Effects of inhaled prostacyclin as compared with inhaled nitric oxide on right ventricular performance in hypoxic pulmonary vasoconstriction

Effects of Inhaled Prostacyclin as Compared With Inhaled Nitric Oxide on Right Ventricular Performance in Hypoxic Pulmonary Vasoconstriction Bernhard Zwissler, MD, Martin Welte, MD, and Konrad Messmer, MD

Objective: Recently, inhalation of prostacyclin (PGI2) has been shown to cause selective pulmonary vasodilation. However, the effects of inhaled PGI2 on right ventricular (RV) performance are still unknown and therefore were compared with those of inhaled nitric oxide (NO). Design: Reported measurements design. Setting: Animal research laboratory. Animals: Six anesthetized, ventilated dogs (28 --+ 2 kg). Interventions: Pulmonary hypertension was induced by decreasing F)O2 to 0.09 - 0.11 ('hypoxic pulmonary vasoconstriction', HPV). Subsequently, a single dose of either NO (50 ppm) or PGIz-aerosol (0.9 -+ 0.3 n g / k g / m i n ) was randomly added to the inspired gas. Measurements and Main Results: Measurementswere performed before induction of HPV and 10 minutes after application and withdrawl of each drug. Central hemodynamics, global RV function, and local RV function (n = 5, sonomicrometry) were assessed. HPV resulted in an increase of pulmonary artery pressure (PAP), pulmonary vascular resistance (PVR), RV stroke work, right coronary artery flow, maximal rate of RV pressure increase (RV dP/dtmax), and maximal velocity of shortening of contractile elements (Vmax). In contrast, RV ejection fraction, RV end-diastolic volume, RV end-diastolic fiber length, and systolic fiber

EDUCTION OF pulmonary artery pressure (PAP) in patients with pulmonary artery hypertension has been shown to improve right ventricular (RV) performance, thereby possibly contributing to a better prognos~s for these patients. 23 Until recently, however, selective pulmonary vasodilatlon has only been achieved by inhalaUon of mtric oxide. 4-6 In a recent study, the authors have provided the first experimental evidence that the intratracheal application of prostacyclin (PGI2)-aerosol--simflar to inhaled mtric oxide (NO)--has the potentml to elicit selective pulmonary vasodilation m hypox~c pulmonary hypertension m dogs, 7 suggesting a potential role of inhaled PGI2 in the treatment of pulmonary hypertension m man. Meanwhile, this hypothesis has been supported by a preliminary chnical report of Walmrath et al, who demonstrated a decrease of pulmonary artery pressure subsequent to the inhalation of PGI2 m three patients with respiratory failure) Because data of the effect of inhaled PGI2 on RV performance are not available, the effects of this new therapeutic concept on the right venmcle have been invesUgated and were compared with those of inhaled NO obtained in a model of mild pulmonary hypertension induced by ventilation with a low inspiratory fraction of oxygen. Global alveolar hypoxia is known to increase pulmonary artery pressure by elicitmg hypoxic pulmonary vasoconstriction and has been used to assess selective pulmonary vasoddation by inhaled NO in several recent studies. 4,9A° METHODS AND MATERIALS

The studies were performed in six foxhounds of either sex (mean body weight 27.6 -+ 1.9 kg). All ammals received care in compliance with the Gulde for the Care and Use o f Laboratory Animals (NIH publication No. 85-23, revised

shortening were unchanged. Both PGI2-aerosol and NO attenuated the HPV-induced increase in PAP and PVR without affecting arterial pressure. NO, but not PGi2, resulted in an increase of RV ejection fraction from 42 to 46% (p < 0.05). Right coronary flow dropped from 29 to 21 mL/min during PGI2 (p < 0.05). RV stroke work, RV dP/ dtmax, and Vmax decreased subsequent to both NO and PGI2, whereas local RV function was not affected. Conclusions: In pulmonary hypertension induced by HPV, PGI2-aerosol and inhaled NO reduced RV afterload and, hence, RV oxygen demand, with only minor changes of stroke volume and cardiac output, indicating an improvement of overall efficiency of RV contraction. RV ejection fraction increased on NO, but not with PGI2. This might be explained by the fact that the reduction of pulmonary vascular resistance during PGI2 amounted to only 65% of the effect of NO. In summary, both inhaled NO and PGIz-aerosol showed beneficial effects on RV performance and may prove helpful in the treatment of acute pulmonary hypertension. Copyright © 1995 by W.B. Saunders Company

KEY WORDS: aerosol, nitric oxide, prostacyclin, pulmonary hypertension, right ventricular performance

1985). The study was approved by the institutional animal care and use committee. After premedication with proplomazine, 1.5 mg/kg, IM [Combelen; Bayer AG, Leverkusen, FRG], anesthesia was induced by intravenous (IV) injectlon of pentobarbital, 20 mg/kg [Nembutal; Ceva, Bad Segeberg, FRG], plritramlde, 0.75 mg/kg [Dipidolor; Janssen, Neuss, FRG], and alcuromum, 0.25 mg/kg [Alloferm; Roche, Grenzbach-Whylen, FRG], and maintained by continuous infusion of pentobarbltal, 5 mg/kg/h. After termination of the surgical preparation, an addmonal infusion of pintramide, 150 ~g/kg/h, and alcuronium, 75 p.g/kg/h, was started. Flmd losses were replaced by IV infusion of Ringer's solution, 5 mL kg/h. A warming pad was used to keep core body temperature between 35.5° and 37°C. The dogs were endotracheally intubated and mechanically ventdated at a rate of 12 cycles per minute using an respiratory fraction of oxygen (FIO2) of 0.5 (Servo 900B, Siemens-Elema, Solna, Sweden). Tidal volume (VT) was adjusted to maintain an arterial partial pressure of carbon dioxide (PCO2) of 35 to 40 mmHg A fluid-filled catheter (PP270; Portex, Hythe, UK) was posmoned in the descending aorta via the left femoral artery to measure mean arterial pressure (MAP). A modified pulmonary artery catheter (Swan-Ganz TD catheter

