The Effects of Propofol on Hemodynamics and Renal Blood Flow in Healthy and in Septic Sheep, and Combined with Fentanyl in Septic Sheep

Abstract

Sepsis is characterized by myocardial depression and systemic vasodilation, both of which are most likely mediated by nitric oxide.Propofol inhibits nitric oxide synthase and may therefore be beneficial in sepsis. On the other hand, renal blood flow, known to be only minimally affected by propofol in healthy subjects, may be drastically reduced in septic individuals, because the renal microvasculature is known to be very sensitive to nitric oxide. In this study, the effects of propofol in healthy and in septic sheep, and in combination with fentanyl, were analyzed and compared with nonanesthetized septic sheep. In healthy sheep, propofol caused only minor hemodynamic changes. In septic sheep, however, hemodynamics deteriorated. Renal blood flow was reduced to 60% +/- 10% of the preseptic baseline and to 39% +/- 4% of the septic value. This reduction was selective, since the cardiac output decreased significantly less. These adverse effects of propofol on hemodynamics and renal blood flow were reduced when propofol was combined with fentanyl.

(Anesth Analg 1996;82:738-43)

Anesthesia induced by propofol decreases heart rate, mean arterial pressure (MAP), cardiac output, and stroke volume ^[1]. Propofol produces vasodilation, but whether a reduction in myocardial contractility contributes to the reduction in MAP is still controversial. Some investigators have found propofol to cause a direct and dose-dependent reduction in contractility ^[1], whereas others exclude any negative inotropic effect of propofol ^[2].

Although propofol is accepted as a safe anesthetic in subjects with stable hemodynamics, little is known about its effects on those with unstable hemodynamics. Mayer et al. ^[3] showed that the hemodynamic effects of propofol are tolerated in healthy dogs, but that hemodynamics deteriorate in dogs with myocardial ischemia. On the other hand, propofol combined with fentanyl is a safe anesthetic regimen for patients with impaired ventricular function ^[4].

Propofol has not been evaluated in septic subjects. Sepsis is characterized by unstable hemodynamics as a result of reduced myocardial contractility and vasodilation, so that additional vasodilation and myocardial depression caused by propofol may be detrimental in sepsis. However, propofol is known to inhibit nitric oxide synthase (NOS) ^[5]. Since nitric oxide (NO) is the final common mediator of septic vasodilation ^[6], the use of propofol in sepsis may lead to vasoconstriction and be beneficial. The renal vasculature is very sensitive to NO ^[7]. Consequently, propofol may reduce renal perfusion in septic subjects, although it is known to have only moderate effects on renal perfusion in healthy subjects ^[8,9]. The effects of propofol on renal blood flow during sepsis have yet to be reported, even though the drug, whether or not combined with fentanyl, is widely used in intensive care units for sedation of critically ill patients, many of whom are septic. Renal malperfusion can trigger renal failure, which is a common finding in sepsis and is correlated with poor outcome ^[10]. Therefore, we studied the effects of propofol on hemodynamics and renal perfusion in healthy and in septic sheep. Since propofol is a poor analgesic and its combination with fentanyl causes less hemodynamic changes ^[11], we also examined the effects of propofol with fentanyl.

Methods

Experiments were performed according to the National Institutes of Health guidelines on the use of experimental animals. The study was approved by our Animal Care and Use Committee. Twenty-five female Merino sheep were surgically prepared for chronic study. After a 24-h fasting period, the animals were anesthetized with halothane (2.5-3.5 vol% in oxygen) through an animal anesthesia mask until depth of anesthesia allowed endotracheal intubation (inner diameter 10 mm; Mallinckrodt, Glens Falls, NY). The sheep were then mechanically ventilated with 1.5-2.5 vol% halothane in oxygen. Respiratory frequency was adjusted to maintain arterial CO₂ levels within normal range, using a tidal volume of 12 mL/kg. Under sterile conditions, femoral arterial and venous catheters were inserted and a Swan-Ganz catheter (Model 93A-131-7F, American Edwards Laboratories, Irvine, CA) was inserted through the right jugular vein into the pulmonary artery. Through a left-sided thoracotomy, a silastic catheter was placed into the left atrium. An ultrasonic flow probe (Transonic Systems, Inc., Ithaca, NY) was positioned on the left renal artery through a laparotomy. Animals were weaned from mechanical ventilation and allowed to awaken and recover for at least 5 days.

When the animals were fully recovered, the catheters were connected to pressure transducers (Statham Gould P23 ID, Oxnard, CA) and a physiologic recorder (Honeywell OMJ9; Electronics for Medicine, Pleasantville, NY). A computer (Model 9529; American Edwards Laboratory, Irvine, CA) was used for cardiac output measurements by thermodilution technique. Mixed-venous and arterial blood gases were analyzed with a model 1302 pH/blood gas analyzer and model 282 CO-Oximeter (Instrumentation Laboratory, Lexington, MA).

