSS Inc., Chicago, Illinois). Mean values of demographics and CVD risk factors were calculated across CGPs by means of one-way analysis of variance. Significance of pair-wise comparisons was adjusted for multiple comparisons using a Bonferroni correction. Categorical variables were compared by a chi-square test. We evaluated the associations of incident and recurrent CVD and the CGP groups using Cox proportional-hazards regression models, adjusting for traditional risk factors, CCA IMT and diameter, and subclinical disease. Atrial fibrillation and creatinine levels were also introduced into the multivariable models, given that previous analyses in the CHS population showed these to be risk factors for CHD or CeVD (21,22). Prediction of subsequent cardiovascular events was analyzed separately in subjects without prevalent CVD (incident events) and in subjects with prevalent CVD at baseline (recurrent events). Participants were censored at the time of incident claudication, death, or loss to follow-up. Unadjusted Cox proportional-hazards regression was used to plot curves for incident CVD or CeVD and compare survival rates among the CGP groups.
RESULTS The population sample consisted of 4,291 (76%) subjects without prevalent CVD at baseline and 1,349 (24%) subjects with prevalent CVD. The baseline characteristics of
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93 Table 1. Baseline Characteristics of the Study Population Age (yrs) Female Family history of MI Current smoking BMI (kg/m2) Fasting glucose (mg/dl) Diabetes Normal IFG New or known diabetes Total cholesterol (mg/dl) HDL cholesterol (mg/dl) History of CHD History of CeVD History of either CHD or CeVD (CVD) Hypertension Controlled hypertension Use of antihypertensive medications SBP (mm Hg) DBP (mm Hg) PP (mm Hg) Creatinine (mg/dl) Atrial fibrillation on ECG Maximum common carotid IMT (mm) Average common carotid IMT (mm) Average common carotid diameter (mm) Carotid stenosis ⬎75% Ankle-arm index ⬍0.9 Normal (CGP1) Arterial concentric remodeling (CGP2) Arterial hypertrophy (CGP3) Arterial hypertrophy with dilation (CGP4)
72.8 (5.6) 3,283 (58.2) 1,632 (28.9) 653 (11.6) 26.7 (4.7) 110.7 (36.0) 3,935 (69.8) 751 (13.3) 894 (15.9) 211.3 (39.2) 54.4 (15.7) 1174 (20.8) 310 (5.5) 1,349 (23.9) 3,280 (58.2) 1,075 (19.1) 2,629 (46.6) 136.41 (21.83) 70.83 (11.34) 65.53 (18.67) 1.06 (0.40) 143 (2.5) 1.06 (0.21) 1.04 (0.26) 9.23 (0.98) 50 (0.9) 656 (11.6) 3,429 (60.8) 248 (4.4) 1,420 (25.2) 543 (9.6)
Entries are mean (SD) or N (%). Total: N ⫽ 5,640. BMI ⫽ body mass index; CeVD ⫽ cerebrovascular disease; CGP ⫽ carotid geometric pattern; CHD ⫽ coronary heart disease; CVD ⫽ cardiovascular disease; DBP ⫽ diastolic blood pressure; ECG ⫽ electrocardiogram; HDL ⫽ high-density lipoprotein; IFG ⫽ impaired fasting glucose; IMT ⫽ intima-media thickness; MI ⫽ myocardial infarction; PP ⫽ pulse pressure; SBP ⫽ systolic blood pressure.
the 5,640 study participants are given in Table 1. The median follow-up was 10.2 years (maximum was 11.1 years). A total of 1,878 (33.3%) participants had coronary events and 835 (14.8%) participants had cerebrovascular events during follow-up. Of the subjects without prevalent CVD, 1,453 (33.9%) experienced a new event during the follow-up time (1,210 [28.2%] CHD event and 720 [16.8%] CeVD event), whereas 749 (56%) of those with prevalent CVD experienced a subsequent event (668 [49.5%] CHD event and 115 [8.5%] CeVD event). The combined CVD end point was used after initial separate analyses with CHD and CeVD showed similar results. Prevalence of CGP. The prevalence of the four measured CGPs in the entire population is listed in Table 2. Subjects with prevalent CVD had a higher frequency of variant CGPs compared with those without prevalent CVD (44.6% vs. 37.5%, respectively, p ⬍ 0.001). Prevalent CVD was associated with a lower prevalence of normal (CGP1) and concentric remodeling (CGP2) and with a higher prevalence of hypertrophy (CGP3) and hypertrophy with dilation (CGP4) compared with subjects without prevalent CVD (Table 3).
