Arterial stiffness and left-ventricular diastolic dysfunction: Guangzhou Biobank Cohort Study-CVD

Abstract

Brachial–ankle pulse wave velocity (baPWV), a marker of arterial stiffness, is an established cardiovascular risk factor of ventricular stiffening. We studied the association of baPWV with left-ventricular (LV) diastolic function in a sub-study of the Guangzhou Biobank Cohort Study. In all, 378 Chinese subjects with a normal ejection fraction (>50%) had baPWV measurement by a noninvasive automatic waveform analyser, carotid intima-medial thickness (IMT) measurement by B-mode ultrasonography and cardiac diastolic function assessment by echocardiography. After adjusting for age, both baPWV and IMT were associated with LV mass index, posterior wall end-diastolic thickness and inter-ventricular end-diastolic thickness, but only baPWV was associated with deceleration time, atrial flow velocity and E/A ratio. Multivariable linear regression model showed that baPWV and mean arterial pressure, but not IMT, were significantly associated with E/A ratio (β=−0.02, P=0.03 and β=−0.36, P=0.02, respectively). The receiver operator characteristic curve showed that baPWV was better than pulse pressure or mean arterial pressure to detect LV diastolic dysfunction (E/A<1.0). Our study suggested that increased baPWV might be an independent risk factor or marker for diastolic dysfunction. Early detection of an intervention on increased baPWV may be important for prevention of cardiac diastolic dysfunction.

Introduction

Pulse wave velocity (PWV) has been reported as a marker of arterial stiffness and widely used in clinical and population-based epidemiological studies.^{1, 2} Brachial–ankle PWV (baPWV) obtained by a noninvasive method using oscillometric technique has been shown to be strongly correlated with increased cardiovascular events³ and be more representative than the traditional invasively derived carotid femoral PWV.⁴

Left-ventricular (LV) diastolic dysfunction is the most common cardiac abnormality associated with hypertension and is a predictor of cardiovascular disease⁵ and all-cause mortality.⁶ A European study on 1274 community-based volunteers showed that 2.8% of the subjects aged 25–35 years and 15.8% aged 65 years or above had LV diastolic dysfunction, and the prevalence of diastolic dysfunction was higher than that of systolic dysfunction and increased in the elderly.⁷ A more recent study in Canberra showed that the prevalence of diastolic dysfunction was 34.7% in 1275 randomly selected residents of Canberra aged 60–86 years.⁸ Both of these studies showed that diastolic dysfunction is common in older populations and often unaccompanied by overt congestive heart failure.^{7, 8}

Emerging data have shown that PWV is associated with ultrasonographically detected LV diastolic dysfunction clinically and epidemiologically.^{9, 10} Although arterial stiffness as measured by baPWV was associated with atherosclerosis and diastolic heart failure in patients with hypertension,⁹ it is less clear whether an increased baPWV relates to the LV diastolic dysfunction in the general population. Furthermore, the association between carotid intima-medial thickness (IMT), which is an established marker of atherosclerosis, and LV diastolic dysfunction has not been reported.

The present cross-sectional study examined the association between electrocardiographically determined LV diastolic dysfunction and baPWV or carotid IMT in a community-based sample.

Methods

Subjects

The Guangzhou Biobank Cohort Study is a collaborative study of Guangzhou 12th People's Hospital, China, the University of Hong Kong and the University of Birmingham, UK. Subjects are recruited from ‘The Guangzhou Health and Happiness Association for the Respectable Elders’ (GHHARE), which is a community social and welfare association aligned with the municipal government. Membership is open to older adults for a monthly fee of about 4 RMB (50 US cents). About 7% of Guangzhou permanent residents aged 50 years or above are members of this association. We have recruited over 30 000 older people from September 2003 to December 2007. Details of participant recruitment and methods in the Guangzhou Biobank Cohort Study have been reported earlier.¹¹

We only included those who were ambulatory and not receiving treatment for life threatening diseases, such as cancer. In a more detailed sub-study on cardiovascular disease, 378 subjects who had a normal ejection fraction (EF>50%) were included in this study from a random sample of 1996 during the Phase 3 of the Guangzhou Biobank Cohort Study (November 2006–September 2007).¹² The study has received ethical approval from the Guangzhou Medical Ethics Committee of the Chinese Medical Association, Guangzhou, China. All participants gave written, informed consent before participating in the study.

