Since the first pioneering observations on selective pulmonary vasodilatory effects of inhaled nitric oxide (iNO) (1,2), its use as a therapeutic strategy, in adult and pediatric cardiac surgical patients with pulmonary hypertension (PH) and impending right ventricular (RV) failure, has been expanding rapidly. In contrast to common IV vasodilators, such as prostaglandins, sodium nitroprusside, and nitroglycerin (NTG; Perlinganit®; Schwarz Pharma, Mannheim, Germany), which lack specificity to the pulmonary vasculature, iNO is not associated with systemic vasodilation, and its use thus reduces the risk of arterial hypotension and RV ischemia. Despite the development of safe techniques for iNO administration and special devices for iNO and nitrogen dioxide (NO2) monitoring, potential iNO toxicity remains an issue of concern (3–5), as reflected by a continuing search for alternative therapeutic modalities. The main concerns regarding iNO toxicity relate to possible direct pulmonary cytotoxicity, formation of NO2 (3–5), and (to a lesser extent) formation of methemoglobin (MetHb) in patients with inadequate MetHb reductase activity (3). The alternative use of inhaled prostacyclin (PGI2) (6) or inhaled prostaglandin E1 (PGE1; Prostin VR®; Pharmacia & Upjohn, Puurs, Belgium) (7) has been proposed, but local toxicity of commercial preparations of PGI2 (glycine buffer, pH 10.5) and PGE1 (ethanol-saline) is unknown. IV adenosine has been shown to produce significant pulmonary vasodilation, an increase in cardiac output (CO), and no reduction of arterial pressure (8), but it is unclear whether the observed decrease in systemic vascular resistance (SVR) was secondary to the increase in CO (indicating absence of systemic effects) or vice versa.
The unresolved issue of possible iNO toxicity stresses the importance of avoiding indiscriminate application and restricting use to patients with a positive benefit/risk ratio (3). In adult patients after cardiac surgery, studies comparing iNO with IV vasodilators are scarce. One report (9) suggests that the use of IV PGI2 in patients with moderate PH after heart transplantation might be superior to iNO, with regard to myocardial performance and organ perfusion. It was our aim to determine in a prospective, randomized crossover study whether iNO is superior to IV PGE1 and NTG, with respect to systemic and pulmonary hemodynamics, oxygen variables, and RV performance in adults with severe PH after cardiac surgery.
With ethical committee approval, written, informed consent was obtained from patients with severe PH as documented during preoperative cardiac catheterization. Definitive selection was based on persistence of PH after cardiac surgery, defined as mean pulmonary artery pressure (MPAP) ≥30 mm Hg and/or pulmonary vascular resistance (PVR) ≥300 dyne · sec · cm−5. An additional prerequisite for study entry was stable postoperative circulatory conditions to ensure steady state, allow repetitive withdrawal of vasodilator substance for reassessment of baseline values, and to avoid the necessity for hemodynamic interventions, such as volume substitution or a change in inotropic support. Patients with mechanical circulatory assistance or with echocardiographic evidence of significant pulmonary or tricuspid valve regurgitation were excluded.
Intraoperatively, IV PGE1 0.1 to 0.2 μg · kg−1 · min−1 was infused to facilitate weaning from cardiopulmonary bypass. The type and amount of inotropic support is shown in Table 1.
The investigation was performed in the intensive care unit within the first 24 h after surgery. PGE1 was discontinued at least 60 min before initial baseline measurements. Patient sedation consisted of either propofol and morphine (11 patients) or midazolam, combined either with alfentanil (2 patients) or fentanyl (1 patient), at constant infusion rates. The patients' tracheas were orally intubated, and their lungs were ventilated with oxygen-enriched air, using a Servo 900C ventilator (Siemens-Elema AB, Solna, Sweden). Ventilatory variables were kept constant, and no hemodynamic intervention, other than application and withdrawal of iNO, PGE1, and NTG, was initiated during the study interval.
