A Comparison of Superficial Versus Combined (Superficial and Deep) Cervical Plexus Block for Carotid Endarterectomy: A Prospective, Randomized Study

Abstract

Carotid endarterectomy may be performed by using cervical plexus blockade with local anesthetic supplementation by the surgeon during surgery. Most practitioners use either a superficial cervical plexus block or a combined (superficial and deep) block, but it is unclear which offers the best operative conditions or greatest patient satisfaction. We compared the two techniques in patients undergoing carotid endarterectomy. Forty patients undergoing carotid endarterectomy were randomized to receive either a superficial or a combined cervical plexus block. Bupivacaine 0.375% to a total dose of 1.4 mg/kg was used. The main outcome measure was the amount of supplemental lidocaine 1% used by the surgeon. Subsidiary outcome measures were postoperative pain score, sedative and analgesic requirements before and during surgery, and postoperative analgesic requirements. Median supplemental lidocaine requirements were 100 mg (range 30–180 mg) in the superficial block group and 115 mg (range 30–250 mg) in the combined block group. These differences were not statistically significant (Mann-Whitney U-test). There was no significant difference in the number of patients needing postoperative analgesia between the groups (11 of 20 in the deep block group versus 8 of 20 in the superficial block group) in the 24 h after surgery. The median time to first analgesia in the superficial block group was 150 min, more than in the combined block group (median time 45 min) but this difference, although large, was not statistically significant (Mann-Whitney U-test). We found no significant differences between the anesthetic techniques studied. All patients reported satisfaction with the techniques.

Implications 

Carotid endarterectomy may be performed satisfactorily by using either superficial or combined block, and it is found that peroperative lidocaine requirements will be the same regardless of which block is used. The decision to use one block or the other might, therefore, reasonably be influenced by the relative safety of the superficial block compared with the combined block, because previous work suggests the deep injection is associated with a more frequent complication rate.

Carotid endarterectomy is often performed under cervical plexus block by using neurologic status as a continuous monitor of cerebral perfusion. Regional block for carotid endarterectomy requires blockade of the cervical nerves C2-4 and may be performed by using a superficial cervical plexus block or a combination of superficial and deep cervical plexus blocks. Local supplementation with 1% lidocaine administered by the surgeon is commonly used peroperatively. However, deep cervical plexus block is associated with potentially serious complications, such as diaphragmatic dysfunction (in 61% of patients), which can lead to respiratory distress (^1,2), and there is a risk of epidural, subarachnoid, or vertebral artery injection of local anesthetic (³). Superficial block alone might prevent these complications.

Although a number of studies have compared regional with general anesthesia for endarterectomy (^4–6), very few have compared the different types of regional anesthesia. Stoneham et al. (⁷) compared the efficacy of superficial block alone with deep block alone and found little difference between the two. There are a number of reasons to extend this work and examine the combined (superficial and deep) block. First, in our institution, as in many others (^3,6,8), a combined block is more commonly performed than deep block alone, and the results of Stoneham et al. (⁷) cannot be simply extrapolated to our situation. Second, when a superficial block alone is performed at our institution [and elsewhere (⁹)], a bupivacaine dose of 1.4 mg/kg (approximately 30 mL of 0.375% bupivacaine) is used; Stoneham et al. (⁷) used a more modest dose of only 20 mL of 0.375% bupivacaine in all their patients (⁷). Finally, Davies et al. (³) have explicitly stated that the combined block produces a more complete block (with better intraoperative muscle relaxation detectable by the surgeons) and a more frequent success rate than superficial block alone, but conceded that there were no hard data to support this notion. It would seem important to investigate this suggestion.

We perform both types of block routinely, and anecdotally, we have found that they appear equally suitable for surgery. The purpose of this study was to identify any objective differences in anesthetic efficacy between the two blocks.

Methods

This study had the approval of the Institutional Review Board of the University of Michigan. Forty consecutive patients undergoing elective carotid endarterectomy using regional anesthesia were randomized by means of a computer-generated random numbers table to two groups: either superficial cervical plexus block alone or a combined superficial and deep cervical block. Patients with known bleeding diathesis, history of allergy to local anesthetics, local sepsis, or known diaphragmatic motion abnormalities were excluded from the study.

