Anesthesiologists often use radial pulse wave appearance to estimate stroke volume and hypovolemia. However, unrecognized differences between central aortic and peripheral arterial pressures may lead to incorrect conclusions. It is usually assumed that the pressure is about the same throughout the arterial tree. This may be correct in regard to mean and diastolic pressures (1), but not to the systolic pressure. In young persons, the systolic pressure is higher in the periphery than in the ascending aorta because of arterial resonance (2).
With increasing age or hypertension, systolic pressure in the ascending aorta increases to levels close to those in peripheral arteries because the reflected energy from the efferent pulse wave will return to the ascending aorta during cardiac ejection, thereby increasing aortic systolic pressure and left ventricular afterload (3).
Through relaxation of smooth muscles in small arteries, nitroglycerin decreases arterial elastance and, hence, pulse wave velocity (4). A decreased velocity implies that the reflected pulse wave comes later during cardiac ejection. Aortic systolic pressure is therefore reduced, and pressure after aortic valve closure during early diastole is increased (5). Indeed, this indicates that nitroglycerin is beneficial to coronary patients not only by reducing preload and relaxing coronary vessels, but also by reducing afterload and by improving coronary perfusion pressure. Because such effects may not be readily apparent when blood pressure is monitored in the radial artery (6), we also measured blood pressure in the ascending aorta in patients scheduled for coronary artery surgery. From a monitoring perspective, our aims were: 1) to find the major hemodynamic determinants for the change in radial pulse wave configuration during nitroglycerin administration; and 2) to describe the increasing difference between radial and aortic systolic pressures when blood pressure is reduced with nitroglycerin. From a pathophysiological perspective, our aim was to analyze to what extent changes in stroke volume, vascular resistance/impedance, and wave reflection are responsible for the increased difference between radial and aortic systolic pressures.
After having confirmed that nitroglycerin has a selective ability to decrease aortic systolic pressure and that this was clearly associated with decreases in stroke volume index (SVI), we included prostacyclin for comparison. In contrast to nitroglycerin, prostacyclin causes arteriolar dilation and usually does not affect SVI. Prostacyclin could thus help in understanding the effects of nitroglycerin.
The study was approved by the hospital’s ethical committee, and informed consent was received from each patient. Eighteen patients scheduled for coronary artery bypass grafting were included (Table 1). Of the chronic preoperative drug therapy, only β-receptor blockers were administered with the premedication. Measurements were performed on the morning before surgery. The patients were premedicated according to age and weight with flunitrazepam orally and morphine and scopolamine intramuscularly. During catheterization, sedation was supplemented with intermittent doses of 1 mg of midazolam IV to assure that all patients were drowsy or sleeping during the study. Two liters of oxygen per minute was supplied nasally, and pulse oximetry was monitored.
Aortic recordings were obtained with an 8 F Cordis tip manometer catheter with a pressure interface (Sentron, Roden, The Netherlands) (7). The catheter has a flat frequency response up to 180 Hz. The catheter was introduced into the ascending aorta from the femoral artery and positioned just proximal to the brachiocephalic trunk. Pulmonary arterial pressure and central venous pressure were obtained with a 7.5 F pulmonary artery flow-directed catheter. Cardiac output was measured by thermodilution. The radial artery was cannulated with a 45-mm Ohmeda (Swindon, UK) arterial cannula (inner diameter 1.0 mm) and connected to external pressure domes (Peter van Berg, Kirchseeon, Germany) with only 10 cm of high-pressure tubing to minimize signal distortion. When tested in our laboratory (8), this pressure set showed a natural frequency of 25 Hz and a damping coefficient of 0.35–0.5. Siemens Sirecust 1281 cardiovascular monitoring system (Siemans Medical Electronics Inc., Danvers, MA) uses a high-pass filter with a cutoff at 16 Hz. This implies that the signal has minimal distortion. All signals were AD-converted on-line at a 200-Hz sampling rate and processed with data acquisition software (Acqknowledge; Biopac).
Care was taken to measure cardiac output (mean of three to four bolus injections equally distributed through the respiratory cycle) during steady-state conditions in the control state and at each steady level of lowered mean arterial blood pressure (MAP). Cardiac output and stroke volume were indexed according to body surface area (CI and SVI, respectively). Systemic vascular resistance index (SVRI) and left ventricular minute and stroke work indices (LVMW and LVSWI, respectively) were calculated according to standard formulae.
