Author + information
- Received April 10, 1998
- Revision received September 18, 1998
- Accepted October 30, 1998
- Published online March 1, 1999.
- ↵*Reprint requests and correspondence: Dr. David L. Wessel, Cardiac ICU Office, Farley 653, Children’s Hospital, 300 Longwood Avenue, Boston, Massachusetts 02115
We compared the ability of inhaled nitric oxide (NO), oxygen (O2) and nitric oxide in oxygen (NO+O2) to identify reactive pulmonary vasculature in pulmonary hypertensive patients during acute vasodilator testing at cardiac catheterization.
In patients with pulmonary hypertension, decisions regarding suitability for corrective surgery, transplantation and assessment of long-term prognosis are based on results obtained during acute pulmonary vasodilator testing.
In group 1, 46 patients had hemodynamic measurements in room air (RA), 100% O2, return to RA and NO (80 parts per million [ppm] in RA). In group 2, 25 additional patients were studied in RA, 100% O2and 80 ppm NO in oxygen (NO+O2).
In group 1, O2decreased pulmonary vascular resistance (PVR) (mean ± SEM) from 17.2 ± 2.1 U·m2to 11.1 ± 1.5 U·m2(p < 0.05). Nitric oxide caused a comparable decrease from 17.8 ± 2.2 U·m2to 11.7 ± 1.7 U·m2(p < 0.05). In group 2, PVR decreased from 20.1 ± 2.6 U·m2to 14.3 ± 1.9 U·m2in O2(p < 0.05) and further to 10.5 ± 1.7 U·m2in NO+O2(p < 0.05). A response of 20% or more reduction in PVR was seen in 22/25 patients with NO+O2compared with 16/25 in O2alone (p = 0.01).
Inhaled NO and O2produced a similar degree of selective pulmonary vasodilation. Our data suggest that combination testing with NO+O2provides additional pulmonary vasodilation in patients with a reactive pulmonary vascular bed in a selective, safe and expeditious fashion during cardiac catheterization. The combination of NO+O2identifies patients with significant pulmonary vasoreactivity who might not be recognized if O2or NO were used separately.
Elevated pulmonary vascular resistance (PVR) complicates the evaluation, clinical course and outcome of patients with congenital heart disease or end-stage pulmonary disease. It is a crucial factor in determining the timing or type of intervention, and has been invoked as the primary determinant of mortality in many lesions (1,2). Opinion varies on what resistance must be achieved with vasodilator testing to insure safe operability for children with congenital heart disease. An increased or fixed elevation in PVR may deny patients the chance of corrective surgery, leaving them susceptible to the development of progressive obliterative pulmonary vascular disease and reduced life expectancy (3,4). Demonstration of pulmonary vasoreactivity in patients with end-stage pulmonary disease may differentiate patients who would benefit from long-term medical therapy (5,6)from those with high, fixed resistance who should be more urgently considered for lung transplantation (3). Safe and expeditious demonstration of maximal pulmonary vasodilation in patients with a reactive pulmonary bed is therefore an important objective.
Many vasodilators have been utilized for diagnostic testing during cardiac catheterization. Systemic vasodilators with their attendant risks of hypotension and increased intrapulmonary shunt may be hazardous (7), especially in patients with ventricular outflow tract obstruction. Breathing oxygen (O2) remains a standard means of pulmonary vasodilator testing in pediatric cardiac catheterization laboratories (8,9). However, failure to respond to acute treatment with O2has been reported in some patients who did indeed have reactive pulmonary vasculature (3,10).
Inhaled nitric oxide (NO) is a selective pulmonary vasodilator with minimal systemic effects and does not increase intrapulmonary shunting. It can be administered easily with O2or room air (RA) during cardiac catheterization by either ventilator or mask. The purpose of this study was to compare the ability of NO and O2to identify patients with a reactive pulmonary vascular bed during cardiac catheterization. We further compared the hemodynamic effects of breathing nitric oxide in oxygen (NO+O2) to breathing O2alone during acute vasodilator testing.
We enrolled patients between January 1992 and December 1996 who had mean pulmonary artery pressure ≥30 mm Hg, PVR >3 U·m2, and were determined during catheterization to require vasodilator testing. We included for analysis 71 patients who had complete hemodynamic measurements to allow calculation of vascular resistances.