From the Departments of Anesthesia and Surgical Research, Ludwrg-Maxtmthans-Umverstty Mumch, Khmkum Grosshadern, Faculty of Medtcme, Mumch, Federal Repubhc of Germany Address reprmt requests to Prtv -Doz Dr Bernhard Zwtssler, Department of Anesthesta, Ludwzg-Maxlmthans-Umverslty, Khmkum Grosshadem, Marchtonmtstr 15, 81366 Munchen, Germany Copyright © 1995 by W B Saunders Company 1053-0770/95/0903-001053 00/0

Journal of Cardlothoractc and Vascu/arAnesthes/a, Vol 9, No 3 (June), 1995:pp 283-289

283

284

93A-431-7.5F G; Baxter Healthcare Corp., Santa Ana, CA) was inserted into the pulmonary artery via the right external jugular vein to measure mean pulmonary artery pressure (PAP), cardiac output (CO), RV ejecnon fraction (RVEF), and RV end-diastolic volume (RVEDV). After median sternotomy and perlcardiotomy, prewarmed, precalibrated tip manometers (PC 370; Mfllar Instruments, Houston, TX) were inserted into the RV via a stab incision at the apex of the right ventricle and into the left venmcle (LV) via the right carond artery. In five dogs, an ultrasonic flow probe (Transonic Systems Inc., Ithaca, NY) was placed around the right coronary artery. For measurement of local RV contraction, a pair of miniaturized (¢ 1.5 to 2 ram), piezoceramic ultrasonic transducers was implanted into the RV inflow tract of the RV free wall as described before, u Owing to technical problems, sonomicrometric data could only be evaluated in five animals. At the end of surgical preparation, the pericardium was closed by a running suture avoiding constraint to the myocardium. Thereafter, the chest was closed, lungs were remflated, and remaining air was removed by a chest drain. Subsequently, the dogs were turned to the left lateral decubitus position. The correct localization of all crystals was verified by inspection post mortem. Nimc oxide was obtained in 50-L cylinders as a precise mixture of 200 ppm NO in nitrogen (Linde AG, UnterschleiBheim, FRG). During inspiration, the NO/N2 mixture was introduced into the respiratory gas using a T tube connected to the endotracheal tube. A magnetic valve (Servo Nebulizer 945; Siemens, FRG) triggered by the electronic regulation system of the ventilator was used for insuffiation of NO excluswely during inspiration. The respiratory NO flow was varied using a precise pressure reduction valve (SP 750.0183.M; Kuhnke, Malente, FRG), and the NO pressure was monitored with an electronic manometer (Type 352-P; Debro GmbH, Meerbusch, FRG). The correct inspiratory NO concentration was produced by varying the NO admixture to the respired gas mixture at constant tidal volumes. An inspiratory NO concentration of 50 ppm was selected for the present experiments because higher inspiratory NO concentrations were shown to produce no additional vasodilatory effect both in sheep with heparine-protamane-induced pulmonary hypertension 5 and in sheep with U46619-1nduced pulmonary hypertension. 6 The F~O2 was maintained constantly during administranon of the NO/N2 mixture by increasing the inspiratory oxygen concentration proximal to the NO insufltat~on port. The inspired oxygen concentration (FIO2) was continuously monitored distal to the NO msufflatlon port (Oxydlg; Drager AG, Liabeck, Germany). Before the experiments, the complete application system (ventilator, pressure reduction valve, electromc manometer) had been calibrated to achieve the desired lnsplratory NO concentration at different ndal volumes and oxygen concentrations using a chemiluminescence N O / N O x analyzer (Beckman, Model 951 A; Beckman Instruments, Munich, FRG). The inspiratory NO concentration at a given system setting varied no more than 3% during repeated calibration measurements. Mean methemoglobin levels increased from 0.8 - 0.2 to 1.2 + 0.3%

