Aortic stiffness as a risk factor for recurrent acute coronary events in patients with ischaemic heart disease

European Heart Journal (2000) 21, 390 396 Article No. euhj.1999.1756, available online at http://www.idealibrary.com on Aortic stiffness as a risk factor for recurrent acute coronary events in patients with ischaemic heart disease C. Stefanadis, J. Dernellis, E. Tsiamis, C. Stratos, L. Diamantopoulos, A. Michaelides and P. Toutouzas Hippokration Hospital, Department of Cardiology, University of Athens, Greece Aims Aortic elasticity is an important determinant of left ventricular performance and coronary blood flow. Moreover, it has been shown that aortic elastic properties deteriorate in patients with coronary artery disease. However, the predictive role of aortic elasticity in the occurrence of coronary events, has not been addressed so far. Therefore, we set out to test prospectively the hypothesis that invasive as well as non-invasive measures of aortic elastic properties, assessed at rest from pressure diameter relationships, could predict the development of recurrent coronary events. Methods and Results Clinical variables and measures of aortic function were assessed in 54 normotensive patients with coronary artery disease. The aortic pressure diameter relationship was derived invasively with a high-fidelity Y shaped catheter (developed in our Institution) for aortic diameter measurements, simultaneously with a Millar catheter for aortic pressure measurements. Aortic root distensibility was assessed by non-invasive techniques. During an average of 3 years follow-up, 12 of 54 patients either developed unstable angina (n=8) or acute myocardial infarction (n=4). By multivariate Cox model analysis, aortic stiffness was the strongest predictor of progression to any end-point (relative risk: 3 24, CI: 1 79 to 5 83; P=0 000). When aortic stiffness was not considered, aortic distensibility was the only independent predictor for acute coronary syndromes (relative risk: 0 37 CI: 0 21 to 0 65; P=0 000). Conclusion In patients with coronary artery disease, aortic elastic properties are powerful and independent risk factors for recurrent acute coronary events. (Eur Heart J 2000; 21: 390 396) 2000 The European Society of Cardiology Key Words: Acute coronary syndromes, aortic stiffness, aortic distensibility, Cox proportional hazards models. See page 342 for the Editorial comment on this article. Introduction Systolic and diastolic blood pressures may carry an increased risk of fatal and non-fatal cardiac events [1,2]. Pulse pressure changes may reflect alterations in the elasticity of large arteries. There is now increasing evidence that high pulse pressure is an independent risk factor for cardiovascular mortality, especially coronary mortality, in different populations [3]. Pulse pressure measurement may help in the evaluation of the individual risk and therefore in the therapeutic decision making. Wide pulse pressure was identified as a predictor of myocardial infarction controlling for other known risk factors by Cox regression [1,3]. Revision submitted 15 June 1999, and accepted 16 June 1999. Correspondence: Christodoulos Stefanadis, MD, FESC, FACC, 9 Tepeleniou Str., 15452 Paleo Psychico, Athens, Greece. 0195-668X/00/050390+07 $35.00/0 However, pulse pressure does not take into account all parameters of aortic elastic properties. Quantitative information on the elastic properties of the aorta can be obtained by means of concomitant determinations of pressure and aortic diameter. In many investigations the state and function of the aorta was assessed noninvasively by pulse wave velocity or by ultrasonically measured pulsatile aortic dimension changes [4,5]. However, these measures have inherent limitations for clinical application because of their dependence on arterial pressure or susceptibility of peripheral muscular arteries to the measuring environment [6]. Recently, we described a new highly accurate method for obtaining aortic pressure diameter relationship in conscious humans [7 11]. With this method [7 11], aortic diameters were acquired using a high-fidelity intravascular catheter developed in our Institution. It incorporates an ultrasonic displacement meter at its distal end. Aortic pressures were 2000 The European Society of Cardiology