189
Table 2. Prevalence of CGP by Gender
CGP1 CGP2 CGP3 CGP4 Total variant CGP
Men
Women
1,413 (59.9) 106 (4.5) 603 (25.6) 235 (10.0) 944 (40.1)
2,016 (61.4) 142 (4.3) 817 (24.9) 308 (9.4) 1,267 (38.6)
Chi-square 1.224: p value 0.269 for comparison of total variant CGP prevalence between men and women. CGP ⫽ carotid geometric pattern.
The demographic and risk factor profiles of each CGP are summarized in Table 4. Mean carotid IMT was higher for each variant CGP compared with CGP1, with a progressive increase in IMT values from arterial concentric remodeling (CGP2) to arterial hypertrophy with dilation (CGP4) to arterial hypertrophy (CGP3). Although carotid IMT was increased in subjects with concentric remodeling (CGP2) compared with CGP1, blood pressure values, glucose, and cholesterol levels were similar in both these patterns, and hypertension prevalence in concentric remodeling (CGP2) was lower than in CGP1. New CVD by CGP. As shown in Figure 1, in subjects without prevalent CVD, the rate of new combined cardiovascular events was not significantly different in CGP1 and concentric remodelling (CGP2), but was higher in arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4) than in normal geometry (CGP1) or concentric remodeling (CGP2). In subjects with prevalent CVD, the rate of new combined cardiovascular end points was much Table 3. Prevalence of Variant CGP by Prevalent CHD, CeVD, and CVD (CHD or CeVD)
CGP1* CGP2* CGP3* CGP4* Total variant CGP*
CGP1† CGP2† CGP3† CGP4† Total variant CGP†
CGP1‡ CGP2‡ CGP3‡ CGP4‡ Total variant CGP‡
No Prevalent CHD (n ⴝ 4,466)
Prevalent CHD (n ⴝ 652)
2,777 (62.2) 210 (4.7) 1,078 (24.1) 401 (9.0) 1,689 (37.8)
652 (55.5) 38 (3.2) 342 (29.1) 142 (12.1) 522 (44.5)
No Prevalent CeVD (n ⴝ 5,330)
Prevalent CeVD (n ⴝ 310)
3,275 (61.4) 240 (4.5) 1,311 (24.6) 504 (9.5) 2,055 (38.6)
154 (49.7) 8 (2.6) 109 (35.2) 39 (12.6) 156 (50.3)
No Prevalent CVD (n ⴝ 4,291)
Prevalent CVD (n ⴝ 747)
2,682 (62.5) 207 (4.8) 1,020 (23.8) 382 (8.9) 1,609 (37.5)
747 (55.4) 41 (3.0) 400 (29.7) 161 (11.9) 602 (44.6)
*Chi-square 17.219, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. †Chi-square 17.020, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. ‡Chi-square 21.883, p value ⬍ 0.001 for comparison of total variant CGP prevalence by baseline disease status. Abbreviations as in Table 1.
