Author + information
- Received August 18, 2009
- Revision received November 6, 2009
- Accepted November 23, 2009
- Published online April 6, 2010.
- Christian Apitz, MD*,‡,
- Janette T. Reyes, RN*,
- Helen Holtby, MD†,
- Tilman Humpl, MD* and
- Andrew N. Redington, MD*,* ()
- ↵*Reprint requests and correspondence:
Dr. Andrew N. Redington, Division of Cardiology, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada
Objectives The purpose of our study was to characterize the hemodynamic and corresponding pharmacokinetic responses to a single dose of oral sildenafil by children with pulmonary arterial hypertension (PAH) undergoing invasive testing.
Background Although used frequently for the treatment of children with PAH, data regarding the acute responses to sildenafil are limited.
Methods Thirty-six patients (mean age 7.5 ± 5.9 years; 24 females) were studied during cardiac catheterization with general anesthesia. Eight of 36 (22%) had idiopathic PAH; the remainder had associated congenital heart disease. Hemodynamics and serum cyclic-guanosine monophosphate levels (cGMP) were evaluated at baseline and after inhaled nitric oxide (NO) (40 ppm). In addition, cGMP and sildenafil levels were measured 30 min after administration of sildenafil (0.5 mg/kg, suspended in 5 ml sterile water) through a nasogastric tube.
Results For the 36 patients, the pulmonary vasodilating capability of oral sildenafil was lower than that of inhaled NO (2.8% vs. 11.6% reduction in pulmonary vascular resistance indexed to body surface area [PVRI], respectively; p = 0.01). However, only 21 of 36 (58%) patients had a detectable sildenafil level. In those with detectable sildenafil levels, the fall in PVRI was greater (−11.6% vs. −19.1%, p = NS). Mean cGMP levels at baseline and after NO were 41.8 ± 20.0 pmol/ml and 83.8 ± 35.5 pmol/ml, respectively (p < 0.0001). Surprisingly, there was no significant increase in cGMP in patients with either undetectable (37.5 ± 29.8 pmol/ml) or detectable (44.4 ± 31.7 pmol/ml) sildenafil levels (p = NS compared with baseline) with sildenafil.
Conclusions Our study demonstrates suboptimal absorption of sildenafil in almost half the children undergoing acute hemodynamic testing. When detectable, there was no statistically significant difference between the fall in PVRI associated with sildenafil and NO despite lower circulating cGMP levels in the sildenafil group. These data should be taken into account when designing acute testing protocols, and assessing the acute response to sildenafil in patients with PAH.
Acute administration of a single oral dose of sildenafil to adults with pulmonary hypertension causes a significant decrease in mean pulmonary artery pressure (mPAP) and pulmonary vascular resistance (PVR) (1,2), and its utility in chronic therapy is now established (3,4).
The effectiveness of sildenafil as a pulmonary vasodilator in children with congenital heart disease (CHD) was first reported in a small case series post-operatively in 1999 (5), and confirmed 4 years later in the first detailed prospective open-label study during cardiac catheterization and post-operatively (6). Using an intravenous preparation in that study, the authors were able to show a similar acute hemodynamic response to that of inhaled nitric oxide (NO), and a direct relationship between cyclic-guanosine monophosphate (cGMP) level and therapeutic response. Although sildenafil is now used frequently for long-term treatment of children with pulmonary arterial hypertension (7), clinical data regarding the acute pharmacokinetic and hemodynamic responses to sildenafil are limited.
Therefore, the purpose of this study was to characterize the responses to fixed dosing of oral sildenafil for children with pulmonary hypertension undergoing invasive hemodynamic testing in the catheterization laboratory.
The study design was open label, prospective, and interventional. The study protocol was approved by the research and ethics review board of the Hospital for Sick Children, Toronto, Canada, and informed signed consent was obtained from the study subjects and their parents.
All patients undergoing cardiac catheterization to assess pulmonary hypertension, defined as mean pulmonary arterial pressure (mPAP) >25 mm Hg or pulmonary vascular resistance index (PVRI) >5 Wood units (WU), indexed for body surface area (BSA), were eligible for inclusion. These patients routinely undergo pulmonary vascular reactivity drug testing in the cardiac catheterization laboratory before decisions about therapy. We excluded patients with hepatic or renal insufficiency and known retinal disease.
