The augmentation index (AI) is the ratio of the ejection

Transcription

1 Augmentation Index Is Elevated in Aortic Aneurysm and Dissection Yasushige Shingu, MD, Norihiko Shiiya, MD, PhD, Tomonori Ooka, MD, PhD, Tsuyoshi Tachibana, MD, PhD, Suguru Kubota, MD, PhD, Satoshi Morita, PhD, and Yoshiro Matsui, MD, PhD Department of Cardiovascular Surgery, Hokkaido University Hospital, Sapporo, Hokkaido, and Department of Biostatistics and Epidemiology, Yokohama City University Medical Center, Yokohama, Kanagawa, Japan Background. The augmentation index, the ratio of the ejection pressure from the heart to the reflection pressure from the arterial system, has recently been recognized as one of the indexes of left ventricular afterload. We studied it in patients with aortic aneurysm and dissection, using carotid artery diameter waveform obtained from an echotracking system. Methods. Forty-six patients were divided into the following three groups based on pathology: group A, 21 patients with thoracic aortic aneurysm; group B, 15 patients with chronic aortic dissection; and group C, 10 patients without any aortic diseases. Using an echo-tracking system on the carotid artery, we measured stiffness parameter, arterial compliance, and the augmentation index. Results. There was no significant difference in stiffness parameter and arterial compliance among the three groups. The augmentation index was significantly higher in groups A and B than group C (22 10%, 22 13% vs 8 17%; p 0.012). Female (p 0.028) and heart rate (p 0.005) were significantly associated with the augmentation index and the significance of aortic diseases was marginal (p 0.056). Conclusions. The carotid augmentation index is elevated in patients with aortic aneurysm and dissection. (Ann Thorac Surg 2009;87:1373 8) 2009 by The Society of Thoracic Surgeons The augmentation index (AI) is the ratio of the ejection pressure from the heart and the reflection pressure from the arterial system. The age-related linear increase of AI was first reported by Kelly and colleagues in 1989 [1]. Elevated AI has been reported to cause an increase in left ventricular (LV) afterload, myocardial mass, and oxygen consumption [2]. The AI increase has also been reported in vascular lesion associated with end-stage renal disease, but its change in aortic aneurysm and dissection has not been reported. We previously reported that LV diastolic function can be severely reduced in patients with type-b chronic aortic dissection that has a double-barrel aorta and a narrowed true lumen. We hypothesized that this was a result of elevated LV afterload caused by chronic aortic dissection [3]. However, because we have found no study that has evaluated LV afterload in patients with aortic dissection, we elucidated changes in the index of LV afterload in patients with aortic aneurysm and dissection. As an index of LV afterload, we used the AI, which is calculated from the carotid artery diameter waveform obtained by an echo-tracking system. Accepted for publication Feb 16, Address correspondence to Dr Shingu, Address: , Nishi 3 Chome, Kita 18 Jo, Kitaku, Sapporo, Hokkaido, , Japan; y-shingu@ mopera.net. Material and Methods Arterial Diameter Waveform Measurements and Calculation of the Arterial Mechanical Parameters We used an echo-tracking system (Alpha 10; Aloka, Tokyo, Japan) with a 7.5 MHz linear array probe to measure the arterial diameter changes with one-sixteenth ultrasound wavelength precision; that is mm. The data are updated at a rate of 1 khz. The steering system enables easy optimization of the beam direction based on the vessel direction. The same echo-tracking system was validated in the report by Niki and colleagues [4]. Subjects rested in a supine position for at least 10 minutes. Data were acquired from the left common carotid artery at about 2 cm proximal to the carotid bulb. When a plaque was present, the position of the echo-tracking gate was changed just proximally or distally. The transducer was held in the hand and stabilized against the patient chest walls throughout the study. Echo-tracking gates were manually set at the high echoic line just outside the intima-media complex (near the edge of the adventitia side) where stable tracking was possible. The arterial diameter was determined as the difference between the anterior and posterior wall