A 16-gauge peripheral IV cannula was sited, and a 20-gauge cannula placed in the contralateral radial artery to facilitate invasive monitoring of arterial blood pressure. Noninvasive blood pressure, five-lead electrocardiography (leads II and V5), and pulse oximetry were also commenced before placement of the block and were continued until discharge from the postanesthesia care unit (PACU).

Placement of cervical plexus blocks was performed by attending anesthesiologists experienced in the technique, all of whom were part of the Vascular Anesthesiology Group of the University of Michigan Medical Center.

Patients in one group underwent superficial cervical plexus block (^9,10). Briefly, a 20-gauge needle was introduced into the skin (after raising a small bleb of 1% lidocaine) at the midpoint of the posterior border of the sternocleidomastoid muscle, and 0.375% bupivacaine was injected along the posterior border in both cranial and caudal directions subcutaneously and superficial and deep to the fascia of the muscle (^3,11). A “fan” injection was also performed subcutaneously from the posterior border of sternocleidomastoid toward the midline of the neck. To allow for some expected supplementation by the surgeons during the operation, we planned the maximum dose to be 1.4 mg/kg (approximately 30 mL of 0.375% bupivacaine). The final volume therefore depended somewhat on the patient’s body weight, and the dose used was recorded.

Patients in the second group received combined cervical plexus block. The deep element of the block was performed by using a single-injection technique at the C4 level. This has been described previously elsewhere (¹²). Briefly, with the patient semirecumbent and the head turned away, the cervical transverse processes were palpated approximately 1 cm posterior to the posterior border of sternocleidomastoid muscle. A 2.5-cm, 25-gauge needle was introduced at right angles to the skin, after raising a small skin bleb with 1% lidocaine, and aimed in a slightly caudad direction. On location of the transverse process 1–2 cm from the skin (or if the patient reported paresthesia in the distribution of the cervical plexus), plain bupivacaine 0.375% was injected after negative aspiration for blood. Sensory loss to pinprick in the C2-4 dermatome distribution was confirmed. The superficial part of the combined block was then performed as described above, again by using 0.375% bupivacaine. The total dose of 0.375% bupivacaine used was designed to be 1.4 mg/kg. Approximately one third of the total dose (i.e., approximately 10 mL of solution) was placed deep, and two thirds of the dose (i.e., approximately 20 mL) was placed superficially (^3,13). The doses used were recorded.

During the insertion of the monitoring lines and placement of the block, the minimum sedation necessary to alleviate patient anxiety with either midazolam or diazepam was administered.

The surgeons were blinded to the type of block used, and additional lidocaine 1% was administered peroperatively by them in 1-mL aliquots (either superficially into skin and subcutaneous tissues or deep into and around the carotid sheath) as considered necessary when the patient reported discomfort. During the operation, the protocol planned to avoid all IV analgesia or sedation. However, it became apparent that occasionally some sedation was necessary, and therefore, midazolam, fentanyl, or propofol in the smallest doses needed to alleviate anxiety were administered. All drugs and doses administered were recorded.

After surgery was complete and on arrival in the PACU, patients were asked to score the level of discomfort they had experienced during the operation by using an 11-point verbal pain scale (¹⁴) administered by a staff member blinded to the type of block used (0 = no pain, 10 = worst pain imaginable). The attending anesthesiologist also asked the surgeons if they could identify the type of block that had been used.

Orders were written for all patients in the first 24 h after operation for IV morphine in 2-mg aliquots and/or 1 g of acetaminophen. This was to be administered at the discretion of the PACU or ward nurses (both blinded to the type of block used). If the patient requested analgesia in the PACU, the time of administration was recorded. Patients were followed up on the ward, and their medical case notes were reviewed to confirm the drug, dose, and time of any analgesic administration on the ward.

In summary, surgeons, PACU staff, and ward staff were blinded to the type of block. By definition and of necessity, anesthesiologists were not blinded to the block they were performing. Data collection of verbal pain scores was conducted by staff blinded to the type of block given.

It was considered reasonable to suppose that a difference between the two groups of 40 mg of lidocaine (4 mL of 1% lidocaine) would be of some clinical importance. From the data of Stoneham et al. (⁷) and our own preliminary experience, we estimated that the standard deviation of the dose of additional lidocaine administered would be no more than 40 mg. Therefore, to achieve a power in this study of 0.8, 17 patients in each group were needed (^15,16). We recruited 20 patients in each group.

To compare the means of the patients’ height, age and weight in the two groups, Student’s unpaired t-test was used.