Patients were studied under general anesthesia with mechanical ventilation with a baseline fraction of inspired oxygen (FiO2) of 0.25 (if not required to be higher for clinical reasons). Anesthesia was induced with sevoflurane, midazolam, and remifentanil. Sevoflurane was discontinued after induction. Rocuronium was used for muscle relaxation. A nasogastric tube was placed, and its position confirmed by fluoroscopy. Measurement of baseline hemodynamics included arterial and venous saturations, blood gases, systemic and pulmonary artery pressures, left atrial (or pulmonary capillary wedge) pressure, and right atrial pressure in the standard manner with fluid-filled catheters. End-tidal carbon dioxide and systemic oxygen consumption were continuously determined using respiratory mass spectrometry. Oxygen saturations were measured by co-oximetry after sampling in the superior vena cava, pulmonary vein, pulmonary artery, and systemic artery. We estimated systemic and pulmonary blood flows from the Fick equation. We calculated systemic and PVRs from standard equations (mean arterial pressure minus mean atrial pressure divided by flow). Blood flow and vascular resistances were indexed to BSA.
Assessment of pulmonary vascular reactivity was undertaken as follows: measurements were made at baseline (at “usual” FiO2) and with FiO2 0.7 (if higher than usual requirements). The patients were then returned to baseline FiO2, and after 10 min, the effect of additional inhaled NO at 40 ppm for 10 min was recorded. The NO was then discontinued, and new baseline hemodynamics were measured after 10 min. Subsequently, a dose of sildenafil (0.5 mg/kg, suspended in 5 ml sterile water) was administered through the nasogastric tube. Measurements were repeated at 30 min.
For the determination of sildenafil and cGMP levels, blood samples of each patient were taken from the pulmonary artery and were then transferred to heparinized polypropylene tubes and centrifuged for 10 min at 4,000 rpm; the supernatant plasma was pipetted into screw-capped polypropylene tubes and stored at −80°C, within 50 min of blood sample collection. Measurements of cGMP levels were recorded at baseline, after 10 min of inhaled NO, and before and 30 min after administration of sildenafil. Plasma samples were analyzed using a commercially available enzyme immunoassay (Amersham cGMP, GE Healthcare UK Ltd., Buckinghamshire, United Kingdom). The sildenafil level was measured 30 min after oral administration of sildenafil. The quantitative analyses for sildenafil and its N-desmethyl metabolite were performed at NMS Labs, Willow Grove, Pennsylvania, using high-performance liquid chromatography with tandem mass spectrometry.
The primary outcome measure was the PVRI and the mPAP at cardiac catheterization. A significant acute response to NO and/or sildenafil was defined as a fall in mPAP and/or PVRI of at least 20% relative to the baseline value (8). The secondary outcome measure was the cGMP level at baseline, after NO, and after sildenafil, as well as the sildenafil level 30 min after oral administration.
Data are presented as mean and standard deviation. Comparisons were performed by nonparametric Mann-Whitney test if the sample groups were not paired, for example, comparison between patients with idiopathic and PAH associated with CHD, or the comparison between responders and nonresponders. Paired t tests were utilized to compare the hemodynamic parameters after each intervention to the corresponding baseline value of each patient and to evaluate differences of the effect on PVR of NO and sildenafil.
We used a linear regression test to examine the correlation between plasma sildenafil concentration and cGMP levels, and fall in PVRI, respectively. Analysis was performed using GraphPad statistical software package (San Diego, California). The null hypothesis was rejected when p < 0.05.
Thirty-six patients (mean age 7.5 ± 5.9 years; 24 females) fulfilled entry criteria and were enrolled in the study protocol. The clinical characteristics of the patients are outlined in Table 1. The diagnosis was idiopathic pulmonary hypertension in 8 of 36 (22%) patients, and 28 (78%) patients had associated CHD. At baseline, the mPAP was 46.4 ± 18.2 mm Hg, and the PVRI was 16.5 ± 10.8 WU × m2 BSA.
Mean pulmonary artery pressure decreased with hyperoxia from 46.4 ± 18.2 mm Hg to 44.4 ± 17.4 mm Hg (p = 0.02), and with inhaled NO from 46.5 ± 18.2 mm Hg to 43.9 ± 17.9 mm Hg (p = 0.01). There was a nonsignificant fall in mPAP after administration of oral sildenafil (46.1 ± 18.4 mm Hg to 44.9 ± 18.7 mm Hg). Pulmonary vascular resistance decreased with hyperoxia from 16.5 ± 10.8 WU × m2 BSA to 12.9 ± 8.1 WU × m2 BSA (p = 0.002) and with inhaled NO from 16.4 ± 10.2 WU × m2 BSA to 14.8 ± 10.4 WU × m2 BSA (p = 0.01). With sildenafil, there was a nonsignificant fall in mean PVR from 15.5 ± 10.0 WU × m2 BSA to 15.1 ± 11.2 WU × m2 BSA. For all 36 patients, the pulmonary vasodilation 30 min after oral sildenafil was lower than that with inhaled NO (−2.8 ± 26.7% vs. −11.6 ± 23.5% PVRI reduction, p = 0.01) (Fig. 1A). However, using the criteria of Rich et al. (8), 31% of patients had a significant hemodynamic response to NO (n = 11), and 28% to sildenafil (n = 10).
