INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

NITRIC OXIDE (NO), an endothelial-derived lipophilic free radical, has been identified as a powerful vasodilator (25). Elevated levels of NO have been reported in experimental animals after bacterial challenge (6) and after administration of endotoxin (14) or proinflammatory cytokines, such as tumor necrosis factor (16) and interleukin-1 (8). Therefore, NO is considered to play a major role in the hemodynamic changes observed in septic shock, with the main effect of NO being regulation of systemic vascular resistance (SVR). Originally, much attention was paid to NO synthase inhibition, because the hyperdynamic changes in the cardiovascular system associated with endotoxemia were assumed to be caused by NO. Accordingly, NO synthase inhibition was evaluated as a means of preventing or correcting those changes. In animal models of sepsis, NO synthase inhibition succeeded in preventing an abnormal reduction in SVR. However, systemic and regional O₂ transport were impaired. In addition, depression of cardiac output and peripheral blood flow was observed. (10, 19, 29) These studies suggest that nonselective NO inhibition causes extreme vasoconstriction and exaggerates pulmonary hypertension.

Because inhaled NO has been reported to attenuate pulmonary vasoconstriction without systemic hemodynamic changes (34), it has been evaluated as a therapeutic agent. A beneficial effect of inhaled NO has been observed in patients with chronic obstructive pulmonary diseases (1), chronic pulmonary hypertension (27), adult respiratory distress syndrome (30), congenital heart disease (28), persistent pulmonary hypertension of the newborn (38), and in animal models with smoke inhalation injury (21, 24). Elevated pulmonary vascular resistance (PVR) in the presence of low SVR is commonly observed in septic patients. The studies conducted at this institute, using a porcine model of acute endotoxemia, have demonstrated that NO inhalation treatment improved cardiovascular and pulmonary function; PVR was markedly decreased, and cardiac work indexes and systemic O₂ delivery were improved (20, 22, 23). Similar studies by others have confirmed these results (4, 32). However, those studies were performed in acute models, with only a brief duration of study, and were terminated within several hours of onset. The risk of long-term NO has not been defined except in small animal models (11). The effects of long-term continuous administration of inhaled NO and possible complications of NO inhalation, including pulmonary injury caused by accumulated NO₂ (a metabolite of NO) and methemoglobinemia, remain unknown.

Moreover, there have been no reports of the effect of inhaled NO on cardiac contractility in sepsis. All previous studies have utilized conventional hemodynamic parameters, such as left ventricular (LV) stroke work, ejection fraction, or maximum change in pressure over time (dP/dt), to measure LV function. Because those indexes depend on both preload and afterload, they are not precise indicators of LV function in an endotoxemic model where preload and afterload can change dramatically. We have developed and previously described an awake swine model of endotoxemia, in which cardiac contractility can be measured directly (13). With this model, the slope (Ees) of the LV end-systolic pressure-volume relationship (ESPVR) can be obtained. This contractile index is known to be insensitive to vascular loading conditions (31). Accordingly, the model can be used to study the pathophysiology of endotoxemia-induced alteration in cardiac function.

Loh et al. (17) reported that inhalation of NO caused a decrease in PVR associated with an increase in LV-filling pressure in patients with heart failure due to LV dysfunction. This suggests an adverse effect of NO inhalation on LV function. We hypothesize that NO inhalation may have a beneficial effect on LV dysfunction in a porcine model of endotoxemia. The purpose of this study was to describe effects of continuous long-term inhalation of NO on systemic and pulmonary hemodynamics and on LV contractility during endotoxemia.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Twelve adult female Yorkshire swine, weighing between 27 and 41.8 kg, were studied. All study protocols were approved by the Animal Use Committee of the US Army Institute of Surgical Research and adhered to the provisions of the Animal Welfare Act. Animals were observed for 1 wk before use to exclude the possibility of preexisting disease. After this observation period, they were fasted for 24 h.