The following data were obtained in this study for each of the two groups: the dose of bupivacaine used for the block; the dose of additional lidocaine administered by the surgeon; the doses of any perioperative sedation or analgesia; the verbal pain score; the type of analgesia; and the time to first analgesia postoperatively. The main outcome measure for this study was the amount of supplemental lidocaine 1% administered peroperatively. The null hypothesis was that the two study groups (superficial and combined blocks) would require identical doses of supplemental lidocaine. The other variables were regarded as subsidiary measures of outcome. For all these variables, statistical significance of the differences between the two groups was assessed by using the Mann-Whitney U-test or χ² (with Yates’ correction for continuity) as appropriate. A level of P < 0.05 was taken as statistically significant.

Results

All patients successfully underwent carotid endarterectomy, and none needed conversion to general anesthetic. In all cases, when asked after the procedure, the surgeons felt unable to give an opinion as to which type of block had been used. All patients expressed satisfaction with the anesthesia and said that they would be willing to repeat the experience under regional anesthesia.

During the study, there were no patient refusals, and only one patient was not recruited into the study as the surgeons eventually planned to operate on both carotid arteries (at different times) during the period of the study. There are statistical reasons not to recruit such patients with planned bilateral interventions (¹⁷).

Patient age, height, weight, and sex ratios were similar between the two groups (Table 1).

Table 1: Physical Characteristics of Patients

Both groups received similar doses of bupivacaine for the blocks (Table 2), with small variations arising from differences in patient size. All patients required sedation for the blocks and during surgery, and the amounts of diazepam or midazolam used were minimal and similar in both groups.

Table 2: Preoperative and Some Perioperative Data

The amount of supplemental lidocaine used was not different between the two groups (P = 0.68; Mann-Whitney U-test;Fig. 1).

Figure 1: Box plot of supplementary lidocaine requirements during carotid endarterectomy for superficial and combined blocks. The horizontal line represents the median dose, the lower and upper limits of the box represent the interquartile range, and the error bars represent the range.

Pain scores for the operation were low and satisfactory in both groups (Table 3). The later postoperative analgesic requirements were not different between the groups. Eleven patients in the deep block group needed some form of analgesia (7 requiring morphine) as compared with 8 in the superficial block group (only 5 requiring morphine). The combined group needed analgesia somewhat earlier (median time to first request being 45 vs 195 min) but this difference, although large, was not statistically significant.

Table 3: Postoperative Data

None of the patients needed placement of a Javid shunt, and there were no postoperative complications that could be attributed directly to placement of the regional block. However, the following were noted in the combined block group: one patient suffered postoperative bradycardia (heart rate 30–35 bpm) requiring treatment with atropine 0.6 mg; one patient suffered expressive dysphasia the day after operation; and one patient complained of transient swallowing difficulty which resolved before discharge to the ward.

All patients were discharged home 24–48 h after their operation (except for one patient in the superficial block group who, after 48 h, underwent elective coronary bypass surgery which had been planned to occur during his hospital stay).

Discussion

The main finding of this study is that carotid endarterectomy may be performed successfully by using either a superficial or a combined (deep and superficial) block. Our main end point (the dose of supplemental lidocaine used intraoperatively) did not differ between the two groups. Our subsidiary measures of effectiveness (pain scores, postoperative analgesia) were also not different between the groups. We could not confirm the suggestion of Davies et al. (³) that surgeons would find that the combined block provides an improved operative field.

These results confirm and extend the findings of Stoneham et al. (⁷), who compared superficial block alone with deep block alone and found the intraoperative requirements for lidocaine 1% to be similar between the blocks (and similar to the values reported in our study). They also confirmed that surgeons were unable to distinguish the type of block used from the operative conditions they experienced. It is interesting that, as in our study, the occurrence of peroperative complications in their study (albeit unrelated directly to the type of block used) was confined to the group in which a deep block had been used. The results of their study appear to differ from our results in only one major respect: in their study, an unusually large number of patients in the superficial block group required morphine postoperatively (14 of 20), compared with only 5 of 20 in our study. One reason for this difference might be that, in accordance with our normal practice, we used approximately 30 mL of 0.375% bupivacaine; Stoneham et al. (⁷) used a standard dose of only 20 mL of 0.375% bupivacaine in all patients. This would suggest that a dose of approximately 30 mL of 0.375% bupivacaine, rather than 20 mL, is a more optimum dose for a successful block.