Aortic stiffness and recurrent coronary events 391 acquired simultaneously and at the same aortic level with a catheter-tip micromanometer [7 11]. Although aortic elastic properties have been assessed in patients with coronary artery disease the predictive role of aortic elasticity has not been studied so far. The purpose of the present study was to assess the effect of aortic stiffness on the long-term risk of recurrent coronary events in patients with newly diagnosed coronary artery disease. Methods Patients Patients with an age range from 50 to 60 years, who underwent cardiac catheterization for evaluation of newly diagnosed coronary heart disease (luminal stenosis >70% in diameter in at least one of the major coronary arteries) were enrolled in the study. Information about arterial hypertension, hypercholesterolaemia, cigarette smoking, diabetes melitus, family history, and body mass index was recorded. Patients with arterial hypertension (systolic arterial pressure >140 mmhg and/or diastolic arterial pressure >90 mmhg), left main coronary artery disease, valvular heart disease, history of previous coronary heart disease, congenital heart disease, dilated cardiomyopathy, chronic obstructive pulmonary disease and diabetes mellitus were excluded prior to entry into the study. Fifty-six patients were enrolled in the study, but two were immediately lost from follow-up. Fifty-four patients were therefore followed. With respect to the clinical state, 14 patients had acute myocardial infarction, 20 unstable angina and 20 stable angina. The patients received standard medical therapy from their physicians for the secondary prevention of coronary artery disease. Thirteen patients underwent revascularization, seven percutaneous transluminal coronary angioplasty and the remaining six, bypass surgery. Patients were either examined at least every 6 months in an outpatient clinic (44% of patients) or interviewed by telephone (56%). In the latter case, family members or the personal physicians were also interviewed. We recorded new coronary events on follow-up (medical records were obtained in 85% of these cases). The protocol was approved by our Institutional Ethics Committee and all patients gave informed consent before participating. Study protocol Studies were performed at 0900h in a controlled room temperature of 20 C. Patients arrived in the catheterization laboratory in the fasting state and underwent diagnostic left heart catheterization and coronary arteriography using a standard percutaneous femoral approach. After diagnostic catheterization, all patients were allowed to relax in the supine position for 30 min. Invasive measurements of aortic diameters and pressures were done as previously described [7 11], simultaneously with aortic root echocardiograms and sphygmomanometric blood pressure measurements. Invasive procedure Briefly, a Y-shaped intravascular catheter, which was developed in our Institution and uses sonometry for the measurement of diameters, was used to measure the proximal descending aorta. At each arm of the catheter, a piezoelectric crystal (5 MHz in frequency, 1 mm, Crystal Biotech) is attached. The technical characteristics of the device include: (1) resolution for assessment of changes in diameter of 10 μm (2) flat ( 5%) frequency response in testing up to 40 Hz, (3) no measurable phase lag between forced oscillations of the device and the signal in the frequency response range (4) minimal loading on the aortic wall (0 45 g per arm when the distance between the arms is 1 cm). Aortic pressures were recorded with a catheter-tip micromanometer (Model SPC-330, Millar Instruments) [7 11]. Echocardiographic study Simultaneously the echocardiographic examination was carried out using Sonos 2500 Hewlett Packard equipment with a 2 5 MHz transducer to determine diameters of the ascending aorta. Each subject was placed in a slight left recumbent position and the ascending aorta was recorded at a level 3 cm above the aortic valve, on M-mode tracings, guided by the two-dimensional echocardiogram in the parasternal long axis view. Internal aortic diameters were measured according to the method previously described in detail [12]. All patients had blood pressure measured in the supine position with a mercury sphygmomanometer. Aortic elastic indexes [7 12], namely aortic strain and aortic distensibility were calculated from the echocardiographically-derived aortic diameters and the clinical blood pressure. The pressure diameter relationship and aortic stiffness constant (Fig. 1), were calculated from the invasively derived recordings according to methods explained in the appendix. Wave reflections were evaluated by measuring the augmentation index by the high-fidelity aortic pressure waveform. Statistical analysis We used one-way analysis of variance for comparisons of means and chi-square for comparisons of proportions. Correlations were evaluated with curve estimation regression statistics and related plots. The Kaplan Meier product limit estimate method was used to determine the rate of progression from study entry to the end-points. These end-points included acute coronary syndromes namely, hospitalization for unstable angina or acute myocardial infarction. When an end-point was reached, the patient was censored from further analysis.