Sildenafil and cGMP levels
Fifteen of 36 (42%) patients had a sildenafil level below the lower threshold for quantification (<1.2 ng/ml). Furthermore, 24 of 36 (76%) patients had a desmethylsildenafil level below the lower limit of quantification (<1.2 ng/ml). The mean plasma concentration of desmethylsildenafil in patients with detectable levels was 41.7 ± 32.4 ng/ml. The mean plasma concentration of sildenafil in those with detectable levels was 69.3 ± 104.9 ng/ml, and 9 of 21 (43%) were responders. In patients with detectable sildenafil levels, the pulmonary vasodilating capability of oral sildenafil was not statistically significantly different from inhaled NO (−11.6 ± 23.2% vs. −19.1 ± 18.9%, p = NS) (Fig. 1B). In patients with an undetectable sildenafil level, there was no detectable pulmonary vasodilating capability 30 min after administration of sildenafil (+9.4 ± 27.1%, p = NS).
Comparing sildenafil responders and nonresponders, the reduction of PVRI with NO was −30.5 ± 20.2% in responders versus −4.3 ± 20.7% in nonresponders, and with sildenafil, it was −29.6 ± 18.5% in responders versus +7.4 ± 21.8% in nonresponders (p < 0.0001, for both comparisons). In terms of absolute measurements, the PVRI before and after sildenafil in the responders was 15.2 ± 8.9 WU × m2 and 10.8 ± 6.9 WU × m2 (p = 0.003), and in the nonresponders it was 15.5 ± 10.6 WU × m2 and 16.8 ± 12.2 WU × m2 (p = NS).
Mean cGMP levels at baseline and after NO were 41.8 ± 20.0 pmol/ml and 83.8 ± 35.5 pmol/ml, respectively (p < 0.0001) (Fig. 2A). Comparing sildenafil responders and nonresponders, we realized higher cGMP levels with NO and a significantly steeper decline after withdrawal of NO in sildenafil nonresponders (p < 0.0001) (Figs. 2B and 2C).
Surprisingly, there was no increase in cGMP with sildenafil in either patients with undetectable (37.5 ± 29.8 pmol/ml) or detectable (44.4 ± 31.7 pmol/ml) sildenafil levels (p = NS compared with baseline). Nonetheless, in the latter group, there was a weak correlation between sildenafil level and fall in PVRI (r = 0.41; p = 0.06) (Fig. 3A) and between sildenafil and cGMP levels (r = 0.5; p = 0.02) (Fig. 3B).
Idiopathic PAH versus PAH associated with CHD
Comparing patients with idiopathic PAH to those with PAH associated with CHD, we found higher sildenafil levels in patients with idiopathic PAH (151.02 ± 179.5 ng/ml vs. 43.8 ± 56.2 ng/ml; p = 0.04), whereas a positive response was more frequently seen in patients with PAH associated with CHD (32% vs. 13%). There was no significant difference in the cGMP response between the 2 patient groups.
This study is the first to detail the hemodynamic and pharmacokinetic responses to enterally administered sildenafil by children undergoing acute pulmonary vasodilator testing. The accurate assessment of PAH and its response to therapy is a critical component of management of children with and without associated structural heart disease. Invasive hemodynamic testing in the catheterization laboratory, usually employing general anesthesia in younger infants, remains the gold standard for diagnosis, assessment of prognosis, and guidance of long-term therapy (9). Testing protocols vary, but most include evaluation of pulmonary vascular responsiveness to “selective” pulmonary vasodilators, such as oxygen and NO and prostacyclin (10–12). While not being advised by international guidelines because of the absence of published data on its utility as a predictor of calcium-channel blocker response, there are several studies assessing the acute effects of orally administrated sildenafil in adults with pulmonary hypertension (2,13–16). Although the drug is almost universally effective hemodynamically, few studies relate these responses to the measured levels of plasma sildenafil or cGMP. The data are even more limited for children. However, a direct relationship between cGMP levels and hemodynamic response was shown in response to both inhaled NO and sildenafil administered intravenously to children undergoing testing in the cardiac catheterization laboratory in a previous study from members of our group (6), and recently some centers have introduced oral sildenafil as an alternative to inhaled NO to test pre-operative operability in children with congenital heart defects (17). We believe it is timely, therefore, to document the responses to enteral sildenafil in children undergoing cardiac catheterization.