190
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Table 4. Demographic and CVD Risk Profile by CGP Status (Means ⫾ SE or %)
Age (yrs) Female (%) Family history of MI (%) Current smoking (%) BMI (kg/m2) Fasting glucose (mg/dl) Diabetes (%) Total cholesterol (mg/dl) HDL cholesterol (mg/dl) Pre-existing CHD (%) Prevalent CeVD (%) Prevalent CVD (%) Hypertension (%) Controlled hypertension (%) Use of anti-hypertensive medications SBP (mm Hg) DBP (mm Hg) PP (mm Hg) Creatinine (mg/dl) Atrial fibrillation on ECG (%) Carotid stenosis ⬎75% (%) Ankle-arm index ⬍0.9 (%) Maximum carotid IMT (mm) Average carotid IMT (mm) Carotid diameter (mm) W/L ratio Vascular mass (mm2)
CGP1 Normal (n ⴝ 3429) (60.8%)
CGP2 Remodelling (n ⴝ 248) (4.4%)
CGP3 Hypertrophy (n ⴝ 1,420) (25.2%)
CGP4 Hypertrophy With Dilation (n ⴝ 543) (9.6%)
72.8 ⫾ 0.1 58.8 31.2 10.4 26.4 ⫾ 0.1 107.9 ⫾ 0.5 13.1 209.8 ⫾ 0.7 55.3 ⫾ 0.3 19.0 4.5 21.8 54.0 19.0 44.2 134.2 ⫾ 0.4 70.6 ⫾ 0.2 63.6 ⫾ 0.3 1.05 ⫾ 0.33 2.3 0.5 9.7 0.96 ⫾ 0.002 0.89 ⫾ 0.002 9.15 ⫾ 0.02 0.198 ⫾ 0.0003 2.70 ⫾ 0.01
72.1 ⫾ 0.3 57.3 29.9 7.3 25.9 ⫾ 0.2 107.1 ⫾ 1.9 11.3 213.7 ⫾ 2.3 56.7 ⫾ 1.1 15.3 3.2 16.5 44.4 15.7 38.7 131.8 ⫾ 1.2 71.1 ⫾ 0.7 60.7 ⫾ 1.1 1.04 ⫾ 0.24 2.9 0 7.7 1.00 ⫾ 0.007 1.01 ⫾ 0.004 8.16 ⫾ 0.03 0.248 ⫾ 0.001 3.41 ⫾ 0.03
72.9 ⫾ 0.2 57.5 32.6 14.0 27.1 ⫾ 0.1 115.8 ⫾ 1.2 22.1 215.4 ⫾ 1.1 52.4 ⫾ 0.4 24.1 7.7 28.2 65.8 19.6 50.2 140.2 ⫾ 0.6 70.9 ⫾ 0.3 69.3 ⫾ 0.5 1.09 ⫾ 0.53 2.7 1.6 17.7 1.265 ⫾ 0.006 1.348 ⫾ 0.007 9.44 ⫾ 0.02 0.286 ⫾ 0.001 6.31 ⫾ 0.08
72.8 ⫾ 0.1 56.7 32.3 14.7 27.8 ⫾ 0.2 117.1 ⫾ 1.9 20.9 209.2 ⫾ 1.7 53.1 ⫾ 0.6 26.2 7.2 29.7 71.2 19.9 56.7 142.5 ⫾ 1.0 72.1 ⫾ 0.5 70.3 ⫾ 0.8 1.07 ⫾ 0.41 3.3 1.7 12.4 1.156 ⫾ 0.006 1.135 ⫾ 0.004 10.26 ⫾ 0.04 0.222 ⫾ 0.001 4.32 ⫾ 0.03
p < 0.05 n.s. n.s. n.s. 3,4 ⬎ 1,2 4 ⬎ 3 ⬎ 1,2 3,4 ⬍ 1,2 3,4 ⬍ 1,2 3 ⬎ 1,4 1,2 ⬎ 3,4 3,4 ⬎ 1,2 3 ⬎ 1,2 3,4 ⬎ 1,2 3,4 ⬎ 1 ⬎ 2 n.s. 3,4 ⬎ 1,2 3,4 ⬎ 1,2 4⬎1 3,4 ⬎ 1,2 3⬎1 n.s. 3⬎1 3 ⬎ 1,2,4 3⬎4⬎2⬎ 3⬎4⬎2⬎ 4⬎3⬎1⬎ 3⬎2⬎4⬎ 3⬎4⬎2⬎
1 1 2 1 1
All p values are adjusted for all six two-way comparisons using Bonferroni’s method. W/L ⫽ wall-to-lumen ratio; other abbreviations as in Tables 1 and 3.
lower in concentric remodeling (CGP2) as compared with CGP1, and not significantly higher in arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4) as compared with CGP1.