Renal blood flow was measured with the ultrasonic flow probe and a flow meter (Model T 101, Transonic Systems, Inc.). Cortical and medullary blood flows in both kidneys were measured with colored microspheres (E-Z TRAC, Interactive Medical Technology, Los Angeles, CA). Approximately 3.75 million colored microspheres, with a diameter of 15 micro meter, were injected into the left atrium. Simultaneously, reference blood was continuously withdrawn from the femoral artery. This withdrawal at a rate of 10 mL/min began 30 s before injection of the microspheres and was stopped 2 min after the injection.

Sheep were randomly assigned to one of four treatments.

1) Six sheep were anesthetized with propofol (propofol, nonseptic). Anesthesia was induced with 10 mg/kg propofol and maintained with 10 mg centered dot kg^-1 centered dot h^-1 for 30 min.

2) Seven sheep received a continuous infusion of live Pseudomonas aeruginosa (2.5 times 10⁶ colony-forming units [CFU]/min) for more than 48.5 h. After 48 h, they were anesthetized with propofol (propofol, septic). The induction dosage was 10 mg/kg, and anesthesia was maintained with 5 mg centered dot kg^-1 centered dot h^-1 for 30 min.

3) Seven sheep received a continuous infusion of live P. aeruginosa (2.5 times 10⁶ CFU/min) for more than 48.5 h. After 48 h, they were anesthetized with fentanyl, combined with propofol (propofol/fentanyl, septic). Anesthesia was induced with 0.05 mg/kg fentanyl, followed by 2 mg/kg propofol, given 5 min later. Anesthesia was maintained for 30 min with a continuous infusion of propofol at a rate of 2 mg centered dot kg^-1 centered dot h^-1.

4) Five sheep received a continuous infusion of live P. aeruginosa (2.5 times 10⁶ CFU/min) for more than 48.5 h. These sheep were neither anesthetized with propofol nor with fentanyl/propofol (unanesthetized, septic).

Depth of anesthesia was satisfactory in all sheep, as judged by muscular relaxation, ability to tolerate endotracheal intubation without swallowing, tolerance of mechanical ventilation, and withdrawal to a moderate pain stimulus (placement of a surgical clamp between the hoofs of the right front leg).

Hemodynamic measurements were taken at baseline, after 48 h of sepsis and 30 min after induction of anesthesia. At the same time points, renal blood flow was measured by the use of the ultrasonic flow probes, and colored microspheres were injected. Sheep in the propofol, nonseptic group were studied only at baseline and after 30 min of anesthesia. Sheep in the unanesthetized, septic group were studied at baseline, after 48 h of sepsis and after 48.5 h of sepsis. The vascular resistance in the kidney cortex (cortical vascular resistance [CVR]) was calculated as the ratio of MAP to cortical blood flow measured by the microsphere technique (mm Hg centered dot mL^-1 centered dot min^-1).

After completion of the study, all sheep were anesthetized with ketamine and received a lethal injection of KCl. At necropsy, both kidneys were removed. Samples of the renal cortex and medulla were taken from both kidneys and were sent together with the reference blood samples to Interactive Medical Technology (Los Angeles, CA) for analysis of renal blood flow.

Statistical analysis between groups was performed by the use of factorial analysis of variance. Within groups, an analysis of variance for repeated measures with post-hoc Dunnett's test was used. Statistical significance was accepted at P <or=to 0.05. Data are presented as mean +/- SEM. Regional blood flows are expressed as percentage of baseline.

Results

In one sheep (propofol/fentanyl, septic), no renal blood flow data were available. In another sheep (propofol, septic), blood flows to the left and right kidney cortex and medulla were quite different (>15%). Both sheep were excluded from further analysis.

Hemodynamic and oxygen transport data are summarized in Table 1; renal blood flow and renal vascular resistance data are summarized in Table 2. Anesthesia with propofol in nonseptic sheep produced only minor hemodynamic changes.

Table 1: Hemodynamics and Oxygen Transport

Table 2: Blood Flow and Vascular Resistance in the Kidney

All sheep treated with a continuous infusion of P. aeruginosa developed a hyperdynamic state of sepsis, characterized by a significant elevation in cardiac output and decrease in MAP, resulting in a marked reduction in systemic vascular resistance index (SVRI). There was a significant reduction in oxygen extraction. Oxygen consumption thus failed to match the probably elevated oxygen demand during sepsis, although oxygen delivery was markedly elevated.

In sheep treated with P. aeruginosa, but not with any anesthetic drug, no hemodynamic changes occurred between 48 and 48.5 h. These sheep showed a continuation of the hyperdynamic state of sepsis.