The first 46 patients (group 1) were studied under the following four study conditions: A) in RA; B) after breathing 100% O2for 15 min; C) after another 15 min in RA, and D) after 15 min breathing NO at 80 parts per million (ppm) in 23% O2(NO+RA). As NO is titrated into a delivery circuit, the delivered fraction of inspired O2(FiO2) is decreased. Therefore a small amount of supplemental O2(23%) was added to NO to avoid administration of a hypoxic gas mixture. The patients in group 1 had a median age of 6.5 years, range 4 months to 59 years.
Twenty-five additional patients (group 2) were studied in the following three conditions: A) in RA; B) after breathing 100% O2for 15 min, and C) after 15 min of inhaling 80 ppm NO in 91% O2. This was the maximal FiO2attainable after dilution with 80 ppm of NO. Patients in group 2 had a median age of 3.5 years, range 5 months to 69 years.
The patients in each group represented a broad spectrum of diagnoses characteristic of a high volume pediatric cardiac catheterization laboratory. Most had unrepaired or previously palliated congenital heart disease, although some had end-stage pulmonary disease (Tables 1 and 2). ⇓Four of 46 patients in group 1 and three of 25 patients in group 2 were mechanically ventilated. The remainder in each group were breathing spontaneously. Sedation was given according to a routine, which included intravenous morphine and midazolam. Partial pressure of carbon dioxide (Pco2) was normal throughout the study in both groups.
Hemodynamic measurements included left atrial, right atrial, pulmonary and systemic arterial pressures during each of the conditions described above. Cardiac output was measured by thermodilution in patients without an intracardiac shunt. In those with shunts, O2consumption was measured (Waters Inc., model MM20, Rochester, Minnesota), and systemic and pulmonary blood flows were calculated using the Fick equation with inclusion of dissolved O2. Errors related to sampling site variances were minimized by ensuring that, for each patient, venous samples were collected at the same site during each of the study conditions.
Delivery and monitoring of NO
Detailed descriptions of the technical aspects of our delivery of NO in both ventilated and spontaneously breathing patients have been published previously (11,12). We used NO gas (Scott Specialty Gases, Plumsteadville, Pennsylvania or BOC Gases, Murray Hill, New Jersey) of medical grade quality, which conformed to Food and Drug Administration standards. In the spontaneously breathing individuals NO was delivered using the titration technique from source tanks with an 800-ppm concentration. Flow rates greater than the patients’ minute volumes were delivered through a one-way inspiratory valve to a face mask. The expired gases were scavenged using a reservoir bag and regulated wall suction. In the seven patients who were mechanically ventilated, ventilator settings were kept constant throughout the study. Nitric oxide, nitrogen dioxide (NO2) and FiO2were continuously monitored from a sampling port at the airway (Thermoenvironmental Instruments Chemiluminescence model 42H, Franklin, Massachusetts or NOxBOX Electrochemical Inhaled NO Therapy Monitor, Bedfont Scientific USA, Medford, New Jersey). Peak measured NO2concentrations were recorded in all patients during delivery of the drug. Because there were no reports of methemoglobinemia during 15-min diagnostic trials of NO at 80 ppm (11), we eventually ceased to routinely measure methemoglobin levels during brief inhalations. Therefore, methemoglobin levels were obtained by cooximetry (CIBA-Corning model 2500, Medfield, Massachusetts) after 15 min in the first 22 of 46 patients in group 1 and not thereafter. Written informed consent was obtained from the patients or their parents under a protocol approved by the Clinical Investigation Committee of Children’s Hospital and submitted to the Food and Drug Administration.
Statistical analysis and calculations
Results are presented as mean values ± SEM. Vascular resistances were calculated using standard equations and were expressed in Wood units corrected for body surface area (U·m2). Groups 1 and 2 were analyzed separately with patients in each group acting as their own controls. Repeated measures analysis of variance was used to look for differences in the measurements over the four study conditions in group 1 and the three conditions in group 2. If differences were found, then the Bonferroni multiple comparisons procedure was used to determine where differences existed. A p value <0.05 was considered significant.
Group 1: comparison of RA, O2, RA and NO+RA
Pulmonary vascular resistance differed across the four conditions, (p < 0.0001) (Table 3). Oxygen decreased PVR from 17.2 ± 2.1 U·m2to 11.1 ± 1.5 U·m2(p < 0.05). Administration of inhaled NO at 80 ppm in RA caused a comparable decrease from 17.8 ± 2.2 U·m2to 11.7 ± 1.7 U·m2(p < 0.05) (Fig. 1). Comparison of the mean percentage decreases from RA to O2(36.9 ± 3.3%) and RA to NO+RA (35.1 ± 3.5%) revealed no difference by paired ttest. Changes in pulmonary artery pressures may not reflect pulmonary vasodilation, because there were patients with intracardiac shunts. Nevertheless, the mean pulmonary artery pressure was significantly lower in both O2and NO+RA compared with RA despite increases in pulmonary blood flow in 21 of 23 patients with intracardiac shunts during treatment.