Demographic and laboratory parameters

A standardized questionnaire was used to assess personal disease history and lifestyle, including smoking status and physical activity according to the International Physical Activity Questionnaire.¹³ The reliability of the questionnaire was tested in 200 subjects with κ values of 0.88 and 0.96 for the two questions about smoking status.¹⁴ Resting seated blood pressure was measured three times using an automated sphygmomanometer (Omron 705CP, Tokyo, Japan), which had been previously validated by comparison with the manual mercury sphygmomanometer,¹⁵ and the last two measurements were averaged for analysis. Mean arterial pressure was defined as (systolic blood pressure+diastolic blood pressure ^*2)/3. Pulse pressure was defined as systolic blood pressure−diastolic blood pressure. Height and weight were measured using standardized procedure that had been described elsewhere¹² and body mass index was assessed. Fasting parameters including plasma glucose and lipids (triglycerides, total, low-density lipoprotein- and high-density lipoprotein-cholesterol) were measured in the Clinical Laboratory of the Guangzhou Number 12 Hospital using standardized procedures.

Intima-medial thickness

Carotid B-mode colour ultrasonographic examination was performed using ALT HDI 3000 mainframe enhanced, linear array scanner (medium frequency, 7.5 MHz) by a specialist physician. The operators were registered ultrasound doctors who had a professional certificate for colour Doppler ultrasound measurement awarded by the Ministry of Health of China. All scans were performed following a predetermined, standardized scanning protocol for the right and left carotid arteries using images of the far wall of the distal 10 mm of the common carotid arteries. Three scanning angles, with the image focused on the posterior wall, were recorded from the angle showing the greatest distance between the lumen–intima interface and the media–adventitia interface. Carotid IMT measurements were performed off-line with the use of automated image analysis software. Details on the procedure used in this study and its reproducibility have been published elsewhere.¹²

Brachial–ankle PWV

Subjects had baPWV measurement in the supine position after 5 min of bed rest using an automatic waveform analyser (BP-203RPE; Colin Medical Technology, Komaki, Japan). This device stored data of the waveforms of both brachium and ankles for a sampling time. The time interval between the wave front of the brachial waveforms and that of the waveforms of ankle was automatically measured, which was defined as T. The path lengths from the suprasternal notch to the elbow (La) and also from the suprasternal notch to the ankle (Lb) were automatically calculated based on the patient's height. Then, baPWV was calculated using the following equation: baPWV (cm s⁻¹)=(Lb−La)/T, and the averaged left and right baPWV were obtained for the data analysis.

Diastolic function parameters

M-mode echocardiograms were obtained as guided by two-dimensional echocardiography using an echocardiographic instrument (General Electric Vivid 7, Milwaukee, WI, USA) equipped with a transducer having a frequency range of 2.5–3.5 MHz. All scans performed following a standardized scanning protocol were recorded for off-line analysis. To reduce inter-observer variability, all scans were analysed by a single experienced physician. The LV mass (LVM) was calculated by Devereux's method.¹⁶ The LVM index (LVMI) was calculated as the LVM divided by the body surface area. The LV ejection fraction (LVEF was derived automatically from the equipment. The peak velocity of early trans-mitral flow (E velocity) and the peak velocity of atrial flow (A velocity) were recorded. E/A was the ratio of E velocity to A velocity. The deceleration time was the interval from the E-wave peak to the decline of the velocity to baseline. Isovolumic relaxation time was determined as the interval between the end of ejection and the onset of LV filling. An E/A ratio of ⩽1.0 was considered as indicating of diastolic dysfunction.¹⁷ All subjects had a normal LVEF (EF>50%) and no subject in this study had a pseudonormal diastolic function (E/A ratio of 1.0 to 1.5 and deceleration time >240 ms).¹⁷ The physicians who performed the assessments above were blinded to other information.

Statistical analyses

All data analysis was conducted using Stata/IC 10.1. Statistical significance was set at a two-tailed P<0.05. Analysis of covariance was used for continuous variables with adjustment of age. Multiple linear regression analysis was applied to evaluate whether the relationship between E/A ratio and baPWV or carotid IMT is independent from other cardiovascular risk factors. Stepwise forward regression models were used to select significant predictors of E/A ratio. Receiver operating characteristic curve was used to compare baPWV, pulse pressure and mean arterial blood pressure in discriminating subjects with LV diastolic dysfunction (E/A ratio of <1.0).

Results

Subject characteristics

The mean (s.d.) age of the 378 subjects were 59.1 (6.1) years; 57.9% were men and 13.0% had an educational level of college or above. Most of them were physically active (70.3%) and did not smoke (18.7% current smokers); 8.0% had type 2 diabetes; 31.8% had hypertension; and only two third of the hypertensive subjects had blood pressure lowering medication (Table 1).

