Cesarean section is one of the most commonly performed operations in the United States. A survey of American hospitals delivering 1500 or more babies a year revealed that 29% of cesarean sections were performed under epidural anesthesia, 33% under spinal anesthesia, and 35% under general endotracheal anesthesia . A similar but more recent study showed that 56% of cesarean sections were performed under epidural anesthesia, 33% under spinal anesthesia, and only 12% under general endotracheal anesthesia1. Yet despite the increasing popularity of epidural anesthesia over general endotracheal anesthesia for cesarean section, published maternal local anesthetic pharmacokinetic data are limited [2-11]. Furthermore, most existing reports are based on venous rather than arterial maternal blood sampling [2-7]. Maternal venous drug concentrations do not necessarily represent arterial levels perfusing the brain, heart, and placenta [12-15].
(1) Gibbs CP. Obstetric anesthesia: USA. Proceedings of the 28th annual meeting of the Society for Obstetric Anesthesia and Perinatology. May 1-4, 1996, Tucson, AZ.
This study was designed to determine the maternal pharmacokinetics of epidural lidocaine and bupivacaine during cesarean section since these drugs are used routinely for obstetrics across the country. Efforts were made to perform the anesthetic procedure exactly as it would be conducted so that the pharmacokinetic data would reflect that likely to be observed in the clinical situation. Therefore, we investigated the maternal pharmacokinetics of epidural lidocaine 2% with 1:200,000 epinephrine and plain bupivacaine 0.5% at cesarean section using maternal (Ma) arterial blood sampling.
With internal review board approval and each patient's written informed consent, 20 pregnant women aged 17-38 yr at or near term and scheduled for cesarean section, were recruited to the study. Each patient ingested two effervescent antacid tablets (Alka Seltzer Gold[R], Miles Inc, Elkhart, IN) and received ranitidine 50 mg and metoclopramide 10 mg intravenously (IV) for acid aspiration prophylaxis.
An 18-gauge cannula was inserted IV and 750-1000 mL of balanced salt solution was infused. Patients were placed in the sitting position and prepared and draped in the usual sterile fashion. The epidural space was identified using loss of resistance to saline through a 17-gauge Tuohy epidural needle at either the L (2-3) or L3-4 interspace. A 20-gauge epidural catheter was placed 5 cm into the epidural space, after which the needle was removed and the catheter was secured. Upon completion of the procedure, the patient was positioned supine but with left lateral tilt on the operating table. Standard monitors were used during both the epidural placement and cesarean section. Arterial blood pressure was monitored every 2 min for the first 30 min after injection of the anesthetic into the epidural space and then every 5 min thereafter.
The dorsum of the wrist was infiltrated with 1-2 mL of 1% lidocaine. The radial artery was palpated and cannulated using a 20-gauge Teflon catheter (Jelco[TM]; Critikon, Tampa, FL). If the patient complained of pain or three attempts were made and failed, the study was aborted.
A 3-mL "test dose" of 1.5% lidocaine with epinephrine 1:200,000 was injected to test for unintended subarachnoid or intravascular misplacement of the epidural catheter. Either 5 mL of lidocaine 2% with epinephrine 1:200,000 (n = 10) or bupivacaine 0.5% (n = 10) were injected epidurally every 5 min until a T2-4 level of sensory anesthesia to pinprick was demonstrated. Patients also received epidural fentanyl 1 micro g/kg 10-15 min before incision.
Ma blood samples (3 mL) were drawn at the following intervals: 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 90, 120, 150, and 180 min after epidural injection of the lidocaine "test dose." Ma and umbilical vein (Uv) blood samples were drawn immediately after birth. Blood samples were centrifuged, decanted, frozen, and stored for later analysis. Patient demographic data, induction to delivery intervals, and estimated blood loss were recorded for each case.
Plasma lidocaine concentrations were determined using a gas-liquid chromatographic method as previously described  substituting bupivacaine as the internal standard. Extracted samples were injected into the gas-liquid chromatograph (Model 3700, Varian Associates, Palo Alto, CA.), equipped with an SPB-1 fused capillary column (Supelco, Bellefonte, PA) and a thermionic specific detector. Peak heights were obtained using an integrator (Model 4290, Varian Associates).
Plasma bupivacaine concentrations were measured by high-performance liquid chromatography (HPLC) according to the method described by Ha et al.  using butorphanol as the internal standard. A variable wavelength ultraviolet detector (Model 486, Waters Associates, Milford, MA) set for 210 nm was used in the chromatography. The HPLC column (micro Bondapak C18, Waters Associates), was equilibrated with a mobile phase of acetonitrile and 0.05 M Na2 HPO4 (30:70 vol/vol, pH 5.8) at 1 mL/min.
