| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Medical Physics, Cardiovascular Research Institute Amsterdam, and 2Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
Submitted 4 August 2004 ; accepted in final form 13 December 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
75% of a normal aortic pressure and heart rate >80 beats/min.
systolic flow limitation; coronary reserve; steal; heart; microspheres; regional conductance
A unique relation between DTF and HR would circumvent the need for expressing data on subendocardial perfusion as a function of DTF. However, although it is known that DTF decreases with increasing HR (17), HR and DTF are not uniquely related. For example, in a canine heart, a decrease of Pc was shown to result in an increase of DTF especially at full vasodilation (20). In patients, the DTF-increasing effects of isoproterenol and dobutamine at constant HR have been demonstrated (4). The absence of a unique relation between HR and DTF on one hand and the dominant role of DTF in subendocardial perfusion on the other hand also follow from the observation that, at the ischemic threshold, DTF rather than HR correlates with the significance of coronary stenosis in patients (13). Although DTF has been recognized as an important variable for subendocardial perfusion, DTF has rarely been presented as an independent variable in experimental studies.
Because systolic compression reduces the volume of blood in the subendocardium, the time needed to replace this volume is dependent on Pc. If diastolic time is too short, the refill will not be completed before the next systole and the resistance will rise disproportionally (16, 26, 27). Consequently, we hypothesized that, with decreasing DTF, subendocardial flow will become zero at a finite value of DTF and that this value decreases with increasing Pc. To measure regional flow, fluorescently labeled microspheres were used in the present study. Because of the limited amount of tracers that can be used in a single experiment, we chose to study the effect of DTF on the flow distribution at preset levels of perfusion pressure. We quantified the values of DTF at which the perfusion over the myocardium is even (DTF1) and at which subendocardial perfusion is predicted to be zero (DTFzf). To demonstrate that DTF and not HR is the determining factor for subendocardial perfusion, DTF was increased at constant HR by means of intracoronary dobutamine administration in additional experiments.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Anesthesia, ventilation, and drugs. Premedication was a mix of ketamine (15 mg/kg) and midazolam (1 mg/kg im). Anesthesia was maintained with a mixture of sufentanil (4 µg·kg–1·h–1, Sufenta Forte) and midazolam (0.8 mg·kg–1·h–1 iv). The animals were ventilated artificially with 50% O2-50% N2O. Adequate ventilation was checked by analyzing exhaled gases continuously and arterial blood samples at least every 1 h. Arterial oxygen saturation was kept over 90%, PCO2 below 40 Torr, and pH between 7.4 and 7.45. Bicarbonate was given, when necessary, to keep base excess above –4.0 mM.
Before the perfusion system was connected, ibuprofen (12.5 mg/kg iv) was given to suppress inflammatory reactions resulting from the perfusion system, as well as heparin (bolus 5,000 IU iv, followed by 5,000 IU/h continuous) to avoid coagulation. During measurements, full dilation of the coronary circulation was induced by adenosine (30–60 µg·kg–1·h–1 ic) infusion. Full dilation was verified from the absence of an increase in coronary flow upon administration of an additional adenosine bolus.
Experimental protocol.
Measurements were started after an equilibration period of 30 min. The pressure in the perfusion system was adjusted to the desired pressure (40–100 mmHg) in the left main stem. In each animal of a group of 16, at this constant Pc, DTF was varied, either by pacing the heart or by stimulating the vagus nerve, aiming at HRs of 60, 90, 120, 150, and 180 beats/min. When, after the adenosine infusion was started, the desired conditions were in a steady state within 2 min, a continuous recording was made of perfusion pressure, coronary flow, Plv, and ECG, and microspheres were injected into the perfusion line over a period of 10 s to determine regional flows. Seven colors of fluorescent microspheres were available (Molecular Probes, Fluospheres blood flow determination kit, 15 µm). After sonification and suspension in 4 ml of blood, 100,000 microspheres per color were injected. The seven fluorescent colors were used in random order. In five additional experiments that followed the same experimental procedure but also explored a different protocol, microspheres were injected at constant HR and perfusion pressure during a control period and after intracoronary administration of dobutamine (0.6–3.0 µg·kg–1·min–1) for the purpose of establishing its effect in relation to this study.