Anesthesia with propofol in this septic state caused hemodynamic deterioration. Cardiac output decreased to baseline levels. Simultaneously, MAP, which was already lowered during the course of sepsis, decreased even further. The SVRI remained unchanged. The decrease in oxygen delivery was not accompanied by an increase in oxygen extraction, leading to a marked reduction in oxygen consumption.

When propofol was combined with fentanyl, these hemodynamic variables were significantly less affected. Cardiac output remained above baseline level. MAP and SVRI, as well as oxygen consumption, remained unchanged.

Renal blood flow was measured in all groups with ultrasonic flow probes and colored microspheres. Healthy, anesthetized sheep (propofol, nonseptic) showed a nonsignificant reduction in renal blood flow (82% +/- 13%). Cortical and medullary blood flow, as well as CVR, showed similar minor changes. Sepsis led to a significant increase in renal blood flow when measured with ultrasonic flow probes. Cortical renal blood flow measured with colored microspheres showed an increase as well, but failed to reach significance Table 2. Medullary renal blood flow was not affected by sepsis. Anesthesia with propofol led to a marked and highly significant reduction in renal blood flow (60.3% +/- 9.6% of baseline; 38.6% +/- 4.3% of value at 48 h of sepsis). The observed reduction was similar when measured with ultrasonic flow probes or with colored microspheres and was accompanied by an increase in CVR to 148.0% +/- 19.1% of preanesthetic septic value (48 h of sepsis). In contrast to renal blood flow, cardiac output was not reduced below baseline. The combination of propofol with fentanyl prevented a decrease of renal blood flow below baseline. However, renal blood flow was reduced to a greater extent than cardiac output (P = 0.025). Unanesthetized, septic sheep showed no change in renal blood flow.

Discussion

Anesthesia in septic subjects is potentially harmful, because anesthetics may induce further vasodilation and myocardial depression. Propofol, a safe anesthetic in healthy subjects, has not been evaluated in sepsis. Propofol may exacerbate hemodynamic deterioration by producing additional vasodilation and myocardial depression. On the other hand, propofol is known to inhibit NOS ^[5]. This enzyme is found in the body in constitutive and inducible forms. Sepsis stimulates synthesis of the inducible form. As a result, NO is produced in large amounts during sepsis and is one of the main mediators of septic vasodilation. Under these circumstances, the ability of propofol to inhibit NOS may prevent further vasodilation. Combining it with fentanyl may further reduce adverse hemodynamic effects of propofol ^[11]. Fentanyl increases the free plasma fraction of propofol ^[12]; thus, the dosage of propofol needed to maintain anesthesia can be markedly reduced. Finally, fentanyl is a useful adjunct for propofol, because propofol is a poor analgesic.

The induction of a sustained hyperdynamic state of ovine sepsis by the continuous infusion of live P. aeruginosa is described in detail elsewhere ^[13]. The sheep develop all major signs of sepsis known to occur in humans, such as a hyperdynamic circulation, a reduction in myocardial performance with concomitant increase in heart rate, reduced renal function, and fever.

In all groups, a satisfactory depth of anesthesia was achieved by the chosen doses of propofol and fentanyl. We determined these doses from experience in our laboratory in producing anesthesia in sheep with these drugs. The induction dose of 10 mg/kg propofol is relatively high, compared with the dose required to produce anesthesia in humans. However, the required induction dose of propofol is known to be species-dependent ^[14] and may be as high as 15 mg/kg in dogs ^[15]. Runciman et al. ^[16] have reported that healthy, Merino sheep required 8 mg/kg for induction and a maintenance dose of 12 mg centered dot kg^-1 centered dot h^-1. Their dosage is similar to that established in our laboratory. Septic sheep required significantly less propofol for anesthesia maintenance, especially in combination with fentanyl, so that we reduced the maintenance dosage in these animals to 5 and 2 mg centered dot kg^-1 centered dot h^-1, respectively.

Propofol in healthy subjects reduces MAP and cardiac output ^[1]. The reason for this cardiovascular depression is a direct vasodilation ^[17], an increased vagal tone, and a sympatholytic action of the anesthetic ^[18]. Runciman et al. ^[16] reported an increase in MAP during anesthesia with propofol in sheep. This difference between studies could not be explained by inadequate or light anesthesia. In our experiments, MAP was reduced in septic, but not in healthy sheep, indicating a lack of compensation in sepsis. The latter may be related to a myocardial depression during ovine sepsis ^[19].

The entire hemodynamic situation deteriorated in septic sheep anesthetized with propofol. In contrast to healthy sheep, sheep given propofol had a significant decrease in cardiac output and marked hypotension beyond the hypotension associated with sepsis, per se. This hypotension is a serious side effect of propofol anesthesia and could not be demonstrated in healthy sheep.

