During pregnancy, the pharmacokinetics of drugs, including local anesthetics, may be affected by changes in body fluid volume and composition, as well as by changes in hemodynamics. Limited data are available regarding the effects of pregnancy on the disposition of bupivacaine, currently the most commonly used local anesthetic in obstetrics. In the only study of which we are aware, there were no significant differences between nonpregnant and pregnant women in the maximum serum concentration, terminal elimination half-life (T1/2beta), or area under concentration-time curve (AUC) after epidural injection of the drug . Distribution volumes and clearance were not determined.
In view of concerns regarding the well known cardiotoxicity of bupivacaine, an alternative-ropivacaine-is being investigated. In our previous study examining the effects of gestation on the pharmacokinetics of ropivacaine, the volume of distribution during the terminal phase of drug elimination (Vdbeta) and during steady state (Vdss) tended to be lower and total body clearance (CL) was significantly lower in pregnant compared with nonpregnant sheep . These gestational changes could be expected to lead to higher blood concentrations of ropivacaine during pregnancy, thus potentially increasing maternal risk of systemic toxicity. Yet, in a recent study conducted in nonpregnant and pregnant sheep given intravenous (IV) infusions of bupivacaine or ropivacaine to assess systemic toxicity, there were no differences in the doses or serum concentrations of ropivacaine (and bupivacaine) required to produce toxic manifestations, which appeared in sequence, viz. convulsions, hypotension, apnea, and circulatory collapse . Further, in pregnant ewes, larger doses of ropivacaine than bupivacaine were necessary to produce systemic toxicity; however, the serum concentrations of the two drugs at the onset of each toxic manifestation were similar.
There are no systematic studies comparing the effects of pregnancy on the pharmacokinetics of ropivacaine and bupivacaine. This information would be important because if a lower volume of distribution and CL during pregnancy were unique to ropivacaine, higher maternal blood concentrations and greater fetal drug exposure could result, compared with those expected with bupivacaine.
The purpose of the present study was to compare the effects of pregnancy on the disposition of ropivacaine and bupivacaine and to derive the pharmacokinetics of ropivacaine and bupivacaine in pregnant and nonpregnant sheep to determine whether these can explain the differences in systemic toxicity noted between the two drugs.
Twelve nonpregnant and 12 pregnant ewes near term (term 148 days) were studied according to a protocol approved by our institutional animal care and use committee. After an overnight fast, polyethylene catheters were inserted into the ewe's common carotid artery and jugular vein via a neck cutdown performed under general endotracheal anesthesia using 2%-3% halothane in oxygen. Antibiotics were administered postoperatively.
At least 2 days were allowed for recovery. On the day of study, the ewe was weighed and contained in a cart with freedom to stand or lie down. Arterial blood pressure and heart rate (cardiotachometer) were recorded on a polygraph. After a control period of 30-60 min, each ewe received ropivacaine or bupivacaine, 6 micro mol/kg (approximately 2 mg/kg), by IV infusion over 15 min. The two drugs were given in random sequence on two separate occasions 48 h apart. This random cross-over design was used because it conserved resources by reducing the number of animals required and because previous studies demonstrated that there is no residual local anesthetic when given 48 h apart [2,4]. The investigators administering the drug were blinded to its identity. Since convulsions or alterations in physiologic variables may affect drug kinetics, an infusion rather than a bolus injection was administered to avoid systemic toxicity. Arterial blood samples were obtained prior to and at the end of infusion and at 1, 2, 5, 10, 15, 30, 60, 90, 120, 150, 180, 240, and 300 min thereafter. Arterial blood pH and gas tensions were determined in duplicate samples (anticoagulated with heparin) obtained prior to and at the end of infusion. All other blood samples were allowed to clot, and serum was separated after centrifugation and frozen at -20 degrees C until drug analysis was performed within 2 wk of collection. Contact with polystyrene plastic or stoppers containing trisbutoxyethyl phosphate ester plasticizer was avoided. Protein binding was determined using an ultrafiltration system (Centrifree [R]; Amicon, Bedford, MA, with YMT membranes) in aliquots of serum obtained at the end of infusion and at 15 and 60 min thereafter and also in serum obtained prior to drug exposure, to which ropivacaine or bupivacaine was added to achieve a concentration of 1 micro g/mL. 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 [5,6]. Serum pH was first adjusted with microliter quantities of 0.1 N hydrochloric acid or sodium hydroxide so that it was equal to the blood pH determined at the approximate time of sampling. Serum was centrifuged for 45 min at 2000g. Previous studies have shown that equilibrium dialysis of serum and this ultrafiltration technique produce comparable results and that there is no binding of local anesthetics to the membrane .