Table 1 Demographic characteristics and biochemical parameters in 378 subjects with normal ejection fraction

Table 2 shows that low E/A (⩽1.0) was significantly associated with increasing age, mean arterial stiffness, pulse pressure, carotid IMT, baPWV, hypertension and medication for hypertension (P from 0.03 to <0.001).

Table 2 Demographic characteristics and biochemical parameters by E/A ratio in 378 subjects with normal ejection fraction

Impact of PWV and IMT on indices of LV diastolic function

Table 3 shows that after adjusting for age, increased baPWV was significantly associated with increased peak velocity of A-wave and LVM, LVMI, posterior wall end-diastolic thickness and inter-ventricular end-diastolic septal thickness, and decreased E/A ratio (P from 0.02 to <0.001). Significant association was also found in carotid IMT with LVM, LVMI, posterior wall end-diastolic thickness and inter-ventricular end-diastolic septal thickness (P from 0.02 to <0.001).

Table 3 Age-adjusted linear regression models of the associations between brachial–ankle pulse wave velocity (PWV) and carotid intima-medial thickness (IMT) with indices of cardiac diastolic function in 378 subjects with normal ejection fraction

Predictors of E/A ratio

Table 4 shows that, in multivariable linear regression models with forced entry of age and baPWV or IMT using stepwise forward selection, baPWV, but not carotid IMT, was a significant predictor for LV diastolic dysfunction, with an increase in 1.0 m s⁻¹ baPWV associated with a decrease in 0.02 E/A (P=0.03).

Table 4 Multivariable linear regression models of the associations between brachial–ankle pulse wave velocity (PWV) and carotid intima-medial thickness (IMT) with E/A ratio in 378 subjects with normal ejection fraction

We also performed a multivariable regression analysis for the association between baPWV/IMT and E/A ratio, adjusting for age, sex, high-density lipoprotein- and low-density lipoprotein-cholesterol, body mass index, height, fasting plasma glucose, mean arterial pressure, education, smoking and medication for hypertension, and the results remained unchanged (regression coefficient=−0.02, P=0.03). No association between IMT and E/A ratio was found (Table not shown).

Figure 1 shows that, compared to pulse pressure and mean arterial pressure, the overall performance (area under curve (AUC)) of baPWV was superior in detecting LV diastolic dysfunction (E/A<1.0) (AUC_pp=0.61 (95% CI 0.55–0.67, P<0.001); AUC_map=0.65 (95% CI 0.59–0.70, P<0.001); AUC_pwv=0.69 (95% CI 0.64–0.74, P<0.001).

Figure 1

Comparison of overall performances of pulse pressure (pp), mean arterial pressure (map) and brachial–ankle pulse wave velocity (PWV). The areas under the curve (AUC) for detection of left-ventricular diastolic dysfunction (E/A ratio<1) were AUC_pp=0.61 (95% CI 0.55–0.67, P<0.001); AUC_map=0.65 (95% CI 0.59–0.70, P<0.001); AUC_pwv=0.69 (95% CI 0.64–0.74, P<0.001).

Discussion

We have shown that baPWV, but not IMT, was significantly associated with LV diastolic function in subjects with a normal EF, independent of other cardiovascular risk factors. Compared to mean arterial blood pressure and pulse pressure, baPWV was a better predictor (cross-sectionally) of diastolic dysfunction, as measured by E/A ratio. Therefore, baPWV, assessed as a screening tool in this study, was simple, reliable, noninvasive and better then pulse pressure or mean arterial blood pressure for the detection of sub-clinical diastolic dysfunction. The validity of the baPWV has been reported earlier, with an inter- and intra-observer coefficient of variation of 8.4 and 10.0%, respectively.¹⁸ To the best of our knowledge, this is the first population-based study in a Chinese community sample showing that baPWV as a simple and noninvasive marker for arterial stiffness was an independent predictor of LV diastolic dysfunction in people with a normal LVEF.