Surgical procedure. Each animal was sedated with atropine (0.08 mg/kg), ketamine (2.2 mg/kg), and xylazine (2.2 mg/kg) administered intramuscularly. Inhalation of a mixture of isoflurane, N₂O, and O₂ followed induction and placement of an endotracheal tube. The animal was placed in the right lateral position. Under sterile conditions, a Tygon catheter was placed in the left common carotid artery to measure mean arterial pressure (MAP) and to sample blood. Also, the left internal jugular vein was cannulated with a flexible Tygon catheter for administration of endotoxin and resuscitation fluid. One radiopaque sheath introducer (8.5-Fr; American Edwards Laboratories, Irvine, CA) was inserted into the left external jugular vein for later insertion of a thermodilution catheter. A thoracotomy was performed through an incision in the fifth intercostal space. A pneumatic occluder (vascular occluder; In Vivo Metric, Healdsburg, CA) was positioned around the inferior vena cava (IVC), and the pericardium was opened. At this point, intravenous lidocaine administration was started (initial bolus was 2 mg/kg, followed by 2 mg/min continuous infusion). A micromanometer (model P5; Konigsberg Instrument, Pasadena, CA) was implanted through a small incision in the apex of the LV to measure intracavitary pressure continuously. Two pairs of ultrasound dimension transducers (Internal Ventricular Transducer-pair, 4.0 mm, ID 3-2, Triton Technology, San Diego, CA) were implanted on the LV endocardium by using a transmyocardial needle puncture technique for measurement of anterior-posterior dimension (D_min) and base-apex dimension (D_maj). The transducers were connected to an oscilloscope (2236A, Tektronix, Wilsonville, OR), and the signals were checked to confirm placement of crystals. All catheters were exteriorized dorsally. The pericardium was left open, and the chest was closed after a drainage catheter was inserted. This chest tube was removed when the animal resumed spontaneous breathing. The animal was monitored until its full recovery from anesthesia. Pentazocine (30 mg iv) was administered every 12 h as needed for pain. Daily cefazolin (500 mg iv) was initiated preoperatively and repeated until the day of the experiment.

Experimental protocol. These studies were performed in conscious animals during a 24-h observation period. Approximately 3 h after the recovery from anesthesia, animals were randomized to either the control group (Control, n = 6) or the NO-inhalation group (NOI, n = 6). All animals received continuous intravenous endotoxin (Escherichia coli 0111: B4; Difco, Detroit, MI) at a dose of 10 µl · kg¹ · h¹ for 24 h and were resuscitated with lactated Ringer solution (20 ml · kg¹ · h¹ for the initial 4 h, 15 ml · kg¹ · h¹ for the next 3 h, and 10 ml · kg¹ · h¹ for the rest of the study). Thirty minutes after endotoxin began, inhalation of NO was initiated in the NOI group and was continued throughout the study. During the study, the tracheostomy tube was connected to a nonrebreathing circuit that consisted of a 5,000-ml reservoir bag and a one-way valve to separate inspired from expired gas. The inspired gas was room air for the Control. The NOI breathed a mixture of room air and O₂ with added NO (237 ppm in N₂, Air Produce, Austin, TX) immediately diluted to an inspired concentration of 20 ppm NO with a fraction of inspired O₂ of 0.21. The residence half-time of NO in the reservoir bag was <30 s, with a fresh gas flow of 10,000 ml/min. Expired gases were scavenged and discarded. Hemodynamic indexes, including contractility, were serially measured, and blood samples were obtained. At the end of 24 h, the animals were euthanized, and the site of crystal implantation was confirmed by necropsy.

