Determination of the Minimum Local Analgesic Concentrations of Epidural Bupivacaine and Lidocaine in Labor Malachy 0. Columb, FRCA, and Gordon Lyons, FRCA St. Jamess University Hospital, Leeds, United Kingdom

The aim of this study was to devise a clinical model to determine the effective concentrations in 50% of pa- tients (EC,,) for bupivacaine and lidocaine in the first stage of labor and define EC,, as the minimum local analgesic concentration (MLAC). This should allow the determination of relative analgesic potency and, subse- quently, the local anesthetic sparing efficacy of other epidural analgesics. Parturients not exceeding 5 cm cer- vical dilation who requested epidural analgesia were enrolled. The two studies involved 81 women (bupiva- Caine y1 = 41, lidocaine n = 40). After a lumbar epidural catheter was placed, 20 mL of the concentration of local anesthetic being tested was given. The concentration was determined by the response of the previous patient

to a higher or lower concentration using double- blinded, up-down sequential allocation. Efficacy was assessed using loo-mm visual analog pain scores with less than 10 mm within 1 h defined as effective. MLAC was determined using the formula of Dixon and Mas- sey. Results show MLAC bupivacaine 0.065% (95% con- fidence interval [CI] 0.045-0.085), MLAC lidocaine 0.37% (95% CI 0.32-0.42), equivalent to 2 and 14 mmol solutions, respectively. Thus bupivacaine was 5.7 times more potent than lidocaine in weighted and 7 times more potent in molar ratios at analgesic EC,,, in the volume of local anesthetic studied.

(Anesth Analg 1995;81:833-7)

T here has been a tendency to reduce the concen- trations of local anesthetics used for epidural analgesia in labor. This has been further accom- plished by the addition of other epidural analgesics such as opioids and clonidine (1,2). There are many studies describing and comparing various recipes for providing epidural analgesia (3-5). It is difficult, however, to determine the contribution of each drug to the overall efficacy of the analgesia. The reason for this is the lack of pharmacodynamic data detailing the analgesic dose-response relationships of epidural local anesthetics. The aim of this research was therefore to devise a clinical model that would allow the estima- tion of the epidural analgesic EC,, of local anesthetics in the first stage of labor. This was defined as the minimum local analgesic concentration (MLAC). The model was then used to determine the relative anal- gesic potencies of bupivacaine and lidocaine.

This work has been presented in part at the Obstetric Anaesthe- tists Association, Derby, United Kingdom, April 1994, and the European Society of Regional Anaesthesia, Barcelona, Spain, May 1994.

Accepted for publication May 22, 1995. Address correspondence to Malachy 0. Columb, FRCA, St.

Jamess University Hospital, Beckett St., Leeds, United Kingdom LS9 7TF.

01995 by the International Anesthesia Research Society 0003-2999/95/$5.00

Using this model, it should be possible to assess the local anesthetic sparing efficacy of the various opioids and other analgesics by their effect on MLAC as the dependent variable. This should facilitate the estima- tion of epidural potency ratios for the different opioids and describe any interactions in terms of synergy, addition, or antagonism.

Methods

After institutional ethical approval, 81 parturients, ASA physical status I and II, at more than 36 wk gestation requesting epidural analgesia for labor pain at 3- to 5-cm cervical dilation were enrolled. Women who had received opioid or sedative medication were excluded from the studies.

After starting an intravenous 0.9% saline infusion, a lumbar epidural catheter was inserted. A midline ap- proach with the patient in the flexed sitting position was used. After skin infiltration with 1 mL of 1% lidocaine the epidural space was located using loss of resistance to saline at the L2-3 or L3-4 level. Loss of resistance to saline was preferred to air for identifying entry to the epidural space to reduce the occurrence of epidural air pockets affecting spread. The volume of saline used was restricted to less than 2 mL to reduce

Anesth Analg 1995;81:833-7 833

834 OBSTETRIC ANESTHESIA COLUMB AND LYONS ANESTH ANALG ANALGESIC EC,, OF EPIDURAL BUPIVACAINE AND LIDOCAINE IN LABOR 1995;81:833-7

the possibility of dilution. The catheter was advanced 3 cm into the epidural space and aspirated. A test dose was omitted for the purposes of the study. The patient was then given a 20-mL bolus over 5 min of the particular concentration of local anesthetic being eval- uated. The concentration received by a particular pa- tient was determined by the response of the previous patient to a higher or lower concentration using the technique of double-blinded, up-down sequential al- location. Patients were monitored using routine non- invasive hemodynamics and tococardiography.

The first study was conducted using bupivacaine and involved 41 women. Lidocaine was used in the second study and involved a further 40 women. The injectate was freshly prepared using saline 0.9% to dilute the local anesthetic to achieve 20 mL of the desired concentration at room temperature (20C).