Continuous patient monitoring included pulse oximetry, electrocardiography (leads II and V5), systemic arterial pressure through a 4F femoral artery catheter, central venous (CVP), and pulmonary artery pressures (PAP) through a triple lumen 7.5F pulmonary artery catheter (PAC), inserted via the right internal jugular vein. Patients with sinus rhythm were preoperatively assigned to a PAC, enabling determination of RV ejection fraction (RVEF) (REVOX™ or REF™; Baxter Healthcare Corp., Irvine, CA), and a corresponding CO computer (Explorer™; Baxter Healthcare Corp.). Patients with chronic atrial fibrillation were assigned to an Intellicath™ (Baxter Healthcare Corp.) or Opticath PAC (Abbott Laboratories, North Chicago, IL), and a Vigilance™ (Baxter Healthcare Corp.) or an Oximetrix R3 SO2/CO system (Abbott Critical Care Systems). CO was estimated using the average of 3 to 5 iced saline injections (all within ±12%), randomly distributed throughout the respiratory cycle. Ten patients with postoperative sinus rhythm or atrial pacing qualified for additional RVEF measurements. The system determining RVEF uses rapid response thermistors for beat-to-beat analysis of the changes in indicator concentration, thus measuring RV systolic and diastolic volumes and calculating RVEF; intracardiac electrocardiography is provided by a ventricular and a pulmonary artery electrode.
iNO 40 ppm, PGE1 0.1 μg · kg−1 · min−1, and NTG, 3 to 5 μg · kg−1 · min−1, were administered in randomized sequence. Measurements were performed after 15 to 20 min of constant vasodilator drug administration. A minimal period of 20 min was allowed for return of variables to baseline after the vasodilator was stopped, and these values represented the individual baseline before the next randomly applied vasodilator.
All measurements were performed in duplicate to confirm steady state and then averaged for statistical comparisons (see below). After study completion, PGE1 administration was continued, and routine care was provided according to the attending physician's discretion.
Nitric oxide (NO) was delivered from a tank containing a concentration (conc.) of 1000 ppm NO diluted in nitrogen (AGA, Pratteln, Switzerland) via a calibrated flowmeter and a nebulizer circuit (Servo Nebulizer 945; Siemens-Elema AB, Solna, Sweden), synchronized with the inspiratory period of the ventilator. NO was delivered into the inspiratory limb of the circle breathing system. Inspired gas was collected close to the endotracheal tube, at 30 cm from the NO inlet to ensure adequate gas mixing. Peak inspired NO and NO2 conc. were continuously analyzed by chemiluminescence (CDL 700AL, Eco Physics AG, Dürnten, Switzerland), and exhaust gas was scavenged via a soda-lime absorber. MetHb conc. (as percentage of total hemoglobin) was determined in duplicate at the end of vasodilator drug exposure, using multiwave length spectrophotometry (IL 482™ CO-Oximeter System or IL 1400™ BG Electrolytes Analyzer; Instrumentation Laboratory, Inc., Lexington, MA).
The following variables were measured: heart rate (HR), mean systemic arterial pressure (MAP), MPAP, CVP, and pulmonary capillary wedge (PCWP) pressure, CO, RVEF (if applicable), arterial and mixed venous pH, partial pressures of carbon dioxide and oxygen, arterial and mixed venous oxygen saturations, hemoglobin, and MetHb. From these measurements, the following variables were calculated using standard formulas: transpulmonary pressure gradient (TPG = MPAP − PCWP), PVR, SVR, PVR/SVR ratio, right coronary artery perfusion pressure (PRCA = MAP − CVP), cardiac index (CI), oxygen delivery index (DO2I), oxygen consumption index (VO2I), oxygen extraction ratio (O2 ext), intrapulmonary shunt fraction (Qs/Qt), and the ratio of arterial partial pressure of oxygen to inspired oxygen fraction (PaO2/FiO2).
Values are presented as median and percentile. Comparison of treatments were performed using the Friedman test, followed by pairwise Wilcoxon's signed rank tests with Bonferroni correction.
For the Friedman test, a P value < 0.05 was considered significant. For the post hoc comparisons (by the Wilcoxon's signed rank test), P values were multiplied by 3 and considered significant if P < 0.05 after multiplication (Bonferroni correction).