Lidocaine and bupivacaine concentrations were ascertained from peak-height ratios with the internal standard and subsequent comparisons to a calibration curve run concurrently with known lidocaine or bupivacaine concentrations. Calibration curves for lidocaine and bupivacaine were linear from 0.1 to 10.0 micro g/mL and exhibited interday coefficient of variation of 1.48 and 4.1, respectively (correlation coefficients, r < 0.99).
Estimates of the beta elimination rate constant and y intercept were determined from the terminal phase data points by linear regression analysis of the logarithm of the plasma concentrations and time. The total area under each curve was estimated by the trapezoidal rule, permitting calculation of the apparent plasma clearances of lidocaine and bupivacaine from the ratio of the administered doses to the areas under the plasma/time curve. Elimination half-lives (t 1/2beta) and apparent volumes of distribution were calculated using a noncompartmental model using standard pharmacokinetic formulae .
Local anesthetic protein binding was determined using a Spectrum Equilibrium Dialyzer (Spectrum Medical Industries. Inc., Los Angles, CA.) equipped with Spectrapore 2 dialysis tubing with a molecular cutoff of 12,000 daltons. Protein binding estimates were obtained by dialyzing a portion of each subject's 30-min arterial blood sample against phosphate buffer at pH 7.4 for 3 h. Concentrations on both sides of the membrane were measured using HPLC or gas liquid chromatography as described earlier. By using plasma from this time point no additional local anesthetic needed to be introduced as enough endogenous drug was available for the determination.
Individual results and, where appropriate, mean +/- SEM are presented. Correlations between various kinetic variables were assessed using regression analysis. Single measurement comparison between groups (e.g., demographic data) was performed using unpaired t-test. Repeated measures analysis of variance with post hoc Bonferroni corrected t-test was used to determine intragroup and intergroup differences for the longitudinal (e.g., hemodynamic) data. Fisher's exact test was used for nonparametric data. A value of P <or=to 0.05 was considered significant.
The patients studied represent a homogeneous group with regard to height and weight, without statistical difference between the lidocaine and bupivacaine groups (Table 1). There was a statistical difference in age between the groups (P <or=to 0.05).
No longitudinal differences in heart rates or blood pressures were detected by repeated measures analysis of variance (Table 2). IV ephedrine was administered to 7 of 10 lidocaine and 5 of 10 bupivacaine patients (not significant, Fisher's exact test). Patients treated for hypotension in the lidocaine group required significantly (P <or=to 0.05) more IV ephedrine than their bupivacaine counterparts (37.5 +/- 5.0 vs 16.0 +/- 3.0 mg., respectively). Estimated blood loss exceeded 500 mL in 5 of 10 and 8 of 10 patients in the lidocaine and bupivacaine groups, respectively (not significant, Fisher's exact test). No patient lost more than 1000 mL of blood. Durations of surgery for the lidocaine and bupivacaine cohorts were statistically similar (56 +/- 4.3 and 53 +/- 3.5 min, respectively).
The pharmacokinetic data observed for lidocaine and bupivacaine are presented in Table 1 and Table 3. Mean Ma lidocaine and bupivacaine concentrations peaked at 30 and 40 min, respectively, and then rapidly declined in a logarithmic fashion (Figure 1). Linear regression analyses were performed between weight adjusted dosage and the peak Ma local anesthetic concentrations (Figure 2). Relationships were not statistically significant for either drug. However, if an outlier value is excluded from the lidocaine group (Patient E, the bracketed symbol in Figure 2, left), statistical significance is obtained (P <or=to 0.05). No correlation was found between peak Ma lidocaine or bupivacaine concentrations and either patient height (r = 0.05 and 0.24, respectively) or weight (r = 0.33 and 0.23, respectively).
Previous attempts to define maternal local anesthetic pharmacokinetics at cesarean section have been based on venous blood sampling [2-7]. Some of these investigators simply pooled multiple-patient, single-point determinations of maternal plasma local anesthetic concentrations from which they extrapolated the pharmacokinetic data . Studies of IV rather than epidural local anesthetic administration have been performed but failed to account for the role of local anesthetic transport out of the epidural space into the blood stream in a two-compartment model . Researchers who did measure arterial local anesthetic concentrations after epidural administration did not use their results to calculate pharmacokinetic data [8-11].
