Secondary Logo

Journal Logo

Analgesia: Research Reports

Ultrasound-Guided Perineural Circumferential Median Nerve Block With and Without Prior Dextrose 5% Hydrodissection

A Prospective Randomized Double-Blinded Noninferiority Trial

Dufour, Eric MD*; Donat, Nicolas MD; Jaziri, Souhail MD; Kurdi, Okba MD; Couturier, Christian MD§; Dreyfus, Jean-François MD, PhD; Fischler, Marc MD*

Author Information
doi: 10.1213/ANE.0b013e31825fa37d

Ultrasound-guided peripheral nerve block involves correct identification of the targeted nerve and its neighboring vascular structures until completion of homogeneous perineural spread guaranteeing nerve block effectiveness.1 Satisfactory local anesthetic (LA) spread may require numerous needle redirections and injections2 with a risk of systemic toxicity due to accidental vascular puncture, even for experienced physicians.3,4 Three reasons can explain this complication: the difficulty in visualizing the needletip during the entire procedure,5 the compression of small veins by the ultrasound probe,4 and the fact that the detection of intravascular injection is only possible after the injection, when there is no discernible hypoechoic bolus.6 Performing a hydrodissection, i.e., a perineural area of spread with dextrose 5% (D5W) may be an additional step in reducing the risk of intravascular injection and systemic toxicity. Indeed, the D5W hydrodissection that separates the nerve from neighboring vessels can be easily seen as a dark pocket in which LA can be subsequently deposited without any needle redirection. However, hydrodissection with D5W may dilute LA and decrease block effectiveness.

Therefore, we conducted a prospective, randomized, double-blind trial to test the hypothesis that onset of sensory block and sensorimotor success rates are not reduced when performing an ultrasound circumferential perineural injection with D5W preceding LA (D5W-LA group), or with LA alone (LA group).


This trial was registered with (NCT01058525). After Ethics Committee approval and after obtaining written informed consent, 100 adult patients with an ASA physical status I–II scheduled for elective ambulatory surgery of the hand or wrist involving structures innervated by the median nerve were prospectively enrolled. Patients unable to speak French, pregnant patients, and those with diabetes mellitus, history of clinical or laboratory evidence of abnormal bleeding, infection at the injection site, allergy to LA, or preexisting central or peripheral neuromuscular disease were not included. A sensory and motor examination of the median nerve of each potential participant was then performed to exclude those with abnormalities (E.D. or N.D.). Patients were allocated to the LA or D5W-LA group using computer-generated randomization.

Nerve Block Procedures

All blocks were performed in the preoperative room by the same investigator (E.D.) with standard monitoring including electrocardiogram, noninvasive arterial blood pressure, and pulse oximetry. A 20-gauge IV catheter was placed in the hand contralateral to the operative site. Intravenous sedation (1 mg of midazolam) was given. After skin sterilization with iodine polyvidone, the ulnar, radial, and median nerves were successively blocked about 4 to 5 cm above the elbow using 3 different puncture sites. The probe, protected by a sterile cover, was applied to obtain a transverse view of each nerve: at the medial aspect of the abducted and externally rotated arm, above the medial epicondyle for the ulnar nerve; at the lateral aspect of the arm internally rotated, above the lateral epicondyle for the radial nerve; and at the medial aspect of the arm abducted at right angle to the body for the median nerve.

After skin infiltration with plain lidocaine 1%, a 30° bevel, 22-gauge, 50-mm-long insulated needle (Stimuplex, B-Braun Medical, SA, France) was advanced using a short-axis in-plane approach and placed adjacent to the targeted nerve, avoiding injection through the epineurium. For the ulnar and radial nerves, the LA solution, i.e., lidocaine 1.5% with epinephrine 1:200,000 prepared with careful air bubble removal, was injected gradually after aspiration tests and with the needletip redirected until an entire circumferential spread was obtained.

An unlabeled syringe containing 6 mL of LA or 6 mL of D5W was used to perform median block in the LA group and hydrodissection in the D5W-LA group. Small aliquots were injected and the needle redirected as needed to entirely surround and mechanically separate the nerve from its surrounding structures. For both groups, the perineural circular spread obtained after this first syringe was judged to be satisfactory only if a hypoechoic image entirely surrounded the median nerve on the basis of a static axial ultrasound view (Fig. 1).7 At the end of the injection with the first syringe, the block procedure was terminated and the needle removed in the LA group, while in the D5W-LA group, the needle was placed between the brachial artery and the nerve, and 6 mL of LA was slowly injected using a second syringe into the perineural pocket with a static needle position for the duration of the injection.