Seventeen patients were initially assigned to the study, but three were excluded because MPAP and/or PVR decreased after surgery, below the limits qualifying for study entry. The remaining 14 patients completed the study protocol. Their characteristics and outcome are shown in Table 1, and the results of hemodynamics, oxygen, and gas exchange variables in Tables 2 and 3, and Figures 1–3. In four patients, the effects induced by iNO, PGE1, and NTG were not fully reversible after drug withdrawal, despite waiting periods of up to 60 min; this, however, does not invalidate the results, because drug sequence was randomized, individual baseline values (i.e., baseline values before subsequent drug exposure) were not significantly different in the study group for any variable, and only changes to these individual baselines were analyzed.
Inhaled NO, PGE1, and NTG were effective in reducing PVR and TPG (P = 0.003, Table 2, Figure 1), and all three vasodilators induced significant reductions of MPAP. iNO induced no significant change in MAP or SVR (Table 2, Figure 1), while PGE1 and NTG led to a similar decrease in MAP (P = 0.003) and a reduction in SVR (P ≤ 0.005). The median PVR/SVR ratio decreased with iNO (P = 0.003), indicating selectivity to the pulmonary vasculature, whereas it was unchanged with PGE1 and NTG. The individual baseline PVR/SVR ratio correlated well with the pulmonary vasodilator response (Figure 2). PCWP was unaltered with iNO, but decreased with both PGE1 (P = 0.01) and NTG (P = 0.003), while CVP was significantly reduced by all three vasodilators (P < 0.015). The CVP decrease was more pronounced with NTG than with PGE1 or iNO (P = 0.003). PRCA did not change with iNO, but significant and similar reductions of PRCA were observed with NTG and PGE1 (P = 0.003).
iNO and PGE1 led (at a constant HR) to a significant increase in CI (P = 0.012 and 0.006, respectively) and Stroke Volume Index (P ≤ 0.005). NTG induced no significant change in CI or HR, RVEF increased with PGE1 (P = 0.015) and was unaltered with iNO and NTG. No significant differences were found when the degree of change of CI and RVEF from baseline was compared between vasodilators.
Serious hypotension occurred with NTG in three patients and with PGE1 in two patients.
Intrapulmonary shunt fraction (Qs/Qt) and PaO2/FiO2 ratio did not change significantly with iNO. In contrast, PGE1 and NTG induced an increase in Qs/Qt (P = 0.006 and 0.014, respectively) and a decrease in PaO2/FiO2 from baseline (P < 0.005), but the changes were not significant when compared with iNO.
With iNO, in contrast to PGE1 and NTG, DO2I increased (P = 0.039), while PGE1, in contrast to iNO and NTG, led to a significant reduction in O2ext (P = 0.045) (Table 3, Figure 3). No significant differences were found when the degree of change of DO2I and O2ext was compared between vasodilators, including NTG.
With iNO, NO2 levels of 2.4 (1.8, 4.2) ppm were detected at FiO2 values of 0.35 to 0.70. In Patient 14, NO2 amounted to 6.4 ppm and, thus, was above the value of 5 ppm acceptable during short iNO exposure (3). Median MetHb levels significantly increased from 0.64% to 1.06% with iNO (P = 0.003) and the maximal value measured was 1.55%. With NTG and with PGE1, MetHb was significantly lower than with iNO (P < 0.05), and the change from baseline was not significant.
In-hospital outcome was favorable in all patients (Table 1), and all were discharged in good condition. Long-term followup revealed 12 survivors; Patient 10 underwent successful bilateral lung transplantation 2 yr after hospital discharge. Two patients died; Patient 3 died after 19 mo because of gastric cancer, and Patient 8 died after 3 mo of an unknown reason.
The main findings of this study are that 1) iNO, PGE1, and NTG were of similar efficacy in reducing PVR, and all patients responded to vasodilator therapy; 2) only iNO led to selective pulmonary vasodilation; 3) iNO was not superior to PGE1 with regard to CI and RV performance; and 4) after study completion, PGE1 administration was continued with favorable in-hospital outcomes.