Our finding suggests that arterial stiffness is independently associated with load-dependent alterations in LV diastolic function. Previous studies in both Asia and Western counties have shown a significant association between diastolic dysfunction and arterial stiffness using different markers, such as carotid–femoral PWV, cardio-ankle vascular index and augmentation index.^{9, 10, 19, 20, 21, 22} Abhayaratna et al.^{10, 20} showed that among older people in Australia, increased arterial stiffness was associated with severe LV diastolic dysfunction, and the strength of the association varied according to the specific measure, with PWV appeared to be superior to central or brachial pulse pressure for the detection of diastolic dysfunction. As most of the studies above were based on cross-sectional data, whether the association between increased PWV and diastolic dysfunction is causal cannot be determined. However, such association is temporally feasible and biologically plausible. The reflected wave travels more rapidly along the arterial tree in those with arterial stiffness, resulting in an increase in aortic systolic pressure and a decrease in aortic pressure during diastole.²¹ The resultant increase in after-load during LV systole and reduction in coronary perfusion during LV diastole may lead to LV hypertrophy and slowing of LV relaxation.²³

A population-based study on 1621 older American men and women showed that increased pulse pressure, as an index of arterial stiffening, was associated with increased incidence of congestive heart failure, and thus suggested that pulse pressure was a predictor of congestive heart failure in elderly people.²⁴ Zile et al.²⁵ reported that objective measurements of LV diastolic function serve to confirm rather than establish the diagnosis of diastolic heart failure. E/A ratio is one of the established marker of cardiac diastolic dysfunction,²⁶ and the changes in E/A ratio seem to precede the changes in the deceleration time of mitral inflow in the early stage of cardiac diastolic dysfunction.⁶ BaPWV is a more valid marker of arterial stiffness and better predictor of diastolic dysfunction compared with pulse pressure and other PWV markers, such as heart–carotid PWV, heart–femoral PWV, carotid–femoral PWV or femoral–ankle PWV.^{27, 28} We have shown that increased baPWV was associated with cardiac diastolic function as indicated by E/A ratio. Therefore, to our knowledge, this is the first community-based study to report that increased baPWV might be a risk factor or marker for diastolic dysfunction.

A previous study in Japan has shown a significant association between E/A ratio and baPWV in 147 patients with hypertension.⁹ We have consistently found that baPWV was a significant predictor (cross-sectionally) of LV diastolic dysfunction in a community sample, independent of multiple potential confounders, such as age, mean arterial blood pressure and lipids. Estimation of baPWV using an automatic oscillometric method to identify early diastolic dysfunction among those with normal EF function may have important pathological and clinical implications. Our findings have provided consistent and complementary information and suggested that, early screening and treatment of arterial stiffness was important for preventing diastolic heart failure, especially in older people. However, the cross-sectional nature of the study does not allow us to determine whether increased baPWV preceded the ecocardiographic detected diastolic dysfunction.

We also found that after adjusting for age, increased carotid IMT was associated with increase LVMI and posterior wall thickness. However, such associations could not be observed after adjusting for other potential confounders. We found no association between E/A ratios and IMT in our population. Previous Western studies in hypertensive patients and with small sample size have shown that carotid IMT was associated with markers of diastolic function^{29, 30, 31, 32, 33} such as isovolumic relaxation time and DT. However, data on the relationship between E/A ratio and carotid IMT were scarce, with only one Japanese study reported an unadjusted insignificant association.⁹We could not determine the precise mechanism underlying the association between diastolic dysfunction and arterial stiffness or atherosclerosis and future longitudinal studies are needed.

Several limitations of this study should be considered. First, the study sample was homogeneous and the association between arterial stiffness and diastolic dysfunction was analysed using cross-sectional data. Thus, further prospective studies are needed to confirm our findings in other populations and to clarify the temporal association between diastolic dysfunction and arterial stiffness. Second, a high percentage of our subjects (31.8%) were hypertensive, with two thirds of them taking blood pressure lowering medication, which to some extent may differentially modify the indices of arterial stiffness and diastolic dysfunction and influence the true relationship between these markers. However, given that hypertension is a common and widespread global epidemic, such limitation can only overcome by stratified analysis in large community-based studies. We have added mean arterial pressure and medication for hypertension in the stepwise multivariate models and the results remained unchanged, but the sample size was not adequate to perform a stratified analysis. Furthermore, our study included an ethnically homogeneous population thus further studies in different ethnic groups taking into account ethnic difference in the prevalence of hypertension are warranted. Third, as this study was not a blinded, interventional trial, we cannot rule out bias in our subjects. However, as the physicians performing all the assessments were blinded to subjects’ identity and other clinical information, diagnosis bias was unlikely. Finally, given that carotid–femoral PWV is a validated marker of arterial stiffening over the central elastic arteries,²¹ further studies are necessary to compare the baPWV with carotid–femoral PWV in predicting diastolic dysfunction.

In conclusion, this study has shown that baPWV, but not IMT, correlated with diastolic dysfunction, and might be a risk factor or marker for diastolic dysfunction, independent of other potential vascular risk factors. Early detection and intervention of increased baPWV may be important for prevention of cardiac diastolic dysfunction. Further studies are needed to confirm these findings in other populations.