Although addition of clonidine to local anesthetics can prolong pain relief after peripheral nerve block, a dose-range effect has not been determined.
Fifty-six outpatients undergoing carpal tunnel release were randomly assigned to receive in a double-blind fashion 45 ml of a mixture containing either 400 mg lidocaine plus saline or 400 mg lidocaine plus 30, 90 or 300 microg clonidine for axillary nerve block. In each group (n = 14), blocks were evaluated at regular time intervals to determine sensory and motor functions in the five nerve regions of the hand and forearm. Also, adequacy of the block for surgery, postoperative pain intensity, and side effects were evaluated.
Compared with saline, each dose of clonidine reduced the onset time of sensory block and extended the field of adequate anesthesia. Ten minutes after injection, 30 microg clonidine was more effective than 90 microg clonidine in producing sensory blockade. Sedation occurred with clonidine 30 and 300 microg. Clonidine reduced the use of supplementary intravenous anesthetic agents for surgery and produced dose-dependent prolongation of analgesia, reaching a mean 770 min (range, 190-1440 min) for the largest dose. Clonidine also produced a dose-dependent decrease in systolic arterial pressure of up to -22.5% (range, -6.0 29.9%) of baseline. With clonidine, 300 microg, three patients had mean arterial pressure of <55 mmHg; four patients had episodes of arterial oxyhemoglobin saturation of <90%, and two others were not discharged because of hypotension.
This study suggests that a small dose of clonidine enhances the quality of the peripheral blocks from lidocaine and limits the classical alpha2-agonist side effects to sedation.
The ability of alpha2-agonists to enhance central [1–5] and peripheral neural blockade, [6–9] when added to local anesthetics, has been demonstrated in animals and humans for more than a decade. Analgesia that follows axillary nerve blocks is prolonged by clonidine from 40–100%, depending on the local anesthetics used, but opinions differ on the incidence of side effects. Whereas Eledjam et al.  and Singelyn et al.  recommend the use of clonidine, others emphasize the possibility of nausea,  bradycardia, [7,9] hypotension,  and marked sedation.  Gaumann et al.  state that 150 micro gram clonidine does not offer advantages compared with 200 micro gram epinephrine because fewer patients are completely pain-free after the block.
Several dose-response studies have been performed using intrathecal [2,3] and epidural clonidine, [4,10] but those concerning axillary nerve blocks have only tested a single dose of clonidine, 150 micro gram, [6–8] or have been limited in scope.  In the study of Buttner et al.,  effects of mepivacaine plus 120 and 240 micro gram clonidine were compared with those of mepivacaine alone, and clear-cut relationships between clonidine dose and duration of block analgesia or side effects were not documented.
This study attempts to define a dose-response relationship for clonidine added to lidocaine for axillary block.
Materials and Methods
After approval by our human investigation and ethics committee (Comite d'ethique du CHRU de Nantes), we studied 56 American Society of Anesthesiologists' (ASA) physical status 1–2 outpatients with no contraindications to brachial plexus block who were scheduled for carpal tunnel repair. All patients had a normal sensory and motor examination in the ulnar, median, radial, medial antibrachial, and musculocutaneous nerve regions of the hand and forearm. Patients taking opioids, calcium channel blockers, clonidine, and related compounds were excluded from the study. Details of the anesthetic technique and details of the study protocol were fully explained at the preoperative visit. Informed consent was obtained from each participating patient. On the morning of surgery, patients were randomly assigned to one of four groups (n = 14 each) receiving either saline or 30, 90, or 300 micro gram of clonidine in 400 mg of lidocaine. The patients and the anesthesiologist were unaware of group assignment. Patients received no premedication or additional medication before or during the block.