Drug concentrations in serum and serum water were determined using gas chromatography . The assay was calibrated to measure concentrations of either local anesthetic as low as 0.02 micro g/mL. The day to day coefficient of variation was typically <5% at 1.00 micro g/mL and <10% at 0.05 micro g/mL throughout the study. The proportion of bound drug was calculated from drug concentrations determined in serum and serum water.
A two-compartment open model was used to describe the serum concentration data. Fitting data to a three-compartment model did not provide additional useful information. Initial estimates of the elimination rate constant (beta) were determined by linear regression analysis, and using the method of residuals , values for the distribution rate constant (alpha) were derived. Initial estimates of the volume of central compartment (Vc) and microconstants (K12, K21, K10) were determined by using standard equations , using a correction factor for infusion . The AUC was determined using the trapezoidal rule. The AUC extrapolated beyond the last sampling interval was calculated by dividing the concentration at 300 min by the slope of the terminal portion of the drug concentration time curve and adding it to the AUC. Pharmacokinetic data were generated from these estimates with a computer program employing an opportunistic, nonlinear regression method (SIMPLEX) based on the algorithm of Nelder and Mead . Serum concentrations of the drug were weighted according to the method of Ottaway , which conferred greater weight to the lower concentrations during the elimination phase.
Differences in blood pressure and heart rate were compared using repeated measures analyses of variance, and Student's t-test for paired data was used to evaluate changes in acid-base status and blood gas tensions. Differences in pharmacokinetic variables were evaluated using an univariate factorial analyses of variance model, which incorporated the particular drug used and the reproductive status of the ewe (sheep nested within reproductive status) as main effects. Because every animal was exposed to each local anesthetic in random order, the type of drug used was considered a random cross-over factor . The statistical significance of the two main effects and their interaction was evaluated with the general linear model feature in the statistical software package Systat [R] (Version 6.0 for Windows [R]; SPSS, Inc., Chicago, IL). P < 0.05 was considered to be statistically significant. All results are expressed as the mean +/- SE.
Pregnant ewes were studied between 130 and 135 days' gestation. Their weights of 68.1 +/- 2.9 kg when given bupivacaine (Bup Pr) and 68.1 +/- 3.0 kg when given ropivacaine (Rop Pr) were significantly greater than those of corresponding nonpregnant ewes, which were 50.6 +/- 2.1 (Bup NP) and 50.2 +/- 2.1 kg (Rop NP), respectively (P < 0.05).
All animals were in good condition throughout the study (Table 1). There were no signs of systemic toxicity, not even at the time of highest circulating drug concentrations. The highest mean total serum concentration for both drugs was found at the end of infusion and was greater in pregnant (Rop Pr 2.79 +/- 0.21 and Bup Pr 2.85 +/- 0.21 micro g/mL) compared with nonpregnant animals (Rop NP 2.16 +/- 0.13 and Bup NP 2.18 +/- 0.17 micro g/mL) (P < 0.05) (Figure 1 and Figure 2). In each animal, the subsequent decline in total serum concentrations could be described by using a biexponential equation. By the end of the sampling period, ropivacaine concentrations had decreased to 0.04 +/- 0.01 (Rop NP) and 0.06 +/- 0.01 micro g/mL (Rop Pr), which were lower than those for bupivacaine (Bup NP 0.07 +/- 0.01 and Bup Pr 0.09 +/- 0.02 micro g/mL) (P < 0.05).
Derived pharmacokinetic indices are listed in Table 2. There were no significant differences between nonpregnant and pregnant ewes in the distribution half-life (T1/2alpha), the T1/2beta, and the Vc for either drug. For both drugs, the volume of distribution during the terminal phase of drug elimination (Vdbeta) and the volume of distribution at steady state (V (d)ss) were lower in pregnant than nonpregnant animals. The CLs of ropivacaine and bupivacaine were also lower in pregnant compared with nonpregnant sheep (Table 2).