To correlate the intermittent thermodilution measurements with the continuous pressure registration, the latter was averaged. The start of arterial systole and the aortic valve incisure were identified by the maxima of the second derivative of the pulse wave. The relative contribution of systolic area to the entire pulse wave was expressed as aortic systolic integral percentage (AoSi%) and radial systolic integral percentage (RaSi%), respectively.
Augmentation index (AI) of the pulse pressure, an index of wave reflection, was measured by taking the ratio of the height of the late systolic peak (S2) to that of the early systolic peak (S1). Diastolic pressure is subtracted from both pressures, i.e.,EQUATION 1 The systolic inflection points were detected as described by Kelly et al. (9) and Takazawa et al. (10) by using the fourth derivative of the pulse. The fourth derivative defines systolic inflection during upstroke in the aorta at the second zero crossing (from above to below) and in the radial pulse wave at the third crossing (from below to above) (Fig. 1). In the radial artery, visual inspection could always be used to confirm detection of the second systolic peak by the fourth derivative. Because, in the aorta, the systolic upstroke of all aortic pulse waves were smooth (Fig. 2, lower panel), visual detection of systolic inflection (11) was impossible. Therefore, the aortic augmentation index (AoAI) was entirely based on the fourth derivative (9).
Maximal rate of rise in the aortic pulse curve (max dP/dt) was used as an index of cardiac contractility, and the delay between the start of systolic upstroke in the aorta and the radial artery (pulse wave delay; PWD) was used as an index of pulse wave velocity in the arm. Effective arterial elastance index (EaI) was calculated as aortic systolic pressure × 0.9 × SVI−1 according to Kelly et al. (12). As a measure of pulsatile load, the difference between EaI and the quotient SVRI/RR interval was calculated (12).
The last third of diastole in the ascending aorta was used for calculation of arterial compliance (CaI) according to Liu et al. (13). The formula used was: EQUATION 2 In this formula, P2 represents diastolic pressure and P1 represents aortic pressure when one third of diastole remains.
Aortic mean pressure (AoMAP) was reduced (0%–5%, 10%–15%, and 20%–30%) in two to four steps with IV infusions of nitroglycerin (1 mg/mL) or prostacyclin (2 μg/mL). In the first nine patients, only nitroglycerin was used. In the subsequent nine patients, the protocol started with prostacyclin; after a washout period of at least 15 min, nitroglycerin was infused, aiming at similar stepwise reductions of MAP with both drugs. Thus, the investigation was designed to study effects, not doses.
All results are given as mean ± SEM. In comparisons within and between groups, Student’s t-test was applied. Stepwise analysis of covariance was used to discriminate the importance of changes in the following dependent variables: aortic-radial systolic pressure gradient (RaSAP-AoSAP), AoSi%, AoAI%, RaSi%, and RaAI%. Independent variables were: AoSAP, AoMAP, RaSAP, RaSAP-AoSAP (when not used as a dependent variable); central venous pressure (CVP), pulmonary capillary wedge pressure (PCWP), ejection time, heart rate (HR), PWD, SVI, SVRI, EaI, Aomax dP/dT, and CaI. When RaSAP-AoSAP was used as a dependent variable, AoAI% and RaAI% (indices of pulse wave reflection) were included as independent variables. A P value <0.05 was considered statistically significant.
A similar mean (range) total decrease in AoMAP was achieved with nitroglycerin (−12 [2 to −53] mm Hg) and prostacyclin (−10 [4 to −28] mm Hg) (Table 2). Nitroglycerin, but not prostacyclin, increased the mean RaSAP-AoSAP gradient from 6 to 16 mm Hg. During nitroglycerin administration, AoSAP can be predicted to be 0.84 × RaSAP + 5 mm Hg (Fig. 3). The corresponding formula without vasodilation in the control state shows AoSAP = 0.98 × RaSAP − 2 mm Hg.
Stepwise analysis of covariance (Table 3) showed that 88% (coefficient of determination) of the nitroglycerin-induced effect on systolic pressures (RaSAP-AoSAP) is explained by changes in SVI, MAP, and EaI. Because the increased RaSAP-AoSAP relates to a decreased AoSAP, the aortic pulse wave was analyzed. Changes in AoSi% are 94% explained by SVI, HR, and SVRI, whereas changes in AoAI are explained by ejection time and CVP. In the periphery, changes in RaSi% are explained by changes in arterial pressure, SVI, and EaI, whereas changes in RaAI are explained by SVI and EaI.