Are there differences in local anesthetic concentrations between the arterial and venous circulations? Tucker and Mather  state that "… large arteriovenous differences in the concentrations of local anesthetics have been observed. Under these dynamic conditions, the time profile of drug concentration in arterial blood will peak earlier than at targets for toxicity in vital organs." Furthermore, "Drug concentrations in peripheral venous blood peak later than those in the brain, particularly if samples are taken from vessels draining a poorly perfused region without significant shunt flow. The significance of arteriovenous drug concentration differences will depend on the rate of change of drug concentration." In the presence of a high thoracic sympathetic block, forearm and hand venous circulation may be sluggish due to compensatory vasoconstriction. Under these circumstances, upper limb arteriovenous local anesthetic concentration differences could be further accentuated. For these reasons, we concur with the finding of others [13-15] that venous concentrations may not accurately reflect arterial concentrations perfusing the brain, heart, and placenta.
The arterial concentrations of local anesthetics are very important, since they can cause seizures and cardiac arrest [19,20]. Circulating lidocaine concentrations that are toxic to animals have been defined . Similar measurements for bupivacaine are scarce . For ethical reasons, analogous human data are lacking. The Ma plasma concentration is also an important determinant of local anesthetic transfer across the placenta . This may be of special concern to the preterm as well as the growth restricted fetus rendered acidemic by labor-related stress .
After an extensive literature review, we believe the present study to be the first to derive maternal pharmacokinetic data for lidocaine and bupivacaine from sequential maternal arterial rather than venous blood sampling. Ma lidocaine concentrations peaked at 31 +/- 2.3 minutes. Serum concentrations in two patients (G and J) given less than the average dose of epidural lidocaine (6.1 +/- 0.6 mg/kg) peaked within 20 minutes of injection. Two other parturients (A and E) attained peak serum levels in excess of those considered toxic to animals. [20,21,24]. No adverse maternal or fetal responses were noted. Presumably the duration of exposure of the brain, heart, and placenta to high peak Ma lidocaine concentrations in these two cases was too brief to cause harm to either mother or child. Alternatively, vital organ concentrations were lower than the Ma levels and therefore not organ toxic.
Braid and Scott  demonstrated a poor correlation between peak venous concentrations of lidocaine and body mass in nonpregnant subjects. Data obtained here support this finding as peak Ma lidocaine concentrations were unrelated to the weight adjusted dosage, even considering when exclusion of the single outlier (Patient E) thus "forcing" the data to suggest that there is significant correlation between the weight-adjusted dose and the maximum arterial lidocaine concentration achieved. Due to the large interpatient variability in the measured the peak arterial lidocaine concentrations, anesthesia care providers probably should not rely on the milligram per kilogram dose of lidocaine given as a reliable predictor of maximum arterial lidocaine concentration that will ultimately result.
Individual lidocaine pharmacokinetic variables in this small population of healthy, pregnant women varied widely (Table 1). Of special interest are the Uv/Ma concentration ratios which were comparable to other reports . These differences in fetal/maternal ratio of local anesthetic concentrations are not unexpected considering that induction to delivery times (lidocaine: 35-74 minutes, bupivacaine: 41-56 minutes), and therefore Ma concentrations, varied greatly.
Ma bupivacaine concentrations peaked at 40.5 +/- 1.7 minutes at levels well below those considered toxic (Table 2). Maximum bupivacaine concentrations did not correlate with weight-adjusted dosage. As with lidocaine, calculated pharmacokinetic variables for bupivacaine varied widely. Uv/Ma ratios differed markedly, again reflecting differences in induction to delivery intervals and the resulting dissimilar Ma drug concentrations at birth. Uv/Ma ratios were similar to those reported previously for bupivacaine .
Other factors, such as fetal pH, can influence fetal/maternal concentration ratios. Cord blood gas analyses were not performed in this study, and therefore fetal pH influences cannot be excluded. However, in light of the elective nature of the surgery and the excellent fetal outcomes, as reflected by Apgar scores, both maternal and fetal pH were assumed to be within normal limits.
These variations can also be due to pregnancy and anesthesia-related factors, for example, competition between local anesthetics and the hormones of pregnancy for hepatic enzyme degradation, fluctuations in maternal organ perfusion invoked by epidural sympathectomy, the effects of seven or more hours of fasting on total body fluid dynamics during pregnancy, and the influence of antacid ingestion on maternal acid base balance.
In summary, neither lidocaine nor bupivacaine peak Ma concentration correlated with their weight-adjusted dosage, although with data exclusion in the case of lidocaine a correlation could be forced. Ma lidocaine concentrations exceeded toxic levels in two patients, but we observed no adverse maternal or fetal effects. Peak Ma bupivacaine concentrations remained well below toxic levels. Derived pharmacokinetic data for both epidurally administered local anesthetics exhibited great interpatient variability. The results suggest that epidural bupivacaine reliably produces acceptable Ma concentrations below the toxic range.