Unadjusted Cox proportional hazard-regression in subjects without prevalent CVD confirmed a similar rate of events for CGP2 and CGP1 and a higher rate of events for CGP3 and CGP4. Event rates started to diverge from
Figure 1. Cumulative rate for incident and/or recurrent events over the 10.2 years follow-up according to carotid geometric pattern (CGP) status. CVD ⫽ cardiovascular disease.
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Figure 2. Unadjusted incidence of coronary heart disease or cerebrovascular disease according to carotid geometric pattern (CGP) status and presence or absence of prevalent cardiovascular disease (CVD).
CGP1 after approximately one year (Fig. 2). In subjects with prevalent CVD, unadjusted Cox proportional hazardsregression showed a much lower rate of recurrent events for concentric remodeling (CGP2) and a slightly higher rate of recurrent events for arterial hypertrophy (CGP3) and arterial hypertrophy with dilation (CGP4). Rates started to diverge very early for concentric remodeling (CGP2) and divergence increased with time (Fig. 2). Multivariable Cox proportional hazard models in only those subjects without prevalent CVD showed that whereas CGP2 did not differ from CGP1, the variant “hypertrophic” carotid geometric patterns (CGP3 and CGP4) persisted as independent predictors of new combined cardiovascular events, after taking into account traditional risk factors, carotid IMT and diameter, and subclinical disease (Table 5). Specifically, hypertrophy (CGP3) was associated with a 25% (hazard ratio [HR] ⫽ 1.25, 95% confidence interval [CI] ⫽ 1.03 to 1.53) higher risk of new CVD, and hypertrophy with dilation (CGP4) with a 43% (HR ⫽ 1.43, 95% CI ⫽ 1.16 to 1.75) higher risk (Table 5). In multivariable Cox proportional hazard analyses limited to subjects with prevalent CVD, the CGP groups were not associated with increased risk of recurrent events after adjustment for traditional risk factors, carotid IMT, and subclinical disease (Table 5). Secondary analyses for more specific end points showed that when angina pectoris was deleted from the definition of CHD the results were similar to those presented in Table 5. The CGP3 and CGP4 were associated with incident myocardial infarction and a combined outcome of either myocardial infarction or congestive heart failure. When CGP was removed from the model for those free of CVD at baseline, and IMT, lumen diameter, and the product (interaction) of the IMT and diameter were included, the interaction term was significant (p ⬍ 0.03). The HR of incident CVD associated with a 1 standard deviation (SD) increment in IMT (0.26 mm) at the mean value of lumen diameter was 1.16 (95% CI 1.10 to 1.23), and the HR of a 1 SD increment in lumen diameter (0.98 mm) at the mean value of IMT was 1.15 (95% CI 1.08 to 1.24).
Scuteri et al. Carotid Geometry and Cardiovascular Events
191
The coefficient of the interaction term was negative, suggesting an increased risk of CVD associated with a 1 SD increase in IMT when the diameter is below the mean and vice versa. When both are above their means, the negative interaction results in a somewhat less than additive effect of wall thickness and lumen diameter on the HR. When CGP and lumen diameter were removed from the model for those free of CVD, IMT was significantly associated with the risk of CVD (HR ⫽ 1.18, 95% CI ⫽ 1.12 to 1.25, p ⬍ 0.001 per IMT SD ⫽ 0.26). Adjusting only for age and gender and using outcomes and exclusions similar to a prior CHS report (with less follow-up and outcomes restricted to myocardial infarction or stroke) by O’Leary et al. (11), the HR associated with a 1 SD increment in maximum carotid IMT was 1.32 (95% CI 1.25 to 1.40), compared with 1.35 (95% CI 1.25 to 1.45) reported by O’Leary et al. (11). Additional adjustment as performed by O’Leary et al. (11) yielded a HR per 1 SD increment in maximum carotid IMT of 1.23 (95% CI 1.15 to 1.31), compared with 1.27 (95% CI 1.17 to 1.38) reported by O’Leary et al. (11).
