Comparative Systemic Toxicity of Ropivacaine and Bupivacaine in Nonpregnant and Pregnant Ewes 

Twenty-four nonpregnant and 24 near-term pregnant ewes were studied under a protocol approved by the Institutional Animal Care and Use Committee.

After an overnight fast, polyethylene catheters were inserted into each ewe's common carotid artery and jugular vein via a neck cutdown performed under general endotracheal anesthesia with 2–3% halothane in oxygen. Antibiotics were administered postoperatively.

On the day of study (7–9 days after surgery), the ewe was weighed and contained in a cart with freedom to stand or lie down. (Each animal had been studied 2–4 days earlier to determine the effects of pregnancy on the pharmacokinetics of ropivacaine and bupivacaine. [16]) Arterial blood pressure and heart rate (cardiotachometer) were recorded continuously on a polygraph. Cardiac rhythm also was monitored using a transvenous intracardiac electrode placed percutaneously via a jugular vein on the day of study. Ewes were randomized to receive an intravenous infusion of either 0.5% bupivacaine hydrochloride or 0.5% ropivacaine hydrochloride at a constant rate of 0.5 mg *symbol* kg *symbol* sup -1 *symbol* min sup -1 until circulatory collapse. The investigators were blinded to the identity of drug infused. Arterial blood samples were obtained in duplicate before infusion and at the onset of toxic manifestations, which appeared in usual sequence: convulsions, hypotension, apnea, and circulatory collapse. Convulsions began at the onset of tonic-clonic seizures. Hypotension was defined as a precipitous decrease in blood pressure of at least 40% from levels that were increased during convulsion. Finally, the disappearance of pulsatile blood pressure indicated circulatory collapse. Blood pH and gas tensions were determined from arterial blood samples anticoagulated with heparin. Nonanticoagulated samples were allowed to clot, and serum was separated after centrifugation and frozen until drug analysis within 2 weeks of collection. Contact with polystyrene plastic or stoppers containing trisbutoxyethyl phosphate ester plasticizer was avoided. Serum rather than plasma was chosen to avoid the artifactual effects of in vitro lipolysis, which are particularly significant in the plasma of pregnant animals. [17,18] Serum pH first was adjusted with microliter quantities of 0.1 Nitrogen hydrochloric acid or sodium hydroxide to be equal to the blood pH determined at the time of sampling. Using an ultrafiltration system (Amicon Centrifree with YMT membranes), serum water was obtained from up to 1-ml aliquots of serum. Centrifugation was for 45 min at 2,000 g. Thereafter, drug concentrations in serum and serum water were determined using gas chromatography. [19] The assay was calibrated to measure concentrations of either local anesthetic as low as 0.02 micro gram *symbol* ml sup -1. Electrocardiograms were analyzed for the presence of malignant ventricular arrhythmias by one of the authors (P.D.A.), a cardiologist, blinded to the animal's group.

The mean doses and serum concentrations of drug (total and unbound) at the onset of each toxic manifestation were compared using analyses of variance factorial. Intergroup comparisons were made using Sheffe's F test. Analyses of variance for repeated measures (Sheffe's F test) were used to evaluate differences between groups for heart rate, mean arterial blood pressure, pH, and gas tensions at each toxic manifestation. Differences between groups in the incidence of malignant ventricular arrhythmias were evaluated using Fisher's exact test. For all analyses, a P value less than 0.05 was considered significant. All results are expressed as the mean plus/minus SE.

The range of gestational ages in the ropivacaine pregnant (Rop Pr) group was 130–138 days, and in the bupivacaine pregnant (Bup Pr) group 130–140 days (term being 148 days). The weights of pregnant ewes, 67.3 plus/minus 3.2 kg (Rop Pr) and 64.6 plus/minus 2.1 kg (Bup Pr), were greater than those of nonpregnant animals given ropivacaine (Rop NP) or bupivacaine (Bup NP), 47.6 plus/minus 1.6 and 49.6 plus/minus 2.2 kg, respectively (P < 0.001). All animals were in good condition before study, mean heart rate, arterial blood pressure, pH, and gas tensions being within normal limits for our laboratory (Table 1and Table 2).

