SUMMARY We have characterized N-13 ammonia as a myocardial blood flow imaging agent suitable for positron-emission computed tomography. However, the mechanisms of uptake and retention of this agent in myocardium are not known, and effects of altered metabolism were not considered. Therefore, we studied the uptake and retention of N-13 ammonia in myocardium under various hemodynamic and metabolic conditions in open-chest dogs. N-13 ammonia was extracted nearly 100% during its initial capillary transit, followed by metabolic trapping that competed with flow-dependent back diffusion. At control flows, the first capillary transit extraction fraction (E) of N-13 ammonia averaged 0.82 ± 0.06. It fell with higher flows by E = 1 -0.607 exp -125/F. Myocardial N-13 tissue clearance half-times were similarly inversely related to blood flow, and ranged from 110-642 minutes. Cardiac work and changes in the myocardial inotropic state induced by isoproterenol and propranolol did not affect E or the tissue clearance half-times. Low plasma pH reduced E by an average of 20%, while elevated plasma pH had no effect. Decreases in flow below control also were associated with a fall in E. Inhibition of glutamine synthetase with L-methionine sulfoximine impaired metabolic trapping of N-13 ammonia and implicates the glutamic acid-glutamine reaction as the primary mechanism for ammonia fixation. The product of E times flow predicts the myocardial N-13 tissue concentrations, which increased by 70% when flow was doubled. Thus, blood flow and metabolic trapping are the primary determinants of myocardial uptake and retention of N-13 ammonia. The relative constancy of metabolic trapping over a wide range of hemodynamic and metabolic conditions demonstrates the value of N-13 ammonia as a myocardial blood flow imaging agent. N-13 AMMONIA* has been characterized as an indicator for the noninvasive visualization of regional myocardial perfusion by positron computed tomography (PCT).' Use of N-13 ammonia has also permitted noninvasive detection of mild, 47% diameter coronary stenosis in the intact dog.2Because fixation of N-13 ammonia in myocardium occurs through metabolic pathways, alterations in both the hemodynamic and metabolic state of the heart could modify the uptake of N-13 ammonia, and hence, limit its value as a flow indicator. In blood, N-13 (NH3) ammonia exists primarily in its ionic species, NH4+, the ammonium ion, which apparently can substitute for K+ on the sodium-potassium transmembraneous exchange system in red blood cells.3 It thus may be actively transported into myocardium. On the other hand, NH3 can diffuse across cell membranes because of its lipid solubility and is rapidly replenished by conversion of NH4+ to NH3 as it leaves the vascular space.4" 5Transmembrane exchange therefore may occur through an active transport mechanism or *The term ammonia is used to refer to the chemical equilibrium of NH3 NH4+ in which the prominent form is NH4+.

Positron emission tomography allows noninvasive assessment of myocardial blood flow and metabolism, and may aid in defining the extent and severity of an ischemic injury. This hypothesis was tested by studying, in chronically instrumented dogs, regional blood flow and metabolism during and after a 3 hour balloon occlusion of the left anterior descending coronary artery. The metabolic findings after ischemia were compared with the recovery of regional function over a 4 week period. N-13 ammonia was used as a blood flow tracer, and C-11 palmitic acid and F-18 deoxyglucose as tracers of fatty acid and glucose metabolism, respectively. Regional myocardial function was monitored with ultrasonic crystals implanted subendocardially. Regional function improved most between 24 hours and 1 week after reperfusion, but was still attenuated at 4 weeks. The slow functional recovery was paralleled by sustained metabolic abnormalities, reflected by segmentally delayed clearance of C-11 activity from myocardium and increased uptake of F-18 deoxyglucose. Absence of blood flow and C-11 palmitic acid uptake at 24 hours of reperfusion correlated with extensive necrosis as evidenced by histologic examination. Conversely, uptake of C-11 palmitic acid with delayed C-11 clearance and increased F-18 deoxyglucose accumulation identified reversibly injured tissue that subsequently recovered functionally and revealed little necrosis. Thus, recovery of metabolism after 3 hours of ischemia is slow in canine myocardium and paralleled by slow recovery of function. Metabolic indexes by positron tomography early after reperfusion can identify necrotic and reversibly injured tissue. Positron tomography may therefore aid in defining the extent and prognosis of an ischemic injury in patients undergoing reperfusion during evolving myocardial infarction.

Regional myocardial blood flow can be measured accurately and noninvasively from serially acquired and reoriented short-axis 13N-ammonia images, thus overcoming limitations inherent to the use of transaxially acquired images and permitting a more complete evaluation of regional blood flows throughout the left ventricular myocardium.

The usefulness of [1l-tC]acetate as a tracerof overall myocardial oxidative metabolism for use with positron emission tomography has been investigated in 12 closed-chest dogs. Myocardial ltC activity clearance kinetics after intravenous administration of [1-"C]acetate in dogs have been determined noninvasively by positron emission tomography. Biexponential fitting of regional myocardial`C time-activity curves was performed to give clearance half-times and fractional distribution. The rate constant kl for the early rapid phase of "C activity clearance was found to correlate linearly with myocardial oxygen consumption (y=0.0156x+0.039; SEE=0.023; r=0.95). kl was approximately 7% lower in septal sectors compared with the left ventricular free wall, suggesting that regional oxygen consumption in the septum was lower; a concomitant regional attenuation of blood flow in the septum relative to the left ventricular free wall was also observed. In dogs using carbohydrates as the predominant fuel, kl oxygen consumption was somewhat more than in dogs using predominantly free fatty acids (0.021±0.002 compared with 0.018+0.002, p <0.01), indicating that increased carbohydrate consumption is associated with a small increase in kl at constant oxygen consumption. It is concluded that measurement of myocardial [1-"Cjacetate kinetics allows noninvasive determination of cardiac oxygen consumption by positron emission tomography and that the technique is relatively insensitive to myocardial fuel selection. (Circulation 1989;79:134-142) P ositron emission tomography (PET) provides a unique opportunity for the noninvasive study of regional metabolism in vivo.1 While metabolic tracers currently in use for PET are

