Correlation of Echocardiographic Findings With Cerebral Infarction in Elderly Adults
The AGES-Reykjavik Study
Background and Purpose—Chronic effects of hypertension may be observed in multiple end organs. Previous reports suggest that cardiovascular morphological features can mirror cerebral infarction. In this cross-sectional analysis of elderly subjects, we investigated the relationship of a comprehensive set of echocardiographic measures with cerebral infarction detected by MRI.
Methods—We compared echocardiographically determined left ventricular (LV) mass, left atrial volume, aortic root diameter, mitral annular calcification, and measures of diastolic function with cerebral infarction determined by MRI using logistic regression in a random sample drawn from the Age Gene/Environment Susceptibility-Reykjavik Study cohort. The model was first adjusted for age and gender, and then for age, gender, and vascular risk factors.
Results—Among 692 subjects aged 75 (standard deviation, 6) years, 28% had at least 1 cerebral infarct. When adjusted for age and gender, the presence of cerebral infarction was modestly related to LV mass (odds ratio [OR], 1.01; 95% confidence interval [CI], 1.00–1.02) and left atrial volume (OR, 1.03; 95% CI, 1.01–1.05), as well as the lowest quartile of early-to-late pulsed Doppler velocity ratio (early-to-late pulsed Doppler velocity ratio <0.75; OR, 1.87; 95% CI, 1.22–2.87). The latter relation remained significant after adjustment for vascular risk factors and LV ejection fraction (OR, 1.82; 95% CI, 1.16–2.86).
Conclusion—Of all echocardiographic parameters, LV filling abnormality as indicated by low early-to-late pulsed Doppler velocity ratio displayed the strongest association with cerebral infarction and this relationship was independent of vascular risk factors. This simple marker of cerebral infarction may be useful when evaluating older patients.
Chronic hypertension leads to concomitant remodeling of the cardiac and vascular systems. Atherosclerotic changes in the cerebrovascular system, ultimately leading to incident stroke, are mirrored by development of left ventricular (LV) hypertrophy, resultant diastolic dysfunction, and heart failure with preserved systolic function. Thus, assessment of ventricular morphology and function may be indicative of the extent of vascular disease in less easily accessible sites such as the brain. Although the association of LV hypertrophy detected by ECG with stroke outcome was described many years ago,1 advanced imaging techniques including 2-dimensional and Doppler echocardiography have largely supplanted the ECG in measuring left ventricular mass (LVM) and provided further insights into associated changes of ventricular function.
Echocardiographic predictors of vascular outcome focused initially on LV morphology. LV hypertrophy was found to reflect the severity and chronicity of systemic hypertension.2 In related fashion, LVM and other cardiovascular measures have been reported to predict stroke risk.3 In fact, in the Framingham Study, for each 50-g increase in LVM, there was ≈1.5-fold increase in relative risk for subsequent cardiovascular events.4 Other echo findings associated with chronic hypertension and stroke include mitral annular calcification height and enlargement of the left atrium (LA) and aortic root.5–8
Much previous epidemiological work investigating the cardio-cerebrovascular relationship was limited by its reliance on single-dimensional or M-mode technology. Two-dimensional and 3-dimensional assessment is now available to better-delineate cardiac morphology, as are sensitive measures of diastolic and systolic ventricular function that may be affected by chronic hypertension. The latter includes pulsed and tissue Doppler measures of ventricular relaxation and filling. In addition, previous work focused on clinical stroke data that necessarily failed to evaluate the presence of silent cerebral infarction. Importantly, MRI screening can detect not only previous clinical but also subclinical cerebral infarction,9 which is consequential because clinical and silent cerebral infarctions have been related to cognitive decline and dementia in older subjects.10–12
We tested the relationship of newer measures of cardiac morphology and function to cerebral infarction detected by MRI in a well-characterized community-based cohort of older subjects (AGES-Reykjavik).13
Patients and Methods
AGES-Reykjavik is an extension of the Reykjavik Study, a community-based cohort established in 1967 to study cardiovascular disease prospectively in Iceland. The rationale and design of AGES-Reykjavik, which is cosponsored by the Icelandic Heart Association and the National Institute of Aging, National Institutes of Health, has been described elsewhere.13 AGES-Reykjavik has been approved by the Icelandic National Bioethics Committee and the National Institute of Aging Institutional Review Board. Between 2002 and 2006, 5764 men and women participated in detailed phenotypic evaluations of cardiovascular, neurocognitive, musculoskeletal, and metabolic systems. Within this cohort, all eligible subjects were offered a cerebral MRI, and 954 subjects were selected randomly for cardiac assessment by echocardiography. The brain and cardiac evaluations were performed within a 1-month interval.