Most studies comparing the effects of iNO versus IV vasodilators in adults were performed before cardiac transplantation (10,11) or in primary pulmonary hypertension (2,12). With regard to adult cardiac surgery, we are aware of only one study, by Kieler-Jensen et al. (9), who evaluated in two groups of patients with moderately elevated PVR after heart transplantation the effects of 5, 10, and 20 ppm iNO and of IV PGI2, PGE1, and sodium nitroprusside. PGI2 was considered to be the ideal IV drug, because it induced the most pronounced reductions in both RV and LV outflow impedance, the least effect on venous capacitance vessels, and a significant improvement in systemic perfusion. iNO was recommended if heart transplantation was complicated by RV failure and systemic hypotension. We studied a patient population with considerably higher PVR and lower RVEF and included all vasodilators (iNO, PGE1, and NTG) in the crossover study design. PGE1 and NTG were chosen as IV vasodilators because of the documented effectiveness of these drugs to treat PH (13–15).
Our study supports previous findings (9) that PGE1 lacks selectivity on the pulmonary vasculature, despite a pronounced first-pass elimination in the lungs (16). PGE1 and NTG induced significant reductions in outflow impedances of both ventricles and a decrease in right and left cardiac filling pressures. In contrast to previous observations in animals with vasoconstrictor-induced (13) and microembolic PH (14), neither drug induced a significant change in PVR/SVR ratio, indicating no difference in pulmonary specificity.
With PGE1, we observed an increase in CO, probably induced by combined reductions of RV and LV outflow impedance and an increase in RVEF, despite a decrease in right coronary perfusion pressure. This is in contrast to the report by Kieler-Jensen et al. (9), who found no increase in CO or RVEF with PGE1. The higher RV outflow impedance and lower RVEF present in our patient population might explain the different observations, although none of our patients presented with RV failure, as evidenced by normal RV filling pressures and normal early postoperative values for CO, DO2I, and O2ext. With NTG, the profound dilating effect on venous capacitance vessels possibly prevented systemic perfusion and myocardial performance improvement. The systemic arterial and venous vasodilatory actions of PGE1 and NTG led to serious hypotension in three patients while on NTG and in two patients while on PGE1. It is reasonable to assume that hypotension could have, at least in part, been corrected by fluid loading, but this was not in accordance with the study protocol.
The effect of iNO on CO varies in the literature. Whereas some investigators found an increase in CO (9), others failed to demonstrate an improvement in systemic perfusion (17,18). In our study, iNO induced a similar rise in CO to PGE1. A significant iNO-induced rise in CO is expected, particularly when RV dysfunction secondary to elevated PVR is present. Two case reports (19,20) demonstrating a dramatic improvement in hepatic (19) and systemic (20) perfusion when iNO was applied to patients with severe RV failure, support this hypothesis.
In contrast to the present and other studies performed in adults after cardiac surgery without significant LV or biventricular dysfunction (9,18), several reports of iNO, given to heart transplant candidates with PH and associated chronic left or biventricular failure, showed a decrease in calculated PVR primarily because of an increase in PCWP (10,11,21). In the presence of severe LV failure, iNO led to the occurrence of pulmonary edema (22). It is currently unclear whether volume shifts caused by selective pulmonary vasodilation, leading to an increase in pulmonary venous return to the failing LV (21,23) or a NO-induced attenuation of the positive inotropic response to β-adrenergic stimulation in humans with LV dysfunction (24), are responsible for the rise in LV filling pressure, but recent studies suggest that a negative inotropic action of iNO is unlikely (23,25). In contrast to iNO, IV drugs, inducing both pulmonary and systemic vasodilation, led to a decrease in PCWP and an improvement of systemic perfusion, in patients with PH and chronic biventricular failure referred to heart transplantation (11,12). These observations suggest that IV vasodilators may be superior in the presence of concomitant LV dysfunction and that caution might be advisable when administering iNO under these conditions.
Our data confirm the findings of Pepke-Zaba et al. (2) that patients who respond to iNO also respond to IV vasodilators and vice versa. If, however, irreversible structural anomalies predominate in patients with primary PH (2,12) or in pulmonary embolization, the capacity to reduce PVR is expected to be only minor or absent. Fullerton et al. (17) reported a lack of response to iNO early after cardiac surgery in patients with high PVR caused by aortic and/or mitral valve disease. In our study, all patients responded to all three vasodilators, but in the two patients with chronic pulmonary embolization (Patients 9 and 10), PVR/SVR ratio increased from 0.55 to 0.66 with PGE1 and from 0.48 to 0.61 with NTG (mean values), indicating more pronounced effects on left than RV impedance.
