Exercise is a common method of physiologic stress testing and is widely used in the diagnosis of coronary artery disease with radionuclide agents. However, for those who are unable to exercise adequately, an alternative procedure is needed. Pharmacologic stress testing with radionuclide agents is an option, which has been used for more than 20 years in the diagnosis of coronary artery disease. Two kinds of agents, coronary vasodilators (adenosine and dipyridamole) and β-adrenergic receptor agonists (dobutamine), are used in pharmacologic stress testing. These agents exert their actions through different mechanisms. Dobutamine increases contractility and heart rate by stimulating β-adrenergic receptors in the heart, thereby resulting in a significant increase in CBF. Adenosine increases CBF via direct vasodilation, and dipyridamole increases the circulating concentration of adenosine by blocking its reuptake and metabolism (Mahmarian and Verani, 1994; McGuinness and Talbert, 1994; Allison et al., 1996; Leppo, 1996; Marwick, 1997; Cerqueira, 2000).

Four types of adenosine receptors (AdoR), A₁,A_2A,A_2B, and A₃, have been identified. It has been well established that the chronotropic, dromotropic, inotropic (in atria), and anti-β-adrenergic effects of adenosine are mediated by A₁ AdoR (Belardinelli et al., 1995; Shryock and Belardinelli, 1997). The vasodilator effect of adenosine is thought to be mediated primarily, but not exclusively, by A_2A AdoR. However, other types of AdoR (e.g., A_2B AdoR) appear to play a significant role in the vasodilation induced by adenosine (McGuinness and Talbert, 1994; Morrison et al., 2002; Talukder et al., 2002).

The coronary vasodilator effect of adenosine is the basis for the use of this nucleoside with radionuclide imaging in the heart to detect underperfused areas of myocardium. However, pharmacologic stress induced by adenosine is associated with a high prevalence of side effects, including dyspnea, chest pain, and atrioventricular nodal block (Mahmarian and Verani, 1994; McGuinness and Talbert, 1994; Marwick, 1997; Cerqueira, 2000). In addition, the hypotension resulting from the adenosine-induced vasodilation in the peripheral circulation is another undesirable effect. Overall, these side effects limit the usefulness of adenosine in pharmacologic stress testing. Because a number of side effects induced by adenosine appear to be mediated by AdoR types other than the A_2A receptor, it is likely that selective agonists of A_2A receptors may be more vascular-specific and cause fewer undesirable effects than adenosine.

The ultrashort duration of coronary vasodilation by adenosine also makes it a less than ideal agent for pharmacologic stress testing, as it is inconvenient to infuse adenosine intravenously for 4 to 6 min during pharmacologic stress testing. Several potent, high-affinity, and selective A_2A AdoR agonists have been developed (Glover et al., 1996; Alberti et al., 1997; Belardinelli et al., 1998; Shryock et al., 1998). However, the clinical use of these compounds may be limited because of their high affinity for the receptor, which yields an excessively prolonged duration of action and may cause significant hypotension. A novel, selective, moderate-affinity A_2A AdoR agonist, the 2-(N-pyrazolyl) adenosine derivative CVT-3146, has been developed (Gao et al., 2001a). Our previous results indicated that CVT-3146 causes a longer and more potent coronary vasodilation in conscious dogs than adenosine, but is not associated with the magnitude of hypotension seen with adenosine (Trochu et al., 2003).

The aims of this study were to compare the vasodilator effects of CVT-3146 and adenosine on coronary and peripheral circulations and their effects on cardiac output in conscious dogs.

Materials and Methods

The experiments were performed in 13 chronically instrumented conscious dogs (weighing from 21 to 33 kg, either sex). The protocols were approved by the Institutional Animal Care and Use Committee of New York Medical College and conform to the Guiding Principles for the Use and Care of Laboratory Animals, from the National Institutes of Health.

Surgical Procedures

The dogs were sedated with Acepromazine (0.3 mg/kg, i.m.) and anesthetized with pentobarbital sodium (25 mg/kg, i.v.).

Effects of CVT-3146 or Adenosine on the Coronary Blood Flow and Cardiac Output (n = 7). The chest was scrubbed for sterile surgery, and a thoracotomy was made in the fifth intercostal space under artificial ventilation. A Tygon catheter (Cardiovascular Instruments, Wakefield, MA) was inserted into the descending thoracic aorta for the measurement of blood pressure. A Doppler ultrasonic flow transducer (Craig Hartley, Houston, TX) was placed around the left circumflex coronary artery for the measurement of CBF. An electromagnetic flow transducer (45 or 50 mm; Carolina Medical Inc., King, NC) was placed around the aorta for the measurement of aortic blood flow as cardiac output (CO). The chest was closed in layers. The catheter and wires were tunneled subcutaneously and exited the skin at the back of the dog's neck.