The manifestations of systemic toxicity occurred in the majority of animals as anticipated: convulsions followed by hypotension, apnea and, finally, circulatory collapse. [3,4,7,15] In two sheep (one nonpregnant, one pregnant) given bupivacaine, apnea and circulatory collapse occurred simultaneously within 1 min after the onset of convulsions. There were no significant differences between nonpregnant and pregnant animals in the doses or serum concentrations of either drug required to elicit toxic manifestations (Figure 1and Figure 2). Seizure activity started at similar doses of ropivacaine and bupivacaine in nonpregnant animals, but in pregnant animals, a greater dose of ropivacaine than of bupivacaine was required to produce convulsions (7.5 plus/minus 0.5 vs. 5.0 plus/minus 0.6 mg *symbol* kg sup -1, respectively, P < 0.05;Figure 1). Serum concentrations of ropivacaine were similar to those of bupivacaine at the onset of convulsions in all four groups of animals (Figure 2). At the onset of seizure activity, heart rate and mean arterial blood pressure increased (Table 1), but arterial pH and gas tensions remained unchanged (Table 2). Pregnant animals given ropivacaine or bupivacaine had lower (20–30%) heart rates at the onset of convulsions compared to corresponding nonpregnant ewes (Table 1).

As infusion of drug continued, other manifestations of systemic toxicity became evident, terminating in circulatory collapse. In nonpregnant ewes, the mean drug doses at circulatory collapse were similar (11.6 plus/minus 1.0 mg *symbol* kg sup -1 for ropivacaine and 8.9 plus/minus 0.6 mg *symbol* kg sup -1 for bupivacaine; power of statistical analysis = 0.84, a posteriori), but in pregnant animals the lethal dose of ropivacaine was significantly greater than for bupivacaine (12.9 plus/minus 0.8 vs. 8.5 plus/minus 1.2 mg *symbol* kg sup -1, respectively, P < 0.05;Figure 1). The corresponding serum concentrations of ropivacaine were similar to those of bupivacaine in both nonpregnant and pregnant ewes (Figure 2). There were no significant differences among the groups in the ratios of the dosages or serum concentrations of local anesthetic resulting in circulatory collapse and those required for convulsive activity (CC/CNS ratios;Table 3).

Pregnancy did not affect the serum protein binding of either drug. With increasing serum concentrations and concomitantly increasing severity of toxic manifestations, there was a progressive decrease in the protein binding of both local anesthetics in nonpregnant and pregnant animals. The serum concentrations of free ropivacaine and bupivacaine at each toxic manifestation were similar in nonpregnant ewes (Figure 3). At apnea and circulatory collapse, there was a significantly greater free concentration of ropivacaine than of bupivacaine in pregnant animals (P < 0.001).

The proportion of animals having a malignant ventricular arrhythmia (ventricular tachycardia or fibrillation) as the terminal event was similar in all groups (P = 0.4). For technical reasons, cardiac rhythm could not be monitored in three experiments.

These results indicate that the systemic toxicity of ropivacaine is not enhanced by ovine pregnancy, but neither is that of bupivacaine. The doses and serum concentrations of both drugs associated with toxic manifestations were similar in nonpregnant and pregnant ewes. The data for ropivacaine are in agreement with our earlier open-label study indicating that the systemic toxicity of ropivacaine was not affected by pregnancy. [15] However, the current results with regard to bupivacaine are surprising because it was previously demonstrated that lethal doses and plasma concentrations were lower in pregnant than in the nonpregnant ewes. [3] The reasons for this discrepancy are unclear but may be related to the lack of blinding and the relatively small sample size in the earlier work. In the current study, the power of statistical analysis (0.84) was adequate to detect previously reported differences between nonpregnant and pregnant sheep in toxic doses and serum concentrations of bupivacaine. Although all animals in the current study were given both drugs intravenously 2–4 days earlier, [16] it is doubtful that this exposure affected the toxicity of either ropivacaine or bupivacaine because equal numbers of sheep in each toxicity group were randomized to receive one of the two drugs.

In nonpregnant ewes, similar doses and serum concentrations of ropivacaine and bupivacaine were associated with the onset of convulsions and circulatory collapse. In pregnant animals, the doses required to produce these toxic manifestations were approximately 40–50% greater for ropivacaine than for bupivacaine, but the corresponding serum concentrations of the two drugs were similar. This was probably due to a shorter elimination half-life and faster clearance for ropivacaine. [16].