DISCUSSION Previous CHS analyses reported a positive association between IMT, a single parameter of carotid structure, and the incidence of new CHD or CeVD in subjects 65 years of age and older who did not have clinical manifestations of CVD (11). In the present study, we have shown that specific carotid structural geometric patterns are associated with the development of new cardiovascular events, independently of age, gender, traditional risk factors, and carotid IMT. The present observation linking large vessel geometry to the risk of CVD, thus adding an additional perspective regarding the impact of IMT as an independent risk factor for CVD. That the association of specific variant CGPs with subsequent cardiovascular events differed in subjects by presence of clinical cardiovascular disease at baseline is noteworthy. The CGP2 did not differ from CGP1 in its association with future events and showed a trend towards lower risk of subsequent cardiovascular events in subjects with prevalent CVD. In contrast, the “hypertrophic patterns,” such as CGP3 and CGP4, were associated with increased risk of incident cardiovascular events in subjects without prevalent CVD. The nonsignificant association of IMT with CVD in the multivariable models reported in the present study does not contradict previous results in the same CHS population (11), because CGP3 and CGP4 represent participants in the upper quartile of the IMT distribution. That both CGP3 and CGP4 are significantly related to outcome supports the association of IMT with a more broadly defined CVD outcome. When CGP and carotid diameter were removed from the model for those free of CVD, the association of carotid IMT with risk of myocardial infarction and stroke was almost identical to that of O’Leary et al.
192
Scuteri et al. Carotid Geometry and Cardiovascular Events
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
Table 5. Predictors of Subsequent Cardiovascular Events A. Among Those Free of CVD at Baseline
Age (per 1 standard deviation ⫽ 5.58 yrs) Female gender (vs. male) Family history of MI (yes vs. no) BMI (per 1 standard deviation ⫽ 4.71 kg/m2) SBP (per 1 standard deviation ⫽ 21.83 mm Hg) Anti-hyptertensive medication use (yes vs. no) Ankle-arm index ⬍ 0.9 (yes vs. no) Creatinine (per 1 SD ⫽ 0.40 mg/dl) Atrial fibrillation on ECG (yes vs. no) Diabetes Glucose intolerant vs. normal New or known diabetes vs. normal Average carotid IMT (per 1 SD ⫽ 0.26 mm) Average carotid diameter (per 1 SD ⫽ 0.98 mm) Carotid structural geometric patterns CGP2 vs. CGP1 CGP3 vs. CGP1 CGP4 vs. CGP1
HR
95% CI
p
1.32 0.85 1.14 0.99 1.19 1.13 1.35 1.09 2.15
1.25–1.40 0.74–0.97 1.02–1.29 0.93–1.05 1.13–1.26 1.01–1.27 1.13–1.60 1.03–1.14 1.59–2.89
⬍ 0.001 0.013 0.027 0.692 ⬍ 0.001 0.039 0.001 0.001 ⬍ 0.001 ⬍ 0.001
1.24 1.56 1.06 1.11
1.06–1.46 1.34–1.81 0.98–1.17 1.03–1.12
1.12 1.25 1.43
0.83–1.50 1.03–1.53 1.16–1.75
0.158 0.006 0.005
B. Among Those With a History of CVD at Baseline
Age (per 1 standard deviation ⫽ 5.58 yrs) Female gender (vs. male) Family history of MI (yes vs. no) BMI (per 1 SD ⫽ 4.71 kg/m2) SBP (per 1 SD ⫽ 21.83 mm Hg) Anti-hyptertensive medication use (yes vs. no) Ankle-arm index ⬍ 0.9 (yes vs. no) Creatinine (per 1 SD ⫽ 0.40 mg/dl) Atrial fibrillation on ECG (yes vs. no) Diabetes Glucose intolerant vs. normal New or known diabetes vs. normal Average carotid IMT (per 1 SD ⫽ 0.26 mm) Average carotid diameter (per 1 SD ⫽ 0.98 mm) Carotid structural geometric patterns CGP2 vs. CGP1 CGP3 vs. CGP1 CGP4 vs. CGP1
HR
95% CI
p
1.15 0.78 1.05 1.02 0.96 1.37 1.32 1.07 1.45
1.07–1.25 0.65–0.94 0.90–1.23 0.94–1.10 0.88–1.04 1.14–1.65 1.08–1.62 1.01–1.12 1.05–2.03
⬍ 0.001 0.007 0.534 0.689 0.258 0.001 0.007 0.018 0.027 ⬍ 0.001
1.18 1.58 1.02 1.13
0.93–1.48 1.32–1.90 0.90–1.16 1.01–1.25
0.64 0.94 1.04
0.37–1.09 0.70–1.26 0.79–1.37
0.729 0.026 0.406
CI ⫽ confidence interval; HR ⫽ heart rate. Other abbreviations as in Table 1.