ZWISSLER, WELTE, AND MESSMER

after NO inhalation and returned to baseline values after termination of NO application. To avoid oxadation of NO to NO2, the critical contact time of NO and O2 during inspiration was kept short by using a short inspiration time (1.65 s) and by introducing NO into the inspiratory gas mixture very close to the endotracheal tube (19 cm). The concentrations of NO2 measured at the endotracheal tube during calibration were below 1 ppm. For inhalation of prostacyclin as aerosol, a jet nebulizer (Servo Nebulizer 945; Siemens, FRG) was connected to a Siemens 900B ventilator. The nebulizer chamber (Intersurgical LTD, Twlckenham, UK) was located 16 cm from the endotracheal tube. The pressure supplied to the nebulizer chamber was monitored with an electronic manometer (Type 352-P; Debro GmbH, Meerbusch, FRG). The driving pressure of the nebulizer decreased during inspiration from 3.39 - 0.04 bar to 2.05 + 0.03 bar and, consequently, the flow rate of the nebulizer decreased from about 15 to 9 L/ram. The nebulizer delivered aerosol particles with a mass median diameter of 3.54 p.m at a constant flow rate of 8 L/mm (geometric SD 1.94 p~m; determined by laser diffraction analysis); at this flow rate, the percentage of aerosolized particles with a diameter of less than 2 ~m is greater than 21% (data provided by the manufacturer: Intersurgical LTD, Twickenham, UK). Mean flow rates above 8 L/rain were chosen because the fraction of parncles with a diameter of less than 2 ~m, which are hkely to settle in the alveolar region of the lung, increases with flow rate. VT and total inspiratory flow rate were maintained constantly during nebulization by decreasing the inspiratory flow from the ventilator. Prostacyclin was supplied as the sodium salt of epoprostenol (Flolan; Wellcome, London, UK) dissolved in 50 M glycine buffer of pH 10.5 at a concentranon of 10 p~g/mL; the solution was prepared on the day of the experiment and stored on ice until use. Before inhalation, prostacyclin was further diluted with normal saline to a concentration of 430 ng/mL, and a total volume of 10 mL of this solution was filled into the nebulizer chamber. To assess the actual amount of prostacyclin nebulized, the total volume of the prostacyclin solution in the nebulizer chamber was measured before and immediately after inhalation (volume nebulized x concentration/duration of inhalation). The concentration of dissolved PGI2 filled into the nebulizer chamber (430 ng/mL) resulted in an application rate of PGIz of approximately 0.9 ng/kg/min, which is considerably lower than the dose commonly used for IV therapy of pulmonary hypertension. The authors reasoned that such a low dose should be effective if there is a selective vasodalating effect of inhaled PGI 2 on pulmonary circulation. All measurements were performed with the dogs in left lateral position. MAP and PAP were recorded (Astromed MT 9500; Astro-Med, Inc., West Warwick, RI) using Statham P23Db transducers (Gould-Statham, Oxnard, CA) referred to the right atrium. Electrocardiogram and ventrlcular pressures were sampled every 4 ms, digitized in real time (A/D ME-26; Meilhaus Electronic, Puchheim, FRG), and stored for evaluanon using a personal computer 386. Data

INHALED PGI2 AND RV PERFORMANCE

285

were analyzed with interactwe software ("Meduse"; H. Zeintl, Heidelberg, FRG). All parameters were evaluated at end-expiration using the average of three consecutive beats. In the RV, maximal and end-diastolic (RVEDP) pressures were assessed. CO, RVEF, and RVEDV were obtained from triplicate thermodilution measurements at end-expiratmn and were averaged (REF-1, EJection Fractaon/Cardiac Output Computer; Baxter Healthcare Corp., Santa Ana, CA). SV = CO" 1,000/HR PVR = (PAP-LVEDP) - 79.9. CO RVSW = SV" PAP • 0.133 - 10 -3 TTIRv = Fsy s " T~ys"HR, where SV = stroke volume, CO = cardmc output, HR -heart rate, PVR = pulmonaryvascular resistance, LVEDP = LV end-diastolic pressure, RVSW = RV stroke work, TTIRv = RV tension-time index, Fsy s = area beneath the systolic RV pressure curve, and Tsys = systolic ejection period. As parameters of global RV contractility, the maximal rate of pressure increase (RV dP/dtma~) and the maximal velocity of contractile element shortening (Vmax) were assessed. Vm~xwas obtained by linear extrapolation (r > 0.9) of the calculated velocity of shortening of contractile elements an the isovolumetric phase, a2 Myocardial segment lengths were measured at enddiastole (Ld,a) and end-systole (Lsys). End-diastole was defined as the beginning of the upstroke of dRVP/dt, end-systole as the point of maximal negative RVdP/dt.a3 In addition, maximum and minimum segment length (Lmax, Lmln) were determined. All values for length were normalazed by assuming Ld,a at control to be 10 mm. To quantify the pattern of local segmental motion, the percentage of Table I

systolic fiber shortening (Ssys) as well as protosystolic (Esys) were defined as follows:

elongation

Ssys ( % )

= (Ldl a -- Lsys) " 1 0 0 / L & a

Esys ( % )

= (Lma x - L&a ) - 1 0 0 / L d ~ a

After termination of surgery, the dogs were isovolumacally hemodiluted with 6% dextran 60 (Makrodex 6%; Schiwa, Glandorf, FRG) to a hematocrit of 30% in order to achieve identical values of hematocrit and hemoglobin m all animals. The withdrawn blood was used to replace blood samples taken for laboratory analysis. Hematocnt and hemoglobin did not change significantly during the experiment. Control measurements were performed 30 minutes after termination of surgery under stable hemodynamic conditaons (control). Thereafter, FIO2 was decreased to 0.09 to 0.11 to achieve a PaO2 of near 40 mmHg. This resulted an pulmonary artery hypertensaon caused by hypoxic pulmonary vasoconstrictaon (HPV) characterized by a ~ 60% increase of PAP and a ~ 200% increase of pulmonary vascular resistance (PVR) (Table 1). The inspired FIOz was monatored, and PaCO2 was maintained constant by adjusting VT. After a period of stable hemodynamlcs, a second measurement was performed (HPV). Subsequently, either nitric oxide at a concentration of 50 ppm or prostacyclin at a concentration of 430 ng/mL was administered in random order. Measurements were recorded after 10 minutes of NO or PGI2 application, respectively (HPV + NO, HPV + PGI) because after that period of time no further changes subsequent to the application of PGI2 or NO have been observed an pilot experiments. Subsequently, the administration of either NO or PGI2 was anterrupted and a further measurement was performed (HPV-NO, HPVPGI) 10 to 15 minutes later. Data are reported as mean __+SD when normally distributed (hemodynamics), otherwise (sonomacrometry) the me-