High-resolution MRI images were acquired on a 1.5-T Signa Twinspeed system (General Electric Medical Systems). The image protocol consisted of axial T1-weighted 3-dimensional spoiled-gradient echo, T2*-weighted gradient echo-type echo planar images, proton density/T2-weighted fast-spin echo, and fluid-attenuated inversion recovery sequences.
Cerebral infarcts were identified by trained radiographers as defects in brain parenchyma with associated hyperintensity on T2 and fluid-attenuated inversion recovery images with a maximal diameter of at least 4 mm, with the exception of cerebellar and brain stem infarcts or infarcts with cortical involvement, which had no size criterion.14 Five percent of all scans were re-read by blinded master readers at Leiden University Medical Center, The Netherlands, to assess inter-reader reliability. A set of scans was also re-read by all readers for presence or absence of parenchymal defects and to calculate average intra-reader and inter-reader reliabilities, which were good (weighted kappa=0.9 and 0.7, respectively).
Standard 2-dimensional pulsed and tissue Doppler imaging was performed in long and short parasternal and 3 apical views with standard equipment (Acuson Sequoia C512). Studies were acquired digitally using established imaging protocols15 during free breathing in the left lateral supine position and stored for off-line analysis (Digiview Workstation; Digisonics). Studies were read qualitatively by an experienced cardiologist. Linear point-to-point and area tracings were made from 2-dimensional studies by ultrasound technicians specially trained in image quantification. All Doppler studies were acquired at sweep speeds of 50 mm/sec. Subjects in atrial fibrillation during the study or with significant valvular heart disease were excluded.
LVM was calculated using the American Society of Echocardiography modification of the Penn formula.16 LA volume was determined by modified Simpson biplane method of discs, which included apical 4- and 2-chamber views.17 The presence and height of mitral annular calcification were measured from 2-dimensional images.5 LV ejection fraction was assessed qualitatively by an experienced cardiologist.
Assessment of Diastolic Function
All LV diastolic filling was assessed during apical 4-chamber imaging. The pulsed Doppler sample volume was placed at the mitral leaflet tips to determine transmitral blood velocities. The leading edge of the transmitral Doppler velocity profile was measured to derive peak early (E) and late atrial phase (A) LV filling and their ratios. Tissue Doppler imaging of mitral annular velocities (E′, A′) were measured from septal and lateral edges of the mitral annulus.
A 10% sample of all studies was selected randomly and reviewed both qualitatively and quantitatively by echocardiographic physicians at the National Heart, Lung, and Blood Institute, National Institutes of Health, who also provided training and quality oversight. Interobserver agreements were good for the overall measures, with Spearman correlation coefficients varying from 0.70 for measurement of LV wall septal thickness to 0.98 for early-to-late pulsed Doppler velocity (E/A) ratio.
We controlled for demographic and vascular risk factors associated with both cerebral infarction and cardiovascular disease. BMI was calculated as weight (kilograms) divided by height squared (meters). History of atrial fibrillation and use of anticoagulant or antiplatelet drugs were noted. Smoking status was determined by self-report and categorized as smoker (current or smoking within the past 12 months) or nonsmoker. Fasting HDL and LDL cholesterol levels were measured. Hypertension was defined as systolic blood pressure >140 mm Hg, diastolic blood pressure >90 mm Hg, or use of an antihypertensive medication. Diabetes was based on self-report, fasting plasma glucose concentration >7 mmol/L, or use of oral hypoglycemic medication or insulin. Previous coronary heart disease was defined as a documented history of coronary artery disease or coronary bypass surgery.