On arrival in the anesthesia induction room, patients were monitored using an electrocardiogram lead II, an automated sphygmomanometer, and a pulse oximeter. A catheter was inserted into a peripheral vein. Axillary perivascular brachial plexus block was then performed by the same and experienced anesthesiologist (P M) using a nerve stimulator. A 5-cm 22-gauge needle (Stimuplex[registered sign], B-Braun Melsungen AG, Germany) was inserted at a point high in the axilla, about 40 mm from the pectoralis major muscle insertion. The patient was supine, with the arm at a 75 [degree sign] angle, abducted, and externally rotated. Each patient received a 40-ml solution of 1% lidocaine (400 mg) combined with clonidine or saline at one of the doses mentioned previously and diluted as needed up to 45 ml with saline solution. Drugs were injected only if a current of less than 0.6 mA elicited visible motor response or paresthesia in the median nerve region. Only the median nerve was stimulated, and the entire volume was injected through stationary needle after identifying the nerve. In case of arterial puncture, the patient was excluded and replaced in the randomization list.
Blocks were evaluated before, 10, 20, and 30 min after injection, and intraoperatively. Sensory block was evaluated in the five regions of nerve distribution in the forearm and hand, using a pinprick test on a three-point scale (0 = normal sensation; 1 = blunted sensation; 2 = no perception) and a temperature test to ether (0 = normal; 1 = loss of temperature perception). Assessment of motor block was derived from the technique described by Bromage  on a three-point scale (0 = normal motor function with full flexion and extension of elbow, wrist, and fingers; 1 = decreased motor strength, with ability to move the fingers only; 2 = complete motor block with inability to move the fingers). At the time of surgery, a 7-cm cuff at 300 mmHg pressure was applied to the arm. Patients were asked to qualify tourniquet pain on a three-point scale (0 = unbearable; 1 = experienced but endurable; 2 = not experienced). The adequacy of the block for surgery was evaluated as complete (total comfort of the patient) or inadequate, i.e., requiring intravenous midazolam (range, 1–5 mg in 1-mg increments) or surgical infiltration of the wound with lidocaine.
In the postrecovery lounge, patients rated their level of pain on a visual analogue scale (VAS) graded from 0 (no pain) to 100 mm (unbearable pain) and pain relief on a five-point verbal rating scale (VRS, 1 = excellent; 2 = very good; 3 = good; 4 = medium; 5 = poor). These evaluations were performed 80 min after injection into the axillary sheath and then every 60 min for the next 3 h. Heart rate (HR), systolic arterial pressure (SAP), diastolic arterial pressure (DAP), arterial oxyhemoglobin saturation (SpO2), and sedation on a five-point scale (0 = wide awake; 1 = drowsy; 2 = dozing intermittently; 3 = mostly asleep; 4 = only aroused by tactile stimulation) were monitored and measured before and every 20 min after the block and postoperatively at the same times as pain evaluation. Patients were discharged from the hospital only when vital signs were stable and normal.
Side effects and their management were defined as follows:(1) hypotension (mean systemic arterial pressure less than 55 mmHg) managed by a 15-mg intravenous bolus of ephedrine;(2) bradycardia (heart rate less than 45 beats/min) managed by an 0.5-mg intravenous bolus of atropine;(3) three or more episodes of SpO2equal to or less than 90% within a 10-min period, warranting oxygen administration via a face mask. In addition, oral paracetamol (500 mg in increments up to 4 g/24 h) was available for pain management in the postrecovery lounge and at home. In the morning after surgery, patients were questioned by telephone about their paracetamol requirements and the potential adverse effects of the block. Duration of analgesia was determined as the time elapsed between drug injection and the first paracetamol request.