Mean systemic arterial pressure, systemic vascular resistance, right atrial pressure, left atrial pressure, heart rate, pH and Pco2did not change with administration of O2or NO. Arterial Pco2(Pao2) increased from 66 ± 3 mm Hg in RA to 278 ± 23 mm Hg with 100% O2; however, there was no significant difference in Pao2between RA and NO+RA (68 ± 4 vs. 73 ± 4 mm Hg).
Using a reduction in PVR of 20% or more as a marker for responsiveness, we compared individual patient results to O2and NO+RA (Fig. 2). Oxygen caused a positive response in 36/46 patients. Of the 10 nonresponders, four responded with a 20% or more decrease to NO. Nitric oxide in RA caused a positive response in 32/46. Of the 14 nonresponders to NO+RA, eight responded to O2. Six patients did not respond to either vasodilator (Table 1).
The peak NO2level was recorded in all 46 patients and was 1.3 ± 0.2 ppm. Methemoglobin measured at the conclusion of the 15-min period of NO inhalation in 22/46 patients was 0.8 ± 0.1%.
Group 2: comparison of RA, O2and NO+O2
Pulmonary vascular resistance differed across the three conditions (p < 0.0001) (Table 4). Pulmonary vascular resistance decreased from 20.1 ± 2.6 U·m2in RA to 14.3 ± 1.9 U·m2in O2and further to 10.5 ± 1.7 U·m2in NO+O2(Fig. 3). Pulmonary vascular resistance was significantly lower in O2compared to RA (p < 0.05). The PVR with combination therapy was statistically lower than that measured in air or in O2(p < 0.05 for both). Oxygen caused a reduction in PVR from baseline of 20% or more in 16 of 25 patients. Of the remaining nine patients, six responded with a 20% or more decrease when inhaling NO+O2. Three patients did not have a positive response to either O2or the combination of O2and NO (see Table 2). Using the McNemar’s test, if responsiveness differed between the two treatments, patients were more likely to respond to NO+O2than to O2alone (p = 0.01). Ten of the 25 patients underwent complete surgical repair of their cardiac defects within 1 month of their catheterization. Those 10 patients had a baseline PVR in RA of 12.9 ± 1.9 U·m2that decreased to 7.1 ± 1.9 U·m2in O2and to 4.1 ± 1.9 U·m2in NO+O2. Each patient undergoing surgical repair had a baseline PVR in RA >6 U·m2. Five of 10 had PVR >6 U·m2in O2, but only one had PVR >6 U·m2in NO+O2. All 10 patients survived and were discharged home on median postoperative day 5.5, range 4 to 29.
Mean pulmonary artery pressure decreased from 63.4 ± 3.7 mm Hg in RA to 57.7 ± 3.5 mm Hg in O2to 50.6 ± 3.5 mm Hg in NO+O2. This occurred despite an increase in pulmonary blood flow in 13 of 15 patients with shunts. The pulmonary artery pressure with NO+O2was significantly lower than that measured in air or in O2(p < 0.05 for both). There was no difference between left atrial pressure in RA and in O2. Left atrial pressure was significantly higher in NO+O2(15.5 ± 1.3 mm Hg) than in RA (13.1 ± 1.1 mm Hg) or O2(12.7 ± 1.0 mm Hg).
Mean systemic blood pressure increased over the three study conditions, from 76.8 ± 2.9 mm Hg in RA to 80.6 ± 2.5 mm Hg in O2to 83.4 ± 2.6 mm Hg in NO+O2. Blood pressure was significantly higher in NO+O2than in RA, but not statistically different from blood pressure in O2. Cardiac index, systemic vascular resistance, right atrial pressure, heart rate, pH and Pco2did not change. Arterial partial pressure of O2was significantly higher both in O2and in NO+O2as compared with RA. Arterial partial pressure of O2was not different in NO+O2compared with O2(302.8 ± 27.9 vs. 277.6 ± 30.4 mm Hg). The peak NO2level during NO delivery was 2.3 ± 0.3 ppm.
