INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Endogenous nitric oxide (NO) produced by the pulmonary vascular endothelial cell modulates basal tone and influences vasoreactivity in the pulmonary vascular bed (1). NO causes vasodilation by stimulating soluble guanylate cyclase in vascular smooth muscle, which increases intracellular cyclic guanosine monophosphate (cGMP) content (4). Smooth-muscle cell cGMP production is countered by hydrolysis by cGMP-specific phosphodiesterases (PDE5) (8). Experimental work suggests that endogenous NO activity (and thus cGMP production) is attenuated in some experimental models of pulmonary hypertension and also in clinical pulmonary hypertension (11). In addition, persistent or increased PDE5 activity contributes to increased pulmonary vascular tone in adult rats exposed to chronic hypoxia (14). Thus, the high resting pulmonary vascular tone and altered vasoreactivity that characterize clinical pulmonary hypertension may reflect decreased endogenous NO production and/or increased or persistent PDE5 activity. In addition, decreased endogenous NO production may contribute to exaggerated hypoxia-induced pulmonary vasoconstriction encountered in some patients at high altitude (15).

Treatment with inhaled NO (iNO) causes selective pulmonary vasodilation in experimental and clinical pulmonary hypertension (15). However, responsiveness to iNO therapy is variable, with some patients having limited or transient hemodynamic improvement. Although multiple mechanisms contribute to decreased responsiveness to iNO, increased cGMP degradation due to persistent or increased PDE5 activity may play a critical role in limiting pulmonary vasodilation by iNO. Whether inhibition of PDE5 activity provides a clinical strategy for reducing pulmonary vascular resistance, enhancing responsiveness to iNO, or attenuating pulmonary vasoreactivity is unknown.

Two PDE5 inhibitors have been well characterized: zaprinast and dipyridamole. These agents have very similar PDE5 enzyme inhibition profiles in vitro, and zaprinast has been shown in several experimental models to prolong and potentiate cGMP-dependent vasodilation (8, 24). However, zaprinast is currently unavailable for clinical use. Dipyridamole has been studied less extensively, partly because it has other putative mechanisms of vasodilation, including inhibition of adenosine reuptake and nonspecific inhibition of cAMP phosphodiesterase (28).

Previous studies suggest that dipyridamole-induced pulmonary vasodilation is not altered by adenosine receptor blockade (31, 32). In addition, in the fetal lamb, inhibition of endogenous NO production with the NO synthase antagonist nitro-L-arginine (L-NA) completely blocks the pulmonary vasodilator response to dipyridamole, but not to adenosine, suggesting that dipyridamole's primary mechanism of action is through its effects on cGMP metabolism (33). Dipyridamole potentiates the vasodilator response to iNO in the normal fetal pulmonary circulation and experimental pulmonary hypertension in vivo, as well as in the isolated rabbit aorta (33).

Because dipyridamole is commercially available, has an excellent safety profile in humans, and is a known PDE5 inhibitor, it is an attractive agent to study in clinical pulmonary hypertension. To test the hypotheses that dipyridamole causes pulmonary vasodilation and augments responsiveness to iNO treatment, we studied the acute hemodynamic effects of this drug, alone and in combination with iNO, in children with severe pulmonary hypertension undergoing cardiac catheterization. Since increased pulmonary vascular reactivity characterizes clinical pulmonary hypertension, we further assessed dipyridamole in patients with severe pulmonary vasoconstrictor responses to acute hypoxia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The following study protocols were reviewed and approved by the Institutional Review Board at the University of Colorado Health Sciences Center prior to patient enrollment. All studies were performed in Denver (elevation = 5,280 ft.).

Protocol 1: Hemodynamic Effects of Dipyridamole and NO in Children with Severe Pulmonary Hypertension

Eleven consecutive patients with a median age of 16 yr (range: 1 to 19 yr) were enrolled in the study after giving informed consent (and assent when appropriate). These patients were referred for cardiac catheterization specifically to evaluate pulmonary vascular responsiveness to vasodilator therapy. The following entrance criteria were used to enroll patients: (1) need for cardiac catheterization to evaluate pulmonary hemodynamics and pulmonary vasodilator responses, as determined by each patient's referring cardiologist; (2) absence of reactive airways disease requiring chronic bronchodilator or antiinflammatory therapy (this exclusion criterion was observed because of dipyridamole's potential for enhancing bronchoconstriction in asthmatic patients [36]; (3) absence of intracardiac shunts; and (4) basal mean pulmonary artery pressure () > 50% mean systemic arterial pressure () during breathing of room air.