Effects of CVT-3146 or Adenosine on the Blood Flow in the Lower Body, Mesentery, and Kidney (n = 6). The dogs were fasted for at least 24 h before the surgery. An endotracheal tube was inserted into the trachea, and the surgery was carried out under spontaneous ventilation. A midline laparotomy was made using sterile surgical techniques. A catheter was inserted into the distal abdominal aorta for the measurement of blood pressure, and an electromagnetic flow transducer (25 mm; Carolina Medical Inc.) was placed around the abdominal aorta just above the iliac bifurcation for the measurement of LBF. Doppler ultrasonic flow transducers were placed around the superior mesenteric artery and left renal artery for measurements of the mesenteric blood flow (MBF) and renal blood flow (RBF), respectively. The abdomen was closed in layers. The catheter and wires were run subcutaneously and exited from the back of the dog's neck. The dogs were allowed to recover from the surgery for 10 to 14 days and were trained to lie on the laboratory table quietly.

Recording From Chronically Instrumented Dogs

Arterial pressure was measured by connecting the previously implanted catheter to a strain-gauge transducer (P23 ID; Statham, Newark, NJ), and mean arterial pressure (MAP) was derived using a 2 Hz low-pass filter. Heart rate was monitored from the pressure pulse interval using a cardiotachometer (Beckman Coulter, Inc., Fullerton, CA).

Effects of CVT-3146 or Adenosine on the Coronary Blood Flow and Cardiac Output. The coronary flow velocity was measured from the ultrasonic flow transducer using a pulsed Doppler flowmeter (System 6; Triton Technology, San Diego, CA). The mean coronary flow velocity was derived using a 2 Hz low-pass filter. CBF (ml/min) was calculated using the formula provided by the vendor. CO (ml/min) was measured from the electromagnetic flow transducer using an electromagnetic flowmeter (Carolina Medical Inc.). Late diastolic coronary resistance (LDCR) and total peripheral resistance (TPR) were calculated as diastolic arterial pressure divided by late diastolic coronary blood flow and as MAP divided by CO, respectively. Mean and phasic signals were recorded at the same time, and a faster speed (25 mm/s) was run to allow calculation of LDCR if needed. LDCR was chosen as the index of coronary vascular resistance since it is independent of the compressive effect of ventricular contraction on coronary microvessels and was calculated as the quotient of late diastolic arterial blood pressure and CBF (Liang and Stone, 1982; Hintze and Vatner, 1984).

Effects of CVT-3146 or Adenosine on the Blood Flow in the Lower Body, Mesentery, and Kidney. LBF was measured from the electromagnetic flow transducer using an electromagnetic flowmeter (Carolina Medical Inc.). The mesenteric and renal flow velocity was measured from the flow transducers using a pulsed Doppler flowmeter (System 6; Triton Technology). Mean blood flow velocities were derived using a 2 Hz low-pass filter. MBF and RBF were calculated using the formula provided by the vendor. Mean resistances in the vascular beds were calculated as MAP divided by the mean blood flow.

Experimental Protocols

On the day of the experiment, the dog was brought to the laboratory and put on an experimental table. The previously implanted devices were attached to the recording equipment. A catheter was inserted into a peripheral vein on the leg and attached to an infusion line so that the drugs could be administered without disturbing the dog. The experiment began after the baseline hemodynamics and blood flow were stable.

Effects of CVT-3146 or Adenosine on the Coronary Blood Flow and Cardiac Output (n = 7). A dose-response curve of CBF or CO to CVT-3146 at doses of 0.1, 0.25, 0.5, 1.0, and 2.5 μg/kg was obtained following intravenous bolus injections. The dose-response curve to adenosine at doses of 10, 25, 50, 100, and 250 μg/kg was obtained as well. Hemodynamics and CBF were allowed to return to the baseline before the next dose was administered. The interval between each dose was 5 to 15 min, depending on the duration of action of the agents.