General characteristics of subjects with and without cerebral infarction were compared using logistic regression. Echocardiographic characteristics of subjects with and without cerebral infarction were analyzed as continuous variables and as quartiles using logistic regression. LVM and LA volume were indexed to body surface area18 to allow comparison independent of obesity.
The relation between E/A ratio and cerebral infarction was analyzed using E/A ratio as a continuous variable and in a quadratic model. Because of a U-shape relation between E/A ratio and mortality with low and high E/A ratios (<0.75 and >1.5) associated with excess mortality,19,20 E/A ratio data were divided into quartiles; the quartile thresholds were 0.74, 0.88, and 1.03. Covariates for these quartiles of E/A ratio were examined using logistic regression for categorical variables and linear regression for continuous variables, adjusted for age and gender for variables other than age and gender.
The overall difference in odds ratio (OR) of cerebral infarction between the quartiles of E/A ratio was analyzed using logistic regression. Then, subjects in the lowest and highest quartiles were each compared with subjects in the 2 middle quartiles. Because the highest quartile included 130 subjects with E/A ratio <1.5 (reported normal range, 0.75–1.5),20 a sensitivity analysis was performed in which the lowest quartile (E/A ratio <0.75) was compared with the 3 other quartiles combined. Last, the lowest and highest E/A ratios (<0.75 and >1.5) were compared with the reported normal range of E/A ratio (0.75–1.5).20
The model was adjusted for age and gender (model I) and then adjusted for BMI, current systolic and diastolic blood pressures, hypertension, smoking status, HDL cholesterol level, diabetes, and previous documented coronary heart disease (model II). Because of the high prevalence of hypertension (79%), measured blood pressures were included in the model. Last, the model was adjusted for LV ejection fraction (model III). In models in which we found a significant association with overall cerebral infarction, we re-ran the analysis using clinical stroke by self-report as the dependent variable.
Data were expressed as mean (standard deviation). The OR and 95% confidence intervals (CI) were computed using SAS 9.1/SAS Enterprise Guide (v4.1). In all analyses, the conventional α-level of 0.05 was used for significance testing.
Of the 954 subjects selected randomly from the AGES-Reykjavik cohort, 122 subjects did not undergo cerebral MRI because of contraindications, incomplete protocol to evaluate infarction, or logistic reasons including refusal or disability. Another 140 subjects were excluded because they had >2 missing echocardiogram values. Thus, the study population consisted of 692 subjects. Compared to those included, excluded subjects were significantly older (77  vs 76 ; P<0.001), more likely to be men (58% vs 53%; P=0.001), and more likely to have cerebral infarction (41% vs 28%; P=0.01).
In the study population, 193 (28%) subjects had MRI evidence of cerebral infarction, but only 42 (6%) subjects described a previous clinical stroke. When adjusted for age and gender, subjects with cerebral infarction were significantly older (78  vs 75 ; P<0.0001) and more likely to be male (55% vs 36%; P<0.0001). They were also more likely to have had a previous documented coronary event (26% vs 11%; P=0.0002) and to have diabetes (21% vs 10%; P=0.0001). Subjects with cerebral infarction had higher systolic blood pressure (146  vs 141 ; P=0.02) and were more likely to use anticoagulant or antiplatelet agents (53% vs 35%; P=0.002).