Eleven patients met the study entrance criteria. One patient had a hypotensive response to dipyridamole necessitating withdrawal from the study. A total of 12 studies were successfully completed without complications on the remaining 10 patients, and are included for data analysis. Table 1 illustrates the ages, diagnoses, and room-air hemodynamics for the study patients. As shown in Table 1, the study population was very heterogeneous. Two patients' diagnoses were difficult to classify and in the table are labeled altitude pulmonary hypertension (PH). These two patients had no evidence of underlying structural heart disease, and both resided at altitudes > 8,000 ft. Since both patients had been diagnosed  7 yr previously with pulmonary hypertension, and since neither had manifested progression of their disease activity over that period, these two patients did not fit the usual clinical profile of pediatric primary pulmonary hypertension. One of these patients (Patient 9) was also included in Protocol 2. 

Patients were initially sedated with intravenous meperidine (1 to 2 mg/kg) and diphenhydramine (1 mg/kg), and subsequently received intermittent doses of midazolam (0.1 mg/kg) and meperidine (0.5 to 1 mg/kg) as needed to ensure patient comfort and analgesia. No patients were intubated at the time of study. Arterial and venous access was obtained through the femoral approach, using the modified Seldinger technique. A size 6 or 7 Fr triple-lumen, flow-directed thermodilution catheter (Baxter Edwards, Irvine, CA) was advanced into a lower branch pulmonary artery under fluoroscopic guidance. A size 4 or 5 Fr pigtail catheter (Cook Catheter, Bloomington, IN) was positioned in the ascending aorta for systemic blood pressure measurements. Transducers were positioned in the midaxillary line and zeroed at atmospheric pressure. Cardiac index (CI) was measured in triplicate by the thermodilution technique (Cardiac Output Computer; Baxter Edwards) and indexed for body surface area (BSA).

Hemodynamic measurements, including , , pulmonary capillary wedge pressure (), and right atrial () pressure, CI, and heart rate (HR) were made while patients breathed room air, 100% oxygen, and 100% oxygen plus inhaled NO (at 20 ppm). Pulmonary vascular resistance index (PVRI) was calculated as follows: PVRI = (  )/CI. Patients were treated with sequential exposure to oxygen and oxygen plus inhaled NO for at least 5 min, or until hemodynamic measurements reached a steady state. After a 10-to-20-min recovery period, measurements of baseline hemodynamics were repeated. In some patients the second baseline measurements were obtained during breathing of 100% oxygen. Dipyridamole (0.6 mg/kg over 15 min) was then administered through either the right atrial (RA) or distal pulmonary arterial (PA) port of the pulmonary artery catheter. The dose of dipyridamole chosen for study had been shown in previous studies to be well tolerated in pediatric patients, without adverse hemodynamic consequences (37). Hemodynamic measurements were recorded immediately after completion of the dipyridamole infusion. Patients were then reexposed to inhaled NO (20 ppm) after loading with dipyridamole.

NO gas (800 ppm in nitrogen) was mixed with oxygen immediately prior to administration via a head hood or mask, with NO and NO₂ concentrations verified by electrochemical sensors or chemiluminescence analyzers (Thermo Environmental Instruments, Franklin, MA). Arterial blood gas measurements were made at baseline and after oxygen and NO exposures. Methemoglobin concentrations were measured at baseline and after the second NO exposure, with a Radiometer OSM3 blood gas analyzer (Copenhagen, Denmark).

Because dipyridamole purportedly raises levels of circulating adenosine, and since aminophylline has adenosine receptor-blocking properties, a dose of aminophylline (5 mg/kg) was available for each patient in case marked hypotension occurred after dipyridamole administration.

Protocol 2: Effect of Dipyridamole on Exaggerated Acute Hypoxia-induced Pulmonary Vasoconstriction

Patients (n = 4; median age: 2 yr) were recruited into this protocol during evaluation of suspected altitude-related pulmonary hypertension (Table 2). Of the four patients in the protocol, two had repaired congenital heart disease (a large primum atrial septal defect [ASD] and cleft mitral valve in one case and a moderate secundum ASD in the other). These two patients resided at 5,500 ft. altitude. The other two patients had structurally normal hearts, resided at altitude > 8,500 ft., and presented with frequent lower respiratory tract infections and chronic growth impairment. One patient had a history of recurrent syncopal episodes. All four patients had evidence of persistent right ventricular hypertrophy by noninvasive assessment.