Effects of CVT-3146 or Adenosine on the Blood Flow in the Lower Body, Mesentery, and Kidney (n = 6). A dose-response curve of LBF, MBF, or RBF to CVT-3146 at doses of 0.1, 0.25, 0.5, 1.0, and 2.5 μg/kg was obtained following intravenous bolus injections. The dose-response curve to adenosine (10, 25, 50, 100, and 250 μg/kg) was obtained following intravenous bolus injections as well.

Materials

Adenosine was purchased from Sigma-Aldrich (St. Louis, MO), and CVT-3146 was synthesized by CV Therapeutics, Inc. (Palo Alto, CA).

Data Analysis

All data are presented as mean ± S.E.M. The statistical significance of differences was determined using a paired t test for the response to each injection of the drugs. The significant differences in the responses between CVT-3146 and adenosine were determined using a one-way analysis of variance followed by Tukey's test. Significant changes were considered to be P < 0.05. A computer-based software package (SigmaStat 2.03; SPSS Inc., Chicago, IL) was used for statistical analysis.

Results

Effects of CVT-3146 or Adenosine on the Coronary Blood Flow and Cardiac Output. Intravenous injections of CVT-3146 or adenosine resulted in a dose-dependent increase in CBF and a small increase in CO as shown in Fig. 1. Following injections of CVT-3146 at doses of 0.1, 0.25, 0.5, 1.0, and 2.5 μg/kg, CBF increased by 35 ± 6, 80 ± 12, 151 ± 22, 173 ± 12, and 205 ± 23%, respectively, from 40 ± 4 ml/min (all P < 0.05). A significant increase in CBF was also observed following injections of adenosine at doses of 10, 25, 50, 100, and 250 μg/kg (58 ± 13, 94 ± 19, 128 ± 7, 158 ± 11, and 163 ± 16%, respectively, all P < 0.05) from 41 ± 5 ml/min. The maximal CBF to CVT-3146 or adenosine was not significantly different (Fig. 1), whereas the ED₅₀ values (potency) of CVT-3146 and adenosine necessary to increase CBF were significantly different (0.45 ± 0.07 μg/kg for CVT-3146 versus 47 ± 7.77 μg/kg for adenosine, P < 0.05).

The second difference in CBF response to CVT-3146 and adenosine was in the duration of coronary vasodilation. As shown in Fig. 2, the duration of coronary vasodilation with adenosine (250 μg/kg) was markedly shorter than that induced by CVT-3146 (2.5 μg/kg). There was an increase in CBF following the injection of adenosine at a dose of 250 μg/kg, but CBF returned to the baseline within 1 min. Following the injection of CVT-3146 at a dose of 2.5 μg/kg, CBF remained at 2-fold above the baseline for more than 2 min. The duration of CBF 2-fold above the baseline for CVT-3146 (2.5 μg/kg) and adenosine (250 μg/kg) was 130 ± 19 s and 16 ± 3 s (P < 0.05), respectively. There were dose-dependent decreases in LDCR following injection of CVT-3146 or adenosine (Fig. 1), and these were not significantly different from each other (P > 0.05).

CVT-3146 and adenosine caused significant increases in CO. CVT-3146 at doses of 0.1 and 0.25 μg/kg resulted in a smaller increase in CO when compared to adenosine at doses of 10 and 25 μg/kg (2.2 ± 0.96 versus 13 ± 1% and 5.0 ± 1.4 versus 16 ± 3%, both P < 0.05). There was no significant difference in the increase in CO induced by higher doses of CVT-3146 or adenosine (P > 0.05). Most importantly, CVT-3146 resulted in a markedly smaller decrease in TPR compared with that induced by adenosine, even though CVT-3146 and adenosine caused a similar peak increase in CBF and a comparable decrease in LDCR. Figure 3 shows the ratio between the decrease in LDCR and TPR for CVT-3146 and adenosine (0.5 versus 50 μg/kg and 1.0 versus 100 μg/kg, respectively, both P < 0.05). The ratio of LDCR/TPR for CVT-3146 was significantly greater than that for adenosine.