Analysis of regional flow.
For flow analysis, a 1- to 1.5-cm slice of myocardium, 2 cm under the base of the heart, was divided into five sections: two papillary regions and three free walls. Each section was subdivided into a subepicardial (2–3 mm), a midmyocardial (variable thickness), and a subendocardial (2–3 mm) layer. The rest of the heart was divided into 1- to 2-g sections to determine total fluorescence. The centrifuge sedimenting method was used to recover the microspheres (32). Regional flow in milliliter per minute per gram of tissue was calculated as the ratio of sample fluorescence over total fluorescence multiplied by total coronary flow as measured in the perfusion line and divided by the sample weight.
Data acquisition and processing. Data were digitized and stored on a personal computer. For a recording to be valid, regular heart periods during time of the microsphere injection were required. Diastole was defined as the period from minimal to maximal derivative of Plv, and DTF was calculated as diastole divided by heart period. The coronary perfusion pressure gradient (Pc) was defined as the difference between the mean pressure in the coronary left main stem and the minimal diastolic Plv (as an estimate of coronary backpressure). Regional conductances were calculated as flow per Pc. Cardiac contractility was expressed as the maximum value of the first derivative of Plv (dPlv/dtmax).
All statistical analyses were done in SPSS. Means ± SD are given where applicable. Differences (ANOVA), regression slopes, and differences between regression slopes [analysis of covariance (ANCOVA)] are statistically significant (P < 0.05) unless mentioned otherwise.
To analyze DTFzf and DTF1, for each goat linear regressions were calculated of DTF and Cendo and the Cendo-to- subepicardial conductance (Cepi) ratio, respectively. Of the significant relations, the intercept at the DTF value of 0, respectively 1, was determined. Finally, a linear regression on the relation was calculated of these intercept values and Pc. In addition, a multiple regression on the pooled data was calculated. Alternative relations between DTFzf and DTF1 with Pc were obtained by substituting 0 for conductance, respectively 1, for subendocardium-to-subepicardium ratio in the regression equation on the pooled data.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
|
|
For each individual experiment, the linear regression of the relation between the Cendo-to-Cepi ratio and DTF was calculated. Four relationships were not significant because of a very limited range of DTF. For the remaining 12 relationships, the DTF for which Cendo-to-Cepi ratio = 1 (DTF1) was determined (see Fig. 2C) and plotted as function of Pc in Fig. 3. The resulting relationship was statistically significant (P < 0.05) and quite similar to the result of multiple regression on all data pooled, including those from the four relations not yielding an individual DTF1 (P < 0.01). Figure 3 also shows all results that we derived from the literature (4 studies). The regression demonstrates that DTF1 diminishes with increasing coronary pressure.
|
|
95 mmHg, the dependence of Cendo on DTF in this pooled data set with dobutamine tallies well with the data obtained by changing HR.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Critique of methods. The use of fluorescent microspheres is a gold standard (8, 32) but limits an experiment to seven interventions. The number of interventions suitable for analysis is further reduced by practical problems like irregular HR during microsphere injection. The limited number of data points per regression diminishes the power of statistical analysis. Therefore, the data were analyzed in two ways: multiple regression on all pooled data and analysis of regression coefficients calculated per animal. There was a slight quantitative difference between the outcome of the two statistical approaches, but this had no consequence for the interpretation of the results in terms of mechanisms.
For the definition of diastolic length, the difference between maximal and minimal time derivative of Plv was chosen based on an earlier study (19). From a signal processing point of view, these points are well defined. One may argue that at higher HR Plv may rise faster than at lower HR, and hence the error may not be constant but HR dependent. However, in that case, dPlv/dtmax would be HR dependent and in our experiments, although significant with a high number of observations, the correlation between the two was very low. The present definition results in a 13% higher DTF than when a level detection was used for Plv with 25% of the Plv pulse as a threshold. Merkus et al. (20) compared this and other definitions of DTF and concluded that the quantification of changes of DTF was hardly affected by its definition.