Influence of Inhaled Nitric Oxide (NO) as Compared With Inhaled Prostacyclin (PGI2) on Hemodynamic Variables (n = 6) Control

HPV

99-+11 98 _+ 10

135_+30"** 88 -+ 19

136-+31 83 -+ 10"

144-+31 84 + 12

139-+33 88 -+ 9

127_+27" 86 -+ 8

140-+28 88 -+ 10

PAP (mmHg) PVR (dyn sec cm -5)

13 -+ 1 110 -+ 125

20 -+ 2 * * * 326 -+ 9 0 * * *

15 -+ 1 " * * 170 - 135"*

20 -+ 3 282 -+ 135

21 -+ 2 314 -+ 113

17 -+ 2 * * * 212 "4- 112"*

21 -+ 3 316 _+ 106

CO (L/min) SV (mL) LVEDP(mmHg) RVPsys (mmHg) RVEDP(mmHg) TTIRV (mmHg • s • mm)

2 8 -+ 0 5 29 _+ 6 88-+35 25 -+ 3 8 --- 5 445 -+ 58

3.6 -+ 0 6* 27 _+ 4 60-+29 32 _+ 5 * * * 4 _+ 4* 649 -+ 7 6 * * *

3.7 -+ 0 5 28 -+ 6 72-+45 26 -- 3* 5 -+ 6 556 - 124"

37 26 71 32 6 710

3.6 -+ 0 6 26 -+ 5 70-+45 34 --+ 6 7 -+ 6 739 _+ 201

3 3 -+ 0 4 26 --- 4 84+42** 28 + 6 * * 7 _+ 6 610 _+ 134"

3.6 -4- 0 5 26 --+ 4 70-+37 33 -+ 6 6 _+ 6 725 -+ 155

HR (beats/mm) MAP (mmHg)

HPV + NO

HPV - NO

-+ 0.5 -+ 4 -+47 - 6 _+ 6 -+ 143

HPV

HPV + PGI2

HPV - PGI2

Data are mean _+ SD. Timepoints of measurements HPV, after reduction of hypoxm pulmonary vasoconstriction and immediately before inhalation of NO or PGI2, respectwely, HPV + NO (HPV + PGI2), 10 to 15 minutes after beginning of the inhalation of nitric oxide (HPV + NO) or prostacychn (HPV + PGI2), respectively, HPV - NO (HPV - PGI2), 15 minutes after withdrawal of NO (HVP - NO) or PGI2 (HPV - PGI2), respectwely, but with HPV maintained Abbreviations' HR, heart rate, MAP, mean arterial pressure; PAP, mean pulmonary artery pressure; PVR, pulmonary vascular resistance, CO, cardiac output, SV, stroke volume, LVEDP, LV end-diastohc pressure; RVP~ys, RV systolic pressure; RVEDP, RV end-dlastohc pressure, TTIRV, RV tension t~me index *,0<005. **p < 0 01, * * * p < 0 001

286

ZWISSLER, WELTE, AND MESSMER

dian and Q1/Q3 quartiles are hsted. Statistical analysis was performed using a repeated measures analysis of variance (rANOVA). In case the F value was significant (p < 0.05), the following time-points were compared by paired t test (hemodynamics) and Wilcoxon's signed rank test (sonomicrometric data), respectively: control versus HPV; HPV versus HPV + NO; HPV versus HPV + PGI2. Bonferroni correction was applied as appropriate. Differences were considered significant forp < 0.05. RESULTS

As shown in Table 1, reduction of HPV caused significant increases of CO (+30%), HR (+31%), PAP (+57%), and PVR (+196%), whereas SV remained unchanged. Both NO and PGlz-aerosol attenuated the HPV-induced increase of PAP ( - 7 6 % and -48%, respectively), PVR ( - 7 3 % and -52%, respectively), and systolic RV pressure ( - 8 3 % and -48%, respectively) without changing CO or SV. Fig 1 demonstrates that HPV had no sigmficant effect on RVEF or RVEDV. Although RVEF was not affected by PGI2, it significantly increased from 42% to 46% (p < 0.05) during NO. RVEDV remained unchanged during both modalities of intervennons. RVEDP was 4.0 _-_ 3.6 mmHg before application of NO and 5.7 - 5.7 mmHg before PGI2 and remained unchanged during application of both substances (Table 1). Fig 2 demonstrates that both parameters of RV contractility measured, V m a x and RV dP/dtma,, increased subsequent to HPV (+70% and +21%, respectively) and were found to be slightly reduced by NO ( - 2 3 % and - 5 % , respectively) and PGI2 ( - 2 6 % and -12%, respectwely). Fig 3 demonstrates that both right ventricular stroke work (RVSW) and QRCAincreased during HPV and were reduced by PGI2. NO application resulted in a 25% decrease of RVSW, whereas the reduction of QRca did not reach statistical mgnificance. There was no sigmficant effect of NO and inhaled PGI2, respectwely, on regional RV fiber length or RV fiber shortening (Table 2). DISCUSSION

independent from cause, each type of pulmonary hypert e n s i o n - b y increasing RV afterload--leads to an increase of RV wall tensmn and myocardial oxygen consumption and often is accompanied by a decrease of RVEFJ 4 The RVEF ,,,

i°

RVEDV I

.......