IV vasodilators may induce hypoxemia by dilating poorly ventilated areas of the lung (26). In contrast, iNO has been shown to improve ventilation-perfusion distributions (27). In our patients, Qs/Qt and PaO2/FiO2 remained unaltered with iNO, and the changes of Qs/Qt and PaO2/FiO2 observed with PGE1 and NTG were not significant compared with iNO. Qs/Qt in our patient population was, however, lower than in patients after correction of congenital heart disease (27) and after heart or lung transplantation (27).
Selective pulmonary vasodilation represents a significant advancement in managing pulmonary hypertension and consequent RV failure after cardiac surgery, but iNO is a potentially toxic gas (3). The United States Occupational Safety and Health administration has set the acceptable time-weighted average iNO value over eight hours at 25 ppm (3), and the environmental limit has been set at 5 ppm for NO and NO2 (3). In our study, patients were exposed to 40 ppm iNO for short terms, because previous investigations in adults after cardiac surgery demonstrated effectiveness and ≤2 ppm NO2 at this level of iNO (18). At 40 ppm iNO, peak NO2 conc. was <4 ppm in 13 patients. In one patient, maximal inspired NO2 was 6.4 ppm at an FiO2 of 0.6, but it is important to note that we measured peak (not mean) inspired NO and NO2 conc. Recent dose-response studies suggest that 20 ppm iNO may be efficacious in adult cardiac surgical patients (9).
In summary, in adult cardiac surgery, PH per se, even if severe, does not necessarily imply postoperative RV failure, and, provided RV function is preserved and coronary artery perfusion not compromised, iNO may not be superior to IV PGE1 with regard to CI and RV performance. This finding might contribute to a more restrictive patient selection for iNO application. Selective pulmonary vasodilation may be indicated, when isolated RV failure is present (28,29), independent of the degree of PH. In patients with additional LV failure, PGE1-induced concomitant reductions of LV filling pressure and outflow impedance might be beneficial. Future studies should focus on better defining patient groups that profit from selective pulmonary vasodilation and on further evaluation of alternatives to NO inhalation.
We are indebted to the nursing staff of the cardiac intensive care unit for their valuable cooperation.
1. Frostell CG, Fratacci M-D, Wain JC, Zapol WM. Inhaled nitric oxide:: a selective pulmonary vasodilator reversing pulmonary vasoconstriction. Circulation 1991; 83: 2038–47.
2. Pepke-Zaba J, Higenbottam TW, Dinh-Xuan AT, et al. Inhaled nitric oxide as a cause of selective pulmonary vasodilatation in pulmonary hypertension. Lancet 1991; 338: 1173–4.
3. Young E. UHC Clinical Practice Advancement Center. Technology report: nitric oxide therapy. Oak Brook, IL: UHC Services Corporation, 1996.
4. Freeman B. Free radical chemistry of nitric oxide: looking at the dark side. Chest 1994; 105: 79S–84S.
5. Weinberger B, Fakhrzadeh L, Heck DE, et al. Inhaled nitric oxide primes lung macrophages to produce reactive oxygen and nitrogen immediates. Am J Resp Crit Care Med 1998; 158: 931–8.
6. Haraldsson A, Kieler-Jensen N, Ricksten S-E. Inhaled prostacyclin for treatment of pulmonary hypertension after cardiac surgery or heart transplantation: a pharmacodynamic study. J Cardiothorac Vasc Anesth 1996; 10: 864–8.
7. Krieg P, Wahlers T, Hartrumpf M, et al. Aerolized prostaglandin E1
in the treatment of pulmonary hypertension: a new strategy in thoracic organ transplantation? Langenbecks Arch Chir 1996;Suppl 1:285–90.
8. Fullerton DA, Jones SD, Grover FL, McIntyre RC Jr. Adenosine effectively controls pulmonary hypertension after cardiac operations. Ann Thorac Surg 1996; 61: 1118–24.
9. Kieler-Jensen N, Lundin S, Ricksten SE. Vasodilator therapy after heart transplantation: effects of inhaled nitric oxide and intravenous prostacyclin, prostaglandin E1
, and sodium nitroprusside. J Heart Lung Transplant 1995; 14: 436–43.