Similar analysis regarding prostacyclin showed that 90% of changes in AoAI are explained by aortic pressure alone (Table 3). Changes in RaAI, RaSAP-AoSAP, AoSi%, and RaSi% are not correlated to hemodynamic changes or relate to eight or more covariates.
Simultaneous recordings in the radial artery show that both drugs reduced the late systolic peak seen in the control state (Fig. 2). However, during infusion of nitroglycerin, late systolic radial pressure was reduced below systolic peak pressure, whereas the systolic pulse wave during prostacyclin retained a slightly double-peaked or rounded appearance. Consequently, the RaAI is far more reduced during a nitroglycerin than a prostacyclin infusion (Table 2). A similar reduction in augmentation is shown in the aorta, although the inflection point between the first and second systolic peak is not visibly discernible (see Methods).
Nitroglycerin was associated with lower SVI and cardiac filling pressures than prostacyclin (Table 2). Concerning indices of resistance/impedance, SVRI and EaI were lower with prostacyclin, but only nitroglycerin reduced the pulsatile load (EaI − [SVRI × RR−1]) and increased arterial compliance. Likewise, only nitroglycerin decreased PWD and Aomax dP/dt.
In this study, we confirmed the clinical observation that a slender pulse wave in the radial artery is associated with a small stroke volume. In our patients, nitroglycerin lowered arterial pressure by decreasing preload and, thus, stroke volume. In contrast, prostacyclin lowered arterial pressure by reducing SVR but did not affect stroke volume.
When nitroglycerin is used, the second systolic peak should be monitored (Fig. 2, upper panel). This peak (of which RaAI is an index) is considered to be the sum of the cardiac pulse and a pulse wave reflected to the radial artery from other parts of the vascular tree (14,15), especially from the lower part of the body (16). The reflection is augmented by increased EaI, as in hypertension and in old age. The radial pulse wave in Figure 2 is a typical example of this. With nitroglycerin, the decrease in RaAI is explained by decreases in SVI and in arterial stiffness (EaI). Most anesthesiologists tend to monitor the systolic fraction of the total pulse (RaSi%). Like RaAI, this is also decreased by stroke volume but is not related to elastance. In contrast, pulse wave attenuation during prostacyclin is not explained by any specific hemodynamic variable, which suggests that prostacyclin-induced changes in stroke volume and wave reflection are small. Because we did not measure pulse wave velocity in the lower parts of the body, we cannot tell how much different timing of wave reflection contributes to the decline in RaAI.
The major pulse wave difference between nitroglycerin and prostacyclin is the large RaSAP-AoSAP gradient with nitroglycerin (Fig. 3). This is explained by decreases in SVI, MAP, and EaI. Reduced pulsatile flow and elastance contribute to the decrease in AoSAP. To what extent is a delay in wave reflection of importance when the two drugs are compared? To study wave reflection, spectral analysis of pressure and flow velocity in the frequency domain is usually applied (5). However, time-domain based variables, as in the present study, can only indirectly reveal properties of wave reflection (12). True EaI is an end-systolic event that reflects impedance to cardiac ejection. Our estimation of EaI (0.9 × AoSAP × SVI−1) closely correlates to true end-systolic measurements (12). End-systole is the period during which wave reflection is pronounced in the ascending aorta. During this period, AoAI has often been used as a variable for wave reflection (10,17). Both EaI and AoAI were reduced by nitroglycerin and by prostacyclin, with EaI somewhat more affected by prostacyclin. However, decreases in AoAI were only explained by preload (CVP) and ejection time during nitroglycerin administration and by aortic pressure during prostacyclin infusion. Notably, the decreased wave speed with nitroglycerin did not explain the reduction in AoAI. Thus, our data indicate that wave reflection is altered by both drugs, but through different mechanisms. Nitroglycerin relaxes larger arteries from the size of the brachial artery (18), but not including arterioles, which is the main site for prostacyclin-induced vasodilation. These pharmacodynamic differences corresponded in our nitroglycerin infusions to a reduced pulsatile load (EaI − [SVRI × RR−1]) and an increased late diastolic CaI, whereas prostacyclin was only associated with reduced SVRI. Although we show indirect evidence for reduced wave reflection with both drugs, the main reason for the increased RaSAP-AoSAP gradient during nitroglycerin-induced hypotension seems to be reduced stroke volume.