The assessment was performed by an observer who was blinded to the concentration being tested. Efficacy was assessed using loo-mm visual analog pain scores (VAPS), where 0 represented no pain and 100 as worst possible pain at 0, 30, and 60 min after the bolus

1.

was injected. Three outcomes were considered:

Effective: VAPS less than 10 mm during con- tractions within 60 min of the injection. A result defined as effective directed a decrement for the next patient. Ineffective: VAPS greater than 10 mm due to nonlocalizing pain, which responded to rescue with a 12-mL bolus of 0.25% bupivacaine or 1% lidocaine as appropriate. A result defined as ineffective directed an increment for the next patient. Reject: VAPS greater than 10 mm due to local- izing pain (segmental, unilateral, perineal) or pain not responding to rescue. A result defined as reject directed that the same concentration be repeated for the next patient.

At 30 min subjects not defined as effective were examined (using pinprick) for evidence of localizing features such as unilateral, segmental, or perineal sparing. Those patients demonstrating localizing fea- tures were defined as rejects. The remainder were given the rescue bolus if VAPS was larger than 30 mm or at patient request. At 60 min any remaining patients who had not requested the rescue bolus and had not yet been defined as effective for the purposes of the study were also given the rescue. Subjects not re- sponding to the rescue bolus were also defined as rejects.

Demographic and obstetric data were collected and are presented as mean (SD) and median (range) as appropriate. The up-down sequences were analyzed using the formula of Dixon and Massey (6) which enabled MLAC with 95% confidence interval (CI) of

Table 1. Demographic and Obstetric Data

Bupivacaine Lidocaine

Age (yr) 25.9 (5.5) 27.5 (4.7) Height (cm) 162 (6.6) 163 (6.0) Weight (kg) 74.1 (11.6) 75 (12.5) Gestation (wk) 39.8 (1.1) 39.6 (1.73) Cervical dilation (cm) 4.5 [3-51 4.5 [3-51 Nulliparous/multiparous 26/4 24/6 Oxytocin 4 3 Initial VAPS (mm) 79 [48-1001 77 [47-1001

Results are expressed as mean (SD), median [range], and count as appropriate.

VAPS = visual analog pain score.

the mean to be derived. The data were also subjected to logit maximum likelihood analysis as a back-up sensitivity test.

Results

There were no significant demographic or obstetric differences in the two studies (Table 1).

Of the 41 women entered into the bupivacaine study, 11 were rejected (Table 21, leaving 30 for anal- ysis. The sequences of effective and ineffective anal- gesia are shown in Figure 1.

MLAC of bupivacaine was 0.065% (95% CI 0.045- 0.085) using the formula of Dixon and Massey (6) and was 0.062% (95% CI 0.052-0.072) by logit analysis, in the volume studied. MLAC 0.065% is equivalent to a 2-mm01 solution.

Of the 40 women entered into the lidocaine study, 10 were rejected (Table 2), leaving 30 for analysis. The sequences of effective and ineffective analgesia are shown in Figure 2.

MLAC of lidocaine was 0.37% (95% CI 0.32-0.42) using the formula of Dixon and Massey (6) and was 0.36% (95% CI 0.31-0.41) by logit analysis, in the vol- ume studied. MLAC 0.37% is equivalent to a 14-mmol solution.

Discussion

The reason for establishing this clinical model was derived from the need to determine the local anes- thetic sparing potential of the various opioids used for epidural analgesia in labor. Although EC,, is usually more relevant clinically, we concluded that estimation of EC,, would provide a more sensitive research tool due to the respective positions of each on the cumu- lative concentration-response curve. The EC,, corre- sponds to the inflection point where the slope is larg- est. Bupivacaine was studied first because it is the most prevalent in use in the United Kingdom and is the most commonly reported in opioid recipe regi- mens. Lidocaine was then studied as the prototypical

ANESTH ANALG OBSTETRIC ANESTHESIA COLUMB AND LYONS 835 1995;81:833-7 ANALGESIC EC,, OF EPIDURAL BUPIVACAINE AND LIDOCAINE IN LABOR

Table 2. Distribution of Rejects

Bupivacaine No. of % wt/vol rejects

Lidocaine % wt/vol

No. of rejects

0.11 1 0.6 1 0.08 1 0.45 1 0.07 2 0.4 2 0.06 3 0.35 3 0.05 3 0.3 2 0.04 1 0.25 1

0.167 MLAC Bupivacaine in Labor

n Effective

0 Ineffective

* 0021 0 10 20 30 40

Patient Number

Figure 1. EC,, bupivacaine as determined by the technique of up- down sequential allocation. Minimum local analgesic concentration (MLAC) is 0.065% which is equivalent to a 2-mmol solution. Error bars represent 95% confidence interval of mean. Testing interval was 0.01%.

