The addition of CO2 or sodium bicarbonate to local anesthetics potentiates their impulse-blocking action on peripheral nerves in vitro [1,2]. Therefore, both carbonation and alkalinization of local anesthetics have the potential to enhance the depth of epidural block.
However, clinical studies have reported conflicting results. Carbonation of lidocaine has been found both to reduce  and to have no influence  on pain during surgery. Alkalinized, compared with hydrochloride, lidocaine produces a greater depression of somatosensory evoked potentials , indicating a deeper sensory block. However, no significant difference has been found between hydrochloride and alkalinized lidocaine in the incidence of pain during surgery . Possible explanations for these inconsistent findings are the relatively rare occurrence of pain during epidural anesthesia, the high heterogeneity of surgical stimulation, and the inherent difficulty in defining pain as a variable during surgery. These factors may render problematic the use of the occurrence of pain as a measure of the quality of epidural block.
The hypothesis of the present study was that adding CO2 or sodium bicarbonate to lidocaine enhances the depth of epidural block. Analgesia was assessed by pain threshold to repeated electrical stimulation . This test was chosen because it enables a reproducible quantification of epidural blockade  and induces central temporal summation . Temporal summation occurs when the repetition of a peripheral stimulus causes increased pain perception as a result of increased and prolonged firing of dorsal horn neurons  (central sensitization). Central sensitization is likely to play an important role in the pathophysiology of many acute and chronic pain syndromes, including postoperative and neuropathic pain . Thus, inhibition of temporal summation may be important for both decreasing the risk of pain during surgery and preventing hyperexcitability states.
The sample size was calculated on the basis of pain threshold after repeated electrical stimulation. We arbitrarily assumed a difference of 30 mA in pain threshold as the minimal value to detect a difference in the depth of block among drugs. Setting alpha = 0.05 and beta = 0.8 and using a SD of 17 mA [observed in a previous investigation ], a significant difference of 30 mA in pain threshold among groups would be detected by a sample size of 8 patients per group. The study was therefore conducted on 24 ASA physical status I or II patients undergoing minor, elective urologic surgery. Exclusion criteria were age <18 yr or >65 yr, history of alcohol abuse or intake of psychotropic drugs, intake of opioids in the last 2 wk, intake of other analgesics or sedatives in the last 24 h, coronary artery disease, or pregnancy. The study was approved by the local ethics committee, and written, informed consent was obtained from all patients.
The investigation was conducted in a randomized, double-blind fashion. Patients were divided into three groups of eight patients according to the type of anesthetic solution they received (see below). Randomization was stratified using the minimization method  according to gender, body weight (<or=to75 kg or >75 kg), and body height (<or=to168 cm or >168 cm) and was performed by drawing lots. One of the authors (MC) performed the epidural puncture and the experimental tests in all patients. Another anesthesiologist prepared and injected the anesthetic solution. MC was not present when the injection was performed, and the other anesthesiologist did not participate in the experimental session.
The anesthetic solutions were kept at room temperature and did not contain epinephrine. They were lidocaine hydrochloride, lidocaine CO2, and lidocaine bicarbonate, prepared as follows: 20 mL of lidocaine hydrochloride 2% (Xylocain[registered sign]; Astra, Dietrkon, Switzerland) plus 2 mL of saline 0.9%, 20 mL of lidocaine CO2 2% plus 2 mL of saline 0.9%, and 20 mL of lidocaine hydrochloride 2% plus 2 mL of sodium bicarbonate 8.4% added immediately before injection, respectively. The carbonated and alkalinized solution were not allowed contact with air to minimize the loss of CO2. This was accomplished by piercing the rubber cap of the lidocaine solution with a needle connected with a three-way stopcock. After injecting 2 mL of saline or sodium bicarbonate, a 2-mL probe was taken for measurement of pH and CO2 concentration. The remaining 20 mL was taken from the other port of the stopcock and used for the epidural block. Syringes were immediately closed with a cap. The 2-mL samples for pH and CO2 determinations were divided into two glass vials sealed with a Teflon-Neoprene septum glass by piercing their caps with a 25-gauge needle.