There were significant differences in pharmacokinetics between the two drugs. In all animals, Vc, Vdbeta, and Vdss were greater and CL lower for bupivacaine than ropivacaine. As a result, the T1/2beta of bupivacaine in nonpregnant and pregnant sheep was longer than that of ropivacaine. The T1/2alpha of bupivacaine was also longer than that of ropivacaine in all animals. Mean residence times (MRT) of bupivacaine were longer than the MRT of ropivacaine.
Both drugs exhibited concentration-dependent protein binding in serum, as in all animals the proportion of drug bound at the highest measured serum concentration (coinciding with the end of drug infusion [Time 0]) was lower than that at 15 and 60 min thereafter (Table 3). The bound fraction for both drugs was lower in nonpregnant compared with pregnant ewes in samples obtained at the end of infusion and 60 min thereafter (Table 3). Compared with ropivacaine, a greater proportion of bupivacaine was bound to serum proteins in both nonpregnant and pregnant animals at the end of infusion and 15 min thereafter and also in the "spiked" control sample. Thus, at Time 0, unbound (free) serum concentrations of ropivacaine in nonpregnant and pregnant ewes were 0.66 +/- 0.4 and 0.59 +/- 0.05 micro g/mL, respectively, and greater than the corresponding concentrations for bupivacaine (0.48 +/- 0.04 and 0.39 +/- 0.03 micro g/mL, respectively) (P < 0.05).
Ovine pregnancy affects the protein binding and pharmacokinetics of IV administered ropivacaine and bupivacaine. For both drugs, the Vdbeta and Vdss were lower in pregnant than nonpregnant sheep. This is consistent with trends found in our earlier open-label study of ropivacaine using a relatively small number of animals . Gestation has an opposite effect on the pharmacokinetics of lidocaine; namely, Vdbeta and Vdss are greater during ovine pregnancy [14,15]. The reason for this difference between the potent long-acting amide local anesthetics and lidocaine is unclear. It could be related to the fact that, as in the current study, the serum protein binding of ropivacaine and bupivacaine (at Time 0 and 15 min) is greater in pregnant than nonpregnant animals. In contrast, although the serum protein binding of lidocaine has not been studied during ovine pregnancy, gestation in women was associated with lower protein binding of the drug . It is possible that ropivacaine and bupivacaine, being more lipid-soluble and protein-bound, have greater affinity for extravascular tissue binding sites also occupied by the steroidal hormones of pregnancy. Competition between these hormones and oxyphenbutazone, an antiinflammatory drug, has been reported . A smaller Vd during pregnancy has also been demonstrated for other drugs such as ritodrine .
Despite the aforementioned differences in Vdbeta and Vdss between nonpregnant and pregnant animals in the present study, their relationship with CL was such that the T1/2beta of neither drug was altered by pregnancy. Similarly, the T1/2beta was not altered by pregnancy in our earlier investigations of lidocaine and ropivacaine [2,15]. In the current study, CL of both ropivacaine and bupivacaine was lower in pregnant than nonpregnant ewes. This may be due to pregnancy-related inhibition of hepatic biotransformation, as both progesterone and estradiol are competitive inhibitors of hepatic microsomal oxidases . Estrogen-induced cholestasis may also impair hepatic drug elimination . Indeed, hydroxylation of bupivacaine but not its further conjugation is inhibited by pregnancy, whereas N-dealkylation is enhanced . A reduction in CL during pregnancy has been reported for other drugs, such as ritodrine and caffeine [18,21]. It should also be noted that pregnant sheep have a greater hepatic blood flow than nonpregnant sheep . This results in an increased CL of lidocaine, a drug that has a high hepatic extraction ratio, but not of ropivacaine or bupivacaine, which have a low to intermediate extraction ratio (30%-50%).
There were also differences in the pharmacokinetics between the two local anesthetics. A greater Vd and lower CL for bupivacaine would explain why its T1/2beta and MRT were approximately 30% longer than those of ropivacaine. The greater lipid solubility of bupivacaine may, in part, be the reason for the differences noted between the two drugs . A shorter T1/2beta and MRT of ropivacaine may explain why during continuous IV infusion to pregnant sheep, greater doses of ropivacaine (compared with bupivacaine) were required to produce toxic manifestations while the serum concentrations of both drugs were similar .