We use DTF as an estimate of the period in which intramural vessels are not compressed, and one may argue that compression starts earlier during the rise of Plv and may still be present during a short period after the minimal value of dPlv/dt. However, then the question arises regarding whether the compression is constant during systole, and this is probably not the case. Analysis of coronary arterial flow waves and the wave of epicardial lymph pressure led to the conclusion that in the beginning and the end of systole compression is related to Plv and in midsystole it is related to contractility (26). Moreover, some recent evidence has been found that cross bridges may relax earlier in the subendocardium that in the subepicardium (37). Obviously, the definition of DTF will determine the absolute value of DTF1 and DTFzf, but it will have little effect on the slopes of their relationships with Pc (Figs. 3 and 4). A more conservative estimate of DTF would result in a lower extrapolated intercept of the DTFzf-Pc relationship with the Pc axis (see Fig. 3). In that case, a systolic and diastolic uncoupling is predicted to occur at a lower value of Pc. However, such a more conservative definition of DTF would not invalidate our conclusion that systole exerts its influence on diastolic subendocardial perfusion in the physiological range of Pc and DTF.
Adenosine is known to be a vasodilator of especially the smallest resistance vessels. We found that, at the dose used in this present study, adenosine dilates all resistance arteries in vitro. Regardless, the adenosine dose was checked for maximal vasodilatory capacity judged by the coronary flow measurements during the experiments. It is not likely that adenosine by itself had an influence on DTF. In the study of Merkus et al. (20), adenosine decreased DTF; however, when DTF was plotted vs. flow, the adenosine effect was indistinguishable from the experiments in which flow was varied without adenosine. The flow effect on DTF may have influenced the HR-DTF relationship. For a certain HR, flow was different between animals because of different perfusion pressures, which may contribute to the scatter of the HR-DTF relationship.
Plv was guarded, and, when it dropped, volume was given intravenously to the animal. At a low perfusion pressure, it was difficult to maintain Plv when HR was increased. Because of this, we could not increase HR to a high value when perfusion pressure was low, as can be inferred from Fig. 2. A reduction in Plv will reduce the compressive forces on the intramural vessels and hence result in an increase of conductance at lower DTF values. In our study, we demonstrate a reduction of conductance at decreasing DTF, but this reduction is an underestimate of the condition in which Plv would remain constant. It should be noted that the conditions of HR and pressure could be set instantaneously and that vasodilation and injection of microspheres were completed within 2 min, preventing a too large challenge to the stability of the preparation.
Relation to other microsphere studies. Although, instead of HR, the duration of diastole has been recognized as a factor that determines myocardial perfusion, previously no studies were designed that varied DTF, as became clear when we extensively surveyed the literature. We compared our data with four earlier studies in which we were able to deduce the DTF1 values, for which several regional flow measurements are necessary at different DTFs. Two of these gave indexes of diastolic time (3, 14); in the other two (2, 11), DTF had to be estimated from HR with the present HR-DTF relation. The data deduced from these four studies have been added to Fig. 3 for comparison. These appeared to be in the same range as the data from the present study.
Interpretation of results. The effect of cardiac contraction on the perfusion of the subendocardium is compensated anatomically by a higher vascular volume of the small arteries in this region (15, 36). This higher volume corresponds to a much lower resistance in the subendocardium than subepicardium when the heart is arrested. During each heartbeat, the contraction reduces intramural blood volume. In the arterial coronary circulation, this change in volume causes a reduction in flow or even retrograde flow, whereas in the coronary veins systolic flow increases (27–29). This variation in volume results in smaller intramural vascular diameters during systole (31), and hence systolic subendocardial resistance will certainly be larger than diastolic resistance. The question is how much of that effect is noticeable in diastole. The swing of intramural volume between systole and diastole depends both on the strength and duration of compression in systole and on the possibility for volume restoration during diastole. Both the duration of diastole and the level of perfusion pressure will promote diastolic volume restoration.
One should keep in mind that a strict separation between the effects of systole and of diastole on subendocardial perfusion is a conceptual simplification of the continuous variations in vascular volume (33, 34). This simplification fits the experimental outcome better at higher perfusion pressures when DTFzf approaches zero at which value subendocardial perfusion is perfectly proportional with DTF (see Fig. 4). At physiological pressures, however, subendocardial flow falls faster with DTF than predicted by a strict separation of systolic and diastolic events because DTFzf is higher than zero.