Patients were not premedicated. Electrocardiogram, noninvasive blood pressure, and hemoglobin oxygen saturation using pulse oximetry were monitored. All epidural punctures were performed at L2-3, in the lateral position, with the median approach, using a 18-gauge Tuohy needle. The epidural space was identified by loss of resistance, injecting no more than 2 mL of 0.9% saline. A catheter was inserted 4 cm cephalad into the epidural space, and the patient was turned to the supine position. A test dose of 5 mL of the anesthetic solution was first injected. After 4 min, the remaining dose was administered in increments of 5 mL every 60 s. During this time, blood pressure was measured every minute. To detect intrathecal injection, patients was asked every minute to raise their legs . Tests were performed in the following order: repeated electrical stimulation, pinprick, cold, motor block, and segmental spread of analgesia.
After the skin had been degreased, bipolar surface Ag/AgCl electrodes filled with electrode gel (interelectrode distance approximately 2 cm) were placed on the right side, at the dermatomes S1 (foot, just distal to the lateral malleolus) and L4 (middle of the patella). A 25-ms, train-of-five, 1 ms, square-wave impulse (perceived as a single stimulus) was delivered from a computer-controlled constant current stimulator (University of Aalborg, Denmark). This stimulus was repeated five times at a frequency of 2 Hz . The current intensity was increased from 1 mA in steps of 1-5 mA until a pain summation threshold was observed. Pain summation threshold was defined as an increase in perception of current intensity during the five stimulations so that the last one or two impulses were perceived as painful. If the threshold was above a maximal current of 100 mA, the threshold was defined as 100 mA. After the epidural catheter had been inserted, patients tried the test for training. Baseline recordings were then made. The test was repeated at the dermatome S1 5, 10, 15, 20, 30, 40, 50, and 60 min after the last bolus of anesthetic solution. At L4, threshold determinations were made at 20, 30, 40, 50, and 60 min. At each time point the test was performed twice, and the mean of the two threshold values was calculated.
Sensitivity to pinprick and cold were tested at S1 and L4, 2 cm from the sites of electrical stimulation, and on the forehead to allow comparison with a non-anesthetized area. Pinprick was performed by pricking the skin twice at an interval of approximately 0.3 s, using a 21-gauge sharp-bevel needle. Cold sensitivity was tested with gel bags (Physiopack; Fisch Laboratories, Vibraye, France) that were kept in a freezer and applied to a 4-cm2 skin surface for 2 s. Response was defined as the presence or absence of sensation of pinprick or cold. A sensation of touch only was defined as absence of pinprick or cold sensation. The tests were performed at the same times as the repeated electrical stimulation. A series of repeated electrical stimulation, pinprick, and cold test lasted 2-3 min.
Motor block was recorded according to the Bromage score  (0 = full flexion of feet and knees, 1 = just able to move knees, 2 = able to move feet only, 3 = unable to move feet or knees) at 20, 30, 40, 50, and 60 min.
The spread of sensory block was assessed on the right side at 30 and 60 min by pinprick as described above. Spread was defined as the number of segments at which no sensation of pinprick was evoked.
At the end of the experiment, patients were transported to the operating room for surgery.
The CO2 concentration and pH of the anesthetic solution were measured using two flow-through cells in series. The cells were of acrylic glass and had an internal volume of 0.3 mL each. They were equipped with a flat front PCO2 electrode (GS-136 COFT; Lazar, Los Angeles, CA) with an electrochemical sensor coupled to a semipermeable membrane, and a flat front pH electrode (FT pH - 1A; Lazar) with a sealed calomel half cell. The electrodes were monitored by using a combined pH/millivoltmeter (MP-60; Lazar). Before each measurement, the electrodes were calibrated according to the instruction manual. After piercing the septum of the vial containing the sample with a 20-gauge needle, the anesthetic solution was sucked into the flow-through cells by a fixed-speed minicartridge peristaltic pump (MS-CA1/860; Ismatec, Glattsbrugg, Switzerland). The flow was then stopped, and the sample was analyzed. In this way, we could avoid any contact of the sample with air, thereby minimizing the loss of CO2. Two samples per patient were analyzed. For statistical analysis, the mean of the two samples was calculated, except for CO2 determination in lidocaine CO2 and bicarbonate solutions, for which the highest value was considered.
Temperatures of the rooms in which the epidural block was performed and in which the anesthetic solution was analyzed were recorded.