We compared the inhaled vasodilators O2and NO in 71 patients during acute vasodilator testing at cardiac catheterization. In 46 patients, 100% O2and inhaled NO at 80 ppm in air produced comparable and selective decreases in mean pulmonary artery pressure and PVR. However, O2or NO used separately failed to identify all patients with a significant capacity for pulmonary vasodilation. The combination of NO (80 ppm) with 91% O2in an additional group of 25 patients produced significantly more pulmonary vasorelaxation compared with O2used alone. In 22/25 patients there was a positive pulmonary vasodilator response during combination therapy compared to only 16/25 when breathing O2alone. None of the 71 patients studied showed any evidence of toxicity from either drug during the brief period of this diagnostic trial. Our data suggest that combination testing with NO in O2provides additional pulmonary vasodilation, can be safely and accurately delivered to patients during diagnostic cardiac catheterization and can rapidly identify patients with pulmonary vasoreactivity. The combination of agents appears to identify patients with significant pulmonary vasoreactivity who might not be recognized if O2or NO were used separately.
Importance of vasodilator testing
The precise stage when pulmonary vascular disease has progressed to a point where surgical repair of congenital heart lesions cannot be safely performed is unknown. Morphologic criteria (2)and pulmonary hemodynamics (13)are useful, but imprecise. Pulmonary vascular resistance calculated to be more than 6 to 8 U·m2has been shown to be associated with poor operative outcome regardless of lung histology (1,4,14). In contrast, patients who respond to vasodilators with a PVR less than 6 to 8 U·m2do well postoperatively (13). Demonstration of a reactive pulmonary bed in patients being evaluated for transplantation has enabled patients to be offered a single organ heart instead of heart–lung block with successful results (15). Patients with elevated resistance but reactive pulmonary vasculature may need more intensive postoperative care and presumably would be excellent candidates for NO therapy in the postoperative period should pulmonary hypertension emerge. Response to acute vasodilator testing in patients with primary pulmonary hypertension is an important marker for survival (3)and may identify patients who would benefit from chronic medical therapy (5,6).
Comparison with other studies
Prior research in children with pulmonary hypertension has shown that O2failed to unmask all reversible pulmonary vasoconstriction (3,10). Prostacyclin administration in patients with pulmonary hypertension breathing O2caused further pulmonary vasodilation. However, prostacyclin can cause systemic side effects including tachycardia and hypotension (16). Previous studies of vasoreactivity in children during cardiac catheterization found variable responsiveness to NO that seemed to parallel the progression of established vascular disease (17). Studies examining the efficacy of NO in O2, including recent work by Allman and colleagues, have suggested differences between the responses to NO, O2and/or the combination of agents (18–20). Each prior study, however, has had insufficient power to establish a significant difference in PVR.
Nitric oxide causes vasorelaxation through a cyclic guanosine monophosphate–mediated pathway. The mechanism of vasorelaxation caused by O2is not clearly known (21). The fact that some patients responded to one agent with significant vasodilation but not the other, and that the majority of patients experienced increased vasodilation with combination therapy compared with O2alone, suggests that the mechanisms may not be identical.
The major recognized toxicities associated with inhaling NO are cytotoxic effects in the lung due to exposure to excess NO2and methemoglobinemia due to the intravascular binding to hemoglobin. Nitrogen dioxide will develop in delivery systems at a rate that is proportional to NO and O2concentrations and contact times between the two gases. When NO was delivered with maximal amounts of O2in this study, NO2levels averaged 2.3 ± 0.3 ppm, below the accepted environmental exposure level of 5 ppm (22). Nitrogen dioxide should be continuously monitored, especially in patients mechanically ventilated with circuits that do not use continuous gas flows. If patients receive prolonged treatment with NO in high concentrations of O2, we recommend reduction in the NO dose to diminish potential dose-related toxicity. There have been no reports of clinically significant methemoglobinemia during brief exposure to NO at doses as high as 80 ppm. Methemoglobin measured at the conclusion of the 15-min period of NO inhalation in 22/46 patients was 0.8 ± 0.1%. This along with previously published results (11)supports the contention that routine measurement of methemoglobin may be unnecessary during brief diagnostic trials of NO.
It is notable that the combination of NO in O2resulted in an increase in left atrial pressure compared with RA or O2alone. Reports have suggested that O2(23)or NO (24)may have deleterious effects to patients with heart failure. In this study no patient demonstrated clinically important pulmonary edema, hemodynamically significant systemic vasoconstriction or decreased cardiac index during the brief administration of O2or NO in O2. Nevertheless, we believe that NO, especially when used with O2, should be carefully monitored in patients with elevated left atrial pressures due to the potential induction of pulmonary edema.
