Secondary Logo

Journal Logo

EACTA Original Article

Instantaneous diastolic pressure–flow relationship in arterial coronary bypass grafts

Kazmaier, S.*; Hanekop, G. -G.*; Grossmann, M.; Dörge, H.; Götze, K.*; Schöndube, F.; Quintel, M.*; Weyland, A.

Author Information
European Journal of Anaesthesiology: May 2006 - Volume 23 - Issue 5 - p 373-379
doi: 10.1017/S0265021505001985



Coronary perfusion pressure (CPP) is calculated by the difference between upstream pressure and effective downstream pressure. Coronary upstream pressure is represented by the mean diastolic aortic pressure. Coronary sinus pressure (CSP) or left ventricular end-diastolic pressure (LVEDP) is generally used as an equivalent of the coronary downstream pressure reflecting the specific characteristics of the coronary anatomy. Coronary vascular resistance (CVR) is calculated from the ratio of CPP and coronary blood flow (CBF).

In a theoretical approach to the pressure–flow (P–F) relationship in arterioles, it was shown that these calculations are a simplification of the actual variable tissue characteristics in the vascular bed of an organ [1]. In experimental investigations the perfusion of isolated extremities ceased at arterial pressures significantly higher than the respective venous pressures [2]. Further experimental trials on diastolic coronary P–F relationships suggested that the effective downstream pressure does not seem to be determined by CSP or LVEDP but by the critical occlusion pressure (COP) of the coronary vasculature, which was considerably higher than LVEDP [3–6]. Moreover, diastolic flow was linearly related to aortic pressure [7]. Dole and colleagues demonstrated that in healthy volunteers COP was 5–10 times higher than LVEDP indicating the presence of a vascular waterfall in the human coronary vascular system [7]. Similar results were found in studies in human beings with coronary artery disease (CAD) by Nanto and colleagues [8,9].

Until now, most intraoperative investigations on flow in coronary artery bypass grafts (CABGs) were primarily performed to evaluate the patency of distal anastomoses [10–12]. Little information is available on the diastolic P–F relationship and on the effective downstream pressure in CABG, although they might have considerable implications with regard to the understanding of flow characteristics following CABG surgery and to the guidance of perioperative treatment. Hence, the main objective of this explorative clinical study was to describe the instantaneous diastolic P–F relationships and to assess COP in CABGs following coronary bypass surgery.


The study was approved by the local institutional review committee and written informed consent was obtained from each patient. Fifteen patients (13 males and 2 females) with CAD were studied following elective CABG surgery. Patients with concomitant valvular heart disease or lack of sinus rhythm were excluded from this study. Antidysrhythmic and antihypertensive medications were continued until the day of surgery. Preanaesthetic medication consisted of 1.0 mg flunitrazepam orally on the evening prior to surgery, and 30 min before transfer to the operating room.

Management of anaesthesia and cardiopulmonary bypass

Induction of anaesthesia was performed with 2.0 μg kg−1 of sufentanil and 0.1 mg kg−1 of pancuronium bromide intravenously to facilitate endotracheal intubation. Anaesthesia was maintained with a balanced technique using 0.5 minimum alveolar concentration (MAC) isoflurane and a continuous infusion of sufentanil (1.0–2.0 μg kg−1 h−1). The patients' lungs were ventilated with a volume-controlled respirator (Cicero, Draeger GmbH, Lübeck, Germany). Respiratory rate, minute volume and FiO2 were adjusted to maintain normoxaemia and normocapnia.

Extracorporeal circulation with standard techniques included venous two-stage cannulation (Medtronic MC2TM, 91246C, 34/46Fr; Medtronic Inc., Minneapolis, USA), central aortic cannulation (Aortic Arch Cannula-Straight/Wire Inlay, 6.5 mm, A232-65; Stöckert Instrumente GmbH, München, Germany) and membrane oxygenation (Hilite® 7000; Medizintechnik AG, Stolberg, Germany). Surgery was performed with cross-clamped aorta and cardioplegic arrest with combined antegrade and retrograde cold blood cardioplegia (Dr. Franz Köhler Chemie GmbH, Alsbach-Hähnlein, Germany) (n = 8) or Histidine Tryptophane Ketoglutarate (HTK) solution (Custodiol®; Dr. Franz Köhler Chemie GmbH, Alsbach-Hähnlein, Germany) (n = 7).