In the current study, we investigated not only the acute hemodynamic effects of a single oral dose of sildenafil, but also its corresponding pharmacokinetics in children. When administered orally in capsule form to adult volunteers, the peak plasma level of sildenafil is seen after 45 to 60 min (18), and this time frame coincides with its clinical effects in patients (13,19). However, pre-clinical studies in juvenile lambs (weighing 16 to 25 kg) have shown the onset of pulmonary vasodilation to occur within 5 min, and maximal vasodilation (with coincident plasma sildenafil levels of 28.8 ± 9.9 ng/ml) to occur 15 min after administration of sildenafil suspension through a nasogastric tube (20). Similarly, after administration as a suspension through a nasogastric tube, an almost immediate hemodynamic response to sildenafil has been reported in infants with PAH (21). In the only previous study of sildenafil levels in children with PAH, Karatza et al. (22) reported mean levels of 109 ± 87 ng/ml 1 h after oral administration in 3 patients. These data are difficult to interpret for the following reasons: individual levels were not given, but there was clearly a wide range, given the standard error of the mean; a sildenafil level was not measured earlier than 1 h after administration; and not only was sildenafil administered orally, but also the exact mode of delivery (tablet/capsule/suspension) was not detailed.
Given these data, and with the aim to minimize the duration of general anesthesia in these often hemodynamically challenged patients, we chose to measure the effects and levels of sildenafil 30 min after administration as a suspension through nasogastric tube. However, our study demonstrates that in the clinical situation, the absorption of sildenafil 30 min after administration through nasogastric tube is quite unpredictable. Indeed, the plasma sildenafil level was undetectable in almost one-half the children, and ranged from 1.2 to 460 ng/ml in those in whom it could be measured. Furthermore, the levels of the N-desmethyl metabolite were also low, suggesting that rapid absorption and conversion by hepatic metabolism was not implicated. There are many other potential causes for these findings, including the use of general anesthesia, opiate analgesia, and transient neuromuscular blockade, all of which may all modify gastric emptying and thus absorption from the gastrointestinal tract. No matter what the cause, these data have important implications for the assessment of effect, and future protocols of use, of enterally administered sildenafil in children undergoing vasodilator testing.
Hemodynamic effect of oral sildenafil
In the absence of the pharmacokinetic data, the hemodynamic response to sildenafil in our total population would appear to be disappointing. However, our results show that when there was a detectable serum level, there was no statistically significant difference in reduction of elevated PVR when sildenafil was compared with inhaled NO. However, the number of patients is relatively small, and with a larger sample size, the trend toward superiority of NO might become significant. Nevertheless, this observation is in agreement with our previous report of the effects of intravenously administered sildenafil (6). Furthermore, in 3 of our patients, there was a response to sildenafil, despite a lack of response to 40-ppm NO. These findings may have important prognostic and therapeutic implications for the individual patient with pulmonary hypertension and are in agreement with similar observations in a previous report of sildenafil effects in adults with PAH (2). In this context, it is of note that when we tested the long-term effects of sildenafil in children with PAH, we found 9 of 14 patients had improved pulmonary hemodynamics at follow-up despite the absence of an acute response to inhaled NO during the primary vasoreactivity testing (7).
Perhaps more unexpected was the finding that a significant reduction in PVRI was observed in patients with an undetectable level of sildenafil. The reasons for this are unclear. While it is possible that there could be bias in data collection, as the operators were not blinded to treatments as in a randomized trial, this was minimized by the rigid time-based protocol that governed hemodynamic data collection, and the fact that changes in blood gas concentrations and oxygen consumption cannot be predicted by the operator and form important elements to the ultimate PVR calculation that was calculated off line. Furthermore, our study was not designed to be, and cannot be interpreted as, a dose-response study. However, this finding does suggest that the therapeutic range for sildenafil might be wider than previously thought, at least for some patients with PAH.