Results are reported as median +/- range and mean +/- SD when appropriate. All analyses were performed on raw data but are presented in some cases as percent changes to ensure clarity. Intergroup parametric data such as patient characteristics, baseline hemodynamics, and amounts of paracetamol were compared using analysis of variance (ANOVA). Intergroup categorical data such as requirement for ephedrine or atropine, number of incomplete sensory blockades, and incidence of desaturation were compared using a contingency table. Interval data between groups were compared by two-way ANOVA for repeated measurements of parametric data. Comparisons of nonparametric data were performed at each time-point by Kruskall-Wallis analysis, followed by a Mann-Whitney U test when only two groups were compared. Pinprick and temperature analyses were made nerve by nerve. To obtain comparisons on the whole upper limb, each patient was assigned a value depending on the grouping of scores observed in each nerve region, which were classified in 20 ranks in descending order. For example, with the pinprick test, a grouping of five regions at score 2 gave the best rank, and a grouping of five regions at score 0 gave the worst rank. Intragroup analyses were performed by Friedman analysis or ANOVA when appropriate. To assess dose-dependent relationships, data were fitted to a nonlinear model. Duration of block analgesia in the four groups was compared by a log-rank test. Data collected in post-recovery lounge after a patient had received intravenous midazolam or supplementary wound infiltration during surgery were not included in analyses. P < 0.05 was considered significant.
Demographics, baseline hemodynamics, and SpO2were not significantly different among the four groups (Table 1). Before the block, all patients had normal sensory perception in the operated limb.
Sensory blockade was more pronounced in the groups receiving clonidine than in the saline group, in the entire hand and forearm and in the musculocutaneous nerve region (Figure 1and Table 2). At 10 min, a dose of 30 micro gram was more effective than 90 micro gram in the entire hand and forearm and in the ulnar nerve region (Figure 1and Table 2). At 20 and 30 min, each dose of clonidine was more effective than saline in the medial antibrachial nerve region. With the temperature discrimination test, blockade of the musculocutaneous nerve was more pronounced at 10 min in the groups receiving clonidine than in the saline group, with no differences between doses of clonidine (Table 2). There were no intergroup differences for motor blocks.
At the time of surgery, no patient requested general anesthesia, although the number of incomplete sensory blocks was different between groups (six patients in the saline group compared with two in the 30- and 300-micro gram clonidine groups and none in the 90-micro gram clonidine group [P = 0.026]). No intergroup differences were observed when patients rated tourniquet pain.
In the postrecovery lounge, VAS and VRS scores were comparable in the four groups at the first two evaluations. At 200 min, VAS scores in each clonidine group and VRS scores in the 90- and 300-micro gram clonidine groups were better than in the saline group. At that time, VAS scores also were lower in the 300-micro gram clonidine group than in the 30-micro gram clonidine group. At 260 min, statistical significance was achieved for a lower VAS and VRS scores in the 300-micro gram clonidine group than in the saline group (Table 3). Time elapsed before the first supplemental analgesic request by patients showed a significant second-order polynomial relationship to the dose of clonidine (Figure 2). The log-rank curves representing the number of patients not requesting analgesia after the block in each group were significantly different (P < 0.01;Figure 3).
After administration of the drugs, 30 and 300 micro gram clonidine resulted in sedation. At 20, 40, and 140 min, sedation was more marked in these groups than in the saline group. At 40 min, patients who received 300 micro gram clonidine were more sedated than those who received 90 micro gram (Figure 4). Episodes of SpO2of less than 90% for 20 s or more were noted at least once in four patients in the 300-micro gram clonidine group (number of episodes = 27) and in none in the other groups (P < 0.01).
After performing the block, there was a significant decrease in SAP in the saline group and in SAP and DAP in the 90- and 300-micro gram clonidine group (Figure 5). The 300-micro gram clonidine group showed a significantly lower SAP than the saline group 40 and 60 min after injection. Significant second-order polynomial relationships were found between clonidine dose and the lowest SAP value observed during the study and between clonidine dose and the lowest DAP value observed during the study (Figure 6). Ephedrine was given at least once for three patients in the 300-micro gram clonidine group, for one patient in the 90-micro gram clonidine group, and for none in the two other groups (not significant). HR did not change significantly in any group. One patient who had received 300 micro gram clonidine felt faint on walking and was not discharged from the hospital. Another patient in this group was not discharged because SAP was less than 80 mmHg at the time of departure.