Measurements were performed 15 min after discontinuation of extracorporeal circulation and prior to reversing heparin anticoagulation. During the short period of measurements, the patient was disconnected from the ventilator to avoid intrapleural pressure changes and volume shifts which might influence P–F relationships. In order to avoid capacitance effects of the arterial vessel on the results of this study, the analysis of P–F relationships only included data from the highest diastolic flow rate in the arterial bypass graft until the end of diastole.

Flow measurements were performed using ultrasound and calculations based on the transit time principle (Cardiomed 4008; Quick-Fit probes (size 2.0–3.0 mm), Medistim, Norway). Flow in the left internal mammary artery bypass (LIMAB) and pressure measurements were simultaneously recorded by the use of A/D converters over a period of 5 s with a sampling frequency of 500 Hz.

Aortic and CSPs were recorded via the standard cannulae for extracorporeal circuit and retrograde blood cardioplegia, respectively (Coronary sinus cannula: retrograde cardioplegia cannula RSH-M014S, 14Fr; Chase Medical, Richardson, Texas, USA). In patients who received antegrade crystalloid cardioplegia only, cannulation of the coronary sinus was carried out directly before discontinuation of cardiopulmonary bypass (CPB). Measurements of LVEDP were performed using a left atrial catheter (Jostra KLAP1751 pressure monitoring catheter, 5.0Fr; Jostra AG, Hirrlingen, Germany) introduced via the upper right pulmonary vein and positioned transmitrally into the left ventricle for the duration of the study period.


The COP was calculated by extrapolating the diastolic segment of the aortic P–F loop to the zero-flow pressure intercept using linear regression analysis. Mean LVEDP was assessed by analysis of the left ventricular pressure curve for each cardiac cycle during the study period (Fig. 1). CSP was calculated as the mean CSP of the entire measurement period. The pulsatility index was calculated by the difference between maximal and minimal flow divided by mean flow.

Figure 1.
Figure 1.:
Original recordings of LIMAB blood flow, blood pressures and ECG. The white area during diastole is the area of interest for the calculation of diastolic P–F relationship. Flow: blood flow in LIMAB to LAD; Paorta: aortic pressure; LVP: left ventricular pressure; ECG: electrocardiogram.


Results are expressed as mean ± standard deviation. Linear regression analysis was performed using flow as dependent variable and aortic pressure as independent variable. Pearson correlation coefficients were calculated in order to test for bivariate correlations. A P-value of <0.05 was considered statistically significant. Statistical procedures were performed with the SPSS/PC+™ statistical software package (Version 11.5, SPSS Inc., Chicago, Illinois, USA).


Data of five consecutive heart beats were analysed in each patient; none of the patients had to be excluded due to missing data or artefacts in measurements of flow, upstream and downstream pressures. Patient characteristics and perioperative data of all patients are given in Tables 1 and 2, respectively. The diastolic time interval of heartbeats which was used for the generation of pressure–flow velocity (P–V) relationships is shown in Figure 1.

Table 1
Table 1:
Patient characteristics data.
Table 2
Table 2:
Perioperative data.

The mean duration of the diastolic measurement interval was 233 ± 59 ms resulting in a mean number of P–V data sets of 582 ± 147 per patient. Mean diastolic aortic pressure, representing the coronary upstream pressure, and mean heart rate (HR) were 60.5 ± 10.0 mmHg and 86.5 ± 7.5 min−1, respectively. Mean CSP and LVEDP were 10.9 ± 3.1 mmHg and 14.4 ± 5.8 mmHg, respectively.

Mean diastolic flow in the arterial bypass graft was 46.6 ± 16.6 mL min−1 with a mean pulsatility index of 2.2 ± 1.2. Flow was linearly correlated to diastolic aortic pressure in all patients with a mean R-value of 0.93 ± 0.08. Mean COP, obtained by extrapolating the linear regression line of diastolic P–F relationship, was 32.3 ± 9.9 mmHg and thus exceeded mean CSP and mean LVEDP by a factor of 3.1 ± 1.1 and 2.6 ± 1.4, respectively. The exemplary calculation of COP is given in Figure 2. Linear P–F regression lines together with respective 95% confidence intervals (CI) and calculated zero-flow intercepts of all measurements are presented in Figure 3. The mean slope of the diastolic P–V relationships was 2.1 ± 1.3 mL min−1 mmHg−1.