Data acquisition and analysis. The data-acquisition system was described previously (33). The crystals were connected to a sonomicrometer (model 120-1001, Triton Technology), and the signals from the crystals were amplified. Signals were continuously monitored on an oscilloscope. The Swan-Ganz catheter and the arterial catheter were connected to the pressure transducers (P23 ID Statham, Gould, Oxnard, CA). Those pressure signals, including that from the LV transducer, were amplified through pressure processors (model 20-4625-526611, Gould). All dimension and pressure signals were attenuated (Vstore; Racal Recorders, Hythe Southampton, England) and digitized (DAS16; Keithley Metrabyte Data-Acquisition System, Taunton, MA). The sampling rate was 200 samples/s. The signals were captured automatically by using custom software for 10 s at 30-s intervals (Compaq Desk Pro 386s). The digitized data were analyzed for various hemodynamic parameters with the same software used for capturing signals. A representative computer display of D_min, D_maj, LV volume (LVV), LV pressure (LVP), and pressure-volume (P-V) loops in one animal during IVC occlusion is shown in Fig. 1.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Representative computer display of left ventricular (LV) anterior-posterior dimension (Dmin), base-apex dimension (Dmaj), LV volume (LVV), LV pressure (LVP), and pressure-volume loop in 1 animal during inferior vena cava occlusion.

Statistical analysis. Data analysis was performed with the use of a Power Macintosh 7100/66 computer (Apple Computer, Cupertino, CA) and a statistical software package (Stat View 4.5, Abacus Concepts, Berkeley, CA). Data are represented as means ± SD. Statistical analysis for hemodynamic parameters and Ees was performed by using the template for a two-factor ANOVA (group, time) with repeated measures on one factor (time). If a significant interaction between group and time was found by ANOVA, post hoc testing was performed with the Scheffé test. P < 0.05 was considered as a statistically significant difference. Additionally, a significant time interaction within groups was evaluated by using the same template to confirm a significant difference. If a significant interaction by time was found, time differences from baseline were assessed by a Bonferroni corrected t-test.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

PVRI increased and arterial PO₂ (Pa_O2) decreased immediately after the initiation of endotoxin but recovered to baseline by 7 h and remained there in both groups (Fig. 2). A significant interaction between group and time was found in PVRI. There was also a significant difference between groups at 1 h by Scheffé test. Pa_O2 decreased initially in both groups, reaching a nadir at 2 h in the Control group and at 4 h in the NOI group. In the NOI group, the Pa_O2 remained near its nadir of 70 Torr for the rest of the study period. In the control group, the Pa_O2 rose from its nadir of 55 Torr but only to the level noted in the NOI group at 24 h. A significant interaction was found and confirmed by significant differences between groups at 1 and 2 h. CI and MAP gradually increased and SVRI decreased in both groups (Fig. 3). These hemodynamic changes were persistent until the termination of the experiment, and they closely resembled human endotoxemia. No significant differences in systemic hemodynamics were found between groups. Persistent tachycardia and pulmonary hypertension were noted in both groups (Table 1). An increase in Ved, which has been observed in human endotoxemia, was also noted. After 7 h, the hemodynamic changes resembled those that occur in humans with endotoxemia. In this 24-h study, no evidence of methemoglobinemia, no abnormal elevation of NO₂ levels, and no disturbance of oxygenation index or pulmonary hemodynamics were noted in any animal.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Serial changes in pulmonary vascular resistance index (PVRI) and arterial PO2 (PaO2) for control group (Control, n = 6) and nitric oxide inhalation group (NOI, n = 6) during continuous endotoxin infusion. Data are means ± SD. * P < 0.05 vs. Control;  P < 0.05 vs. baseline.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Serial changes in cardiac index (CI), mean arterial pressure (MAP), and systemic vascular resistance index (SVRI) for Control group (n = 6) and NOI group (n = 6) during continuous endotoxin infusion. Data are means ± SD.