10. Semigran MJ, Cockrill BA, Kacmarek R, et al. Hemodynamic effects of inhaled nitric oxide in heart failure. J Am Coll Cardiol 1994; 24: 982–8.
11. Kieler-Jensen N, Ricksten S-E, Stenqvist O, et al. Inhaled nitric oxide in the evaluation of heart transplant candidates with elevated pulmonary vascular resistance. J Heart Lung Transplant 1994; 13: 366–75.
12. Sitbon O, Brenot F, Denjean A, et al. Inhaled nitric oxide as a screening vasodilator agent in primary pulmonary hypertension:: a dose-response study and comparison with prostacyclin. Am J Respir Crit Care Med 1995; 151: 384–9.
13. Prielipp RC, Rosenthal MH, Pearl RG. Vasodilator therapy in vasoconstrictor-induced pulmonary hypertension in sheep. Anesthesiology 1988; 68: 552–8.
14. McLean RF, Prielipp RC, Rosenthal MH, Pearl RG. Vasodilator therapy in microembolic porcine pulmonary hypertension. Anesth Analg 1990; 71: 35–41.
15. Sorkin EM, Brogden RN, Romankiewicz JA. Intravenous glyceryl trinitrate (nitroglycerin): a review of its pharmacological properties and therapeutic efficacy. Drugs 1984; 27: 45–80.
16. Piper PJ, Vane JR, Wyllie JH. Inactivation of prostaglandins by the lung. Nature 1970; 225: 600–4.
17. Fullerton DA, Jaggers J, Wollmering MM, et al. Variable response to inhaled nitric oxide after cardiac surgery. Ann Thorac Surg 1997; 63: 1251–6.
18. Girard C, Lehot JJ, Pannetier JC, et al. Inhaled nitric oxide after mitral valve replacement in patients with chronic pulmonary artery hypertension. Anesthesiology 1992; 77: 880–3.
19. Gatecel C, Mebazaa A, Kong R, et al. Inhaled nitric oxide improves hepatic tissue oxygenation in right ventricular failure:: value of hepatic venous oxygen saturation monitoring. Anesthesiology 1995; 82: 588–90.
20. Girard C, Durand PG, Vedrienne C, et al. Inhaled nitric oxide for right heart failure after heart transplantation. J Cardiothorac Vasc Anesth 1993; 7: 381–5.
21. Loh E, Stamler JS, Loscalzo J, Colucci WS. Cardiovascular effects of inhaled nitric oxide in patients with left ventricular dysfunction. Circulation 1994; 90: 2780–5.
22. Bocchi EA, Bacal F, Aulor JO Jr, et al. Inhaled nitric oxide leading to pulmonary edema in stable severe left heart failure. Am J Cardiol 1994; 74: 70–2.
23. Hare JM, Shernan SK, Body SC, et al. Influence of inhaled nitric oxide on systemic flow and ventricular filling pressure in patients receiving mechanical circulatory assistance. Circulation 1997; 95: 2250–3.
24. Hare JM, Loh E, Creager MA, Colucci WS. Nitric oxide inhibits the positive inotropic response to beta-adrenergic stimulation in humans with left ventricular dysfunction. Circulation 1995; 92: 2198–203.
25. Hayward CS, Kalnins WV, Rogers P, et al. Effect of inhaled nitric oxide on normal human left ventricular function. J Am Coll Cardiol 1997; 30: 49–56.
26. Radermacher P, Bantak P, Becker H, Falke KJ. Prostaglandin E1
and nitroglycerin reduce pulmonary capillary pressure but worsen ventilation-perfusion distributions in patients with adult respiratory distress syndrome. Anesthesiology 1989; 70: 601–6.
27. Bender KA, Alexander JA, Enos JM, et al. Effects of inhaled nitric oxide in patients with hypoxemia and pulmonary hypertension after cardiac surgery. Am J Crit Care 1997; 6: 127–31.
28. Wagner F, Dandel M, Günther G, et al. Nitric oxide inhalation in the treatment of right ventricular dysfunction following left ventricular assist device implantation. Circulation 1997; 96(Suppl II): II-291–6.
29. Bhorade S, Christenson J, O'Connor M, et al. Response to inhaled nitric oxide in patients with acute right heart syndrome. Am J Respir Crit Care Med 1999; 159: 571–9.