Cmid depends on DTF; hence, these intramural vessels are also subjected to compressive forces. However, the sensitivity of conductance changes for DTF was less than at the subendocardium (0.197 vs. 0.126). Also the sensitivity of conductance for pressure was less at the midmyocardium than at the subendocardium (0.00049 vs. 0.00074). A dependence of conductance on pressure could only be established for the subendocardium and the midmyocardium and not for the subepicardium. A larger dependence of resistance vessel diameter on perfusion pressure in the subendocardium compared with the subepicardium was found from direct in vivo observations (21) and is consistent with the larger pressure dependence of Cendo.
The absence of a positive correlation between Cepi and Pc cannot be interpreted as absence of microvascular diameter change on a change in Pc. By defining the local conductance as the ratio between flow and Pc, we assume that the myocardial layers are in parallel and act independently. However, transmural arteries do posses a significant resistance (9), and, given a constant DTF, the strong increase of flow in the deeper layers may result in a subepicardial perfusion pressure that is lower than Pc. This holds true for the deeper layers as well. However, because of the strong effect of Pc in the deeper layers, one may conclude that Pc reduces the flow-impeding effect of cardiac contraction such that pressure drop over the transmural vessels is more than compensated.
In the set of 16 experiments aimed at establishing the relationship between flow distribution and DTF with the Pc gradient as a parameter, the relationship between HR and DTF was well defined. The result is that subendocardial flow correlated with HR to a similar degree as with DTF and poses the question regarding why a correlation with DTF should be preferred. However, in our five additional experiments in which we, at constant HR and perfusion pressure, administered dobutamine, it was clearly found that subendocardial flow increased with increasing DTF at constant HR. This increase in subendocardial flow occurred despite a more than twofold increase in contractility, which induced a significant decrease in systolic left main stem flow. This demonstrates that the time for restoring subendocardial blood volume by an increased DTF can sufficiently compensate for the increased compressive force related to the increase of contractility (18). It is not unlikely that the clinical beneficial effect of dobutamine is explained in part by an improved perfusion of the subendocardium by an increase in diastolic time. The improved energetic state of the heart does not only stimulate contractility but also relaxation (12).
Implications. The present results are important for the validation of models. However, we did not intend the purpose of this study to be a review of the models developed so far. Our experimental study supports the concept that a decreased perfusion pressure results in a slower refill of subendocardial vessels during diastole, as predicted by the nonlinear intramyocardial pump model of Bruinsma et al. (5).
Our study provides a rationale for several clinical observations. At the threshold of ischemia, DTF rather than HR correlates with stenosis diameter (13). Exercise values of diastolic time observed in patients with syndrome X were found to be shorter than normal for comparable HRs (30). These clinical findings are in agreement with the present finding on the relationship between DTF and Cendo. Obviously, the suggested mechanisms will be better applicable to situations of well-defined large artery stenoses than the more complicated cases of diffuse or microvascular disease. However, it seems rational to assume that a larger DTF is beneficial in these cases as well.
In the clinical setting, a reduction in perfusion pressure is most often the result of a pressure drop over a stenosis. Guide wire technology allows for the measurement of flow and pressure, separately or combined, distal of a stenosis during a catheterization procedure (25). Criteria have been developed for the prediction of inducible ischemia based on indexes derived from such measurements (22–24). It was found that a flow velocity reserve (resting flow/hyperemic flow) of smaller than two or a pressure drop over the stenosis during adenosine-induced hyperemia of more than 25% of aortic pressure corresponds to inducible ischemia, as measured by a variety of methods varying from single-photon-emission computed tomography to exercise ECG. One may assume that subendocardial underperfusion occurs in these conditions.