Datta et al.  have recently reported minimum differences in peak plasma concentrations for ropivacaine or bupivacaine after epidural administration for cesarean section but, consistent with our findings, they found bupivacaine to have a much longer apparent elimination half-life. However, pharmacokinetic studies of the two drugs after epidural administration may not be reliable for two reasons. The lipid solubility of ropivacaine is much less than that of bupivacaine . Epidural fat may act as a depot for bupivacaine more than it does for ropivacaine, thus facilitating systemic absorption of ropivacaine . On the other hand, ropivacaine decreases epidural blood flow, whereas bupivacaine acts as an epidural vasodilator . The results of our current study involving IV administration clearly indicate that the longer T1/2beta of bupivacaine (compared with ropivacaine) is due to a larger Vdbeta and slower CL for bupivacaine.
Our protein binding data for pregnant sheep are in close agreement with findings obtained in parturients. In the aforementioned study by Datta et al., the mean protein-bound fractions at delivery were 90% and 94% for ropivacaine and bupivacaine, respectively . Unfortunately, no comparisons were made with nonpregnant patients. The results of our current study indicate that the serum protein binding of ropivacaine and bupivacaine is greater in pregnant than nonpregnant ewes at serum concentrations that would be expected to occur during uneventful obstetrical epidural anesthesia. Furthermore, the proportion of ropivacaine bound to serum proteins was lower than that for bupivacaine in all animals. Thus, the free serum concentration of ropivacaine at the end of infusion (Time 0) was greater than that for bupivacaine. In contrast, in another study, there was no difference between nonpregnant and pregnant ewes in the protein binding of either drug at serum concentrations high enough to result in toxicity . This difference could be due to saturation of high-affinity, low-capacity protein binding sites at the higher concentrations, leaving only low-affinity, high-capacity sites available for drug binding.
With the exception of the control sample, serum protein binding was determined in individual ewes at predetermined intervals after drug administration rather than after adding the drug (spiking) to sera obtained from unexposed animals. Also, because the free fraction of amide local anesthetic may vary with pH , each individual protein binding determination was made at the blood pH actually measured in the ewe.
It is unlikely that metabolites of ropivacaine or bupivacaine affected protein binding in the current study since there was no significant difference in protein binding between the control (spiked) and in vivo samples at comparable serum concentrations of drugs. Further, both local anesthetics are metabolized slowly, and protein binding was determined in the initial 60 minutes after administration to the ewe. Similarly, in another study involving lidocaine, metabolites were not found to affect protein binding of the parent compound .
Protein binding may also be increased by major surgery in sheep. Rutten et al.  demonstrated that after thoracotomy, protein binding of drug did not return to control until five to seven days postoperatively. It is unlikely that a neck cutdown performed in 15 minutes could have altered protein binding of either local anesthetic in the current study. Regardless of any residual surgical effects, all animals were treated in exactly the same manner, thus allowing for valid intergroup comparisons.
In conclusion, the pharmacokinetics of both ropivacaine and bupivacaine are altered during ovine pregnancy in a similar way. If these data are applicable to humans, an unintended intravascular injection of either drug could be expected to result in higher peak total serum concentrations in the pregnant than nonpregnant patient, but the drug levels would dissipate at similar rates in both individuals. Thus, the pharmacokinetic alterations occurring during pregnancy do not confer a specific advantage of one drug over the other with regard to unintended IV injection. However, differences between the two drugs, particularly in T1/2beta and MRT, may make ropivacaine more desirable for use in obstetric anesthesia.
1. Pihlajamaki K, Kanto J, Lindberg R, et al. Extradural administration of bupivacaine: pharmacokinetics and metabolism in pregnant and non-pregnant women. Br J Anaesth 1990;64:556-62.
2. Santos AC, Pedersen H, Sallusto JA, et al. Pharmacokinetics of ropivacaine in nonpregnant and pregnant ewes. Anesth Analg 1990;70:262-6.
3. Santos AC, Arthur GR, Wlody D, et al. Comparative systemic toxicity of ropivacaine and bupivacaine in nonpregnant and pregnant sheep. Anesthesiology 1995;82:734-40.