Table 1 shows the association of cerebral infarction with echocardiographic parameters. In model I, adjusted for age and gender, LVM and LA volume indices were significantly higher in participants with cerebral infarction, but cerebral infarction was not significantly related to aortic root diameter, LV ejection fraction, or lateral E/E′ ratios, whether analyzed as continuous variables or as quartiles (quartile results not shown). Similarly, the presence of mitral annular calcification was not associated with cerebral infarction. In model II, adjusted for age, gender, and vascular risk factors, both LVM and LA volume indices became marginally significant. For a 10% increase in LVM or LA volume, the OR were 1.08 (95% CI, 1.00–1.16) and 1.06 (95% CI, 1.00–1.12), respectively. When subjects with a history of atrial fibrillation were excluded, LVM and LA volume remained marginally significant (OR, 1.01; 95% CI, 1.00–1.02; OR, 1.01; 95% CI, 1.00–1.39; model II), respectively.
Whether analyzed as a continuous variable or in a quadratic model, E/A ratio was not significantly related to the presence or absence of cerebral infarction (OR, 0.90 [0.33] vs 0.95 [0.31]; P=0.13 for continuous variable; quadratic results not shown). When E/A ratio was divided into quartiles, as specified in the Methods, increasing E/A ratio quartile was associated with younger age, reduced prevalence of diabetes, and lower diastolic blood pressure (Table 2). The second and third quartiles did not differ significantly except for serum glucose (5.92 [1.43] vs 5.63 [0.71]; P=0.01).
There was a significant overall relationship between E/A ratio quartiles and cerebral infarction (P=0.04). When compared with the 2 middle quartiles, the lowest E/A ratio quartile (<0.75) was significantly related to cerebral infarction in model I, model II, and model III (OR, 1.82; 95% CI, 1.16–2.86; Table 3).
When compared with the 3 other quartiles, E/A ratio <0.75 was also significantly related to cerebral infarction (OR, 1.98; 95% CI, 1.35–2.90; model III; Supplemental Table available online at http://stroke.ahajournals.org).
When compared with the normal range of E/A ratio (0.75–1.5), E/A ratio <0.75 was consistently significant in all models (OR, 2.03; 95% CI, 1.37–2.99; model III; Supplemental Table).
For subjects with clinical stroke by self-report (n=42), the lowest quartile of E/A ratio had a similar point estimate but was not significant in any of the models (OR, 1.97; 95% CI, 0.95–4.08; model I). The highest E/A quartile was not significantly different from the 2 middle quartiles in unadjusted and adjusted models (Table 3) or from the normal range of E/A ratio of 0.75 to 1.5 in any model (OR, 1.51; 95% CI, 0.68–3.35; model I; Supplemental Table). When subjects with a history of atrial fibrillation were excluded, the results were unchanged.
In this cross-sectional analysis, there was a robust association between a low E/A ratio (<0.75) and cerebral infarction independent of age, gender, vascular risk factors, and LV ejection fraction. Even though LVM and LA volume, 2 well-known associations of chronic hypertension, were associated with cerebral infarction independently of age and gender, when the model was adjusted for vascular risk factors, the association became marginally significant. Using more advanced techniques, these results confirm some previous M-mode-based predictors of clinical stroke, including LA size and LVM. However, some, including aortic root size and mitral annular calcification measurements as determined with 2-dimensional point-to-point measurements, were not associated with cerebral infarction. Because of the spatial ambiguity associated with M-mode assessment, these new results seem more plausible but will require confirmation in future 2- and 3-dimensional evaluations. Another possibility is that the high prevalence of vascular risk factors in this elderly cohort obscured any independent association of aortic root diameter and mitral annular calcification with cerebral infarction.
Many earlier studies used M-mode echo data, and the diagnosis of stroke was based on clinical findings with or without cerebral imaging by CT or MRI.4,5,21–23 However, LVM has been related to cerebral infarction detected by MRI in blacks in whom prevalence of combined clinical and silent cerebral infarct was higher than that of clinical stroke alone (20% vs 3%), and these findings are consistent with ours.9
Previous studies have already shown that abnormalities of LV early and late filling velocities (ie, low or high E/A ratios) are associated with increased all-cause mortality.20 Low E/A ratio (<0.6) was associated with higher all-cause and cardiac mortality in the Strong Heart Study of American Indians, but it was not an independent predictor after adjustment for covariates.19 The Strong Heart Study also described higher mortality in younger subjects (57  years) with a restrictive filling pattern indicated by an E/A ratio >1.5. In this AGES-Reykjavik cohort, there were only 33 subjects with E/A ratio >1.5, so no conclusion can be drawn about any associations with the higher value of this parameter. This can also explain the absence of quadratic relation of E/A ratio with cerebral infarction.