Figure 2.
Figure 2.:
Relationship between aortic pressure (Paorta) and blood flow in LIMAB to LAD during five consecutive heart beats. Linear regression is performed on the diastolic part of the P–F loop. Extrapolation of the zero-flow pressure intercept defines the COP.
Figure 3.
Figure 3.:
Mean linear regression lines with 95% CI in 15 patients reflecting the inter-individual scatter of COP (zero-flow pressure intercept) and of the slope of P–F relationships. Flow: blood flow in LIMAB to LAD; Paorta: aortic pressure.

During the measurement period neither COP, the slope of the diastolic P–V relationship nor mean diastolic flow in the arterial bypass graft showed a significant correlation to CSP, LVEDP or HR (Fig. 4). Furthermore, these variables did not depend on the type of cardioplegia.

Figure 4.
Figure 4.:
Correlation between HR and COP.


CSP and LVEDP are commonly used as the downstream pressure to calculate CPP. The determination of the diastolic pressure at zero flow as a measure of the COP in internal mammary artery bypass (IMAB) grafts following CABG surgery revealed that both CSP and LVEDP systematically underestimate the effective downstream pressure for CABG.

Previous investigations focusing on COP in native coronary circulation used the extrapolated zero-flow pressure to define the effective downstream pressure of the coronary circulation. This method is sound since basic physiology predicts that blood flow ceases if the difference between the upstream and the downstream pressure in a vascular tree equals zero; thus, the arterial pressure at zero flow represents the effective downstream pressure of organ blood flow [1].

Dole and colleagues investigated diastolic coronary P–F velocity relationships in the left coronary artery by means of intravascular ultrasound in conscious men without CAD [7]. They found a linear relation between pressure and flow velocity with a mean slope of 0.35 ± 0.12 cm s−1 mmHg−1 before and 0.8 ± 0.48 cm s−1 mmHg−1 after inducing vasodilatation with angiographical contrast medium. Even during vasodilatation, the mean slope of the regression lines in their study was about half as steep as with our data indicating a higher vascular resistance. Aside from methodological differences, the discrepancies in slope may be explained by the fact that the patients in our study were anaesthetized with a balanced technique using isoflurane resulting in decreased sympathetic activity and coronary vasodilatation. This explanation is also supported by findings of Dole and colleagues who showed that COP decreased by 27% after vasodilatation [7]. However, mean COP in our study was about 17% lower than mean COP after vasodilatation in the study of Dole and colleagues [7].

P–F relationships in our patients were assessed in IMAB grafts and might considerably differ from flow characteristics in epicardial segments of native coronary arteries.

Results similar to those of Dole and colleagues were found by other groups in experimental studies as well as in clinical trials in human beings without CAD [8,9,13]. All studies demonstrated a linear relationship between pressure and flow velocity. Slope increased and COP decreased after vasodilatation. In all of these studies, intracoronary injections of adenosine triphosphate were given to achieve maximum vasodilatation. The administration of adenosine triphosphate did not only depress vasomotor activity but also induced an atrioventricular blockade followed by an increased duration of the diastole, which was up to 10 times longer when compared to our data. It remains questionable, whether the results assessed during maximal coronary vasodilatation and atrioventricular blockade reflect physiological conditions in human coronary circulation.

In most of the studies mentioned above, coronary flow was measured indirectly by assessing intracoronary blood flow velocity. Flow velocity is linearly related to flow only if the diameter of the vessel, in which the measurements are performed, remains constant. It was demonstrated in several studies that the cross-sectional area of epicardial coronary vessels is nearly independent from pressure and flow [14,15]. Thus, differences in results can be hardly explained by the different measurement techniques, but more likely by different study conditions.

In our study, the flow was directly assessed using ultrasound and the transit time principle. The relative accuracy of ±2% for measurements with the flow probes used in this study has been validated [11,16]. Measurements were performed in each patient during a 5-s period of apnea. As the heart frequency of all patients was above 60 beats min−1, five consecutive heart beats could be evaluated in every patient. Moreover, temporal resolution was sufficient due to a sampling frequency of 500 Hz resulting in 117 ± 29 paired P–F data per heart beat. The absence of intrapleural pressure changes which would have altered right ventricular afterload and left ventricular preload resulted in stable haemodynamic conditions with a small intra-individual beat-to-beat variance of data. These stable haemodynamic conditions are shown in Figure 2, which depicts the P–F relationships for five complete cardiac cycles.

Until now, information on the relationship between diastolic upstream pressure and flow in coronary bypass grafts has been lacking. Most previous investigations studying flow in CABGs were aimed at evaluating the patency of the graft and the distal anastomosis, respectively [10,17–19]. These studies demonstrated that under conditions of stable haemodynamics, a high flow rate and a low pulsatility index of the flow curve are indicators of graft patency. In our study, apart from flow values and a low pulsatility index of 2.2 ± 1.2, the slope of the P–F relationship indicated a good patency of IMAB grafts.