4. Gill JL. Design and analysis of experiments in the animal and medical sciences. Vol. 2. Ames, IA: Iowa State University Press, 1978:301.
5. Brown JE, Kitchell BB, Bjornsson TD, Shand DG. The artifactual nature of heparin-induced drug protein-binding alterations. Clin Pharmacol Ther 1981;30:636-43.
6. Ridd MJ, Brown KF, Moore RG, Nation RL. Drug plasma binding and non-esterified fatty acids: methodologic considerations. Int J Pharmac 1982;11:11-20.
7. Arthur GR, Denson DD, Coyle DE. Effect of sodium thiopental on the serum protein binding of bupivacaine. Reg Anesth 1984;9:171-4.
8. Nancarrow C, Runciman WB, Mather LE, et al. The influence of acidosis on the distribution of lidocaine and bupivacaine into the myocardium and brain of the sheep. Anesth Analg 1987;66:925-35.
9. Gibaldi M, Perrier D. Pharmacokinetics. New York: Marcel Dekker, 1975:281-92.
10. Gibaldi M, Perrier D. Pharmacokinetics. New York: Marcel Dekker, 1975:55.
11. Loo JCK, Riegelman S. Assessment of pharmacokinetic constants from postinfusion blood curves obtained after i.v. infusion. J Pharm Sci 1970;59:53-5.
12. Nelder JA, Mead R. A simplex method for function minimization. Comput J 1965;7:308-13.
13. Ottaway JH. Normalization in the fitting of data by iterative methods: application to tracer kinetics and enzyme kinetics. Biochem J 1973;134:729-36.
14. Bloedow DC, Ralston DH, Hargrove JC. Lidocaine pharmacokinetics in pregnant and nonpregnant sheep. J Pharm Sci 1980;69:32-7.
15. Santos AC, Pedersen H, Morishima HO, et al. Pharmacokinetics of lidocaine in nonpregnant and pregnant ewes. Anesth Analg 1988;67:1154-8.
16. Wood M, Wood AJJ. Changes in plasma drug binding and alpha1-acid
glycoprotein in mother and newborn infant. Clin Pharmacol Ther 1981;29:522-6.
17. Weiner M, Siddiqui AA, Shahani RT, Dayton PG. Effect of steroids on disposition of oxyphenbutazone in man. Proc Soc Exp Biol Med 1967;124:1170-3.
18. Caritis SN, Lin LS, Venkataramanan R, Wong LK. Effect of pregnancy on ritodrine pharmacokinetics. Am J Obstet Gynecol 1988;159:328-32.
19. Davis M, Simmons CJ, Dordoni B, et al. Induction of hepatic enzymes during normal human pregnancy. J Obstet Gynaecol Br Common 1973;80:690-4.
20. Harrison LI, Gibaldi M. Influence of cholestasis on drug elimination: pharmacokinetics. J Pharm Sci 1976;65:1346-8.
21. Aldridge A, Bailey J, Neims AH. The disposition of caffeine during and after pregnancy. Semin Perinatol 1981;5:310-4.
22. Rosenberg PH, Kytta J, Alila A. Absorption of bupivacaine, etidocaine, lignocaine and ropivacaine into n-heptane, rat sciatic nerve, and human extradural and subcutaneous fat. Br J Anaesth 1986;58:310-4.
23. Datta S, Camann W, Bader A, VanderBurgh L. Clinical effects and maternal and fetal plasma concentrations of epidural ropivacaine versus bupivacaine for Cesarean section. Anesthesiology 1995;82:1346-52.
24. Dahl JB, Simonsen L, Mogensen T, et al. The effect of 0.5% ropivacaine on epidural blood flow. Acta Anaesthesiol Scand 1990;34:308-10.
25. Denson DD, Coyle DE, Thompson GA, et al. Bupivacaine protein binding in the term parturient: effects of lactic acidosis. Clin Pharmacol Ther 1984;35:702-9.
26. McNamara PJ, Slaughter RL, Pieper JA, et al. Factors influencing serum protein binding of lidocaine in humans. Anesth Analg 1981;60:395-400.
27. Rutten AJ, Mather LE, Plummer JL, Henning EC. Postoperative course of plasma protein binding of lignocaine, ropivacaine and bupivacaine in sheep. J Pharm Pharmacol 1992;44:355-8.