All meanings, we know, depend on the key of interpretation. —George Elliott

In this issue of JASE , Lima et al. report that reduced myocardial blood flow (MBF) velocity reserve and left atrial (LA) diameter independently predict long-term outcomes in patients with nonischemic dilated cardiomyopathy. The investigators studied 195 patients with reduced left ventricular (LV) ejection fractions (<35%) and no coronary artery disease using Doppler and myocardial contrast echocardiography (MCE). After a median follow-up period of 29 months, there were 45 events (43 deaths and two cardiac transplantations). The authors found that LA diameter added incremental prognostic value to clinical factors (including LV ejection fraction) and that further addition of MBF velocity reserve almost doubled the predictive value of the combined model. Coronary blood flow (CBF) reserve measured from the left anterior descending coronary artery and myocardial blood volume (MBV) measured on MCE were not independent predictors of these hard events. Although the findings are robust, the explanations offered for them are not satisfactory. In this brief discussion, I propose some plausible reasons for these interesting observations.

Let us start with LA size. Lima et al. measured LA diameter from the parasternal long-axis view. The left atrium is akin to a pillow, in which in pathologic conditions, the area can increase more than the elevation (thickness). Thus, either LA area measured in the apical four-chamber view or LA volume calculated from the apical two-chamber and four-chamber views better represents LA size than LA diameter measured from the parasternal view. The argument provided by the investigators that LA diameter was used because it is routinely measured, whereas the other dimensions are not, is at best a flimsy one. Because this was purportedly a research study, they could easily have made the additional measurements. Even when it was measured suboptimally, however, LA size was still an independent predictor of outcomes, because in the absence of atrial fibrillation, it reflects the history of LA pressure. It is a much more robust measure of LA pressure than Doppler measures of so-called diastolic dysfunction.

Patients who did not have events, in addition to having smaller left atria, had smaller LV sizes and volumes (in both end-diastole and end-systole) and higher LV ejection fractions on univariate analysis (Table 4 in Lima et al. ). They also had clinically less advanced disease, as defined by New York Heart Association class (Figure 3 in Lima et al. ). It is likely, therefore, that when multivariate analysis was performed, LA size singly reflected a combined sum of all these other abnormalities, which are all individually associated with chronically elevated LA pressure.

Chronically elevated LA pressure can also affect CBF. CBF is determined by the coronary driving pressure (CDP), which normally is the difference between the aortic and right atrial (or coronary sinus) pressures. However, under abnormal conditions, LV diastolic and critical closing pressures can reduce the CDP. Let us first consider LV diastolic (or LA) pressure that exerts force on the myocardium (more on the endocardium than the epicardium) and determines the intramyocardial pressure that, in turn, exerts external compression on intramyocardial blood vessels. Normally, intramyocardial pressure contributes about 10% to 15% to total coronary vascular resistance, and it is customary to ignore it and calculate CDP as the difference between aortic and right atrial pressures.

Myocardial compressive forces can be likened to a Starling resistor, described by Ernest Starling in an isolated heart preparation. He invented a device that consisted of an elastic, fluid-filled, collapsible tube that was placed in an air-filled chamber. By changing the pressure in the air-filled chamber, Starling was able to change the resistance of the tube. He showed that when the resistance in the tube was high because of higher chamber pressure, flow no longer depended on the outflow pressure (right atrial pressure in the case of the heart). Thus, when LV diastolic or LA pressure is high, CDP is the difference between the aortic and LA (or LV diastolic) pressures rather than the aortic and right atrial pressures.

Another important determinant of CDP is the closing pressure, also called the zero-flow pressure. Flow through a tube is determined by the pressure gradient across the inflow and outflow. Flow stops when there is no pressure gradient driving fluid through the tube. In the case of the LV myocardium, flow stops when the pressure in the distal coronary artery is approximately 15 to 20 mm Hg. We have shown that the reason there remains a pressure gradient despite zero flow is that, unlike in a tube, as CDP falls, capillaries collapse from the surrounding tissue pressure in proportion to the drop in CDP. Thus, capillary resistance increases as flow decreases, until it eventually stops flow. In this situation, there is no flow despite there being pressure in the system.

Can these effects (myocardial compressive forces and closing pressure) affect MBF and CBF reserve? It has been experimentally demonstrated that in the presence of adenosine, when LV diastolic pressure is increased, MBF decreases, and the endocardial/epicardial MBF ratio reverses. At the same time, the closing pressure increases. These findings have important clinical implications in that when LA pressure is elevated, the myocardium is more likely to undergo ischemia during periods of stress. Because in the study by Lima et al. , LA size was independent of MBF velocity reserve for predicting hard events, obviously high LA pressure is detrimental to human health independent of its effect on coronary physiology, for the reasons listed above.