A two-factor ANOVA was performed on the Ees values at baseline and at 4, 7, 12, 18, and 24 h. There was no significant difference between groups (P = 0.580). There was a significant difference within groups with respect to time (P < 0.001) and a significant interaction between group and time (P = 0.004). A significant interaction means that the trends were not parallel. Two-tailed paired t-tests were performed on Ees values within groups between baseline and times 4 h. There were no significant differences within the NOI (P > 0.05). Within the Control, the P values for comparisons between baseline and 4, 7, 12, 18, and 24 h values were 0.858, 0.010, 0.006, 0.013, and 0.001, respectively.

After applying a Bonferroni correction for 10 comparisons, the critical P value was 0.05/10 = 0.005, so there was a significant difference between baseline and 24-h values of Ees. When the data were normalized to baseline values for graphic analysis, Ees was lower than baseline in the Control from the seventh hour until the termination of the study (Fig. 4). This impairment of contractility was not found in the NOI. Serial alteration in V₀ was indicated by changes from baseline in Fig. 4. V₀ was relatively stable in both groups, and there was no significant difference between groups. In light of these values of V₀, changes in Ees appear to be primarily influenced by cardiac contractility.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Serial changes in %baseline of slope of LV end-systolic pressure-volume relationship (Ees; top) and changes from baseline in x-intercept of LV end-systolic pressure-volume relationship (V0; bottom) for Control group (n = 6) and NOI group (n = 6) during continuous endotoxin infusion. Data are means ± SD. * P < 0.05 vs. Control;  P < 0.05 vs. baseline.

In representative animals, Ees decreased from 7.83 at baseline to 4.05 at 24 h after endotoxin in the Control, whereas Ees remained at baseline in the NOI (Fig. 5). An increase in Ved in both groups was noted.

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Representative LVP-LVV loops at baseline (A and B) and 24 h after endotoxin (C and D) obtained from animal 41 in Control group (A and C) and from animal 71 in NOI group (B and D).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In a human adult respiratory distress syndrome trial (34), a low concentration (1.25 ppm) of inhaled NO was enough to improve oxygenation, whereas concentrations of methemoglobin and NO₂ increased if inhaled NO exceeded 20 ppm. The most appropriate concentration of inhaled NO to ameliorate the cardiac dysfunction during endotoxemia is unclear, but our results demonstrated that inhalation of 20 ppm NO for 24 h was safe and sufficient to prevent impairment of cardiac contractility in this model. We chose inhalation of 20 ppm NO because it was the concentration which, when administered for 48 h in an ovine model of smoke-inhalation injury, improved pulmonary hypertension and oxygenation without complications. Because NO₂ concentration is determined by the concentration of NO and O₂ in the gas blender (7), we utilized a relatively low concentration of NO. Furthermore, to minimize complications, we have utilized the NO-administration system which was developed at our laboratory (24). This delivery system minimizes the transit time of gas in the delivery circuit and precisely controls the concentration of NO that is delivered and exhaust gases, while the monitoring system provides online analysis of NO, NO₂, and O₂ concentrations (36).

In our endotoxemia model, alteration of Ees was biphasic. At 1 and 2 h, Ees was increased in both groups, and there was no significant difference between groups. We have previously reported (13) that systemic hemodynamic changes during continuous endotoxin infusion in this conscious swine model are biphasic. The early hemodynamic changes, which occur in this model in the first 2 h after the continuous endotoxin infusion is begun, do not simulate human hemodynamics commonly observed during endotoxemia. Thus it is not appropriate to evaluate inhaled NO therapy by making comparisons with cardiovascular parameters in the acute phase in our model. After 4 h, Ees progressively decreased in the Control but returned to baseline levels and remained there in the NOI. Beyond 4 h, other LV contractile indexes, such as dP/dt or ejection fraction, became progressively elevated. Each of these indexes is sensitive to the inotropic state and HR, as well as cardiac preload and afterload, which are all markedly altered after endotoxemia. Although there was a consistent increase in those indexes at 24 h, this may not reflect myocardial contractile status per se (9). Ees is influenced by HR in addition to inotropic status. Because there was no difference between groups in the later phase of the experiment with respect to HR, an increase in HR cannot explain the differences observed in Ees at 7 h. Depressed Ees in the Control group persisted until the end of the study, when other systemic cardiovascular parameters resembled human endotoxemia. This observation supports the hypothesis that NO prevents LV dysfunction.