We previously suggested that the greater toxicity of bupivacaine reported in pregnant as compared to nonpregnant ewes could be related to lower serum protein binding during pregnancy. [3,4,15] In contrast, neither the toxicity nor the availability of free mepivacaine or ropivacaine was found to be increased in pregnant ewes. [4,15] The current study demonstrated no differences between nonpregnant and pregnant ewes in the proportion of free bupivacaine or, for that matter, ropivacaine at the onset of toxic manifestations. There were important changes in the methodology used in the current study that may have enhanced the reliability of the current data. Previously, the pH of serum samples was adjusted to be physiologic for sheep (7.50), [4,15] whereas, in the current study, protein binding was determined at the actual blood pH measured at the onset of each toxic manifestation. This is important because the free fraction of amide local anesthetics is increased with acidosis. [20] Second, serum protein binding was measured in individual ewes involved in the toxicity study, rather than after adding drug to sera obtained from unexposed animals. Others have suggested that in vitro measurements of free bupivacaine substantially underestimate the concentration of circulating drug that is available for transfer into tissues in vivo. [21].

It is unlikely that metabolites of ropivacaine or bupivacaine altered protein binding of the parent drug in the current study. Both local anesthetics are metabolized slowly, and death occurred relatively quickly, usually within 30 min of infusion. Further, at least in the case of another amide local anesthetic, lidocaine, metabolites were found not to affect protein binding of the parent compound. [22] Finally, the surgical preparation was unlikely to affect serum protein binding of drug because studies were conducted at least 7 days later. [23] Regardless of any residual effects of surgery, all animals were treated in the same manner, thus allowing valid intergroup comparisons.

In vitro measurements of transmembrane potentials in ventricular muscle or Purkinje fiber preparations obtained from oophorectomized rabbits treated with progesterone (or vehicle) have shown that bupivacaine, but not lidocaine or ropivacaine, produces greater depression of V^maxin tissues obtained from progesterone-treated as compared to untreated animals. [24,25] The depression of V^maxwould predispose an animal to reentrant type ventricular arrhythmias. [24] The results of the current study challenge extrapolation of these observations to the intact animal because we found no difference in the incidence of malignant ventricular arrhythmias between nonpregnant and pregnant sheep exposed to ropivacaine or bupivacaine. There was also no difference between the two drugs both in nonpregnant and in pregnant sheep. The use of intracardiac rather than surface electrodes improved detection and resolution of cardiac rhythm during convulsions in the current study, and the electrocardiograms were interpreted by a cardiologist unaware of the animal's group. The similarity of ropivacaine and bupivacaine in the incidence of malignant ventricular arrhythmias noted in this study is contradictory to findings in dogs. [11] This discrepancy may in part be explained by method of drug administration (infusion vs. bolus) and species difference.

All ewes were in good condition before the start of the study. The lower heart rates recorded in pregnant animals at the onset of convulsions, regardless of drug allocation, could be due to attenuation of cardiovascular responses to adrenergic stimulation during pregnancy. [26] Changes in acid-base status and blood gas tensions during drug infusion were related to the progression of toxic manifestations in all animal groups.

The margin of safety, defined as the ratio of the dose (or serum concentration) causing circulatory collapse to the dose (or serum concentration) resulting in convulsive activity (CC/CNS ratio), was similar in all groups and consistent with earlier work from our laboratory. [3,15] However, using CC/CNS ratios to evaluate drug safety may be misleading. In the current study, although these ratios were similar for both drugs, greater doses and higher free serum concentrations of ropivacaine than bupivacaine were tolerated by pregnant sheep.

Extrapolation of these findings to the clinical situation of unintended intravascular injection of local anesthetic may be restricted by our methodology. Whereas, in our study, toxic manifestations were clearly delineated in the majority of animals, and circulatory collapse occurred after 20–30 min of infusion, in the reported clinical cases, [5,6] a bolus injection resulted in immediate signs of systemic toxicity, including cardiac arrest. However, it is currently recommended to fractionate the injected dose of local anesthetic so that induction of a block may require 15 min or longer. A slower rate of administration will allow more extensive redistribution of drug so that toxic manifestations may occur at larger doses but lower blood concentrations of drug. [27] Nonetheless, the current results are consistent with data obtained in dogs given bolus injections of local anesthetics. [11] In these animals, greater doses of ropivacaine, as compared to bupivacaine, were required to produce cardiovascular collapse. In conclusion, the systemic toxicity of ropivacaine or bupivacaine is not enhanced by gestation in sheep. Furthermore, greater doses of ropivacaine, as compared to bupivacaine, are needed to produce toxic manifestations in pregnant sheep.