positions. The maximum and minimum diameters (Dmax and Dmin, mm) and AI (%) were presented automatically. The calculation formula for AI was as follows: (D2 D1)/(Dmax Dmin) 100 (Fig 1). Increased arterial stiffness causes an increase in reflected waves from peripheral reflecting sites to the 2009 by The Society of Thoracic Surgeons /09/$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1374 SHINGU ET AL Ann Thorac Surg AUGMENTATION INDEX IN AORTIC DISEASES 2009;87: heart during systole when the ventricle is ejecting blood. This mechanism augments ascending aortic pulse pressure, which increases arterial wall stress and elevates LV afterload. The contribution of wave reflection to ascending aortic pulse pressure and an estimation of systemic arterial stiffness can be obtained using the augmentation index [2]. Five consecutive beats were averaged to obtain representative waveforms. By entering the blood pressure data of the left upper arm as measured with a sphygmomanometer, the indices of stiffness parameter ( ) and arterial compliance (AC, mm 2 / kpa) were calculated automatically and calculated as follows: ln(ps/pd)/[(dmax/dmin) 1], AC (Dmax Dmax Dmin Dmin)/[4(Ps Pd)], where Ps and Pd are systolic and diastolic pressure, respectively. The stiffness parameter is the slope of the exponential function between the relative arterial pressure and the distention ratio of artery (the arterial diameter at a given pressure). This index characterizes the entire deformation behavior of the vascular wall, being independent of the intraluminal pressure within the physiologic range [5]. The echographic study and the use of the clinical records for research were approved by the institutional Ethics Review Board. All patients gave informed consent. Fig 1. Representative diameter change waveform and calculation formula of augmentation index. Diameter of first inflection point is defined as D1 and the next peak as D2. (AI augmentation index; Dmax and Dmin maximum and minimum diameter.) Study 1: Reproducibility of Arterial Diameter Waveform Measurements To confirm the reproducibility of arterial diameter waveform measurements, interobserver and intraobserver variability was studied in five healthy volunteers without cardiovascular diseases and who were not on any medication. All were nonsmoking males whose mean age was 23 1 years. All refrained from caffeine for at least 12 hours before the study. Interobserver variability was evaluated by two observers who measured each subject consecutively. Intraobserver variability was studied by one observer performing two sessions on different days. Each subject was studied at almost the same hour and the interval of the two sessions was 7 days. The average value of the three measurements in each session was compared. Study 2: Correlation of Arterial Diameter Waveform Measurements With Aortic Diseases Dmax, Dmin, AI,, and AC were measured in 46 patients admitted to our hospital from April 2006 to March The mean age was (from 33 to 87) years. Thirty-six (78%) were male and 10 (22%) were female. The body mass index (BMI) was Thirty-eight (83%) had hypertension, 5 (11%) had diabetes mellitus, and 12 (26%) had hyperlipidemia. Thirteen (28%) were smokers, and 8 (17%) regularly drank alcohol. These 46 patients were divided into three groups based on pathology. Group A consisted of 21 (46%) patients with thoracic aortic aneurysm with or without abdominal aneurysm. Group B included 15 (33%) patients with chronic aortic dissection (4 type A and 11 type B). All the type A dissections did not involve the carotid vessels in this study. Group C consisted of 10 (22%) patients without any aortic diseases. The primary diagnosis of the Table 1. Patient Characteristics Characteristics Total Population (n 46) Group A (n 21) Group B (n 15) Group C (n 10) p Value Women, n (%) 10 (22) 4 (19) 5 (33) 1 (10) Hypertension, n (%) 38 (83) 19 (90) 12 (80) 7 (70) Diabetes mellitus, n (%) 5 (11) 2 (10) 1 (7) 2 (20) Hyperlipidemia, n (%) 12 (26) 6 (29) 2 (13) 4 (40) Smoking, n (%) 13 (28) 9 (43) 3 (20) 1 (10) Age, years a b SBP, mm Hg DBP, mm Hg Heart Rate, bpm SVR, dynes sec cm Body mass index, kg/m Hematocrit, % By Turkey s multiple comparison between a and b, the p value was bpm beats per minute; DBP diastolic blood pressure; SBP systolic blood pressure; SVR systemic vascular resistance.