392 C. Stefanadis et al. 120 Aortic pressure (mmhg) 109 98 87 76 y = 0 1003e 0 698x R 2 = 0 999 65 20 9 21 3 21 7 Aortic diameter (mm) Figure 1 Graph showing the aortic pressure diameter relationship. Aortic pressure diameter data obtained during the ventricular ejection, which corresponds to the ascending limb of the loop, were used for the calculation of aortic stiffness constant. 22 1 22 5 Kaplan Meier product limit estimate curves were compared by the log-rank test to provide univariate assessment of the relationship of clinical risk factors and haemodynamic and functional variables to the primary end-points [13]. For all analyses, continuous variables were partitioned according to statistical terciles (i.e. the group was divided into thirds according to the distribution of the variable of interest, with no a priori assumptions or biases relative to cut points) [14]. Variables that were significantly (P<0 05) related to end-points were entered into a multivariate forward stepwise Cox proportional hazards models to confirm their independent predictive value [15]. Results Clinical parameters and measures of aortic performance are presented in Table 1. During follow-up (average 3 years) 12 of 54 patients reached an end-point. From these, eight developed unstable angina and four acute myocardial infarction. The event free survival rate at 58 months of follow-up was 78% (Fig. 2). No patient died from a non-cardiac cause or developed symptoms of ischaemia of the brain or limbs. Pulse pressure was increased, aortic strain and distensibility were decreased and aortic stiffness constant was increased (Table 1). There were no differences between the three groups with respect to all other variables. An excellent inverse correlation between the aortic stiffness constant and aortic distensibility was found if data from all patients were plotted altogether (y=0 98 1 27, r=0 92, P=0 0000, Fig. 3). Univariate predictors of end-points Univariate analysis revealed that aortic stiffness constant, aortic distensibility, and revascularization procedures were significantly associated with the development of recurrent acute coronary syndromes. Multivariate analysis for independent contribution to end-point prediction Although several variables provided significant prediction by univariate analysis, when a multivariate Cox proportional hazards model was constructed, the only independent predictor of risk of end-points was the aortic stiffness constant. The relative risk for any cardiac end-point was 3 24 with 95% confidence interval, 1 79 to 5 83; P<0 001. When aortic stiffness was not considered, all independent prognostic information for acute coronary syndromes was carried by the non-invasive aortic root distensibility alone (relative risk: 0 37 with 95% confidence interval 0 21 to 0 65; P<0 001). Ratio free of all cardiac end-points 0 9 0 8 0 7 0 6 0 4 0 3 0 2 0 1 0 10 20 30 40 50 Months Figure 2 Kaplan Meier analysis of survival of progression to end-points in patients with coronary artery disease. 60

Aortic stiffness and recurrent coronary events 393 Table 1 constant Baseline characteristics of the study patients according to aortic stiffness Characteristic 0 29 6 n=18 Aortic stiffness constant (mm 1 ) 7 0 73 n=18 0 74 3 95 n=18 P value Age (years) 54 3 55 2 55 3 ns Female sex, no. (%) 2 (11) 2 (11) 2 (11) ns Body-mass index (kg. m 2 ) 24 9 2 3 25 5 2 4 25 0 2 2 ns High serum cholesterol, no. (%) 5 (28) 6 (33) 6 (33) ns Cigarette smoking, no. (%) 7 (39) 7 (39) 8 (44) ns Myocardial infarction, no. (%) 5 (28) 8 (44) 7 (39) ns Unstable angina, no. (%) 7 (39) 6 (33) 7 (39) ns Stable angina, no. (%) 6 (33) 4 (22) 4 (22) ns New events, no. (%) 2 (11) 7 (39) 4 (22) <0 05 1 Vessel disease, no. (%) 8 (44) 9 (50) 5 (28) ns 2 Vessel disease, no. (%) 5 (28) 6 (33) 7 (39) ns 3 Vessel disease, no. (%) 5 (28) 3 (17) 6 (33) ns Left anterior descending, no. (%) 13 (72) 13 (72) 14 (78) ns Left circumflex, no. (%) 10 (56) 8 (44) 12 (67) ns Right coronary artery, no. (%) 10 (56) 9 (50) 11 (61) ns Revascularization procedures, no. (%) 2 (11) 7 (39) 4 (22) ns Beta-blocker, no. (%) 5 (28) 6 (33) 7 (39) ns ACE inhibitors, no.