Effects of CVT-3146 or Adenosine on the Blood Flow in the Lower Body, Mesentery, and Kidney. Following injections of adenosine at doses of 10, 25, 50, 100, and 250 μg/kg, LBF increased by 26 ± 7, 45 ± 14, 51 ± 9, 73 ± 17, and 74 ± 11%, respectively, from 0.68 ± 0.05 l/min (all P < 0.05); LVR decreased by 22 ± 3, 30 ± 6, 36 ± 4, 42 ± 6, and 46 ± 5%, respectively, from 165 ± 12 mm Hg/l/min. CVT-3146 at doses of 0.1 and 0.25 μg/kg did not affect LBF (1.0 ± 1.0 and 5.2 ± 1.9%, respectively, both P > 0.05) and LVR (–1.8 ± 1.0 and –6.2 ± 3.1%, respectively, both P > 0.05). The administration of CVT-3146 at doses of 0.5, 1.0, and 2.5 μg/kg increased LBF by 17 ± 3, 21 ± 2, and 33 ± 6% (P < 0.05), respectively, from 0.70 ± 0.05 l/min, and LVR decreased by 14 ± 2, 20 ± 3, and 22 ± 4% (all P < 0.05), respectively, from 158 ± 13 mm Hg/l/min. The vasodilation by CVT-3146 in the lower body was significantly smaller than that induced by adenosine (Figs. 4 and 5).

As shown in Figs. 4 and 5, the administration of CVT-3146 or adenosine caused significant increases in MBF and a decrease in mesenteric vascular resistance (MVR). In the mesentery, there was no significant difference in vasodilation induced by CVT-3146 or adenosine. MBF increased by 18 ± 4, 28 ± 8, 48 ± 8, 69 ± 6, and 88 ± 14% from 216 ± 31 ml/min (all P < 0.05) following respective injections of CVT-3146 at doses of 0.1, 0.25, 0.5, 1.0, and 2.5 μg/kg; MVR decreased by 15 ± 3, 18 ± 5, 32 ± 4, 40 ± 3, and 48 ± 3%, respectively, from 0.58 ± 0.10 mm Hg/ml/min (all P < 0.05). Adenosine at doses of 10, 25, 50, 100, and 250 μg/kg increased MBF by 36 ± 8, 46 ± 10, 66 ± 9, 72 ± 10, and 84 ± 5%, respectively, from 211 ± 26 ml/min (all P < 0.05); MVR decreased by 26 ± 5, 29 ± 5, 41 ± 4, 43 ± 5, and 49 ± 2%, respectively, from 0.57 ± 0.09 mm Hg/ml/min (all P < 0.05) following injections of adenosine. There were no significant differences in the increase in MBF or the decrease in MVR in response to CVT-3146 administration compared with adenosine.

As shown in Fig. 6, adenosine resulted in a dose-dependent decrease in RBF and a dose-dependent increase in RVR. Adenosine at doses of 10, 25, 50, 100, and 250 μg/kg decreased RBF by 46 ± 7, 54 ± 5, 71 ± 6, 80 ± 5, and 85 ± 4%, respectively, from 246 ± 27 ml/min (all P < 0.05) and increased RVR by 109 ± 30, 125 ± 26, 309 ± 80, 545 ± 174, and 683 ± 197%, respectively, from 0.49 ± 0.09 mm Hg/ml/min (all P < 0.05). In contrast, CVT-3146 had no effect on RBF or RVR (Fig. 6). At the highest dose of CVT-3146 (2.5 μg/kg), a small but significant decrease in RBF (11 ± 4%, P < 0.05 compared with the baseline) was observed.

Effects of CVT-3146 or Adenosine on Blood Pressure and Heart Rate.Table 1 shows changes in MAP and HR in response to CVT-3146 and adenosine in all 13 dogs. There was a dose-dependent decrease in MAP following injections of adenosine. After injections of adenosine at doses of 10, 25, 50, 100, and 250 μg/kg, MAP decreased by 9 ± 2, 12 ± 2, 18 ± 3, 21 ± 3, and 35 ± 5 mm Hg, respectively, from 103 ± 3 mm Hg (all P < 0.05). CVT-3146 at doses of 0.1, 0.25, 0.5, 1.0, and 2.5 μg/kg resulted in significantly smaller decreases in MAP at 3 ± 1, 4 ± 1, 7 ± 1, 8 ± 1, and 14 ± 2 mm Hg, respectively, from 102 ± 4 mm Hg (all P < 0.05) compared with adenosine at doses of 10, 25, 50, 100, and 250 μg/kg. There was a significant increase in HR following injections of CVT-3146 or adenosine; however, as shown in Fig. 7, the patterns of tachycardia induced by CVT-3146 and adenosine were significantly different. Adenosine caused a short duration of tachycardia, whereas CVT-3146 resulted in a longer lasting tachycardia.