Ischemically injured reperfused myocardium is characterized by increased 18F-fluorodeoxyglucose uptake as demonstrated by positron emission tomography. To elucidate the metabolic fate of exogenous glucose entering reperfused myocardium, D-[6-14C] glucose and L-[U-13C] lactate were used to determine glucose uptake, glucose oxidation and the contribution of exogenous glucose to lactate production. The pathologic model under investigation consisted of a 3 h balloon occlusion of the left anterior descending coronary artery followed by 24 h of reperfusion in canine myocardium. The extent and severity of myocardial injury after the ischemia and reperfusion were assessed by histochemical evaluation (triphenyltetrazolium chloride and periodic acid-Schiff stains). Thirteen intervention and four control dogs were studied. The glucose uptake in the occluded/reperfused area was significantly enhanced compared with that in control dogs (0.40 +/- 0.14 versus 0.15 +/- 0.10 mumol/ml, respectively). In addition, a significantly greater portion of the glucose extracted immediately entered glycolysis in the intervention group (75%) than in the control dogs (33%). The activity of the nonoxidative glycolytic pathway was markedly increased in the ischemically injured reperfused area, as evidenced by the four times greater lactate release in this area compared with the control value. The dual carbon-labeled isotopes showed that 57% of the exogenous glucose entering glycolysis was being converted to lactate. Exogenous glucose contributed to greater than 90% of the observed lactate production. This finding was confirmed by the histochemical finding of sustained glycogen depletion in the occlusion/reperfusion area. The average area of glycogen depletion (37%) significantly exceeded the average area of necrosis (17%). These data demonstrate enhanced and sustained activity of the nonoxidative glycolytic pathway after a prolonged occlusion with reperfusion in canine myocardium. Because glycogen stores remain depleted, exogenous glucose becomes an important myocardial substrate under these pathologic conditions.

Phasic red cell velocity and diameters of coronary arterioles, capillaries, and venules were measured in the beating turtle and dog heart using high-speed cinematography with transillumination of the left ventricle. In the turtle, arteriolar red cell velocity was diminished during systole, but during diastole arteriolar inflow increased, especially during the rapid and the slow filling period. Capillary and venule red cell velocity was in-. creased during systole, particularly at the time of ejection; however, during diastole red cell velocity declined and the lowest values occurred during isovolumic relaxation. In dog arterioles, capillaries, and venules, the pattern of red cell velocity was similar. Thus, in the turtle and dog, the peak arteriolar red cell velocity occurred in unison with left coronary artery inflow, and the capillary and venule flow pattern followed that of the coronary sinus. The diameters of arterioles, capillaries, and venules in the turtle ventricle all declined about 34% during systole; similar results were obtained in the dog. Capillary arrangement appeared to be predominantly parallel and cocurrent; however, capillary loops with countercurrent flow were occasionally observed. The data on microvascular phasic red cell velocity are consistent with the macroobservations of reduced coronary artery inflow and enhanced coronary sinus outflow during ventricular contraction. The results demonstrate that the shift in the flow pattern occurs at the transition from arterioles to capillaries. KEY WORDS red cell velocitycardiac cycle arteriolar flow capillary circulation phasic capillary flow venule flow ventricular capillaries countercurrent flow• Rebatel (1) stated in 1872 that "entrance of blood into the smaller coronary vessels during systole is difficult" and that "in diastole the capillaries which have been empty of blood as a result of systolic compression, fill again from larger coronary vessels." Subsequently, numerous other investigators have confirmed the observations that coronary inflow occurs primarily during diastole and that outflow in the coronary sinus takes place mainly during systole (2-8). Consequently, there must be a point where this 180° phase shift occurs (9). However, the exact location of this shift has not been observed directly. In the atrium phasic changes in

N-13 ammonia mimics certain properties of microspheres. It rapidly clears from blood into myocardium where it becomes fixed in proportion to myocardial blood flow. Used with positron emission tomography as a means for quantifying in vivo myocardial indicator concentrations, N-13 ammonia may be useful for noninvasive determination of myocardial blood flow with the arterial reference sampling technique. This possibility was examined in 27 experiments in 10 chronically instrumented dogs at control, high and low blood flows. Myocardial blood flow was calculated in vivo from the myocardial N-13 tissue activity concentrations derived from serial cross-sectional images of the heart, the 2 minute arterial input function and the withdrawal rate of arterial blood. These calculations were compared with blood flow determined by the standard microsphere technique. Blood flow determined in vivo with N-13 ammonia and positron emission tomography correlated with microsphere blood flow by y = -36.2 + 1.53x -0.0027x2 (r = 0.94 with a standard error of the estimate of 16 ml/min per 100 g). For flows from 44 to 200 ml/min per 100 g, the relation between in vivo and in vitro measured myocardial blood flow was nearly linear but reached a plateau at flows higher than 200 ml/min per 100 g. These results indicate that in dogs, blood flow in the physiologic range can be quantified in vivo with N-13 ammonia and positron emission tomography.

No abstract