(11), despite the longer follow-up period of the present analysis. In addition, modeling of both IMT and lumen diameter showed that both are predictive of incident CVD, extending the results of the previous paper. Our findings can also raise the issue of whether the CGP classification provides additional information on the relationship between carotid IMT and diameter. Because IMT is used to calculate both vascular mass and W/L, interpretation of mathematical combinations of these terms is difficult. Qualitative CGPs may offer a significant advantage in this regard. CGP3 and CGP4 capture participants in the upper quartile of IMT; CGP3 includes subjects with high IMT and normal or narrowed vessels and CGP4 subjects with high IMT and larger lumens. It is thus not surprising that when the CGP groups are modeled, IMT is no longer significant. Diameter remains significant because its effect is
never directly modeled in the numerator of any of the defined CGPs. It is difficult to speculate about potential mechanisms by which specific CGPs may be associated with cardiovascular events. The altered risk of CVD associated with variant CGPs may relate, in part, to differences in functional characteristics of CGPs that have previously been defined, that is, to specific functional characteristics impacted by the interplay of IMT and diameter (10). Although all deviant CGPs presented a thicker IMT and higher circumferential wall stress than normals (CGP1), concentric remodeling (CGP2) was characterized by similar central and brachial BP levels, pulse wave velocity and augmentation index, and higher (protective) shear stress compared with CGP1. Conversely, beyond their specific profiles, both hypertrophy (CGP3) and hypertrophy with dilation (CGP4) were char-
JACC Vol. 43, No. 2, 2004 January 21, 2004:187–93
acterized by a stiffer vasculature (higher pulse wave velocity and augmentation index, lower distensibility) and a lower shear stress than normals. Thus, although we cannot conclude at present that CGP classification can be a diagnostic tool, clinically superior to carotid IMT or diameter measurement, the results of the present study associate the relevance of vessel geometry with the risk of cardiovascular events in older persons. It has long been recognized that the geometry and functional properties of large vessels becomes altered by aging per se, that is, in the absence of clinically defined vascular disease such as hypertension or atherosclerosis. Studies in healthy volunteers of the Baltimore Longitudinal Study (BLSA) cohort, uniquely and rigorously screened to exclude hidden and clinical coronary artery disease, have defined age-associated arterial changes that occur in the absence of clinical disease. These changes include an increase in IMT and vascular lumen, and reduced compliance (increased stiffness) (23). Thus, interactions between the subclinical vascular aging process and pathophysiologic mechanisms that underlie disease processes, including atherosclerosis or hypertension, may determine the clinical threshold, severity, and prognosis of specific vascular diseases (24,25). In summary, variant carotid geometric patterns are associated with the development of new cardiovascular events, independent of age, traditional risk factors, and various measures of subclinical vascular disease, including carotid IMT. Specific carotid geometric patterns associated with a higher risk of developing new cardiovascular events, such as CGP3 or hypertrophy and CGP4 or hypertrophy with dilation in subjects free of CVD at baseline. Therefore, the underlying processes affecting rearrangement of the geometry of large arteries should be explored for potential insights they may provide in the diagnosis and prevention of CVD. Reprint requests and correspondence: Dr. Angelo Scuteri, Laboratory of Cardiovascular Science, National Institute on AgingNIH, 5600 Nathan Shock Drive, Baltimore, Maryland 21224. E-mail: [email protected].
REFERENCES 1. Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling. N Engl J Med 1994;330:1431–8. 2. Folkow B. Physiological aspects of primary hypertension. Physiol Rev 1982;62:347–504. 3. Mulvany MM, Aalkjaer C. Structure and function of small arteries. Physiol Rev 1990;70:921–61. 4. Schriffin EL. Reactivity of small blood vessel in hypertension: relationship with structural changes. Hypertension 1992;19 Suppl II: II1–9. 5. Baumbach GL, Heistad DD. Remodeling of cerebral arterioles in chronic hypertension. Hypertension 1989;13:968 –72.