I

I

Vm*~ [~t/,]

CON'r~OL

~

HpV

Fig 2 Influence of inhaled NO as compared with inhaled prostacyclin (PGla) on parameters of RV contractility. Data are mean + SD (n = 6l; The dots indicate a significant difference versus the preceding value; *p < 0 05, * * p < 0.01.

decrease of RV function may cause a reduction of CO and systemic oxygen delivery by impairing pulmonary blood flow and may compromise LV performance by LV dlastohc tamponade. 15 Finally, a decrease of arterial blood pressure may occur with reduction of RV coronary driving pressure leading to irreversible RV failure and cardiogenic shock. 16'17 This pathophysiologic sequence might explain why mortality in patients suffering from pulmonary hypertension, eg, induced by adult respiratory distress syndrome (ARDS), is notably higher than in patients with normal PAP. 18't9 Hence, it is generally accepted that a reduction of pulmonary artery pressure and Improvement of RV function are major therapeutic goals in patients with pulmonary hypertensionJ Since the first description of the dilatory effects on pulmonary and systemic vessels of intravenous prostacychn in dogs, 2° the drug has been successfully applied as a pulmonary vasodilator m a m m a l models of acute pulmonary hypertension 27 and In patients with pulmonary hypertension of various causes. 21,22However, at doses efficient in reducing PVR, IV prostacyclin is known to reduce systemic vasodilation and hypotension. 6,23 Recently, the authors have provided the first experimental evidence that the mtratracheal application of PGIz-aerosol elicits selective pulmonary vasodilation in a canine model of pulmonary artery hypertension. 7 Subsequently, inhaled PGI 2 has been shown to reduce PAP in three patmnts with ARDS. 8 Nevertheless, the effects of inhaled PGI2 on RV performance are unknown and therefore have been studied in a suitable animal model.

Model of PulmonaryHypertension Pulmonary artery hypertension induced by hypoxlc pulmonary vasoconstriction has been used as a model for evaluatRVSW

°°.

Q RCA

4o

No

pGiz

m~'to-h

2.

[

'0 t Fig 1. Influence of inhaled NO as compared with inhaled prostacychn (PGI2) on RVEF and RVEDV. Data are mean -+ SD (n = 6). The following timepoints of measurement have been compared by Student's t test: control versus HPV, HPV + PGI 2 versus HPV and HPV + NO versus HPV, respectively The dot indicates a significant difference versus the preceding value, *p < 0,05

Fig 3 Influence of inhaled NO as compared with inhaled prostacyclm (PGla) on RVSW and right coronary blood flow (QRcA) Data are mean -+ SD (n = 6) (QRcA /7 = 5); The dot indicates a significant difference versus the preceding value, *p < 0.05, * * p < 0,01, * * * p < 0.001

INHALED PGI2 AND RV PERFORMANCE

287

Table 2, Influence of Inhaled Nitric Oxide (NO} as Compared With Inhaled Prostacyclin (PGla) on Local Function of the Right Ventricular Inflow Tract (n = 5)

Lm,n (mm)

Lm~×(mm) Lsw(mm) Ln.a(ram) Em,x (%) S~w (%)

Control

HPV

HPV + NO

HPV - NO

HPV

HPV + PGI2

HPV - PGI2

92 (9 1/9 4) 10 1 (10 1/10 1) 95 (9 4/9 6) 10 ( ) 08 (0 7/1 1) 48 (43/6 2)

93 (8 9/9 4) 10 0 (9.9/10 1) 93 (9 1/9 4) 10 (9 9/10 0) 08 (0 0/1 1) 87 (5 8/8 8)

g0 (8 8/9 4) 10 1 (9 9/10 1) 94 (9 4/9 4) 99 (9 8/10 0) 0.9 (0 0/1 7) 61 (5 2/6 9)

92 (9,1/9 4) 99 (9 9/10 2) 94 (9 2/9 5) 99 (9 9/10 1) 0.9 (0 0/1 1) 64 (4 9/10 2)

9.2 (9 1/9 4) 10 1 (9 9/10 2) 94 (9 3/9 4) 10 0 (9 9/10 1) 09 (0 7/1 1) 64 (5 8/7 9)

9.3 (8 9/9 3) 9.9 (9 9/10 3) 93 (9 1/9 4) 99 (9 9/9 9) 17 (0 0/23) 6.7 (5 6/7 9)

93 (8.9/9.4) 10 0 (9 9/10 1) 93 (9 1/9 4) 10 (9 9/10.0) 0.8 (0 0/1 1) 84 (5 6/8 7)