In the present study, the aortic pulse waves were all monophasic without clearly visible systolic inflection. This appearance did not change with either nitroglycerin or prostacyclin (Fig. 2). We suggest that lack of clear inflections during systolic upstroke (11) depend on the relatively older age (14) and stiff arterial vasculature of our patients. This implies that wave reflection in elderly patients with serious coronary disease is so rapid and pronounced that earlier parts of cardiac ejection are also affected. The reflected wave is totally integrated into the systolic part of the aortic pulse wave and cannot be visualized even when wave reflection is delayed pharmacologically. It is therefore not surprising that a nitroglycerin-induced decrease in stroke volume masks the importance of a delayed wave reflection on the decrease in AoSAP.
In summary, from a monitoring perspective, we showed that conclusions about central hemodynamic conditions derived from the radial arterial pulse wave should be arrived at with caution. The typical changes induced by nitroglycerin are strongly related to decreases in stroke volume and EaI, and to an increased difference in RaSAP-AoSAP. This difference can be easily predicted because it follows a linear equation. In contrast, however, alterations in the radial pulse wave during prostacyclin-induced hypotension do not have simple hemodynamic correlates. Furthermore, we provided indirect evidence that the decrease in AoSAP during a nitroglycerin infusion in the elderly patient is more dependent on a reduction in stroke volume than on a delay in wave reflection.
We thank Mrs. Marita Ahlquist for excellent technical assistance and Mr. Tommy Johnsson, senior lecturer, Department of Statistics, Gothenburg University, for statistical advice and valuable criticism.
1. Pauca AL, Wallenhaupt SL, Kon ND, Tucker WY. Does radial artery pressure accurately reflect aortic pressure? Chest 1992; 102: 1193–8.
2. O’Rourke MF. What is blood pressure? Am J Hypertens 1990; 3: 803–10.
3. O’Rourke MF, Kelly RP. Wave reflection in the systemic circulation and its implications in ventricular function. J Hypertens 1993; 11: 327–37.
4. Safar ME, Lévy, BI. The response of large arteries to antihypertensive treatment. II. Pharmacological aspects. In: O’Rourke MF, ed. Arterial vasodilation: mechanisms and therapy. London: E. Arnold, 1993: 127–66.
5. Yaginuma T, Avolio A, O’Rourke M, et al. Effect of glyceryl trinitrate on peripheral arteries alters left ventricular hydraulic load in man. Cardiovasc Res 1986; 20: 153–60.
6. Kelly RP, Gibbs HH, O’Rourke MF, et al. Nitroglycerin has more favourable effects on left ventricular afterload than apparent from measurement of pressure in a peripheral artery. Eur Heart J 1990; 11: 138–44.
7. Aubert AE, Vrolix M, De Geest H, Van de Werf F. In vivo comparison between two tip pressure transducer systems. Int J Clin Monit Comput 1995; 12: 77–83.
8. Gardner RM. Direct blood pressure measurement: : dynamic response requirements. Anesthesiology 1981; 54: 227–36.
9. Kelly R, Hayward C, Avolio A, O’Rourke M. Noninvasive determination of age-related changes in the human arterial pulse. Circulation 1989; 80: 1652–9.
10. Takazawa K, Tanaka N, Takeda K, et al. Underestimation of vasodilator effects of nitroglycerin by upper limb blood pressure. Hypertension 1995; 26: 520–3.
11. Murgo JP, Westerhof N, Giolma JP, Altobelli SA. Aortic input impedance in normal man: : relationship to pressure wave forms. Circulation 1980; 62: 105–16.
12. Kelly RP, Ting CT, Yang TM, et al. Effective arterial elastance as index of arterial vascular load in humans. Circulation 1992; 86: 513–21.
13. Liu Z, Brin KP, Yin FC. Estimation of total arterial compliance: : an improved method and evaluation of current methods. Am J Physiol 1986; 251 (3 Pt 2): H588–600.
14. O’Rourke M. Arterial stiffness, systolic blood pressure, and logical treatment of arterial hypertension. Hypertension 1990; 15: 339–47.
15. Nichols WW, Nicolini FA, Pepine CJ. Determinants of isolated systolic hypertension in the elderly. J Hypertens Suppl 1992; 10: S73–7.
16. Karamanoglu M, Gallagher DE, Avolio AP, O’Rourke MF. Functional origin of reflected pressure waves in a multibranched model of the human arterial system. Am J Physiol 1994; 267 (5 Pt 2): H1681–8.
17. O’Rourke MF, Kelly RP, Avolio A. The arterial pulse. Philadelphia: Lea and Febiger, 1992.
18. Simon AC, Levenson JA, Levy BY, et al. Effect of nitroglycerin on peripheral large arteries in hypertension. Br J Clin Pharmacol 1982; 14: 241–6.