Scuteri et al. Carotid Geometry and Cardiovascular Events
193
6. Mourad JJ, Girerd X, Boutouyrie P, et al. Opposite effects of remodeling and hypertrophy on arterial compliance in hypertension. Hypertension 1998;31:529 –33. 7. Burke GL, Evans GW, Riley WA, et al. Arterial wall thickness is associated with prevalent cardiovascular disease in middle-aged adult: the Atherosclerosis Risk In Communities (ARIC) Study. Stroke 1995;26:386 –91. 8. Poli A, Tremoli E, Colombo A, et al. Ultrasonographic measurements of the common carotid artery wall thickness in hypercholesterolemic patients: a new model for the quantitation and follow-up of preclinical atherosclerosis in living human subjects. Atherosclerosis 1988;70:253– 61. 9. O’Leary DH, Polak JF, Kronmal RA, et al. Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study. Stroke 1992;23:1752–60. 10. Scuteri A, Chen CH, Yin FCP, et al. Functional correlates of central arterial geometric phenotypes. Hypertension 2001;38:1471–5. 11. O’Leary DH, Polak JF, Kronmal RA, et al. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. N Engl J Med 1999;340:14 –22. 12. Tell GS, Fried LP, Hermanson BH, et al. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol 1993;3:358 –66. 13. Price TR, Psaty B, O’Leary D, et al. Assessment of cerebrovascular disease in the Cardiovascular Health Study. Ann Epidemiol 1993;3: 504 –7. 14. Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991;1:263–6. 15. Bouthier JD, Delucas N, Safar ME, et al. Cardiac hypertrophy and arterial distensibility in essential hypertension. Am Heart J 1985;109: 1345–52. 16. Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and ascertainment of cardiovascular events: the Cardiovascular Health Study. Ann Epidemiol 1995;5:278 –85. 17. O’Leary DH, Polak JF, Wolfson SK Jr., et al. Use of sonography to evaluate carotid atherosclerosis in the elderly: the Cardiovascular Health Study. Stroke 1991;22:1155–63. 18. Nichols WW, O’Rourke M. MacDonald’s Blood Flow in Arteries. London: Edward Arnold, 1990:283–329. 19. Girerd X, Mourad JJ, Copie X, et al. Noninvasive detection of an increased vascular mass in untreated hypertensive patients. Am J Hypertens 1994;7:1076 –84. 20. Manolio TA, Kronmal RA, Burke GL, et al. Short-term predictors of incident stroke in older adults—the Cardiovascular Health Study. Stroke 1996;27:1479 –86. 21. Psaty BM, Furberg CD, Kuller LH, et al. Traditional risk factors and subclinical disease measures as predictors of first myocardial infarction in older adults: the Cardiovascular Health Study. Arch Intern Med 1999;159:1339 –47. 22. Fleg JL, Gerstenblith G, Zonderman AB, et al. Prevalence and prognostic significance of exercise-induced silent myocardial ischemia detected by thallium scintigraphy and electrocardiography in asymptomatic volunteers. Circulation 1990;81:428 –36. 23. Lakatta EG, Levy D. Arterial and cardiac aging: major shareholders in cardiovascular disease enterprises. Part I: aging arteries: a “set up” for vascular disease. Circulation 2003;107:139 –46. 24. Nagai J, Metter EJ, Earley CJ, et al. Increased carotid artery intimalmedial thickness in asymptomatic older subjects with exercise-induced myocardial ischemia. Circulation 1998;98:1504 –9. 25. Brooks WW, Bing OHL, Conrad CN, et al. Captopril modifies gene expression in hypertrophied and failing hearts of aged spontaneously hypertensive rats. Hypertension 1997;30:1362–8.
APPENDIX For a full list of the participating CHS Investigators and Institutions, please see the January 21, 2004, issue of JACC at www.cardiosource.com/jacc.html.