E/A ratio describes flow velocities in early and late diastole. Besides describing phasic shifts of LV filling to late diastole, a low E/A ratio may reflect generalized changes in the vascular system, including alterations in laminar flow and a cascade of adverse effects on platelet aggregation and endothelial cell function.24 These maladaptive vascular changes, characterized by heightened oxidative stress and increased proinflammatory and prothrombotic states, could contribute to endothelial dysfunction.25,26 Thus, the mechanisms underlying the relationship between E/A ratio and stroke may be multifactorial.
E/E′, an indicator of LV passive stiffness,27 has been associated with ischemic stroke in subjects with atrial fibrillation,28 but it was not related to cerebral infarction in AGES-Reykjavik. However, this earlier study reported larger ranges and higher mean values for E/E′,28 suggesting stiffer ventricles than were present in AGES-Reykjavik, which could account for the differing results.
The study has a number of limitations. The subjects are the survivors who are still alive 25 years after the Reykjavik Study was initiated, and so the relation of factors to a lethal disease can be underestimated. This limitation could explain why LVM and LA volume have only a modest association with cerebral infarction. Similarly, recall bias could affect characterization of covariates, but this effect should be equally distributed among the study groups. A second limitation is the cross-sectional design, which, in comparison with most previous studies of echocardiographic markers and stroke, does not allow determination of risk, identification of predictors, or sequence of events. The sample size is modest and power may have been reduced by exclusion of subjects without adequate echo measures.
This study demonstrates that even in older subjects, LV mass, LA volume, and low E/A ratio were associated with cerebral infarction detected by MRI. However, the association between low E/A ratio and cerebral infarction appears to be independent of concurrent cardiovascular risk factors, such as hypertension. This relationship is novel in that previous studies have associated cardiac morphology, not diastolic function, with stroke. Thus, E/A ratio, a simple marker of clinical and subclinical cerebral infarction, could be especially useful in elderly patients because a low ratio in patients with cognitive decline and dementia may suggest an etiology of cerebral ischemia.12
The authors thank the participants of the study and the Icelandic Heart Association clinic staff for their invaluable contribution
Sources of Funding
This study was supported by a contract from the National Institutes of Health (N01-AG-1-2100), National Institute on Aging Intramural Research Program, the Hjartavernd (the Icelandic Heart Association), and the Althingi (the Icelandic Parliament).
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.590430/DC1.
- Received May 14, 2010.
- Revision received June 7, 2010.
- Accepted June 30, 2010.
Salonen JT, Puska P, Tuomilehto J, Homan K. Relation of blood pressure, serum lipids, and smoking to the risk of cerebral stroke. A longitudinal study in eastern finland. Stroke. 1982; 13: 327–333.
Savage DD, Drayer JI, Henry WL, Mathews EC Jr, Ware JH, Gardin JM, Cohen ER, Epstein SE, Laragh JH. Echocardiographic assessment of cardiac anatomy and function in hypertensive subjects. Circulation. 1979; 59: 623–632.
Fox ER, Taylor HA Jr, Benjamin EJ, Ding J, Liebson PR, Arnett D, Quin EM, Skelton TN. Left ventricular mass indexed to height and prevalent MRI cerebrovascular disease in an African American cohort: The atherosclerotic risk in communities study. Stroke. 2005; 36: 546–550.