In septic sheep that received a combination of propofol and fentanyl, hemodynamics were also depressed, but did not deteriorate to the extent seen in the sheep receiving propofol alone. This is in contrast to the findings of van Aken et al. ^[20], who found a further myocardial depression when fentanyl was added to propofol. However, this group gave propofol immediately after fentanyl, not waiting for fentanyl to achieve its maximum anesthetic effect, and they did not reduce the dose of propofol. When propofol was given three minutes after the fentanyl bolus, neither beneficial nor harmful effects could be observed in another study ^[21]. Propofol given with a five-minute delay after a large bolus of fentanyl improved hemodynamic stability in septic sheep, as shown in our study. This effect is most likely related to the fact that the large dose of fentanyl achieved a significant effect after five minutes and lasted for the experimental anesthesia period of 30 minutes, thus allowing us to reduce the dosage of propofol needed to maintain adequate anesthesia. This finding is in concert with other studies that have shown an improvement in hemodynamic stability in healthy subjects because of the addition of fentanyl to an anesthetic regimen ^[11].

The combination of fentanyl and propofol also seems to be beneficial in terms of oxygen consumption, because it reduced oxygen consumption less than propofol alone. This reduction in oxygen consumption may be caused by a decrease in nutritive blood flow or an increase in precapillary shunting, both possibly causing tissue hypoxia. Assuming that oxygen consumption correlates with outcome in sepsis, a significant reduction of oxygen consumption by propofol seems to be another crucial side effect, although the reduction in oxygen consumption in our study is less than others have demonstrated ^[16]. However, the decrease in oxygen consumption may be related to a reduced oxygen demand rather than a reduced oxygen extraction and thus may even be beneficial to organs at risk for hypoperfusion. Further studies are needed to characterize this phenomenon.

Sepsis as a general inflammatory process often results in multiorgan failure, especially kidney malfunction. Propofol, although reported to affect hemodynamics by negative inotropy and vasodilation, is thought to have a minimal effect on renal blood flow in healthy subjects ^[9,16]. Like any other anesthetic compound, propofol may decrease renal function ^[8,22]; however, in healthy subjects, the depression in renal blood flow appears to be less than that during anesthesia with volatile anesthetics ^[8].

Although propofol combined with fentanyl led to a reduction in renal blood flow when measured with an ultrasonic flow probe or colored microspheres, it did not reduce renal blood flow below preseptic levels. When propofol alone was given to septic sheep, renal blood flow decreased to 39% +/- 4% of the blood flow after 48 hours of sepsis (equivalent to 60% +/- 10% of baseline blood flow). This reduction in renal blood flow cannot only be related to the reduction in cardiac output, because cardiac output decreased significantly less than did renal blood flow. This selective reduction in renal blood flow is most likely caused by the extraordinary sensitivity of the renal microvasculature to NO; reduced perfusion pressure plays an important role as well. The impact of this selective reduction in renal blood flow on renal function could not be analyzed within the short anesthesia period of 30 minutes, but should be part of further investigations. The combination of propofol with fentanyl prevented this marked reduction in renal blood flow, most likely because of the reduced maintenance dosage of propofol used under this anesthetic regimen. However, when given with fentanyl, propofol caused a decrease in renal blood flow when compared with the septic state alone. In nonseptic sheep anesthetized with propofol alone, no reduction in renal blood flow occurred. Petros et al. ^[23] showed that propofol increases the production of NO only when endothelial cells are present. Consequently, the constitutive form of NOS, which is responsible for the regulation of the basal vascular tone under physiologic conditions, is activated when propofol is given to healthy subjects, thus preventing renal blood flow from decreasing. In sepsis, however, the constitutive form of NOS is inactivated ^[24,25], but the inducible form of NOS is activated and responsible for the tremendous production of NO during sepsis. Recently, O'Donnell et al. ^[26] found that propofol in vitro stimulates the constitutive form of NOS, but blocks the inducible form of NOS. This would explain the in vivo results obtained in our study, wherein propofol caused renal vasoconstriction in septic but not in nonseptic sheep. Because of its high sensitivity to NO, the renal microvasculature is ideal for verifying in vivo the results obtained by O'Donnell et al. in vitro.

In summary, propofol proved to be a safe anesthetic when used in healthy sheep, but showed potentially harmful side effects when given to septic sheep. Sedation or anesthesia with propofol alone in septic patients should be carefully evaluated, since the reduction in renal blood flow in combination with the reduced MAP may result in renal failure. Furthermore, the marked and significant reduction in oxygen consumption may be harmful, especially in septic patients, because an anesthesia-related reduction in oxygen demand, which occurs in healthy individuals, may not occur during the general inflammatory response of sepsis. The combination of propofol with fentanyl reduced all of these side effects during sepsis.