Discussion

The coronary vasodilator effect of adenosine is the basis for its use in pharmacologic stress testing. Unfortunately, pharmacologic stress induced by adenosine is often associated with a high incidence of side effects including dyspnea, chest pain, and atrioventricular nodal block (Mahmarian and Verani, 1994; McGuinness and Talbert, 1994; Marwick, 1997). The atrioventricular nodal block is mediated by A₁ AdoR. Some evidence suggests that A_2B AdoR may play a role in increasing airway resistance (Fozard and Hannon, 1999). This implies that the A_2B AdoR might be responsible for the dyspnea that occurs in pharmacologic stress testing. Because a number of side effects induced by adenosine appear to be mediated by AdoR subtypes other than the A_2A receptor, it is presumed that selective agonists of A_2A receptors would cause fewer undesirable effects.

CVT-3146 is a novel A_2A AdoR agonist. Our previous results have indicated that, in conscious dogs, CVT-3146 is a more potent coronary vasodilator than adenosine (Trochu et al., 2003). The present study demonstrates that the administration of CVT-3146 or adenosine results in a dose-dependent increase in CBF and a dose-dependent decrease in LDCR. Although the maximal increases in CBF were not significantly different, the ED₅₀ values of CVT-3146 and adenosine necessary to increase CBF were 0.45 ± 0.07 μg/kg and 47 ± 7.77 μg/kg, respectively (P < 0.05), indicating that as a coronary vasodilator, CVT-3146 is 100-fold more potent than adenosine.

CVT-3146 or adenosine also caused dose-dependent decreases in TPR; however, the decrease in TPR following the administration of CVT-3156 was significantly smaller than that induced by adenosine (Fig. 1). The smaller decrease in TPR by CVT-3146 may account for the smaller hypotension caused by CVT-3146 compared with adenosine (Table 1). Furthermore, the ratio between decreases in LDCR and TPR observed with CVT-3146 were markedly greater than those for adenosine (Fig. 3), confirming that CVT-3146 is a more selective coronary vasodilator than adenosine.

The duration of the coronary vasodilation induced by pharmacologic stress testing agents is an important determinant of their usefulness. Because adenosine has an ultrashort duration of action, it is usually administered intravenously for 4 to 6 min to achieve the desired coronary vasodilation during stress testing (Mahmarian and Verani, 1994; McGuinness and Talbert, 1994; Allison et al., 1996; Leppo, 1996; Marwick, 1997; Cerqueira, 2000). Several potent, high-affinity, selective A_2A AdoR agonists, such as CGS-21680 and WRC-0470, have been developed (Glover et al., 1996; Alberti et al., 1997; Belardinelli et al., 1998; Shryock et al., 1998; Gao et al., 2001a). Our previous results have shown that these two agents have excessively prolonged coronary vasodilator effects in the isolated rat heart (Gao et al., 2001a), which may limit the clinical usefulness of these agents in pharmacologic stress testing because of their potential to cause significant hypotension. Although the duration of coronary vasodilation caused by CVT-3146 is longer than that seen with adenosine (Fig. 2), there was no significant difference in the magnitude of the peak increase in CBF produced by the two agents. The duration of CBF 2-fold above the baseline was 130 ± 19 s for CVT-3146 (2.5 μg/kg) and 16 ± 3 s for adenosine (250 μg/kg) (P < 0.05). These results clearly indicate that CVT-3146 yields a longer duration of coronary vasodilation than adenosine. Furthermore, CVT-3146 may be administered by a single injection during radionuclide myocardial perfusion imaging, thereby simplifying the procedure. Our recent study in humans has revealed that CVT-3146 can cause a dose-dependent coronary vasodilation following a bolus intravenous injection. The profile of coronary vasodilation by CVT-3146 in humans is very similar to that found in conscious dogs in the present study (Kerensky et al., 2002). This provides another piece of evidence to support that CVT-3146 could be administered by a single injection. Increased potency and longer duration of the action make CVT-3146 a much better candidate for pharmacologic stress testing than adenosine.

In addition to coronary vasodilation, the present results indicate that the administration of CVT-3146 or adenosine causes vasodilation in the mesentery and the lower body. There was no significant difference in the vasodilation response to CVT-3146 in the mesentery compared with adenosine. However, CVT-3146 caused a smaller increase in LBF than adenosine (Figs. 4 and 5). The increase in LBF seen with adenosine was ∼2-fold greater as compared to that induced by CVT-3146 and may account for marked hypotension and reduced TPR by adenosine, even though both drugs caused equivalent increases in CBF. The mechanism for the differential vasodilation in the lower body induced by CVT-3146 and adenosine is likely related to the distribution and/or density of AdoR. It has been demonstrated that there are A₁, A_2A, and A_2B AdoR on vascular smooth muscle and endothelial cells in human skeletal muscle (Lynge and Hellsten, 2000). Adenosine can activate all three types of receptors, thereby causing a greater increase in LBF. CVT-3146 can only activate A_2A AdoR, thus resulting in a smaller increase in LBF. This, in turn, reduces the magnitude of hypotension following injection of CVT-3146 compared with adenosine, which would be advantageous in pharmacologic stress testing.