Because the four patients in Protocol 2 lived at varying altitudes, catheterization was planned to evaluate pulmonary hemodynamics and specifically to assess the pulmonary vascular response to low ambient oxygen concentrations. This was felt necessary in order to determine risk for continued residence at altitude, and also to provide objective data to support a recommendation for relocating to a lower altitude.

Three of the four patients had a normal response to hyperoxia, with a decrease in to  18 mm Hg. One patient had a mild increase of on 100% oxygen ( = 22 mm Hg), and thus received iNO as in Protocol 1. Sixteen percent oxygen was administered to simulate the physiologic effects of altitude. Clinical procedures, including sedation, catheter placement, and measurement of hemodynamic variables, were performed as described for Protocol 1. Patients were administered a hypoxic gas mixture by head hood (FI_O2 = 0.16; balance nitrogen) until reached a steady state (10 to 15 min). One hundred percent oxygen was then administered until hemodynamic parameters recovered, and the patient was then restored to breathing room air. Dipyridamole was administered as described earlier, followed by repeat hypoxic exposure to hypoxia. The second hypoxic exposure was maintained until peripheral oxygen saturation, as measured by pulse oximetry, fell to a level measured during the first hypoxic exposure or less. Hemodynamic measurements were made at the first room-air baseline, after the first hypoxic exposure to hypoxia, at the second room-air baseline, after dipyridamole administration, and after the second exposure to hypoxia.

Statistical Analysis

All data are reported as mean ± SEM. Statistical analysis was done with SuperAnova software (Statview, Berkeley, CA). Comparisons between study points over time were made through a univariate repeated measures analysis of variance (ANOVA) with differences defined by post hoc linear contrast analysis. For the comparison of two discrete variables, a paired t test was used. A value of p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Protocol 1: Hemodynamic Effects of Dipyridamole and iNO in Patients with Severe Pulmonary Hypertension

Thirteen studies were performed on 11 patients. One patient, a 16-yr-old girl with trisomy 21 who had had repair of atrial and ventricular septal defects at 14 mo of age, had a marked decrease in pulmonary and systemic arterial pressures (73 to 40 mm Hg [] and 95 to 57 [] mm Hg, respectively) following administration of dipyridamole. This patient received 20 ml/kg 0.9% saline, and was also given 5 mg/kg aminophylline. Following this therapy, pressures returned to baseline, although the study was not continued. Twelve studies were completed on the remaining 10 patients, and are included for data analysis.

Figure 1 illustrates the effects of 100% oxygen, 100% oxygen + iNO, and 100% oxygen + iNO + dipyridamole on pulmonary and systemic pressures and PVRI. As shown, 100% oxygen decreased from values measured while breathing room air; the addition of iNO (20 ppm) led to a further decrease in and PVRI. remained constant during treatment with 100% oxygen and iNO. The addition of dipyridamole did not cause further pulmonary vasodilation, but did result in a decrease in .

View larger version (14K): [in this window] [in a new window]   Figure 1.   Effects of 100% oxygen, 100% oxygen + iNO, and 100% oxygen + iNO + dipyridamole (Dip) on , , and PVRI in children with severe pulmonary hypertension (Group 1). In comparison with the response to 100% oxygen, iNO caused a further decrease in and PVRI without affecting . The addition of dipyridamole did not cause further pulmonary vasodilation, but did result in a decrease in .

Table 3 summarizes hemodynamic variables at different study points. Inhaled NO decreased and PVRI without affecting CI or . Inhaled NO increased in the study group as a whole (from 17 ± 3 mm Hg to 20 ± 3 mm Hg; p < 0.05), but this was entirely due to the effect of iNO in patients with a measured  > 12 mm Hg at baseline (Figure 2). Dipyridamole reduced PVRI to a similar degree as did iNO (Table 3 and Figure 3). Although some patients had a decrease in following dipyridamole administration did not change, in the group as a whole, and the decrease in PVRI after dipyridamole was primarily due to an increase in CI, which resulted from a 19% increase in HR and a 16% increase in stroke volume (SV) (p < 0.05 versus baseline for each variable). Unlike iNO, dipyridamole was not selective for the pulmonary circulation, equivalently decreasing and calculated SVRI as well as PVRI. Despite the effect of dipyridamole on systemic hemodynamics, only one patient had a decrease in blood pressure significant enough to require intervention.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Effects of iNO on in patients with elevated at baseline (> 12 mm Hg) versus those with normal ( 12 mm Hg). increased during iNO therapy in patients with high baseline values (from 24 ± 2 mm Hg to 31 ± 2 mm Hg; p < 0.05), but did not change in patients with normal (10 ± 1 mm Hg to 9 ± 1 mm Hg).