Age, weight, height, spread of epidural analgesia, pH, and CO2 concentration of the anesthetic solution in the three groups were compared by using Kruskal-Wallis one-way analysis of variance on ranks. Pairwise multiple comparisons among groups were performed using the Student-Newman-Keuls method.
Multiple regression  was used to analyze the pain summation threshold to repeated electrical stimulation. The statistical model had the following structure: Equation 1 where y is the pain summation threshold, i refers to the patient, t refers to time of measurement, CO2 is the dummy variable for lidocaine CO2, BIC is the dummy for lidocaine bicarbonate, time is time of measurement, and uit is a random term responsible for the probabilistic nature of the relationship. The coefficients beta2 and beta3 capture the effects of lidocaine CO2 and bicarbonate on the pain summation threshold at all times relative to lidocaine hydrochloride. If beta4 is positive and beta7 is negative, y increases until a given time and then decreases. The terms (CO2i)(time) and (BICi)(time) and the relative coefficients beta5 and beta6 capture the interactions of the drugs with the time trend. For instance, a positive beta6 means that lidocaine bicarbonate induces a faster onset of y compared with lidocaine hydrochloride. The terms (CO2i)(time)2 and (BICi)(time)2 capture the different concavity of y under lidocaine CO2 and bicarbonate, respectively. Thus, a negative beta9 means that the time profile for y is more concave under lidocaine bicarbonate than under lidocaine hydrochloride.
Pinprick and cold were analyzed by using logistic regression analysis , in which absence of pinprick (cold) sensation was the dependent variable and the types of local anesthetic (i.e., the corresponding dummies) were the explanatory variables. Only data collected at 20, 30, and 40 min were considered.
Motor block was analyzed by using multiple regression analysis , in which the Bromage score was the dependent variable and the types of local anesthetic (i.e., the corresponding dummies) were the explanatory variables.
In all statistical analyses, a P value <0.05 was considered significant. The software used was STATA, version 4.0.
One patient reported dizziness and sensation of strange taste immediately after injection of the test dose. Aspiration of blood through the catheter revealed intravascular placement. The catheter was therefore removed and the puncture repeated at the same level. The subsequent injection was uneventful. No other cases of intravascular or intrathecal injection were observed. We found no difference in patient characteristics among the three groups (Table 1).
Pain summation thresholds were significantly higher, and the onset of action was significantly faster, after lidocaine bicarbonate compared with after lidocaine hydrochloride. No significant differences were found between lidocaine CO2 and hydrochloride (Figure 1). Thresholds for all anesthetic solutions were significantly higher at L4 than at S1 (P < 0.0001).
We found no statistically significant difference in pinprick sensation between lidocaine CO2 and hydrochloride at any dermatome, or between lidocaine bicarbonate and hydrochloride at S1 (Figure 2). The effect of lidocaine bicarbonate on pinprick at the dermatome L4 could not be analyzed because all patients reported absence of pinprick sensation at L4 after lidocaine bicarbonate (logistic regression cannot be performed when an explanatory variable predicts the outcome variable perfectly). However, the frequency of absence of pinprick sensation at L4 for lidocaine bicarbonate seems very similar to that observed after lidocaine hydrochloride (Figure 2).
The use of lidocaine CO2 was associated with a lower frequency of absence of cold sensation at L4 compared with lidocaine hydrochloride. There was no statistically significant difference in cold sensation between lidocaine CO2 and hydrochloride at S1, or between lidocaine bicarbonate and hydrochloride at any dermatome (Figure 3).
The motor block was significantly more pronounced after lidocaine bicarbonate compared with lidocaine hydrochloride, whereas no significant difference was found between lidocaine hydrochloride and CO2 (Figure 4). There were no significant differences in the spread of epidural analgesia among the three groups (Table 1).
The pH and CO2 concentration were higher for lidocaine CO2 and bicarbonate compared with lidocaine hydrochloride. The pH was significantly higher in the bicarbonate than in the CO2 solution (Table 1). Median values (ranges) of the temperature of the rooms in which the epidural blocks were performed and in which the pH and CO2 concentration were analyzed were 23.0 (20.3-25.0)[degree sign]C and 23.0 (21.0-24.0)[degree sign]C, respectively.