Another important finding of the present study is that adenosine caused dose-dependent renal vasoconstriction, whereas CVT-3146 had little effect on the renal circulation (Fig. 6). This is in keeping with previous studies (McCoy et al., 1993; Shryock and Belardinelli, 1997; Pelueger et al., 1999), in which the administration of adenosine yielded afferent arteriolar constriction, an effect mediated by A₁ AdoR. Our findings are consistent with the idea that CVT-3146 does not activate A₁ AdoR (Gao et al., 2001a), and the lack of renal vasoconstriction by CVT-3146 is yet another benefit of using this agent in pharmacologic stress testing. The administration of adenosine over 4 to 6 min may yield prolonged vasoconstriction in the kidney, which is a potentially undesirable effect. More importantly, renal vasoconstriction by adenosine is invisible and difficult to recognize and may impair renal function. Elderly patients and those with coronary arterial disease whose renal function is already diminished may be at the greatest risk for adenosine-induced renal vasoconstriction. Because CVT-3146 has little or no effect on renal blood flow, it is unlikely to induce renal dysfunction when used in pharmacologic stress testing.

Some studies showed that adenosine could increase sympathetic nerve activity in humans, thereby causing direct tachycardia (Biaggioni et al., 1991; Lucarini et al., 1992; Engelstein et al., 1994). The present results have also shown that there is a significant increase in heart rate or tachycardia following injections of CVT-3146 or adenosine. However, as shown in Fig. 7, the patterns of the tachycardia induced by CVT-3146 and adenosine were significantly different. This suggests that the mechanism(s) responsible for tachycardia caused by CVT-3146 and adenosine may be different. Adenosine-induced tachycardia in conscious dogs was of short duration and mostly due to the baroreflex effect caused by hypotension. This reflex tachycardia overcomes bradycardia, which is the direct depressant effect of adenosine on HR and is mediated by A₁ AdoR (Gao et al., 2001b). Our previous work showed that vagal tone might play an important role in adenosine-induced tachycardia in normal conscious dogs, because the blockade of muscarinic receptors or bilateral vagotomy could unmask adenosine-induced bradycardia in normal, conscious dogs (Hintze et al., 1985; Belloni and Hintze, 1987; Belloni et al., 1989). Moreover, adenosine resulted in bradycardia in conscious dogs after the development of pacing-induced heart failure, in which vagal tone is reduced (Belloni et al., 1992). In contrast, our current results show that CVT-3146-induced tachycardia lasts much longer than that caused by adenosine, especially at higher doses (1.0 and 2.5 μg/kg). The tachycardia lasted up to 4 to 5 min following the injection of a 2.5 μg/kg dose of CVT-3146 (Fig. 7). Whether CVT-3146-induced tachycardia observed in the present study is due to the increased sympathetic nerve activity remains to be determined.

It has been reported that the increased endogenous adenosine induced by dipyridamole could cause hyperventilation in conscious humans (Engelstein et al., 1994), which is mediated by peripheral chemoreceptor activation since peripheral chemoreceptor suppression (with hyperoxia) abolished hyperventilation induced by dipyridamole. Adenosine-induced hyperventilation can result in changes in CBF and HR indirectly. We did not measure ventilation in our present study quantitatively, although hyperventilation was observed in some dogs during the administration of CVT-3146 or adenosine. This is one of the limitations of our study.

In summary, our study indicates that CVT-3146 is a 100-fold more potent coronary vasodilator than adenosine, producing a longer duration of coronary vasodilation that may be useful in pharmacologic stress testing. CVT-3146 also causes a smaller decrease in total peripheral resistance and a smaller increase in LBF than that induced by adenosine, suggesting CVT-3146 is more selective for coronary than for peripheral vasodilation. This study also demonstrates that CVT-3146 causes no renal vasoconstriction. These features make CVT-3146 a promising candidate for pharmacologic stress testing with myocardial perfusion imaging using radionuclides for the diagnosis of coronary artery disease.