View larger version (15K): [in this window] [in a new window]   Figure 3.   Effects of dipyridamole and iNO on hemodynamic variables expressed as percent change from baseline. Dipyridamole (0.6 mg/kg) caused significant changes in , CI, and PVRI and SVRI without changing or . In contrast, iNO (20 ppm) caused selective pulmonary vasodilation and an increase in .

The response to dipyridamole in combination with iNO was variable (Figure 4). Half of the study patients ("responders") had more than a 20% additional decrease in PVRI when iNO was administered with dipyridamole by comparison with responses to iNO alone. In responders, the effect of dipyridamole in combination with iNO was additive, since both agents independently reduced PVRI.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Effect of 100% oxygen + iNO and 100% oxygen + iNO + dipyridamole (Dip) on and PVRI in 10 individual patients with severe pulmonary hypertension. Fifty percent of the patients had  20% further decrease in PVRI with the addition of dipyridamole (open squares). In two of these patients (#), the decrease in resistance was selective for the pulmonary circulation.

Arterial blood gas values did not differ between the first and second iNO exposures. Serum pH and values during breathing of room air, 100% oxygen, and 100% oxygen + iNO, and 100% oxygen + iNO + dipyridamole were 7.38 ± 0.01, 7.39 ± 0.01, 7.39 ± 0.02, 7.38 ± 0.02 (p = NS), and 73 ± 4 mm Hg, 266 ± 60 mm Hg, 227 ± 56 mm Hg, and 248 ± 64 mm Hg (p = NS for 100% oxygen versus 100% oxygen + iNO and 100% oxygen + iNO + dipyridamole), respectively. Methemoglobin levels did not change from baseline following the brief exposures to inhaled NO (0.7 ± 0.05 g/dl versus 0.8 ± 0.08 g/dl).

Protocol 2: Effect of Dipyridamole on Exaggerated Acute Hypoxia-induced Pulmonary Vasoconstriction

Table 4 illustrates hemodynamic values at each study point. In the study group subjected to Protocol 2, baseline was mild to moderately elevated during breathing of room air, and decreased to almost normal values during administration of 100% oxygen (34 ± 5 mm Hg [baseline] to 20 ± 1 mm Hg [O₂]; p < 0.05). Initial exposure to acute hypoxia (FI_O2 = 0.16) resulted in a rapid rise in and PVRI to suprasystemic levels in all four patients. Hemodynamic variables returned to baseline values during recovery from acute hypoxia. When administered under normoxic conditions, dipyridamole caused parallel decreases in (from 36 ± 2 mm Hg [baseline] to 26 ± 2 mm Hg [treatment]; p < 0.05) and (from 80 ± 9 mm Hg to 67 ± 10 mm Hg; p < 0.05). Dipyridamole did not change the Rp/Rs ratio, but increased CI (3.9 ± 0.6 units to 6.0 ± 1.2 units; p < 0.05). The increase in CI resulted from a 22% increase in HR and 23% increase in SV (101 ± 5 beats/min to 123 ± 2 beats/min for HR and 39 ± 6 to 48 ± 10 for SV; p < 0.05 for HR only). Repeat exposure to acute hypoxia after dipyridamole administration attenuated the acute hypoxic pressor response, with , PVRI, and the Rp/Rs ratio increasing to only 76%, 54%, and 67% of hypoxia control values (p < 0.05 versus hypoxia control values). In contrast with initial exposure to hypoxia, and PVRI remained below systemic levels during exposure to hypoxia after administration of dipyridamole. did not differ between the first and second exposures to hypoxia (46 ± 2 mm Hg and 43 ± 2 mm Hg, respectively).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inhaled NO has proven efficacy as a selective pulmonary vasodilator, providing a novel approach to the treatment of clinical pulmonary hypertension (16). Since iNO lowers and PVRI by increasing pulmonary vascular smooth-muscle cell cGMP content, it is possible that incomplete responses to iNO reflect failure of the pulmonary vascular smooth-muscle cell to sustain cGMP levels (4). One mechanism of such failure might be increased cGMP degradation by PDE5. Thus, agents that inhibit PDE5 might provide an approach to treating pulmonary hypertension. In the present study, we tested dipyridamole, a PDE5 inhibitor, in patients with pulmonary hypertension and altered pulmonary vasoreactivity. Our goals in studying this agent were twofold: (1) to evaluate the effect of dipyridamole on pulmonary hemodynamics and iNO-induced (and thus cGMP-dependent) pulmonary vasodilation; and (2) to assess the effect of dipyridamole on exaggerated, hypoxia-induced pulmonary vasoconstriction. In patients with severe pulmonary hypertension during rest, we found that iNO (20 ppm) and dipyridamole (0.6 mg/kg) equivalently reduced PVRI, although the effect of dipyridamole was nonselective and resulted almost entirely from an increase in CI. For the study group as a whole, potentiation of iNO-induced pulmonary vasodilation with dipyridamole could not be demonstrated; however, 50% of patients had a  20% further decrease in PVRI with combined treatments in comparison with either agent alone. In four patients with markedly abnormal hypoxia-induced pulmonary vasoconstriction, the administration of dipyridamole effectively attenuated the increase in and PVRI upon reexposure to hypoxic gas, without adversely affecting systemic pressure.