Data are medmn and Q1/Q3 = quartds The following tlmepomts of measurement have been compared by Wdcoxon's signed rank test control versus HPV, HPV + PGI2 versus HPV, HPV + NO versus HPV, respectively (for explanation of the timepomts of measurement see legend to Table 1) No slgmficant differences were found Abbrewations" Lm,n, minimal segment length, Lma×, maximal segment length; Lsvs, systohc segment length, L~,~, dlastohc segment length, Esvs, protosystohc segment elongation, Ssv,, systohc segment shortening

ing the cardiorespiratory effects of inhaled compounds, eg, nitric oxide, in several recent studies 4.9,10 There is good evidence that hypoxic pulmonary vasoconstriction notably contributes to the development of pulmonary hypertension in ARDS because an acute improvement of oxygenation (eg, by extracorporeal lung assist device) immediately attenuates or even reverses pulmonary hypertension in most of these patients (unpublished observations) In the present study, the decrease of the respiratory oxygen fraction to 0 1 reduced PaO2 to 40 mmHg and resulted in an increase of PAP by 57% and of PVR by 196%, respectively. As a consequence of increased RV afterload, a considerable increase of RV tension time index and right coronary blood flow and RVSW occurred, indicating a notable increase of RV oxygen demand in this model. The fact that RVEF was fully maintained in the presence of unchanged RV preload (as estimated by RVEDV, RVEDP, and end-dmstohc fiber length) despite elevated afterload suggests that RV contractility increased subsequent to induction of hypox]c pulmonary vasoconstriction and pulmonary hypertension. Although the authors did not assess plasma catecholamine levels in their experiments, increased RV contractility as elicited by the increase of RV dP/dtmax and Vm~xprobably was caused by a substantial sympathoadrenergic discharge consequent to systemic hypoxemla. Although an increased sympathoadrenerg]c tone together with decreased arterml and mixed venous oxygen tension complicate the interpretation of the cardiovascular effects of drugs, tachycardia and hypoxemia are frequently found chmcally in patients suffering from severe ARDS. Furthermore, neither the inspiratory oxygen fraction nor the arterial or mixed venous oxygen tension did change notably after induction of pulmonary hypertension. This strongly suggests that the observed changes of central hemodynamics and RV performance subsequent to the administration of PGI2 or NO, respectively, are caused by a specific drug effect and do not represent artifacts of the model used.

RV Performance Dunng Inhaled PGI2 The main finding of this study was that in the model described, inhaled PGIz significantly decreased the RV tension-time index, right coronary blood flow, and RVSW work without any change of RVEF or CO. The tension-time index according to Sarnoff is accepted as a valid indtrect measure of ventricular oxygen demand m the RV, 24 and changes of myocardial perfusion do reliably reflect changes of metabolic activity and oxygen demand in the heart. 25 Hence, the data strongly indicate that the oxygen demand of the RV was considerably reduced during "low-dose" inhalation of PGI2, probably caused by the decrease of RV afterload and a concomitant fall of RVSW. The decrease of RV oxygen demand occurred with only minor changes of HR, SV, and CO, indicating an improvement of the overall efficiency of RV contraction. Similarly to inhaled PGIz, inhaled NO was found to reduce puhnonary vascular tone, thereby lowering RV afterload and RV oxygen demand. This agrees with the observation of a considerable fall of RVSW, RV blood flow, and RV contracttlity (as elicited by the fall of Vma~and RV dP/dtmax) during inhalation of NO. There are two reasons probably explaining the reduction of RVdP/dtmax and Vm~ during inhalation of NO or PGI2. First, both parameters of contractility are known to be (at least in part) load dependent. A decrease of RV afterload by inhalation of NO or PGI2 therefore may have reduced both parameters in the absence of changes of intrinsic right ventricular contractility. Second, the decrease of RV afterload and pulmonary vasodilation most probably has reduced the degree of sympathoadrenergic stimulation during HPV. Hence, less lnotropy caused by sympathetic stimulation was necessary to maintain CO, SV, and systemic pressures. The finding that RVEF increased subsequent to NO, but not to PGI2, might be explained by the fact that the reduction of PVR during PGI2 amounted to only 65% of the effect of NO. Interestingly, the reduction of PAP on NO

288

or inhaled PGI2 did not increase CO or SV in these experiments. This probably was because pulmonary artery hypertension was modest m this model (mean PAP 20 mmHg) and did not result in RV dilatation or overt RV failure because RV filling (as estimated by RV enddiastolic pressure and end-diastolic volume) remained unchanged throughout the whole study. Therefore, further studies in more severe pulmonary hypertension are needed to elucidate the effects of inhaled PGI2 on the failing right ventricle. Local RV Function Data of the effects of inhaled vasodilators on local RV function have not yet been pubhshed. Although only five out of six experiments could be evaluated in the present study, these results suggest that neither inhaled NO nor inhalation of PGI2 had any influence on regional segment length or on regional segment shortening of the RV free wall in this model of moderate pulmonary hypertension. The fact that protosystolic segment elongation (Esys) was low at control and remained basically unchanged throughout the experiment strongly indicates that ischemla of the RV free wall did not occur at any timepolnt of measurement. This agrees well with the data obtained from intraventricular pressure measurement or fast response thermodilution technique, which showed no change of RVEDP, RVEDV, right ventrlcular end-systolic volume (RVESV) or RVEF. Two conclusions may be drawn from the analysis of regional RV function. The fact that neither RV enddiastolic pressure, RVEDV, nor end-diastolic segment length has been affected suggests that there was no change of global or local RV compliance subsequent to inhalation of NO or PGI2, respectively, during mild pulmonary hypertension. Such changes have been reported to occur subsequent to the application of drugs applied for cardiac support in case of pulmonary hypertension 26 or subsequent to ventdation with positive end-expiratory pressure, ll Clearly, care has to be taken when diastolic segment lengths are interpreted in terms of local preload because a single pair of crystals, caused by nonuniform fiber direction in the RV free wall, will not simultaneously and accurately reflect the length of all myocardial fibers In a given region. However, because the crystals were positioned along the main axis of fiber shortening of the RV inflow tract, I° it is assumed that end-diastolic segment length as measured in the present study represents local preload of the majority of functaonally active myocardial fibers. Second, the fact that systohc segment shortening remained unchanged in the presence of a reduction of PAP provides evidence that not only global RV work (as indicated by a fall of RVSW), but also local segmental work, was reduced by inhalation of both NO and PGI2. Ltmitations of the Present Study Several limitations of the study have to be considered. The present experiments are the very first in which the