In contrast to data on graft patency, the clinical and scientific value of diastolic P–F relationships and the calculation of COP in coronary bypass grafts is still a matter of discussion due to the lack of data. In the first study, Shimada and colleagues demonstrated a close negative correlation between COP and residual myocardial viability after angioplasty in patients with acute myocardial infarction [20]. In a group of similar patients, Furber and colleagues found a close correlation between short- and long-term myocardial outcome and deceleration time of diastolic coronary flow velocity [21]. In our study in patients undergoing coronary bypass grafting, the COP showed a significant negative correlation with the deceleration time of diastolic flow as presented in Figure 5. Thus, the COP may provide immediate, intraoperative information on the long-term function of the CABG and possibly on myocardial outcome. The analysis of diastolic P–F relationship may provide important information for understanding the mechanical determinants of CBF and may be useful to describe coronary bypass graft function in combination with flow characteristics of the native coronary vessel. The results of our study demonstrate that the effective downstream pressure of CABGs to the left anterior descending (LAD) is consistently higher than the pressure in the coronary sinus resulting in a considerable systematic overestimation of CPP. Substituting LVEDP for CSP as a measure of downstream pressure might be a basically more meaningful approach, as the left ventricular pressure may more clearly reflect the influence of extravascular compression on the coronary vasculature of the left ventricle than CSP. On the other hand, LVEDP does not exactly reflect the pressure in myocardial tissue and, in addition, ignores the influence of arteriolar vasomotor tone which is another major determinant of COP. Accordingly the IMAB zero-flow pressure determined in our study exceeded LVEDP by a factor of 2.6.

Figure 5.
Figure 5.:
Correlation between diastolic deceleration time (DDT) and COP.

Beside these theoretical considerations our observational study also has clinical implications. First, determining coronary COP from arterial grafts enables a more valid calculation of CPP, which may be particularly important in perioperative episodes of arterial hypotension. Second, the assessment of regional CVR from the slope of the diastolic coronary P–F relationships (representing the coronary vascular conductance which is inversely related to CVR) is independent of coronary downstream pressure. Third, this technique enables a better differentiation between changes of extravascular pressure or vasomotor tone (primarily affecting COP) and alterations of vascular diameter (primarily affecting the slope of P–F relationships). This may be of particular clinical interest in situations of impaired graft flow or when one needs to evaluate the influence of vasoactive drugs and anaesthetics during cardiac surgery.

Similarly, the effects of different types of cardioplegia on myocardial tissue characteristics and, consequently, on CBF after discontinuation of extracorporeal circulation are clinically relevant [22,23]. In patients with aortic valve replacement, Jin and colleagues found different coronary zero-flow pressures between warm and cold blood cardioplegia [24]. In our investigation on IMAB grafts, no differences in COP were found between patients who received combined antegrade/retrograde cold blood cardioplegia and patients who received antegrade crystalloid cardioplegia with HTK solution.

In summary, the results of the present study demonstrate that in cardiac surgical patients undergoing CABG procedures, the diastolic P–F relationship in arterial grafts can be analyzed and extrapolated to yield coronary zero-flow pressure as a measure of COP and the slope of the diastolic P–F segment as a measure of coronary vascular conductance which is inversely related to regional CVR. We observed that under balanced anaesthesia both CSP and LVEDP considerably underestimate COP and thus lead to a systematic overestimation of CPP. These findings demonstrate a vascular waterfall phenomenon in blood flow in CABGs. Our results further suggest that the slope of the diastolic P–F relationships provides a more rational approach to distinguish changes in regional CVR from changes in coronary downstream pressure than conventional calculations of CVR.