Adding sodium bicarbonate to lidocaine enhances the depth of epidural blockade. Both analgesia as assessed by repeated electrical stimulation and motor block as assessed by Bromage score are more profound after lidocaine bicarbonate compared with lidocaine hydrochloride. The increase in pH increases the extraneural amount of nonionized local anesthetic, which is the form that diffuses through the lipid phase of the neural membrane . CO2 produced by the addition of bicarbonate (Table 1) and bicarbonate per se reduce the margin of conduction safety of the neural membrane . Moreover, CO2 penetrates into the nerve, where it may determine trapping of the active cationic form of local anesthetic by acidifying the axoplasm .
We found no difference between lidocaine CO2 and hydrochloride in the depth of epidural block. There was no significant difference in CO2 concentration between the CO2 and bicarbonate solution, whereas the pH was significantly higher in the latter solution (Table 1). This suggests that the degree of alkalinization of the anesthetic solution may be an important factor in enhancing the depth of epidural blockade.
Repeated electrical stimulation induces temporal summation, i.e., increased pain perception during nociceptive stimulation . Temporal summation is attenuated but not blocked by isoflurane , ketamine , alfentanil , epidural bupivacaine , and clonidine . It is completely inhibited by spinal anesthesia . The present study confirms that temporal summation is a very potent pain mechanism, which is difficult to block even with epidural anesthesia (Figure 1). The higher pain summation threshold observed after lidocaine bicarbonate suggests that the addition of sodium bicarbonate to local anesthetics may increase the ability of epidural analgesia to prevent central sensitization and hyperexcitability states.
As in previous studies [5,20-22], a faster onset of block with lidocaine bicarbonate, compared with lidocaine hydrochloride, was observed (Figure 1). Failure of other investigations [3,23] to detect a difference in the onset of block may be the result of the use of pinprick and cold for assessing the time course of epidural blockade. As explained below, these tests are not adequate for assessing the efficacy of epidural anesthesia.
Pain thresholds were significantly higher at L4 than at S1 (Figure 1). This confirms previous evidence that the segment S1 may be more difficult to block with epidural anesthesia , perhaps as a result of the wider size of the S1 nerve root .
Pinprick and cold tests are used routinely in clinical practice and frequently for research purposes to assess regional block. Interestingly, most of the results observed with repeated electrical stimulation (Figure 1) were not confirmed by pinprick and cold tests (Figure 2 and Figure 3). Pinprick and cold are very weak stimuli. Even in the absence of pinprick and cold sensation, pain can be evoked by surgical  or experimental  stimuli during epidural anesthesia. Thus, anesthetic solutions characterized by different potencies may manifest the same ability to block pinprick and cold sensation. Unlike quantitative methods, such as threshold determinations, information obtained by pinprick and cold are usually limited to two responses, i.e., presence or absence of sensation. As a result, these tests are probably inadequate for detecting differences in the analgesic potency of local anesthetics. Conversely, repeated electrical stimulation enables a quantification of analgesia, and the induction of temporal summation may render this test closer to clinical pain.
In conclusion, we found that that adding sodium bicarbonate to lidocaine enhances epidural analgesia and increases inhibition of temporal summation. This practice may therefore decrease the risk of inadequate surgical analgesia and improve prevention of central sensitization and hyperexcitability states. Sodium bicarbonate hastens the onset of epidural analgesia and enhances the degree of motor block. Adding CO2 to lidocaine does not seem to confer the same advantages. Pinprick and cold are inadequate tests for comparing drugs for epidural anesthesia. Quantitative methods inducing temporal summation provide more reliable information.
The analyzer for CO2 and pH measurements was provided by Astra Pharmaceutica.
1. Gissen AJ, Covino BG, Gregus J. Differential sensitivity of fast and slow fibers in mammalian nerve. IV. Effect of carbonation of local anesthetics. Reg Anesth 1985;10:68-75.
2. Wong K, Stricharzt GR, Raymond SA. On the mechanisms of potentiation of local anesthetics by bicarbonate buffer: drug structure-activity studies on isolated peripheral nerve. Anesth Analg 1993;76:131-43.
3. Gosteli P, Van Gessel E, Gamulin Z. Effects of pH adjustment and carbonation of lidocaine during epidural anesthesia for foot or ankle surgery. Anesth Analg 1995;81:104-9.