ZWlSSLER, WELTE, AND MESSMER

influence of inhaled PGI2 on RV performance has been quantified and only a single dose of PGI2 has been used for that purpose. The concentration of dissolved PGI2 filled into the nebulizer chamber (430 ng/mL) resulted in an application rate of PGI2 of approximately 0 9 ng/kg/mm, which is considerably lower than the dose commonly used for IV therapy of pulmonary hypertension. The authors reasoned that such a low dose should be effective, if there is a selective vasodilating effect of inhaled PGI2 on pulmonary circulation. However, it has to be stated clearly that up to now neither the minimal effective dosage of inhaled prostacyclin nor a dose-response curve has been established. Therefore, it cannot be excluded at present that a higher concentration of the PGI2 solution might exert a more pronounced effect on pulmonary vascular tone and parameters of RV performance. A further drawback of the present study is that the concentrations of NO and PGI2 used for inhalation resulted in a different degree of pulmonary vasodilatation. It could be argued that it would have been preferable to compare equally effective doses of NO and inhaled PGI> However, by contrasting a single, low concentration of inhaled PGI2 with a high concentration of NO, the authors intended to compare the pulmonary vasodilating capacity of a very new modality of treatment (inhaled PGI2) to the maximal pulmonary vasodilating capacity of a well-characterized compound of known potency m the model used. Clearly, however, further studies are needed in the future where equlpotent doses of PGI2 and NO are matched. The degree of pulmonary hypertension as induced in this model was only mild if judged by clinical measures (ie, by the absolute height of PAP) However, it should be kept m mind that in contrast to other models of pulmonary artery hypertension (eg, application of the thromboxane analog U46619, pulmonary microembolization), the systemic oxygen tension was critically low (40 mmHg) in these experiments. Hence, although an increase of PAP by an average of 7 to 8 mmHg may be estimated as not relevant clinically under conditions of normoxemia, a similar increase may represent a considerable burden for the RV during systemic hypoxemia. This view is based on the authors' experimental experience that any further reduction of F~O2 below 0.1 thereby decreasing PaO2 below 40 mmHg may result in overt cardiac failure and RV dilatation, indicating a borderline ratio between myocardial oxygen demand and oxygen supply in this model even in the absence of severe pulmonary hypertension (unpublished results) The function of the RV free wall was measured only within the RV inflow tract, which may not have been representative tbr the entire RV free wall. Divergent changes of regional function within different regions of the RV free wall have been described to occur in other models of severe pulmonary hypertension 28 as well as during ventilation with positive end-expiratory pressure. 11 However, on the other hand, fiber shortening as obtained from only one segment of the RV free wall was shown to reliably reflect local events (eg, lschemia) within the RV free wallf1-9 Finally, to avoid problems regarding the interpretation of

INHALED PGI2 AND RV PERFORMANCE

289

Vmax and R V dP/dtm~x at changing loading conditions and heart rate, it would have been desirable to examine the end-systolic pressure-volume (P-V) relanonship as a measure of contractility. However, the assessment of P-V curves in the R V remains difficult owing to methodologic problems in accurately measuring instantaneous volume in the irregular R V chamber. In summary, both inhaled PGIz-aerosol and inhaled N O have beneficial effects on R V performance in a canine model of m o d e r a t e pulmonary hypertension induced by

hypoxic pulmonary vasoconstriction. It is suggested that both compounds may also prove helpful in the treatment of severe pulmonary hypertension comphcated by R V failure. However, a dose-response relationship for inhaled PGI2 is still lacking and has to be established in further studies. ACKNOWLEDGMENT

The authors thank Mrs Diane Wagner for her expert technical assistance and Dr Grosse-Wichtrup (Wellcome GmbH, Burgwedel, FRG) for generously providing Epoprostenol (Flolana).