It is important to consider some of the potential limitations to our study. Although it is not possible to blind anesthesiologists to the block they are performing, it might be desirable for a second (blinded) anesthesiologist to conduct the operation thereafter. However, this would potentially require twice as many staff devoted to the study, and this would be difficult if not impossible in most centers. Lack of blinding is unlikely to have had a significant impact on the results, because the main outcome measure was supplemental lidocaine, which was administered by the (blinded) surgeon. It is interesting to note that the single previous study of this kind (⁷) also did not blind the anesthesiologists, and in all previous studies comparing general versus regional anesthesia for carotid endarterectomy (perhaps as of necessity), anesthesiologists have not been blinded (^4–6,18).

Another potential criticism is that we used the single-injection technique for the deep block (¹²). Many practitioners would use three sites of injection (C2, C3, and C4 levels) and use much larger volumes of solution (up to 21 mL) for the deep block (^3,19). There have been no comparisons of the different methods of deep cervical block, so it is difficult to argue that one method provides a better block than another. A study comparing the two methods would be of interest. Similarly, there are no data regarding the optimum volumes of local anesthetic to use for the deep block. However, good data exists on the adequate volume to use. In their detailed study of the anatomy of the cervical plexus, Winnie et al. (¹²) explicitly stated that adequate anesthesia could be achieved with 10 mL of solution, and Mehta and Juneja (²⁰) confirmed this using a nerve stimulator with as little as 10 mL for the deep block. It appears, therefore, that our method of combined block was suitable.

We have been careful to document all sedation received by patients in the perioperative period by review of the charts. We noted that, occasionally, diazepam had been administered before the patient came to the operating room and felt it important to document these doses. It is therefore a potential limitation to the study that benzodiazepine sedation was not rigorously standardized to one drug. However, we were careful to titrate our subsequent sedation with midazolam, which is shorter acting, so that whichever combination of drugs was used, the dose was equipotent with respect to sedative effect. This is perhaps easier to achieve in carotid endarterectomy than in other operations, because it is so important to retain verbal contact with the patient throughout surgery. Table 2 shows that the total doses of sedatives used were very small and equal in both groups. It is perhaps interesting to compare our very modest sedation with that of a previous study (⁷) that reports a mean diazepam dose for this procedure of 22 mg, which is extremely large.

We used a verbal pain scale because previous work has suggested that this is suitable for immediate postoperative pain, whereas the visual analog pain score has been developed for chronic pain (¹⁴), but our evaluation of overall postoperative patient satisfaction might have been more detailed. This would be of interest, given our observation of a large (but nonsignificant) difference in the mean time to first analgesia between the two groups, in favor of the superficial block group (Table 3). If we assume that a difference in time to first analgesia of three hours is clinically relevant, and our estimate of the standard deviation of this variable is 290 minutes, we estimate that to achieve a power of 0.8, at least 43 patients in each group would be needed to assess whether superficial or combined block provides the better postoperative analgesia. Now that the efficacy of superficial block for the operation has been established by our study, it might be important for future studies to concentrate on other postoperative factors as the primary end points for analysis.

In some ways, our results are somewhat surprising, because one might expect a higher rate of success with the combined block because this is a “belt and braces” technique. The question therefore arises as to whether “adding” the deep element to a superficial block provides any extra benefit. From the results presented here, the answer is that it does not, and a superficial block alone appears to be sufficient.

The deep block is associated with several complications, which can include diaphragmatic paralysis (¹), acute respiratory distress (²), and frank intraarterial injection of local anesthetic into the vertebral artery (despite a negative aspiration test for blood) (³). We have been unable to find a single report in the literature of a serious complication resulting from use of the superficial block alone. In their series of 305 carotid endarterectomies, Lee et al. (⁹) reported only three instances of transient vocal cord paralysis and one instance of temporary 12th nerve palsy.

Our result, taken with the result of Stoneham et al. (⁷), would appear to have an important implication for clinical practice. It is generally accepted that the superficial block is easier to perform, easier to teach, and associated with fewer potential complications, and yet it is as effective as the deep or combined block. It is important, therefore, to consider whether our continued use of the deep block with its associated potential complications can be justified, when it appears to offer no added benefit to patients. The number of patients used in our study, however, was relatively small (40 patients), and larger studies are needed to confirm the apparent efficacy of the superficial block and more clearly to define its limitations.