1. Permutt S, Riley RL. Hemodynamics of collapsible vessels with tone: the vascular waterfall. J Appl Physiol 1963; 18: 924–932.
2. Jackman AP, Green JF. A theoretical description of arterial pressure–flow relationships with verification in the isolated hindlimb of the dog. Ann Biomed Eng 1990; 18: 89–101.
3. Bellamy RF. Diastolic coronary artery pressure–flow relations in the dog. Circ Res 1978; 43: 92–101.
4. Uhlig PN, Baer RW, Vlahakes GJ, Hanley FL, Messina LM, Hoffman JI. Arterial and venous coronary pressure–flow relations in anesthetized dogs. Evidence for a vascular waterfall in epicardial coronary veins. Circ Res 1984; 55: 238–248.
5. Farhi ER, Klocke FJ, Mates RE et al. Tone-dependent waterfall behavior during venous pressure elevation in isolated canine hearts. Circ Res 1991; 68: 392–401.
6. Ejaz T, Takemae T, Kosugi Y, Hongo M. The high zero-flow pressure phenomenon in coronary circulation: a simulation study. Front Med Biol Eng 2002; 11: 335–340.
7. Dole WP, Richards KL, Hartley CJ, Alexander GM, Campbell AB, Bishop VS. Diastolic coronary artery pressure–flow velocity relationships in conscious man. Cardiovasc Res 1984; 18: 548–554.
8. Nanto S, Masuyama T, Hori M, Shimonagata T, Ohara T, Kubori S. Zero flow pressure in human coronary circulation. Angiology 1996; 47: 115–122.
9. Nanto S, Masuyama T, Takano Y, Hori M, Nagata S. Determination of coronary zero flow pressure by analysis of the baseline pressure–flow relationship in humans. Japan Circ J 2001; 65: 793–796.
10. D'Ancona G, Karamanoukian HL, Ricci M, Bergsland J, Salerno TA. Graft patency verification in coronary artery bypass grafting: principles and clinical applications of transit time flow measurement. Angiology 2000; 51: 725–731.
11. Groom R, Tryzelaar J, Forest R et al. Intra-operative quality assessment of coronary artery bypass grafts. Perfusion 2001; 16: 511–518.
12. Walpoth BH, Bosshard A, Kipfer B, Berdat PA, Althaus U, Carrel T. Failed coronary artery bypass anastomosis detected by intraoperative coronary flow measurement. Eur J Cardiothorac Surg 1998; 14 (Suppl 1): S76–S81.
13. Klocke FJ, Weinstein IR, Klocke JF et al. Zero-flow pressures and pressure–flow relationships during single long diastoles in the canine coronary bed before and during maximum vasodilation. Limited influence of capacitive effects. J Clin Invest 1981; 68: 970–980.
14. Douglas JE, Greenfield Jr JC. Epicardial coronary artery compliance in the dog. Circ Res 1970; 27: 921–929.
15. Klocke FJ, Ellis AK, Orlick AE. Sympathetic influences on coronary perfusion and evolving concepts of driving pressure, resistance, and transmural flow regulation. Anesthesiology 1980; 52: 1–5.
16. Beldi G, Bosshard A, Hess OM, Althaus U, Walpoth BH. Transit time flow measurement: experimental validation and comparison of three different systems. Ann Thorac Surg 2000; 70: 212–217.
17. D'Ancona G, Karamanoukian HL, Salerno TA, Schmid S, Bergsland J. Flow measurement in coronary surgery. Heart Surg Forum 1999; 2: 121–124.
18. Wolfe JA. The coronary artery bypass conduit: II. Assessment of the quality of the distal anastomosis. Ann Thorac Surg 2001; 72: S2253–S2258.
19. Schmitz C, Ashraf O, Schiller W et al. Transit time flow measurement in on-pump and off-pump coronary artery surgery. J Thorac Cardiovasc Surg 2003; 126: 645–650.
20. Shimada K, Sakanoue Y, Kobayashi Y et al. Assessment of myocardial viability using coronary zero flow pressure after successful angioplasty in patients with acute anterior myocardial infarction. Heart 2003; 89: 71–76.
21. Furber AP, Prunier F, Nguyen HC, Boulet S, Delepine S, Geslin P. Coronary blood flow assessment after successful angioplasty for acute myocardial infarction predicts the risk of long-term cardiac events. Circulation 2004; 110: 3527–3533.
22. Franke UF, Korsch S, Wittwer T et al. Intermittent antegrade warm myocardial protection compared to intermittent cold blood cardioplegia in elective coronary surgery – do we have to change? Eur J Cardiothorac Surg 2003; 23: 341–346.
23. Elvenes OP, Korvald C, Myklebust R, Sorlie D. Warm retrograde blood cardioplegia saves more ischemic myocardium but may cause a functional impairment compared to cold crystalloid. Eur J Cardiothorac Surg 2002; 22: 402–409.
24. Jin XY, Gibson DG, Pepper JR. The effects of cardioplegia on coronary pressure–flow velocity relationships during aortic valve replacement. Eur J Cardiothorac Surg 1999; 16: 324–330.


© 2006 European Society of Anaesthesiology