4. Cole CP, McMorland GH, Axelson JE, Jenkins LC. Epidural blockade for cesarean section comparing lidocaine hydrocarbonate and lidocaine hydrochloride. Anesthesiology 1985;62:348-50.
5. Benzon HT, Toleikis JR, Dixit P, et al. Onset, intensity of blockade and somatosensory evoked potential changes of the lumbosacral dermatomes after epidural anesthesia with alkalinized lidocaine. Anesth Analg 1993;76:328-32.
6. Arendt-Nielsen L, Brennum J, Sindrup S, Bak P. Electrophysiological and psychophysical quantification of central temporal summation of the human nociceptive system. Eur J Appl Physiol 1994;68:266-73.
7. Curatolo M, Petersen-Felix S, Arendt-Nielsen L, et al. Temporal summation during extradural anaesthesia. Br J Anaesth 1995;75:634-5.
8. Price DD. Characteristics of second pain and flexion reflexes indicative of prolonged central summation. Exp Neurol 1972;37:371-87.
9. Coderre TJ, Katz J, Vaccarino AL, Melzack R. Contribution of central neuroplasticity to pathological pain: review of clinical and experimental evidence. Pain 1993;52:259-85.
10. Pocock SJ. Clinical trials. Chichester: John Wiley & Sons, 1984:80-7.
11. Colonna-Romano P, Padolina R, Lingaraju N, Braitman LE. Diagnostic accuracy of an intrathecal test dose in epidural analgesia. Can J Anaesth 1994;41:572-4.
12. Bromage PR. Epidural analgesia. Philadelphia: WB Saunders, 1978;142-7.
13. Altman DG. Practical statistics for medical research. London: Chapman and Hall, 1991:325-64.
14. Strobel GE, Bianchi CB. The effects of pH gradients on the action of procaine and lidocaine in intact and desheated sciatic nerves. J Pharmacol Exp Ther 1970;172:1-17.
15. Petersen-Felix S, Arendt-Nielsen L, Bak P, et al. The effects of isoflurane on repeated nociceptive stimuli (central temporal summation). Pain 1995;64:277-81.
16. Arendt-Nielsen L, Petersen-Felix S, Fischer M, et al. The effect of N-methyl-D-aspartate antagonist (ketamine) on single and repeated nociceptive stimuli: a placebo-controlled experimental human study. Anesth Analg 1995;81:63-8.
17. Petersen-Felix S, Arendt-Nielsen L, Bak P, et al. Psychophysical and electrophysiological responses to experimental pain may be influenced by sedation: comparison of the effects of an hypnotic (propofol) and an analgesic (alfentanil). Br J Anaesth 1996;77:165-71.
18. Curatolo M, Petersen-Felix S, Arendt-Nielsen L, Zbinden AM. Epidural epinephrine and clonidine: segmental analgesia and effects on different pain modalities. Anesthesiology 1997;87:785-94.
19. Curatolo M, Petersen-Felix S, Arendt-Nielsen L, Zbinden AM. Spinal anaesthesia inhibits central temporal summation. Br J Anaesth 1997;78:88-9.
20. Di Fazio CA, Carron H, Grosslight KR, et al. Comparison of pH-adjusted lidocaine solutions for epidural anesthesia. Anesth Analg 1986;65:760-4.
21. Ackermann WE, Denson DD, Juneja MM, et al. Alkalinization of chloroprocaine for epidural anesthesia: effects of pCO2
at constant pH. Reg Anesth 1990;15:89-93.
22. Capogna G, Celleno D, Laudano D, Giunta F. Alkalinization of local anesthetics: which block, which local anesthetics? Reg Anesth 1995;20:369-77.
23. Verborgh C, Claeys M, Camu F. Onset of epidural blockade after plain or alkalinized 0.5% bupivacaine. Anesth Analg 1991;73:401-4.
24. Galindo A, Hernandez J, Benavides O, et al. Quality of spinal extradural anaesthesia: the influence of spinal nerve root diameter. Br J Anaesth 1975;47:41-7.
25. Hogan Q. Size of human lower thoracic and lumbosacral nerve-roots. Anesthesiology 1996;85:37-42.