(%) 8 (44) 8 (44) 9 (50) ns Calcium-channel blockers, no. (%) 4 (22) 5 (28) 3 (17) ns Aspirin, no. (%) 8 (44) 9 (50) 10 (56) ns Left ventricular ejection fraction (%) 46 1 4 3 46 4 4 2 46 4 4 4 ns Heart rate (beats. min 1 ) 62 9 60 9 64 11 ns Aortic pressure (mmhg) Systolic 111 4 10 7 113 7 7 0 116 6 8 5 ns Diastolic 68 6 10 4 69 2 8 2 67 1 11 4 ns Pulse 42 8 4 6 44 5 5 4 49 5 5 1 <0 001 Aortic diameter (mm) Systolic 21 95 2 72 22 57 2 49 21 97 2 46 ns Diastolic 20 46 2 70 21 23 2 34 21 27 2 18 ns Strain (%) 7 2 6 2 3 2 <0 001 Distensibility (cm 2.dyn 1.10 6 ) 2 67 0 68 2 12 3 2 0 64 <0 001 Stiffness constant (mm 1 ) 0 42 0 08 0 66 0 05 1 56 1 22 <0 001 Augmentation index 0 45 0 05 0 47 0 05 0 48 0 06 ns Aortic pressures, diameters, strain and distensibility referred to values obtained non-invasively by echocardiogram and mercury sphygmomanometer. 4 5 Distensibility (cm 2.dyn 1.10 6 ) 4 0 3 5 3 0 2 5 2 0 1 5 y = 0 9831x 1 268 R 2 = 0 8474 0 1 5 2 0 2 5 3 0 3 5 4 0 Stiffness constant (mm 1 ) Figure 3 An excellent inverse correlation is found between distensibility and aortic stiffness constant when data from all patients are plotted altogether.

394 C. Stefanadis et al. Ratio free of all cardiac end-points 0 9 0 8 0 7 0 6 0 4 0 3 0 2 0 1 0 10 20 30 40 50 Months Figure 4 Relationship of aortic stiffness at study entry to occurrence of any cardiac end-point (unstable angina and acute myocardial infarction) during follow-up. The whole population has been divided statistically into terciles. Aortic stiffness boundaries for each tercile are = 0 29 6 mm 1 ; =7 0 73 mm 1 ; =0 74 3 95 mm 1 ; P=0 001. A stiffness constant of 0 74 mm 1 was the lower boundary of the tercile comprising the patients at higher risk (Fig. 4). Patients with aortic stiffness constant higher than this value manifested a 44% risk of recurrent acute coronary events during 58 months of follow-up. The lowest risk tercile was bounded by an aortic stiffness constant of 6 mm 1. Patients with lower aortic stiffness showed a 5 6% risk of progression to any cardiac end-point during the follow-up period. Similarly aortic distensibility boundaries for each tercile and the relationship of aortic distensibility to occurrence of any cardiac end-point are shown in Fig. 5. Discussion 60 Ratio free of all cardiac end-points 0 9 0 8 0 7 0 6 0 4 0 3 0 2 0 1 0 10 20 30 40 50 Months Figure 5 Relation of aortic root distensibility derived non-invasively at study entry to occurrence of any cardiac end-point during follow-up. The whole population has been divided statistically into terciles. Aortic distensibility boundaries for each tercile are =2 42 3 98 cm 2. dyn 1.10 6 ; =1 58 2 42 cm 2. dyn 1. 10 6 ; =0 14 1 58 cm 2. dyn 1.10 6 ; P=0 016. Our data support our hypothesis that the development of recurrent acute coronary syndromes can be predicted among patients with coronary artery disease on the basis of aortic elastic properties. This finding is consistent with the results of previous studies, which also concluded that a high pulse pressure in men is an independent predictor of cardiovascular and especially coronary mortality in both hypertensives and normotensives [1,3]. Our findings are also consistent with studies in which arterial alterations, as determined from carotid elastic modulus, are strong independent predictors of cardiovascular mortality in patients with end-stage renal disease undergoing haemodyalisis [16]. This study adds to those of previous investigators first by demonstrating the relative prognostic power of aortic stiffness constant in comparison with more conventional indices of aortic function and second, by demonstrating that aortic stiffness can predict acute coronary events in patients with coronary artery disease. Thus, we found that aortic stiffness constant was the best single predictor of acute coronary syndromes. The relative utility of the aortic stiffness constant is consistent with our previous reports in hypertensive patients studied after diltiazem administration [9], and in patients during pacing-induced tachycardia [11]. Aortic stiffness constant is a more specific index of aortic elasticity. It takes into account data only from the ejection period. The initial ejection phase mainly depended on the elastic fibres to express the elastic properties of the aortic wall. This study confirmed that aortic distensibility in patients with coronary artery disease can be obtained non-invasively with a high degree of accuracy [12]. This method may be of value for prognosis as well