Levels of cGMP and responses
The major unexpected finding of our study was the lack of a significant change in cGMP levels both in the group with and in the group without measurable sildenafil levels, and largely irrespective of hemodynamic response. In patients with detectable levels of sildenafil, the fall in PVRI was similar to that observed with inhaled NO. There was a marked difference in cGMP responses, however. With inhaled NO, there was a highly significant rise in cGMP levels, whereas there was no change with sildenafil. While there was a loose, and statistically significant, relationship between both sildenafil and cGMP levels and the degree of response of PVRI, the levels of cGMP were markedly lower than that observed with inhaled NO in the same patients. This finding clearly suggests a difference in response or mechanism of action (in terms of cGMP effects) between sildenafil and NO. In a previous study of 13 adult patients (69% idiopathic pulmonary hypertension) Michelakis et al. (2) observed equal potency of oral sildenafil and inhaled NO despite a lower cGMP response to sildenafil (24 pmol/ml vs. 35 pmol/ml), although, interestingly, they were unable to show a correlation between cGMP level and reduction of PVR.
There are several possible explanations for these, and our, findings. First, given the nature of our protocol, it is possible that sildenafil was administered before the full effects of inhaled NO were dissipated. Although there was a period of 10 min between cessation of NO inhalation and sildenafil administration (ample time for the direct effect of NO to have been lost), it is possible that circulating cGMP levels remained raised. However, not only would it be highly unlikely that sufficient absorption of sildenafil could influence the further decline in cGMP levels early after withdrawal of NO, but also a further 30 min elapsed before the potential effects of sildenafil were then assessed. Given also that there was no difference in cGMP levels between the responders and nonresponders, this makes an interaction between NO and sildenafil as an explanation for fall in PVR seen with sildenafil in the absence of measurable levels unlikely, but it cannot be excluded. It is also possible that systemic levels of cGMP do not reflect intracellular cGMP levels within the vascular smooth muscle, given the direct and indirect (through phosphodiesterase type 5) effects of inhaled NO and sildenafil, respectively. Going along with this, and touched upon earlier, the therapeutic response to sildenafil may be different to that of inhaled NO, and may vary from person to person, in a less predictable way to that of NO. It is known for example that different disease states lead to different effects on the levels and function of intracellular phosphodiesterase type 5 itself (23). Finally, there may be a hitherto undescribed mechanism of action of sildenafil in PAH that is not associated with the classical NO-cGMP pathway. Our study was not designed to address these possibilities, and clearly more studies are needed to define the pharmacokinetics, dose response, and mechanisms of actions of sildenafil in these patients.
Our study demonstrates suboptimal absorption of sildenafil in almost half the children undergoing acute hemodynamic testing. When detectable, the effect of sildenafil on pulmonary vasodilation was not significantly different from that of inhaled NO despite marked differences in cGMP levels, implying differences in response or mechanism of action of the 2 therapies. Finally, response to sildenafil was also documented despite undetectable levels, suggesting the therapeutic range for some children might be wider than previously described for adults.
The study was supported by the Innovations Fund of the Labatt Family Centre, Toronto, Canada. Dr. Apitz was the recipient of a research scholarship of the Deutsche Herzstiftung eV, Frankfurt, Germany. Dr. Humpl is an advisor/consultant for Pfizer and Actelion. Drs. Humpl and Redington are joint senior authors.
- Abbreviations and Acronyms
- body surface area
- cyclic-guanosine monophosphate
- congenital heart disease
- fraction of inspired oxygen
- mean pulmonary arterial pressure
- nitric oxide
- pulmonary arterial hypertension
- pulmonary vascular resistance
- pulmonary vascular resistance index
- Wood units
- Received August 18, 2009.
- Revision received November 6, 2009.
- Accepted November 23, 2009.
- American College of Cardiology Foundation
- Michelakis E.,
- Tymchak W.,
- Lien D.,
- Webster L.,
- Hashimoto K.,
- Archer S.
- Sastry B.K.,
- Narasimhan C.,
- Reddy N.K.,
- Raju B.S.
- Humpl T.,
- Reyes J.T.,
- Holtby H.,
- Stephens D.,
- Adatia I.
- Haworth S.G.
- Barst R.J.,
- McGoon M.,
- Torbicki A.,
- et al.
- Galiè N.,
- Torbicki A.,
- Barst R.,
- et al.
- McLaughlin V.V.,
- Archer S.L.,
- Badesch D.B.,
- et al.
- Wilkens H.,
- Guth A.,
- Konig J.,
- et al.
- Wilkins M.R.,
- Wharton J.,
- Grimminger F.,
- Ghofrani H.A.