1.2 - MLAC Lidocaine in Labor

5 W Effective

8 1.0-w 0 Ineffective

Fi n .- 2 L ;

0.8- n

z n E 0 0.6- n

0 0.2 I I

0 10 20 3b 40 Patient Number

Figure 2. EC,, of lidocaine as determined by the technique of up- down sequential allocation. Minimum local analgesic concentration (MLAC) is 0.37% which is equivalent to a 14.mmol solution. Error bars represent 95% confidence interval of mean. Testing interval was O.l%, reduced to 0.05%.

amino-amide. In addition, the analgesic potency ratio could be determined and will be discussed later. Sev- eral aspects of the methodology require further discussion.

Table 3. Concentration-Response Point Estimates

Effective concentration Bupivacaine Lidocaine Potency

(%) % wt/vol % wt/vol ratio

95 0.129 0.52 4.0 90 0.115 0.49 4.3 75 0.091 0.43 4.7 50 MLAC 0.065 (2 mmol) 0.37 (14 mmol) 5.7 25 0.039 0.31 7.9 10 0.015 0.25 16.7

Pomt estnnates are derived from the EC,, expemnental result usrng the table of normal dewates Potency ratlo descrrbes the werght for werght potency of buprvacarne relative to hdocalne.

MLAC = nummum local analgesr concentratmn

The up-down sequential allocation technique, rather than random allocation, was chosen due to the ease with which it estimates the mean of a sample. By starting from effective concentrations and approach- ing the EC,, from above, the number of women sub- jected to potentially inadequate analgesia is mini- mized. This technique has been used to determine dose-response pharmacodynamics for both inhala- tional and intravenous anesthetics (7-9).

Parturients who had exceeded 5 cm cervical dila- tion were excluded in order to minimize the loss of women from the study due to the approach of the second stage of labor with the onset of perineal pain outside the TlO to Ll distribution. Omission of a lidocaine test dose was essential to eliminate any potential analgesic effect due to this. A 20-mL bolus was used to give each concentration being tested every possible chance to achieve adequate spread. The use of a 12-mL bolus in the sitting position has been shown to result in an upper sensory level only to TlO in some patients (10).

The estimation of analgesic EC,, in a clinical setting required that we be exacting and only accept VAPS less than 10 mm as being effective. This was to ensure that MLAC would imply very effective analgesia. Brownridge (11) has shown that parturients request further intervention during epidural analgesia only when the VAPS exceeds 30 mm. It was particularly important to reject those epidural injections with any evidence of inadequate spread as these would obvi- ously bias the results. This was to ensure, as far as was practicable, that those injections designated as ineffec- tive were due to failure of the particular concentration being tested as demonstrated by the response to res- cue with higher concentrations. It is, of course, possi- ble that true effective or ineffective boluses might have been rejected, but as the assessment was blinded to concentration, each concentration in turn had the same criteria applied to it and any bias applied equally. The incidence of reject epidural injections was similar for both local anesthetics, in that 25% demon- strated some evidence of inadequate spread, although

836 OBSTETRIC ANESTHESIA COLUMB AND LYONS ANALGESIC EC,, OF EPIDURAL BUIIVACAINE AND LIDOCAINE IN LABOR

ANESTH ANALG 1995;81:833-7

Concentration-Response Plots

Figure 3. The experimental result, minimum local analgesic concentration (EC,,) is shown with 95% confidence interval of mean. Derived point esti- mates are plotted to demonstrate the concentration- response relationship. The differential slopes imply that the potency of bupivacaine relative to lidocaine increases as concentration decreases.

I I I I I I

0.0 0.1 0.2 0.3 0.4 0.5 0.6

Concentration %wlv

this did not necessarily imply unsatisfactory analgesia as far as the patient was concerned.

The formula of Dixon and Massey (6) was used to derive MLAC and 95% CI from the results of the sequential allocation. This technique analyzed the least common outcome, those tests defined as ineffec- tive, which assumed a normal distribution and are shown in Figures 1 and 2. Logit maximum likelihood analysis has been used by other workers to analyze similar up-down data so we applied this as a back-up sensitivity test (8). We believe that the former test is the more appropriate, even though the latter is gener- ally less conservative with the data. Logit analysis usually is applied to randomly, rather than sequen- tially, allocated data which may result in bias, partic- ularly at the extremes of the logit curve (12,13).