as in long-term follow-up of patients at a high risk of acute coronary events. Distensibility, however, changes from the aortic root to the periphery and, in certain cases, may be quite different at different levels of the aorta. Our results demonstrate that recurrent acute coronary episodes can be predicted both invasively and noninvasively. Thus, although the aortic pressure diameter relationship is the most accurate determinant of segmental aortic elastic properties, non-invasively derived aortic root distensibility is a sufficient alternative non-invasive index of aortic elastic properties. It is well known that aortic elastic properties are affected by the risk factors for ischaemic heart disease, such as age [17], hypertension [9], hypercholesterolaemia [18], and smoking [8,10]. Aortic stiffness is increased with the presence of these risk factors. The findings of the present study indicate that aortic stiffness is an index of the cumulative effect of all risk factors on the elastic properties of the arterial wall. Consequently, aortic stiffness constant can predict the recurrence of coronary events. Aortic stiffness measurement may help in the evaluation of the individual risk and therefore in 60

Aortic stiffness and recurrent coronary events 395 the therapeutic decision making during secondary prevention in patients with ischaemic heart disease. Limitations The main finding of this study is that coronary events may be predicted by aortic stiffness constant. However, there are multiple known risk factors which affect and modulate outcome. Thus, although the aortic stiffness can be used as an additional risk factor this index is by no means a substitute for established risk factors that increase the risk for acute coronary events. An excellent inverse correlation was found between aortic stiffness constant in the proximal descending and aortic distensibility in the ascending aorta, in spite of the fact that the segments studied were different and the existence of altered techniques. This relationship indicates that there is a proportional progression of aortic stiffness at different aortic segments in all patients. Thus, the limitation that both indexes represent focal elastic properties of the aorta is counterbalanced. Despite the several limitations of our study, these findings support the use of our technique in patients with coronary artery disease and acute coronary syndromes and without detectable known risk factors. In this population, aortic stiffness may be an additional risk factor that adversely affects the prognosis in these patients. Furthermore, the impact and efficacy of several medications on aortic impedance has not been studied in detail in post myocardial infarction patients. The noninvasive technique can be widely and repeatedly applied in this setting. In conclusion, patients with coronary artery disease are at increased risk of recurrent events if they have increased aortic stiffness. A clinical application of these findings is to suggest an aortic stiffness measurement in patients with coronary artery disease. Combining evidence from major trials should give the most precise estimate for risk stratification in the management of patients with coronary artery disease. Appendix Estimation of aortic elastic properties [7 12] Non-invasive aortic strain and distensibility Aortic strain was calculated as the ratio: Aortic strain=(systolic diastolic aortic diameter)/diastolic aortic diameter. Cross-sectional distensibility of the ascending aorta was calculated with the following formula: Aortic Distensibility=2 strain/aortic pulse pressure. Aortic pressure diameter relationship Aortic pressure diameter relation was obtained by plotting the pressure vs diameter of digitized data by means of commercially available computer software (Fig. 1). Aortic stiffness constant (Fig. 1) Aortic pressure diameter data obtained during the ventricular ejection, which corresponds to the ascending limb of the loop, were used for the calculation of aortic stiffness constant. The rate of aortic blood pressure changes (dp/dt) was instantaneously calculated and simultaneously recorded with the high-fidelity pressure. Pressure and diameter data during the ascending limb of the loop, starting at the time when the dp/dt curve reached zero baseline (at the beginning of the