Schneider JA, Wilson RS, Cochran EJ, Bienias JL, Arnold SE, Evans DA, Bennett DA. Relation of cerebral infarctions to dementia and cognitive function in older persons. Neurology. 2003; 60: 1082–1088.
Saczynski JS, Sigurdsson S, Jonsdottir MK, Eiriksdottir G, Jonsson PV, Garcia ME, Kjartansson O, Lopez O, van Buchem MA, Gudnason V, Launer LJ. Cerebral infarcts and cognitive performance: Importance of location and number of infarcts. Stroke. 2009; 40: 677–682.
Harris TB, Launer LJ, Eiriksdottir G, Kjartansson O, Jonsson PV, Sigurdsson G, Thorgeirsson G, Aspelund T, Garcia ME, Cotch MF, Hoffman HJ, Gudnason V. Age, gene/environment susceptibility-Reykjavik study: Multidisciplinary applied phenomics. Am J Epidemiol. 2007; 165: 1076–1087.
Rosano C, Sigurdsson S, Siggeirsdottir K, Phillips CL, Garcia M, Jonsson PV, Eiriksdottir G, Newman AB, Harris TB, van Buchem MA, Gudnason V, Launer LJ. Magnetization transfer imaging, white matter hyperintensities, brain atrophy and slower gait in older men and women. Neurobiol Aging. 2010; 31: 1197–1204.
Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MS, Stewart WJ. Recommendations for chamber quantification: A report from the american society of echocardiography’s guidelines and standards committee and the chamber quantification writing group, developed in conjunction with the European association of echocardiography, a branch of the European society of cardiology. J Am Soc Echocardiogr. 2005; 18: 1440–1463.
Dubois D, Dubois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med. 1916; 17: 863–871.
Bella JN, Palmieri V, Roman MJ, Liu JE, Welty TK, Lee ET, Fabsitz RR, Howard BV, Devereux RB. Mitral ratio of peak early to late diastolic filling velocity as a predictor of mortality in middle-aged and elderly adults: The strong heart study. Circulation. 2002; 105: 1928–1933.
Verdecchia P, Porcellati C, Reboldi G, Gattobigio R, Borgioni C, Pearson TA, Ambrosio G. Left ventricular hypertrophy as an independent predictor of acute cerebrovascular events in essential hypertension. Circulation. 2001; 104: 2039–2044.
Gardin JM, McClelland R, Kitzman D, Lima JA, Bommer W, Klopfenstein HS, Wong ND, Smith VE, Gottdiener J. M-mode echocardiographic predictors of six- to seven-year incidence of coronary heart disease, stroke, congestive heart failure, and mortality in an elderly cohort (the cardiovascular health study). Am J Cardiol. 2001; 87: 1051–1057.
Kizer JR, Wiebers DO, Whisnant JP, Galloway JM, Welty TK, Lee ET, Best LG, Resnick HE, Roman MJ, Devereux RB. Mitral annular calcification, aortic valve sclerosis, and incident stroke in adults free of clinical cardiovascular disease: The strong heart study. Stroke. 2005; 36: 2533–2537.
Berk BC, Abe JI, Min W, Surapisitchat J, Yan C. Endothelial atheroprotective and anti-inflammatory mechanisms. Ann NY Acad Sci. 2001; 947: 93–109;discussion 109–111.
Guazzi M, Arena R. Endothelial dysfunction and pathophysiological correlates in atrial fibrillation. Heart Br Cardiac Soc. 2009; 95: 102–106.
Boo YC, Jo H. Flow-dependent regulation of endothelial nitric oxide synthase: Role of protein kinases. Am J Physiol. 2003; 285: C499–C508.
Kasner M, Westermann D, Steendijk P, Gaub R, Wilkenshoff U, Weitmann K, Hoffmann W, Poller W, Schultheiss HP, Pauschinger M, Tschope C. Utility of Doppler echocardiography and tissue doppler imaging in the estimation of diastolic function in heart failure with normal ejection fraction: A comparative Doppler-conductance catheterization study. Circulation. 2007; 116: 637–647.