Clinical experience with dipyridamole in the treatment of pulmonary hypertension is limited. Fullerton and colleagues recently reported their findings with 10 adult patients immediately after cardiopulmonary bypass and left sided valve replacement, who were treated with the combination of iNO (40 ppm) and dipyridamole (0.2 mg/kg) (40). In these patients, combined treatment selectively reduced and PVRI, whereas neither agent alone had an effect on baseline hemodynamics. Fullerton and colleagues did not state whether this response was universally seen, and it is possible that some patients in their study group did not respond to combined treatment, which is what we encountered in our patients with severe pulmonary hypertension. Other factors that might account for our differing results include the clinical heterogeneity and increased severity of pulmonary hypertension in our patient population and the doses of iNO and dipyridamole used. In our study group, two of 10 patients had a selective decrease in and PVRI, with a further reduction of the PVRI/ SVRI ratio with combined therapy; both of these patients had pulmonary hypertension from left atrial and pulmonary venous hypertension, as in Fullerton and colleague's population. It is thus possible that dipyridamole in combination with iNO is most effective in patients with pulmonary hypertension resulting from longstanding left atrial hypertension, although as demonstrated in our patients and by other investigators, iNO administration in the presence of left ventricular dysfunction may lead to further elevation of filling pressures (41). Ivy and associates recently reported their findings with seven patients who had undergone repair of congenital heart disease, demonstrating marked, sustained rebound pulmonary hypertension following discontinuation of low-dose iNO (44). Six of the seven patients received dipyridamole (0.6 mg/kg), after which rebound pulmonary hypertension and its detrimental hemodynamic consequences were completely abolished. In these six patients, dipyridamole did not affect baseline hemodynamics, although again this was a distinctly different patient population than ours. The results obtained by Ivy and associates are consistent with the results we obtained in our four patients with marked hypoxia-induced vasoconstriction, in whom we demonstrated marked attenuation in pulmonary vasoconstrictor responses.

Although dipyridamole has been used clinically for over 30 yr, its therapeutic use has been mainly as an antiplatelet agent (28). Early animal studies suggested that dipyridamole modified pulmonary vasoconstriction and reduced morphologic changes in the pulmonary vascular bed resulting from acute and chronic hypoxia (29). These early studies utilized dipyridamole, because of its recognized antiplatelet activity, to identify the role of platelets in hypoxia-induced pulmonary hypertension. Subsequent work showed that dipyridamole's effects in the pulmonary circulation were not mediated by platelet inhibition (45). More recently, dipyridamole has been recognized as a potent PDE5 inhibitor, although other putative mechanisms for its vasodilating effect have been described. These include interference with adenosine clearance (resulting in increased circulating levels of adenosine, a potent vasodilator) and inhibition of cAMP phosphodiesterases (10, 28). Results of the present study do not allow clarification of dipyridamole's predominant vasodilator activity, and time constraints in the catheterization laboratory did not allow us to compare dipyridamole with a cGMP-independent vasodilator in our two study populations. Work with ovine models suggests that dipyridamole's activity in the pulmonary circulation is not attenuated by adenosine receptor blockade, and is completely blocked by inhibition of NO synthase (31). These experimental findings indirectly suggest PDE5 inhibition as the main mechanism by which dipyridamole produces vasodilation.