ascending limb of the loop) and ending at peak+dp/dt, were fitted to the exponential function: P=b e a.d, where P was the instantaneous aortic pressure and D the aortic diameter. The least-squares method was used for calculation of a and b, where a is the elastic aortic stiffness constant (mm 1 ), which determines the slope of the exponential curve, and b is the elastic constant (mmhg). Wave reflections Wave reflections were evaluated by measuring augmentation index defined as the ratio: (pressure from inflection point to late systolic peak)/(pulse pressure). Inflection point (P i ) and late systolic peak were defined by using the fourth derivative of pressure. The first zero crossing of the fourth derivative (in a direction from above to below the zero line) corresponded to the beginning of the pressure wave upstroke. The second zero crossing in the same direction corresponded to the P i. This study was supported by a grant from the Hellenic Heart Foundation. References [1] Lichtenstein MJ, Shipley MJ, Rose G. Systolic and diastolic blood pressures as predictors of coronary heart disease mortality in the Whitehall study. Br Med J 1985; 291: 243 5. [2] Mitchell GF, Moye LA, Braunwald E et al., for the SAVE Investigators. Sphygmomanometric determined pulse pressure is a powerful independent predictor of recurrent events after myocardial infarction in patients with impaired left ventricular function. Circulation 1997; 96: 4254 60. [3] Madhavan S, Ooi WL, Cohen H, Alderman MH. Relation of pulse pressure and blood pressure reduction to the incidence of myocardial infarction. Hypertension 1994; 23: 395 401. [4] Lang RM, Cholley BP, Korcarz C, Marcus RH, Shroff SG. Measurement of regional elastic properties of the human aorta. A new application of transesophageal echocardiography with automated border detection and calibrated subclavian pulse tracings. Circulation 1994; 90: 1875 82. [5] Cholley BP, Shroff SG, Sandelski J et al. Differential effects of chronic oral antihypertensive therapies on systemic arterial circulation and ventricular energetics in African-American patients. Circulation 1995; 91: 1052 62. [6] Hirai T, Sasayama 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: 78 86. [7] Stefanadis C, Stratos C, Vlachopoulos C et al. Pressurediameter relation of the human aorta. A new method of determination by the application of a special ultrasonic dimension catheter. Circulation 1995; 92: 2210 19. [8] Stefanadis C, Tsiamis E, Vlachopoulos C et al. Unfavorable effect of smoking on the elastic properties of the human aorta. Circulation 1997; 95: 31 8. [9] Stefanadis C, Dernellis J, Vlachopoulos C et al. Aortic function in arterial hypertension determined by pressure-diameter relation: Effects of diltiazem. Circulation 1997; 96: 1853 8.

396 C. Stefanadis et al. [10] Stefanadis C, Vlachopoulos C, Tsiamis E et al. Unfavorable effects of passive smoking on aortic function in men. Ann Intern Med 1998; 128: 426 34. [11] Stefanadis C, Dernellis J, Vavuranakis M et al. Effects of Ventricular Pacing-Induced Tachycardia on Aortic Mechanics in Man. Cardiovasc Res 1998; 39: 506 14. [12] Stefanadis C, Stratos C, Boudoulas H, Vlachopoulos C, Kallikazaros I, Toutouzas P. Distensibility of the ascending aorta in coronary artery disease and changes after nifedipine administration. Chest 1994; 105: 1017 23.. [13] Glanz SA. Primer of biostatistics. New York: McGraw-Hill Companies, Inc, 1981: 373 99. [14] Borer JS, Hochreiter C, Herrold E et al. Prediction of indications for valve replacement among asymptomatic or minimally symptomatic patients with chronic aortic regurgitation and normal left ventricular performance. Circulation 1998; 97: 525 34. [15] Armitage P, Berry G. Statistical methods in medical research, 3rd edn. London: Blackwell Science Ltd, 1994: 484 9. [16] Blacher J, Pannier B, Guern AP, Marchais SJ, Safar ME, London GM. Carotid arterial stiffness as a predictor of cardiovascular and all-cause mortality in end-stage renal disease. Hypertension 1998; 32: 570 4. [17] Nichols WW, O Rourke MF. McDonald s blood flow in arteries. Theoretical, experimental and clinical principles, 4th edn. London; Arnold, 1998: 418 76. [18] Pitsavos C, Toutouzas K, Dernellis J et al. Aortic stiffness in young patients with heterozygous familial hypercholesterolemia. Am Heart J 1998; 135: 604 8.