It is common in clinical practice to make potency comparisons on a weight-for-weight basis rather than a molar basis, although the latter is more in keeping with proper pharmacodynamics. The results show that bupivacaine is 5.7 times more potent than lido- Caine on a weighted basis and 7 times more potent on a molar basis at EC,,. Using the table of normal devi- ates (Za), values for the lOth, 25th, 75th, 90th, and 95th percentiles can be calculated from the formula (6):

mean + Z9su

The derived point estimates are shown in Table 3. The cumulative concentration-response plots for both local anesthetics are shown in Figure 3. The graph shows that the plots are not parallel. This implies, therefore, a concentration-dependent variation in relative po- tency ratios. At EC,, the weighted potency ratio is 4:1, in agreement with current understanding (14,15). It is interesting, therefore, that bupivacaine becomes more potent with respect to lidocaine as the concentrations

are reduced. This has obvious implications when com- parisons are being made between different epidural local anesthetics with respect to side effects such as unwanted motor blockade. Comparisons should be made at equianalgesic concentrations, rather than at presumed equivalent somatic sensory denervation, because it is analgesia that is the aim, not indiscrimi- nate deafferentation. The concept of differential block- ade applies to epidural analgesia for both postopera- tive and labor pain (16). The need exists therefore for continuing research to determine the pharmacody- namic concentration-response relationships for the various drugs and recipe combinations being used to provide epidural analgesia. By using these types of data it should be possible to optimize the combina- tions to provide the greatest effect at the minimum disadvantage.

In conclusion, we have devised a clinical model to determine MLAC of local anesthetics in the first stage of labor. This research shows that bupivacaine is 5.7 times more potent than lidocaine in weighted and 7 times more potent in molar ratios at EC,,, when used for epidural analgesia in labor for the parturients in these studies. We are now in the process of systemat- ically evaluating the local anesthetic sparing efficacies of the various epidural analgesics in current use.

The authors gratefully acknowledge Andrew Vail, Medical Re- search Institute, University of Leeds, for valuable statistical advice.

References 1. Chestnut DH, Owen CL, Bates JN, et al. Continuous infusion

epidural analgesia during labor: a randomized, double-blinded comparison of 0.0625% bupivacainc/0.0002% fcntanyl versus 0.125% bupivacaine. Anesthesiology 1988;68:754-9.

ANESTH ANALG 1995;81:833-7

OBSTETRIC ANESTHESIA COLUMB AND LYONS 837 ANALGESIC EC,, OF EPIDURAL BUIIVACAINE AND LIDOCAINE IN LABOR

2. OMeara ME, Gin T. Comparison of 0.125% bupivacaine with 0.125% bupivacaine and clonidine as extradural analgesia in the first stage of labour. Br J Anaesth 1993;71:651-6.

3. Kehlet H, Dahl JB. The value of multimodal or balanced analgesia in postoperative pain treatment. Anesth Analg 1993; 771048-56.

4. Miller AC, De Vore JS, Eisler EA. Effects of anesthesia on uterine activity and labor. In: Shnider SM, Levinson G, eds. Anesthesia for obstetrics. 3rd ed. Baltimore: Williams & Wilkins, 1993:61-3.

5. Lema MJ, Sinha I. Thoracic epidural anesthesia and analgesia. Pain Digest 1994;4:3-11.

6. Dixon WJ, Massey FJ. Introduction to statistical analysis. 4th ed. New York: McGraw-Hill, 1983:428-39.

7. Powell H, Morgan M, Sear JW. Pregnanolone: a new steroid intravenous anaesthetic. Dose-finding study. Anaesthesia 1992; 47287-90.

8. Inomata S, Watanabe S, Masakazu T, Okada M. End tidal sevoflurane concentration for tracheal intubation and minimum alveolar concentration in pediatric patients. Anesthesiology 1994;80:93-6.

9. Stanski DR. Monitoring depth of anesthesia. In: Miller RD, ed. Anesthesia. 4th ed. New York: Churchill Livingstone, 1994: 1133-8.

10. Milligan KR, Cramp I, Schatz L, et al. The effect of position and obesity on the spread of epidural analgesia. Int J Obstet Anesth 1993;2:134-6.

11. Brownridge I. Epidural analgesia in the first stage of labour. Curr Anaesth Crit Care 1991:2:92-100.

12. Wetherill GB, Levitt H. Sequential estimation of points on a psychometric function. Br J Math Stat Psycho1 1965;18:1-10.

13. Van Hemelriick I. Muller I. Van Aken H. White PF. Relative potency of elCandlone, propofol and thiopental for induction of anesthesia. Anesthesiology 1994;80:36-41.

14. Covino BG. Pharmacology of local anaesthetic agents. In: Nunn JF, Utting JE, Brown BR, eds. General anaesthesia. 5th ed. London: Butterworths, 19891036-8.

15. Vickers MD, Morgan M, Spencer PSJ. Drugs in anaesthetic practice. 7th ed. Oxford, UK: Butterworth-Heinemann, 1991:207.

16. Stevens RA, Bray JG, Artuso JD, et al. Differential epidural blockade. Reg Anesth 1992;17:22-5.