1. Gibbs CP, Krischer J, Peckham BM, et al. Obstetric anesthesia: a national survey. Anesthesiology 1986;65:298-306.
2. Lund PC, Cwik JC. A correlation of the differential penetration and the systemic toxicity of lidocaine, mepivacaine and prilocaine in man. Acta Anaesthesiol Scand Suppl 1996;23:475-82.
3. Braid DP, Scott DB. Dosage of lignocaine in epidural block in relation to toxicity. Br J Anaesth 1966;38:596-602.
4. Raj PP, Rosenblatt R, Miller J, et al. Dynamics of the local anesthetic compounds in regional anesthesia. Anesth Analg 1977;56:110-7.
5. Cole PC, McMorland GH, Jenkins LC, et al. Epidural blockade for cesarean section comparing lidocaine hydrocarbonate and lidocaine hydrochloride. Anesthesiology 1985;62:348-50.
6. Tucker GT, Mather LE. Pharmacology of local anaesthetic agents, pharmacokinetics of local anaesthetic agents. Br J Anaesth 1975;47:213-24.
7. Ramanathan J, Bottorff M, Jeter JN, et al. The pharmacokinetics and maternal and neonatal effects of epidural lidocaine in preeclampsia. Anesth Analg 1986;65:120-6.
8. Mazze RI, Dunbar RW. Plasma lidocaine concentrations after caudal, lumbar epidural, axillary block and intravenous regional anesthesia. Anesthesiology 1966;27:574-9.
9. Brum AGL, van Kleef JW, Gladines MPRR, et al. Epidural anesthesia with lidocaine and bupivacaine: effects of epinephrine on the plasma concentration profiles. Anesth Analg 1986;65:1281-4.
10. Ohno H, Watanabe M, Saitoh J, et al. Effect of epinephrine concentration on lidocaine disposition during epidural anesthesia. Anesthesiology 1988;68:625-8.
11. Mather LE, Tucker GT, Murphy TM, et al. Effect of adding adrenaline to etdocaine and lignocaine in extradural anaesthesia. II. Pharmacokinetics. Br J Anaesth 1976a;48:989-94.
12. Tucker GT, Mather LE. Properties, absorption, and disposition of local anesthetic agents. In: Cousins MJ, Bridenbaugh PO, eds. Neural blockade in clinical anesthesia and management of pain. 2nd ed. Philadelphia: JB Lippincott, 1988:59-110.
13. Misra U, Pridie AK, Bower S, et al. Arterial and peripheral venous concentrations of bupivacaine following combined sciatic and femoral 3-in-1 block [abstract]. Br J Anaesth 1994;72:A.139.
14. Nolte H, al Saydali B, Weissenberg W. The concentration of free lidocaine in arterial, central venous and peripheral vein plasma following intravenous injection. Reg Anesth 1990;13:29-35.
15. Bachman MB, Biscoping J, Adams HA, et al. The significance of the sampling site in determination of plasma levels of local anaesthetics using 0.75% bupivacaine. Reg Anesth 1990;13:16-20.
16. Tucker GT. Determination of bupivacaine (Marcaine) and other anilide-type local anesthetics in human blood and plasma by gas chromatography. Anesthesiology 1970;32:255-60.
17. Ha H, Funk B, Gerber H, Follath F. Determination of bupivacaine in plasma by high-performance liquid chromatography. Anesth Analg 1984;63:448-50.
18. Gillespie WR. Noncompartmental versus compartmental modelling in clinical pharmacokinetics. Clin Pharmacokinet 1991;20:253-62.
19. Albright GA. Cardiac arrest following regional anesthesia with etidocaine or bupivacaine. Anesthesiology 1979;51:285-7.
20. Reynolds F. A comparison of the potential toxicity of bupivacaine, lignocaine and mepivacaine during epidural blockade for surgery. Br J Anaesth 1971;43:567-71.
21. Liu PL, Feldman HS, Giasi R. Comparative CNS toxicity of lidocaine, etidocaine, bupivacaine and tetracaine in awake dogs following rapid iv administration. Anesth Analg 1983;62:375-9.
22. Johnson RF, Herman N, Arney TL, et al. Bupivacaine transfer across the human term placenta. A study using the dual perfused human placental model. Anesthesiology 1995;82:459-68.
23. Morishima HO, Pedersen H, Santos AC, et al. Adverse effects of maternally administered lidocaine on the asphyxiated preterm fetal lamb. Anesthesiology 1989;71:110-5.
24. Mather LE, Cousins MJ. Local anesthetics and their current clinical use. Drugs 1979;18:185-205.
25. Reynolds F, Taylor G. Maternal and neonatal blood concentration of bupivacaine. Anaesthesia 1970;25:14-23.