The patient population studied was quite heterogeneous. However, this accurately reflects the typical spectrum of patients presenting for vasodilator testing during cardiac catheterization. Subgroup analysis of patients with congenital heart disease showed no differences in response compared to the group as a whole. Patients with lung pathology analyzed separately showed similar results, but numbers were too small to form conclusions. Subgroup analysis of patients with left to right shunts did not reveal any difference in response compared to those without shunts. The definition of responder and nonresponder is arbitrary, but a 20% change is often used in drug testing as a marker of responsiveness. There was no apparent predictive marker in patients who responded to one agent but not the other. It may be that repeated exposure to O2or NO would minimize differences between responders and nonresponders. Nonetheless, a single exposure to a drug is the common catheterization protocol. Limited information exists concerning optimal dosing of NO, with some investigators showing maximal vasodilation at doses as low as 2 ppm (19), and others demonstrating a dose–response relationship up to 80 ppm in a similar population (18). This study was designed as a brief diagnostic trial in a catheterization laboratory to determine the most effective and inclusive method of identifying patients with pulmonary vasoreactivity. Accordingly, 80 ppm was used during this brief testing with the appreciation that, if delivered for prolonged periods, it may be associated with dose-related increased toxicity. This study was not designed to demonstrate differences in long-term patient outcomes or clinical value of vasodilator testing. Maximal vasodilatory capacity may be of limited clinical value in some patients. Nevertheless, as a result of information acquired during combination therapy, some patients were offered surgery who did not respond to NO or O2alone; all patients survived.
Individually, NO and O2produced significant and comparable selective pulmonary vasodilation in a heterogeneous group of patients presenting to cardiac catheterization for pulmonary vasodilator testing. However, neither agent used separately identified all patients with the capacity to relax their pulmonary vascular bed. The combination of NO+O2caused significantly greater pulmonary vasodilation compared to O2and identified patients who had pulmonary vasoreactivity that was not appreciated during O2breathing alone. This study suggests that combination testing with NO+O2provides additional pulmonary vasodilation in patients with a reactive pulmonary vascular bed in a specific, safe and expeditious fashion during cardiac catheterization. Nitric oxide in O2distinguishes patients with significant pulmonary vasoreactivity who might not be identified using either agent separately.
We are grateful to John F. Keane, MD for his critical review of the manuscript and to Kimberlee Gauvreau, ScD for her expert assistance with statistical analysis.
☆ Dr. Atz is supported by an award from the National Institutes of Child Health and Human Development and a grant from the United States Food and Drug Administration. Dr. Wessel is supported by grants from the United States Food and Drug Administration and the Research Endowment of Children’s Hospital.
Presented in part at the 45th Annual Scientific Sessions of the American College of Cardiology, Orlando, Florida, March 1996.
- fraction of inspired oxygen
- nitric oxide
- nitrogen dioxide
- partial pressure of carbon dioxide, arterial
- partial pressure of carbon dioxide
- parts per million
- pulmonary vascular resistance
- room air
- Received April 10, 1998.
- Revision received September 18, 1998.
- Accepted October 30, 1998.
- American College of Cardiology
- Hoffman J.I.,
- Rudolph A.M.,
- Heymann M.A.
- Rabinovitch M.,
- Haworth S.G.,
- Castaneda A.R.,
- Nadas A.S.,
- Reid L.M.
- Houde C.,
- Bohn D.J.,
- Freedom R.M.,
- Rabinovitch M.
- Bush A.,
- Busst C.M.,
- Haworth S.G.,
- et al.
- Raffy O.,
- Azarian R.,
- Brenot F.,
- et al.
- Marshall H.W.,
- Swan H.J.C.,
- Burchell H.B.,
- Wood E.H.
- Betit P.,
- Adatia I.,
- Benjamin P.,
- Thompson J.E.,
- Wessel D.L.
- Adatia I.,
- Perry S.,
- Landzberg M.,
- Moore P.,
- Thompson J.E.,
- Wessel D.L.
- Bush A.,
- Busst C.,
- Booth K.,
- Knight W.B.,
- Shinebourne E.A.
- Roberts J.D.,
- Lang P.,
- Bigatello L.M.,
- Vlahakes G.J.,
- Zapol W.M.
- Cornfield D.N.,
- Reeve H.L.,
- Tolarova S.,
- Weir E.K.,
- Archer S.
- ↵(1988) NIOSH recommendations for occupational safety and health standards. MMWR 37:21.
- Haque W.A.,
- Boehmer J.,
- Clemson B.S.,
- Leuenberger U.A.,
- Silber D.H.,
- Sinoway L.I.
- Semigran M.J.,
- Cockrill B.A.,
- Kacmarek R.,
- et al.