REFERENCES

1. Pepke-Zaba J. Hlgenbottam T, Dlnh-Xuan A, Wallwork J' Inhaled mtrlc oxide as a cause of selective pulmonary vasodilation in pulmonary hypertension. Lancet 338.1173-1174, 1991 2 Selld6n H, Wlnberg P, Gustafsson LE, et al Inhalation of nitric oxide reduced pulmonary hypertension after cardiac surgery in a 3.2-kg infant Anesthesiology 78 577-580, 1993 3 D'Ambra MN, LaRala PJ, Philbm DM, et al' Prostaglandln E1 A new therapy for refractory right heart failure and pulmonary hypertension after mitral valve replacement J Thorac Cardlovasc Surg 89:567-572, 1985 4. Frostell C, Frataccl MD, Wain J, et al. Inhaled nitric oxade. A selective pulmonary vasodilator reversing hypoxIc pulmonary vasoconstriction. Circulation 83'2038-2047, 1991 5 Frataccl M-D, Frostell C, Chen T-Y, et al" Inhaled nitric oxlde--A selective pulmonary vasoddator of heparln-protamlne vasoconstriction in sheep Anesthesiology 75.990-999, 1991 6 Rossaint R, Falke KJ, Lopez F, et al Inhaled nitric oxide for the adult respiratory distress syndrome N Engl J Med 328.399-405, 1993 7 Welte M, Zwlssler B, Habazettl H, et al' PG12-aerosol versus nitric oxide for selective pulmonary vasodllatlon in hypoxlc pulmonary vasoconstriction. Eur Surg Res 25 329-340, 1993 8. Walmrath D, Schneider T, Pilch J, et al: Aerosollsed prostacychn in adult respiratory distress syndrome. Lancet 342.961-962, 1993 9 0 w a l l A, Davilen J, Sollevl A' Influence of adenosine and prostacychn on hypoxm-Induced pulmonary hypertension in the anaesthetized pig Acta Anaesthesiol Scand 35'350-354, 1991 10 Archer S, Rlst K, Nelson DP, et al Comparison of the hemodynamlc effects of nitric oxide and endothehum-dependent vasoddators in mtact lungs. J Appl Physiol 68.735-747, 1990 11. Zwlssler B, Forst H, Messmer K' Local and global function of the right ventricle In a canine model of pulmonary mlcroembohsm and olelc acid edema' Influence of ventilation with PEEP Anesthesiology 73:964-975, 1990 12 Mason DT, Spann JF, Zelis R Quantification of the contractile state of the intact human heart Am J Cardlol 26:248257, 1970 13 Morris JJ, Pellom GL, Hamm DP, et al. Dynamic right ventncular dimension Relation to chamber volume during the cardiac cycle J Thorac Cardiovasc Surg 91.879-887, 1986 14. Sibbald WJ, DrIedger AA, Myers ML, et al. Blventrlcular function in the adult respiratory distress syndrome Hemodynamlc and radlonuchde assessment, with special emphasis on right ventricular function Chest 84 126-134, 1983 15 Ferllnz J. Right ventrlcular function in adult cardiovascular disease Prog CardIovasc Dis 25 225-267, 1982

16 Vlahakes GJ, Turley K, Hoffman JIE: The pathophyslology of failure in acute right ventrlcular hypertension Hemodynamic and biochemical correlations Circulation 63:87-95, 1981 17 Mclntyre KM, Sasahara AA' The hemodynamac response to pulmonary embolism in patients without prior cardiopulmonary &sease. Am J Cardlol 28'288-295, 1971 18. Oliphant LD, Sibbald WJ' Right ventncular performance In the adult respiratory distress syndrome, m Vincent JL, Suter PM (eds) Update m Intensive Care and Emergency Medicine New York, Spnnger, 1987, pp 25-37 19. Vidal Melo MF, Lenz C, Forst H, et al' Relevance of multivariate patterns of routine cardiopulmonary measurements for early prognosis in the adult respiratory distress syndrome. Theor Surg 6'63-70, 1991 20. Kadowltz PJ, Chapnik BM, Feigen LP, et al' Pulmonary and vasodilator effects of the newly &scovered prostaglandin, PG12 J Appl Physlo145'408-413, 1978 21. Rubln LJ, Groves BM, Reeves JT, et al Prostacychn reduced acute pulmonary vaso&lation in primary pulmonary hypertension. Circulation 66.334-338, 1982 22 Rubln LJ, Mendoza J, Hood M, et al: Treatment of primary pulmonary hypertension with continuous intravenous prostacychn (epoprostenol) Results of a randomized trial. Ann Intern Med 112'485-491, 1990 23. Radermacher P, Santak B, Wust H J, et al. Prostacychn for the treatment of pulmonary hypertension in the adult respiratory distress syndrome, effects of pulmonary capillary pressure and ventilation-perfuslon distribution. Anesthesiology 72:238-244, 1990 24 Fixler DE, Archle JP, Ullyot D J, et al: Effects of acute right ventrlcular systolic hypertension on regional myocardial blood flow in anesthetized dogs. Am Heart J 85.491-500. 1973 25 Sonntag M, Schrader J: Spatial heterogeneity of myocardial perfusion and metabolism. Europ J Physio1420' 116, 1992 (suppl 1) 26 Vincent JL, Reuse C, Kahn RJ: Effects on right ventrlcular function of a change from dopamlne to dobutamme in critically ill patients Crit Care Med 16'659-662, 1988 27 Meier GD, Bove AA, Santamore WP, et al Contractile function in canine right ventricle. Am J Physiol 239 H794-H804, 1980 28. Zwissler B, Forst H, Messmer K' Acute pulmonary mlcroembohsm induces different changes of preload and contraction pattern m the canine right ventricle Car&ovasc Res 24'285-295, 1990 29 Tommke H, Urabe Y, Ohzono K, et al: Homogeneous distribution and pressure dependent reduction of coronary blood flow in right ventncular free wall during coronary hypoperfuslon in anaesthetised open chest dogs. Car&ovasc Res 23:31-39, 1989