In a previous study, Tsagalou et al. reported that the depressed CBF reserve seen in nonischemic dilated cardiomyopathy was related to the reduced capillary density and diameter in these patients. When a coronary vasodilator is administered, the resistance offered by intramyocardial arterioles and venules is minimized, and the capillaries (which have no smooth muscle and are hence not affected by the vasodilator) become the bottleneck to hyperemic flow. Because capillaries are laid in parallel, the smaller the number of capillaries, the higher the resistance, and hence the lower the CBF reserve. Furthermore, because resistance is related to the fourth power of the radius, a small decrease in capillary diameter can result in a major decrease in CBF reserve.

The best method to assess capillary density and size is by measuring MBV, as can be easily accomplished using MCE. Why then was the MBV fraction measured by Lima et al. not different between patients who survived and those who did not? The problem lies with the MCE method used. These investigators used real-time low–mechanical index MCE, which, compared with intermittent high–mechanical index imaging, is less sensitive in terms of measuring myocardial signals emanating from the microbubbles (because the bubbles are not destroyed but only made to resonate). Consequently, to obtain adequate myocardial signal, more contrast must be given, resulting in saturation of the LV cavity. That is, at the dose of the contrast agent required for adequate myocardial opacification using this method, there is no relation between microbubble concentration in the LV cavity and acoustic backscatter from it. Thus, when myocardial signal is normalized to that of the LV cavity, it does not provide an accurate assessment of MBV fraction. Although MBF velocity measurement is also affected by LV cavity microbubble concentration, it is far less sensitive to it than the MBV fraction measurement.

Another reason for decreased MBF and CBF reserve in patients with dilated cardiomyopathy is increased resting MBF resulting from increased myocardial oxygen consumption because of higher LV wall stress, greater LV mass, increased preload, and resting tachycardia. The increase in resting MBF is mediated through coronary vasodilation. There is additional coronary vasodilation to maintain a constant capillary hydrostatic pressure in the presence of reduced CDP. Thus, the ability to increase MBF and CBF by the same magnitude as in normal subjects when exposed to a coronary vasodilator or stress is attenuated when resting MBF is higher than normal.

Finally, Lima et al. also measured CBF reserve in the left anterior descending coronary artery, and although it was significantly different between patients with and without events, unlike MBF velocity reserve, it was not a predictor of events on multivariate analysis. The reasons given by the investigators for this disparity between the predictive powers of MBF velocity and CBF reserve are not wholly convincing. They state that CBF reserve was measured only in a single coronary artery, whereas MBF velocity reserve represented a sum of values from 11 myocardial segments. Because dilated cardiomyopathy is not a regional disease, CBF reserve from any artery should reflect the underlying pathology. The second reason they give is that for CBF measurements, they measured only velocity and not flow, because they did not measure the coronary vessel dimension. Epicardial coronary vessel dimension increases only marginally from flow-induced vasodilation, and so not measuring it should not affect CBF reserve appreciably. For MCE, the investigators also report a prognostic value only for MBF velocity and not MBF (the product of MBF velocity and MBV). It appears that because their MBV fraction measurement was erroneous, their MBF measurement by MCE was not accurate either and failed to predict events.

If one examines Table 2 in Lima et al. ‘s report, one finds a high degree of concordance between MBF velocity reserve and CBF reserve. Because these patients did not have coronary artery disease and hence collateral flow did not contribute to MBF, the relation between CBF and MBF should be 1:1. The most likely explanation for the disparity in the prognostic value of MCE and Doppler is that it is difficult to obtain an acceptable incident angle between the Doppler beam and the long axis of the left anterior descending coronary artery in a substantial number of patients. The very high success rate of MCE and the very low observer error rate for this technique reported by the investigators are also different from those reported by many other investigators who have been using this technique clinically for a long time.

In summary, because of the relatively large number of patients and a high event rate, Lima et al. were able to obtain meaningful prognostic results in patients with nonischemic dilated cardiomyopathy, for which they should be congratulated. Measurement of LA area or volume and quantification of MBF velocity and MBV by use of intermittent high–mechanical index MCE could have made the study even stronger. A thorough understanding of cardiac and coronary physiology is always helpful in the interpretation of results and provides insights into what otherwise would remain a mere observation without pathophysiologic and mechanistic underpinnings. To end with a quotation from Clifford Stoll, “Data are not information; information is not knowledge; knowledge is not understanding; and, understanding is not wisdom.”