3 Ann Thorac Surg SHINGU ET AL 2009;87: AUGMENTATION INDEX IN AORTIC DISEASES Table 2. Reproducibility of Arterial Waveform Measurements Variables Dmax Dmin AI Interobserver correlation coefficient (p value) 0.96 ( 0.001) 0.94 (0.003) 0.86 (0.017) Intraobserver correlation coefficient (p value) 0.85 (0.009) 0.84 (0.010) 0.70 (0.043) Interobserver percentage error, % Intraobserver percentage error, % AI augmentation index; Dmax and Dmin indicate maximum and minimum diameter. patients in this group was coronary artery disease in 5 individuals and pneumonia in 5. Those with significant aortic valve stenosis and regurgitation or wall motion asynergy and low left ventricular function were excluded from this study. All patients were in stable preoperative states. We also excluded patients whose vascular systems had previously been operated on to exclude the effect of prosthetic grafts. The characteristics of the patients are shown in Table 1. Systemic vascular resistance (SVR) was calculated by simple means (SVR 80 mean arterial pressure/cardiac output measured by transthoracic echocardiography). Although SVR was larger in groups A and B, there was no statistical significance. As blood pressure control medication, angiotensin converting enzyme inhibitor was used in 3, none, and 2 patients (p 0.230), angiotensin receptor blocker was used in 8, 8, and 2 patients (p 0.245), beta blocker was used in 10, 6, and 3 patients (p 0.643), and Ca blocker was used in 17, 10, and 5 patients (p 0.207) in groups A, B, and C, respectively. Statistical Analysis Statistical analysis was performed with SPSS version 10.0 software (SPSS Inc, Chicago, IL). Values were expressed as mean SD. A one-way analysis of variance followed by Turkey s multiple comparison was used to compare the continuous variables, and a 2 test was used for frequency comparisons among the groups. To evaluate the reproducibility of the arterial diameter waveform measurements, correlation analysis was performed, and the intraclass correlation coefficient between the two measurements and the percentage error, which was derived as the absolute difference divided by the initial measurements, was presented. To identify the factors correlating with, AC, and AI, the following factors were entered into univariate analysis by the Student t test or correlation analysis: aortic disease (aneurysm and dissection), female, age, hypertension, diabetes mellitus, hyperlipidemia, BMI, heart rate, and systolic blood pressure. We adapted multiple regression analysis for those parameters whose p values were less than 0.2. A p value less than 0.05 was considered statistically significant. Results Study 1: Reproducibility of Arterial Waveform Measurements The interobserver and intraobserver intraclass correlation coefficients were 0.96 and 0.85 for Dmax, 0.94 and 0.84 for Dmin, and 0.86 and 0.70 for AI, respectively. All correlations were statistically significant. The percentage error of Dmax, Dmin, and AI was %, %, and % for interobserver variability, and %, %, and % for intraobserver variability (Table 2). Study 2: Correlation of Arterial Diameter Waveform Measurements With Aortic Diseases The patient characteristics were not significantly different among the three groups except for age (Table 1), which was significantly lower in group B than in group A (57 18 years vs 72 6 years; p 0.001). The AI was significantly higher in groups A and B than in group C (22 10%, 22 13% vs 8 17%; p 0.012) Although the patients in group B were significantly younger than those in group A, group B s AI was comparable with that of group A. The false lumen diameter in group B (20 13 mm) did not significantly correlate to AI (p 0.638). The extent of dissection (extended to abdominal aorta in 13 out of 15 cases) and the presence of thrombus in the false lumen (7 out of 15 cases) did not also affect AI (p Table 3. Echo-Tracking Data Variables Total Population (n 46) Group A (n 21) Group B (n 15) Group C (n 10) p Value Dmax, mm Dmin, mm AC, mm2/kpa AI, % a b c By Turkey s multiple comparison between a and c, and b and c, the p value was and 0.021, respectively. AC arterial compliance; AI augmentation index; stiffness parameter; Dmax and Dmin indicate maximum and minimum diameter.