In the present study, dipyridamole's vasodilator effects were not specific for the pulmonary circulation, though this is not surprising, since cGMP modulates vascular tone in both pulmonary and systemic vessels. The fact that dipyridamole reduced PVRI by increasing CI (and not by decreasing ) makes it difficult to interpret whether the decrase in PVRI represents true pulmonary vasodilation or recruitment of pulmonary arterioles without further vasodilation. Given the marked increase in pulmonary blood flow with dipyridamole in the setting of severe pulmonary hypertension, the failure of to increase suggests a component of pulmonary vasodilation. Again, time constraints in the catheterization laboratory did not allow us to compare the pulmonary vascular response to dipyridamole with that of a more pure inotropic agent such as dobutamine, in an effort to answer this mechanistic question. It is also unclear whether the increase in CI following administration of dipyridamole resulted from a reduction in afterload or a direct inotropic/chronotropic effect on the myocardium.

Our four patients with exaggerated hypoxic pulmonary vasoreactivity constituted a somewhat unique population. Although normal values for acute hypoxia-induced increases in and PVRI have not been well established, the consensus of several reports suggests that a normal response to acute hypoxia consists of a 20% to 30% increase in , from 10 to 17 mm Hg at baseline to 15 to 22 mm Hg during hypoxia (46). Marked pulmonary pressor responses to acute hypoxia, to suprasystemic levels, as encountered in our four patients, have been reported in only a few unique circumstances in the pediatric age range, and our patients do not readily fit into any diagnostic category (50). Our rationale for studying dipyridamole in these patients stems from experimental work suggesting decreased endogenous NO and cGMP production as being partly responsible for acute hypoxia-induced pulmonary vasoconstriction (54). It thus follows that PDE5 inhibition, by increasing cGMP levels, might attenuate this response, which is consistent with our findings. As demonstrated by other investigators, our findings in the four patients in Protocol 2 lend further support to the clinical utility of dipyridamole in blunting reactivity in disease states characterized by abnormal pulmonary pressor responses (44, 55, 56).

The present study has several limitations. First, because of the relative rarity of severe pulmonary hypertension in pediatric patients, our patient population was very heterogeneous in terms of the demographics and etiology of pulmonary hypertension. It is possible that upregulation of PDE5 activity occurs in some but not all forms of pulmonary hypertension; in such patients, combined therapy with iNO and a PDE5 inhibitor is most likely to be effective. Second, we did not randomize the order of drug administration, and it is possible that some of the effects attributed to combined therapy were merely the result of repetitive iNO dosing. However, augmentation of pulmonary vasodilation with repeated short exposures to iNO has not been described, and experimental evidence suggests that the opposite effect may occur (57). Since dipyridamole's pharmacokinetic profile as a PDE5 inhibitor is not known, and because of time constraints in the catheterization laboratory, we did not feel that giving dipyridamole first was a reasonable option. Last, we did not measure cGMP levels, and it is possible that dipyridamole's effects resulted from other mechanisms. In general, the dose of dipyridamole used in this study was well tolerated, with only one of 15 patients having a decrease in systemic pressure necessitating intervention and withdrawal from the study. However, these patients were hemodynamically stable at the time of study. Whether dipyridamole is safe in unstable postoperative cardiac patients, neonates with right-to-left extrapulmonary shunting, children with intracardiac shunts, or patients with respiratory disease is uncertain. Certainly in the latter population dipyridamole might worsen ventilation/perfusion matching and systemic oxygenation. We conclude that dipyridamole may have clinical utility in some patients with severe pulmonary hypertension who demonstrate partial responsiveness to iNO therapy. In addition, dipyridamole may be useful in attenuating excessive pulmonary vasopressor responses characteristic of some pulmonary hypertensive states. We speculate that dipyridamole is most likely to be effective in disease states associated with limited cGMP production or increased cGMP degradation, especially when combined with iNO. As more potent and selective PDE5 antagonists are developed for clinical use, further studies are warranted to determine the role of PDE5 inhibition in treating pulmonary hypertension.