According to our Fig. 3, the Cendo-to-Cepi ratio drops under 1 at a pressure of 75 mmHg when DTF is lower than 0.575, which corresponds to HR of 85 beats/min, as can be deduced from our Fig. 5. Hence, if HR increases over that value of 85 beats/min, subendocardial flow will drop under subepicardial flow. In itself, the drop of subendocardial flow under subepicardial flow in hyperemic conditions does not imply the occurrence of subendocardial ischemia. Subepicardial perfusion could be still abundant, and subendocardial perfusion could be sufficient. There may also be differences in regional oxygen consumption, although when differences are reported studies demonstrate a higher oxygen consumption at the subendocardium than subepicardium (10). In any case, with a subendocardial perfusion lower than in the subepicardium, the subendocardium is at higher risk than the outer layers of the heart.
The fact that ischemia can be induced in patients at a perfusion pressure of 75 mmHg seems at odds with the experimental findings in dogs where ischemia in the circumflex territory was only found when coronary pressure was dropped to 38 mmHg at an HR of 100 beats/min and to 52 mmHg at an HR of 200 beats/min (7). Most likely, this difference in results is due to collateral flow in dogs, which was unlikely to be present in our study because goats have few collaterals and the whole left main stem was perfused. Hence, the relationship between DTF1 and Pc seems very relevant for the clinical condition, especially in the absence of sufficient collateral development.
The importance of a more accurate description of DTF and Pc-dependent coronary conductance may be clear from clinical studies in which DTF was shown to be an indicator of ischemia (13, 30) but also because of the discrepancy on these issues that surfaces from clinical studies, where some believe that these effects are nonexistent (1) and others feel to be able to demonstrate these physiological mechanisms in humans (25, 35).
Conclusion. DTF is the determining parameter for subendocardial perfusion at a given Pc, and the modeled relation is in agreement with experimental and clinical human data. To sustain coronary perfusion over a range of HRs compatible with exercise, DTF is a necessary variable to be regulated by physiological responses. In the fully vasodilated heart, subendocardial hyperemic perfusion decreases more than proportionally with DTF, and this nonlinearity is more pronounced at lower Pc. In addition, DTF1 depends on Pc. This indicates the relevance of monitoring DTF in the critically diseased heart.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
1-adrenergic tone. Am J Physiol Heart Circ Physiol 266: H795–H810, 1994.This article has been cited by other articles:
|
|
 |
|
  J A Spaan Coronary flow is not that simple! Heart, May 1, 2009; 95(9): 761 - 762. [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kolyva, B.-J. Verhoeff, J. A. E. Spaan, J. J. Piek, and M. Siebes Increased diastolic time fraction as beneficial adjunct of {alpha}1-adrenergic receptor blockade after percutaneous coronary intervention Am J Physiol Heart Circ Physiol, November 1, 2008; 295(5): H2054 - H2060. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. J. de Beer, O. Sorop, D. A. Pijnappels, D. H. Dekkers, F. Boomsma, J. M. J. Lamers, D. J. Duncker, and D. Merkus Integrative control of coronary resistance vessel tone by endothelin and angiotensin II is altered in swine with a recent myocardial infarction Am J Physiol Heart Circ Physiol, May 1, 2008; 294(5): H2069 - H2077. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Westerhof, C. Boer, R. R. Lamberts, and P. Sipkema Cross-talk between cardiac muscle and coronary vasculature. Physiol Rev, October 1, 2006; 86(4): 1263 - 1308. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Spaan and M. Siebes Is myocardial contrast echocardiography ready to assume 'gold standard' status for quantification of collateral flow in humans? Eur. Heart J., July 1, 2006; 27(13): 1627 - 1627. [Full Text] [PDF]
|
|
|
|
|
|
O. Yalcin, F. Aydin, P. Ulker, M. Uyuklu, F. Gungor, J. K. Armstrong, H. J. Meiselman, and O. K. Baskurt Effects of red blood cell aggregation on myocardial hematocrit gradient using two approaches to increase aggregation Am J Physiol Heart Circ Physiol, February 1, 2006; 290(2): H765 - H771. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
) with linear regression (solid line). Dashed line was obtained by multiple regression of ratio of subendocardial to subepicardial conductance against DTF and Pc and in the result substituting 1 for ratio of subendocardial to subepicardial conductance. In addition, data points as calculated from Bache and Cobb (
), and Austin et al. (
) and a dotted line from Flynn et al. (
) are shown (see discussion.
) and human data (
) (