4 1376 SHINGU ET AL Ann Thorac Surg AUGMENTATION INDEX IN AORTIC DISEASES 2009;87: Table 4. Univariate and Multivariate Analysis for Augmentation Index Variables Univariate Analysis Multivariate Analysis p 95% CI p Aortic disease to Female to Age to Hypertension Diabetes mellitus Hyperlipidemia Body mass index Heart rate to Systolic blood pressure stiffness parameter; CI confidential interval. and p 0.344, respectively). There was no significant difference in and AC among the groups, although group B s AC was slightly higher than the others (Table 3). In the multiple regression analysis, the following factors significantly influenced AI: female (, 9.3; 95% confidence interval [CI], 1.1 to 17.6; p 0.028) and heart rate (, 0.3; 95% CI, 0.58 to 0.11; p 0.005). The significance of aortic diseases (aneurysm and dissection) was marginally significant (, 8.5; 95% CI, 0.2 to 17.0; p 0.056). Influence of age upon AI did not reach statistical significance (p 0.065) (Table 4). On the other hand, the independent predictor was age for and AC (, 0.2; 95% CI, 0.1 to 0.3; p and, ; 95% CI, 0.02 to 0.005; p 0.001, respectively). Fig 2. Relationship between AI and age. By regression analysis in groups A and C, the coefficient of age was (95% confidential interval: to 1.462) and its p value was less than (AI augmentation index; Πgroup A; group B; e group C. Comment The present results show that the AI obtained from the carotid artery was significantly elevated in patients with aortic aneurysm and dissection. These results suggest that left ventricular afterload is higher in these populations. Aortic Morphology and LV Afterload The value of AI was almost the same between the patients with aneurysm and those with dissection, although the latter were significantly younger than the former. On the other hand, carotid arterial stiffness ( ) and compliance (AC) were strongly correlated with age. This may be explained by the fact that AI is not a value reflecting atherosclerosis in the carotid artery, but the value reflects the stiffness of the distal part of the arterial tree. Although old age has been reported to be correlated with elevated AI, its correlation with AI was not significant in the present study. This may be explained by the inclusion of the group B (aortic dissection) patients who have high AI despite being younger. Indeed, when we excluded the group B patients, the correlation between AI and age was significant (Fig 2). Although the question remains undetermined whether elevated AI is a consequence or basis of aortic dissection, the present results support our previous hypothesis that impaired LV diastolic function in patients with chronic double-barrel aortic dissection resulted from elevated LV afterload. Recently, Borlaug and colleagues [6] reported that, as an index of late-systolic afterload, AI has the greatest influence on early diastolic myocardial velocity (E ) on tissue Doppler echocardiography, which is an index of early diastolic ventricular relaxation. Their results also support our previous hypothesis. Reliability of the Echo-Tracking System Concerning AI measurement of AI, Kelly and colleagues [1] used a noninvasive transcutaneous tonometer containing a Millar micromanometer. Radial arterial pressure wave measurement by tonometry can be used as a substitute for central aortic pressure waveform [7]. Niki

5 Ann Thorac Surg SHINGU ET AL 2009;87: AUGMENTATION INDEX IN AORTIC DISEASES and colleagues [4] reported the use of an echo-tracking system to measure arterial diameter change. We used arterial diameter waveforms obtained by this echotracking system as a substitute for central aortic pressure waveform. The similarity of pressure waveforms and arterial diameter waveforms has been reported in animals by several authors. In the human carotid arteries, Sugawara and colleagues [8] reported a linear relationship between the two waveforms throughout the cardiac cycle. As a measuring point, the common carotid artery was selected because it provides a similar pressure wave pattern with central aortic pressure. Although AI in radial and carotid arteries was slightly lower than that of the aortic root, AI measured at these peripheral arteries showed linear increase with aging [1]. The interobserver and intraobserver correlation coefficients of vascular diameter and AI were significantly high, and their variability was less than 6% and 22%, respectively. The major reason for the relatively high intraobserver variability of AI seems to be the physiologic changes in the subjects between the two study days. In the second part of this study, all patients were in