Our study evaluated the dose-related effects of clonidine added to local anesthetics in peripheral nerve blocks. The three doses tested varied over a 10-fold range. Compared with lidocaine alone, each dose of clonidine reduced block onset, extended the sensory block, and improved its efficacy at the time of surgery. These effects were noted with the lowest dose studied.
In the previous studies of clonidine in brachial plexus nerve blocks, [6–9] neither the onset time nor the efficacy were influenced by clonidine. Discrepancies between these studies and ours may be explained by differences in the mixture injected (1% plain lidocaine vs. 0.25% bupivacaine  and vs. 1% mepivacaine plus epinephrine  or plus NaHCO3) or the technique used to perform the block (single injection vs. multiple injections).  The tests used to evaluate sensory and motor blocks were less extensive, and their power to differences was limited. The use of potent local anesthetic mixtures, which provide reliable, dense anesthesia also will make detection of any clonidine effect more difficult. Similarly, stimulation of the musculocutaneous nerve and all three major nerves of the brachial plexus yields a greater success rate than stimulation of the median nerve.  This may obscure any differences caused by clonidine.
Because clonidine added to mepivacaine, as opposed to a similar dose of clonidine subcutaneously injected, has been found to be efficient for prolonging brachial plexus block,  it is likely that clonidine acts locally. Many hypothetical mechanisms may be proposed. (1) Local vasoconstriction: Based on previous pharmacokinetic studies, [7,10] it is clear that clonidine does not act like epinephrine; i.e., by inducing local vasoconstriction, thus by slowing local anesthetic absorption. In our study, onset time and extent of the block were enhanced by clonidine, which is at variance with a mechanism of local vasoconstriction. (2) Facilitated lidocaine blockade: Gaumann et al. demonstrated in an isolated, desheathed rabbit vagus nerve model, very low-dose clonidine increased the C-fiber blockade from lidocaine. [13,14] Thus, a drug interaction between lidocaine and clonidine is possible, even though the nerve-tissue concentrations of clonidine needed to produce tonic and phasic blocks in vitro (10 sup -3 M) are at least 1,000 times greater than those which could be obtained in clinical applications. (3) Spinal action: As perineuronal opioids,  it is possible that clonidine acts spinally rather than locally, i.e., after drug transport into epidural and subarachnoid spaces involving movements between endoneural space and subpial space,  slow retrograde axonal transport (less than 13 mm/h) by binding proteins,  and single diffusion along the nerve. Nevertheless, it is questionable any of these explanations could result in a therapeutic concentration in cerebrospinal fluid and the spinal cord 10 min after injection of low-dose clonidine because injection of a 40-ml solution by the axillary route has been found to be confined to a virtual cavity of which the upper end always was located on substantial distance from the spinal cord. 
Like other dose-related studies on the interactions of clonidine with local anesthetics, [2,3] our results argue for a dose-dependent prolongation of the analgesic effects that follows neural blockade. Unlike these studies, they indicate that an increased clonidine dose was accompanied by a greater number of adverse effects. In view of the central action of the drug after its absorption from the axillary nerve sheath and the results of pharmacokinetic studies of clonidine in a presumed comparable range of plasma clonidine concentrations, it is logical that effects such as decreased blood pressure  and an abnormal ventilatory pattern,  but also analgesia duration,  should be related to the dose administered. Surprisingly, a biphasic dose response occurred for sedation in our study, although the unchanged vigilance after 90 micro gram is likely a chance event in a small number of subjects as clonidine has previously been shown to produce dose-dependent sedation. [24,25]
To conclude, our study suggests that a small dose of clonidine enhances the quality of the peripheral blocks from local anesthetics (lidocaine) and limits the alpha2-agonist side effects to the sedation. The best dose to use clinically is between 30 micro gram and 90 micro gram.