Similar protective effects of inhaled NO in other animal models of a myocardial ischemia and reperfusion injury have been reported by Weyrich et al. (37). The mechanism of the protective effect of inhaled NO on LV dysfunction is speculative, because the cause of LV dysfunction during endotoxemia remains undefined. O₂-derived free radicals are considered by many to be a cause of myocardial injury. One source of O₂ radicals is activated neutrophils that have been stimulated by various cytokines or other mediators released by endotoxin. Although NO has been documented to prevent neutrophil adhesion in venula (15) and to attenuate the oxidative burst (3), there is no agreement on whether exogenous NO administration has a beneficial effect on leukocyte superoxide production during endotoxemia. In other studies, we have noted an increase in O₂ radical production, beginning 3 h after endotoxin infusion in the anesthetized swine model (unpublished data). Other factors, including coronary insufficiency, circulating myocardial depressant factor, and structural disorder of the myocyte or myocardial edema are proposed as possible causes of LV impairment. None of them has been confirmed, and whether they are affected by exogenous NO is unknown.

Also of interest in this study are the serial changes in Ved. In our previous study (13), the same amount of fluid was administered to a swine model with and without endotoxin. In the animals that did not receive endotoxin, MAP increased but Ved did not change significantly. In the animals that did receive endotoxin, Ved was elevated. There was no difference in MAP between groups. Because the volume administered was the same in both groups, the increase in Ved was caused by endotoxin infusion, not by volume loading. This increase in Ved in both groups during the late phases of the experiment is consistent with findings in a previous clinical study of septic shock (26). Several other reports (12, 35) indicate that end-diastolic compliance is impaired during sepsis. These changes, which suggest myocardial depression, have been related to the overproduction of NO. In this study, NO administration did not have extreme vasodilative effects, confirming the relative safety of long-term NO inhalation. Burkhoff et al. (2a) have reported that the ESPVR is considered nonlinear and that an increase in Ved could cause a reduction in the measurement of Ees. In our study, Ved does not appear to account for the difference in Ees between groups, because Ved increased in both groups and there is no significant difference in Ved between groups.

PVRI reached a peak, and Pa_O2 reached a nadir, at 1 h after endotoxemia in the Control. PVRI returned to normal levels by 12 h and remained there throughout the study period. Pa_O2 rose toward normal but remained at 79% of baseline values from 8 h onward. Only before 4 h were there significant differences in HR, PVRI, and Pa_O2 between the NOI and Control groups. In the NOI group, the rate and magnitude of the rise of PVRI were reduced, and the magnitude of the decrease in Pa_O2 was reduced. The magnitude of pulmonary hypertension and impairment of oxygenation during the later phase were so modest that there was no observable effect of NO inhalation. Thus, in our model, we have shown that continuously inhaled NO failed to improve pulmonary vasoconstriction and oxygenation at 24 h. Variable effects of inhaled NO on high MPAP have been reported in humans and swine (2, 18). Effects of inhaled NO, which may depend on the duration of endotoxemia and the severity of pulmonary hypertension, may account, in part, for this discrepancy. The time-related variability may be particularly pertinent to the variable results in human studies, because it is difficult to determine the time when endotoxemia begins in patients.

In summary, in a chronic model of endotoxemia, the reduction of pulmonary hypertension and improvement of oxygenation associated with continuous NO inhalation were limited to the acute phase. Continuous inhalation of 20 ppm NO for 24 h did not lead to pulmonary complications. Inhalation of NO prevented reduction of Ees during the chronic (7-24 h) phase in this awake porcine model of endotoxemia. We conclude that continuous NO inhalation is a reasonable therapeutic alternative for use in improving LV function in human endotoxemia.