steady states after a few days of admission and the variability of AI may be much smaller. The small interobserver variability suggests the reliability of this system. Limitation The major limitation of this study is its small study population. The control group (group C) without aortic disease should have been matched for gender and heart rate to exclude these effects on AI. Strictly speaking, the reflected wave in the carotid artery is derived not only from the descending and abdominal aorta but also from the cerebral arteries. We cannot separate these two factors with this system, where the starting point of the reflected wave is not determined by simultaneous pulse Doppler wave velocity but by the automatic detection of inflection. As Niki and colleagues [4] reported, the simultaneous measurement of pulse wave velocity is ideal to precisely detect the inflection point. Summary Carotid augmentation index is elevated in patients with aortic aneurysm and dissection, suggesting increased LV afterload. Further study will be required to determine the relationship between the elevated AI and the LV diastolic function observed in patients with aortic dissection in our previous study. References Kelly R, Hayward C, Avolio A, O Rourke M. Noninvasive determination of age-related changes in the human arterial pulse. Circulation 1989;80: Nichols WW, Singh BM. Augmentation index as a measure of peripheral vascular disease state. Curr Opin Cardiol 2002;17: Shingu Y, Shiiya N, Mikami T, Matsuzaki K, Kunihara T, Matsui Y. Left ventricular diastolic dysfunction in chronic aortic dissection. Ann Thorac Surg 2007;83: Niki K, Sugawara M, Chang D, et al. A new noninvasive measurement system for wave intensity: evaluation of carotid arterial wave intensity and reproducibility. Heart Vessels 2002;17: Hirai T, Sasayma S, Kawasaki T, Yagi S. Stiffness of systemic arteries in patients with myocardial infarction. A noninvasive method to predict severity of coronary atherosclerosis. Circulation 1989;80: Borlaug BA, Melenovsky V, Redfield MM, et al. Impact of arterial load and loading sequence on left ventricular tissue velocities in humans. J Am Coll Cardiol 2007;50: Takazawa K, Tanaka N, Takeda K, Kurosu F, Ibukiyama C. Underestimation of vasodilator effects of nitroglycerin by upper limb blood pressure. Hypertension 1995;26: Sugawara M, Niki K, Furuhata H, Ohnishi S, Suzuki S. Relationship between the pressure and diameter of the carotid artery in humans. Heart Vessels 2000;15: INVITED COMMENTARY I read with great interest the article of Shingu and colleagues [1], demonstrating an elevated augmentation index (AI) in thoracic aortic aneurysms and thoracic aortic dissections. The concept of AI is not very well known to many cardiac surgeons, but it has received intense study and attention in the last decade of the cardiac physiology literature. As the incident systolic pressure wave generated by cardiac ejection encounters zones in impedance, such as arterial bifurcations and obstructive lesions, part of the wave is reflected backward, summing with the systolic wave, something that is noticed as a notch to the ascending part of the blood pressure waveform (Fig 1). In young patients who do not have significant atherosclerotic disease and preserved ejection fraction, this notch is almost at the top of the systolic part of the waveform, because this backward-moving wave reaches the pressure sensor much later. After a coronary artery bypass procedure in a young patient, it is common, and very familiar to cardiac surgeons, during the first hours in the intensive care unit, to observe a dicrotic blood pressure waveform, which is something commonly referred as hyperdynamic circulation (Fig 1; Left). This is in contrast to patients of advanced age, hypertension, and left ventricular heart failure; in these cases, the increased wave velocity causes the reflected wave to reach back to the heart earlier, in mid-systole, considerably increasing the late-systolic load, impairing cardiac ejection, and diastolic relaxation (Fig 1; Right). Normal elastic arterial tissue absorbs more energy and reflects and propagates a pressure wave much slower than a stiff atherosclerotic vessel. The magnitude of this reflected pressure wave can be quantified by the AI (Fig 1). The AI is essentially proportional to peripheral afterload. High AI means high arterial stiffness and afterload. Consequently, high AI is associated with increased myocardial mass, increased myocardial oxygen consumption, and increased left ven by The Society of Thoracic Surgeons /09/$36.00 Published by Elsevier Inc doi: /j.athoracsur