Figure 1
Figure 1:
Typical image of hydrodissection with 6 mL of dextrose 5% before local anesthetic injection (hypoechoic image entirely surrounding the median nerve). The needle is indicated by the small arrows.

Block procedure time was defined as the time taken to inject the contents of the first syringe. Thus, it does not include the time to perform the LA injection of the contents of the second syringe in the D5W-LA group, which lasted a few seconds because it does not require needle mobilization. If real-time nerve swelling was observed at any time during the procedure, we considered it an intraneural injection, and the needle was immediately removed. The needle was then repositioned outside the presumed epineural sheath, and the remaining volume was injected. Patients with presumed intraneural median nerve injection were excluded from analysis. In case of incomplete circumferential spread with the first syringe, the protocol injection was continued, and patients were also excluded.

If necessary, the musculocutaneous nerve and the medial cutaneous forearm nerve were blocked at the forearm, 3 cm below the elbow crease, by skin infiltration.

Median Nerve Block Clinical Evaluation

Evaluation of sensory and motor median nerve block, by an observer blinded to the patient's group, was performed every 5 minutes for 30 minutes after needle removal.

Sensory and motor assessments were performed using cold and light touch tests and the index flexion test. Sensory block was evaluated by contralateral comparison of cold sensation elicited by ice and light touch sensation performed on the palmar aspect of the 2 distal phalanges of the index finger and on the thenar eminence. Sensory rating was quantified as follows for each test: normal sensation = 0, reduced sensation = 1, or total loss of sensation = 2. Motor examination was evaluated by asking the patient to flex the 2 distal index phalanges and scored as follows: no loss of strength = 0; reduced strength in comparison with the contralateral index finger = 1; total immobility = 2. Patients with a baseline sensory or motor abnormality (i.e., score 1 or 2) were excluded from the study. The primary criterion of blocking the median nerve was considered successful if light touch was completely suppressed (score of 2) on the palmar aspect of the 2 distal phalanges of the index finger, an area specific to the median nerve.8

At the end of median nerve block assessment and before surgery, the block was supplemented at the wrist if any potential surgical territory was not completely anesthetized. In case of pain during surgery, an additional infiltration with plain lidocaine 1% was performed by the surgeon.

Postoperative Course

A combination of paracetamol 1 g, ketoprofen 50 mg, and tramadol 50 mg every 6 hours was used for postoperative analgesia depending on pain intensity. Before discharge, patients were asked to note the time at which sensory and motor function of the first 3 fingers returned to normal, and were told that they would be telephoned the next day. Block duration was defined as the interval between time to sensory and motor recovery and the end of block performance.

Routine neurological examination (including evaluation of pain, anesthesia, paresthesia, and weakness of the operated limb) was performed by the surgeon in the outpatient clinic 3 days and 1 month postoperatively.

Statistical Analysis

Descriptive statistics for quantitative variables include mean, SD, median, minimum, and maximum. For qualitative variables, counts and proportions are provided. In view of the expected nonnormality of sample values, a randomization analog of the t test was used for noninferiority testing. Fisher exact test and Chen's calculation of confidence intervals were used for proportions. For the primary criterion (light touch block), a t test based on Welch's procedure and a Mann–Whitney test were also calculated and led to the same conclusion as the randomization t test (footnote in Table 2). A P value <0.05 was considered statistically significant.

Table 2
Table 2:
Onset Times for Sensory and Motor Median Nerve Block, Block Procedure Times, and Block Durations

The primary objective of the study was to establish whether the time to light touch block for the index finger after hydrodissection was not significantly longer than the current routine procedure. A consensus group agreed that the clinically significant critical range was longer than 5 minutes but shorter than 10 minutes, which led to a recommendation of 7 minutes. Because a pilot study that compared the effectiveness of an extraneural circumferential lidocaine injection around the median nerve at the elbow among 70 patients with and without carpal tunnel syndrome had shown an SD of 11 minutes in patients without carpal tunnel syndrome, the sample size for α = 0.05 and β = 0.80 was calculated at 43 patients per group. To cope with attrition, we decided that up to 100 patients would be recruited. The group also suggested that a 1-minute difference would be critical for the new procedure duration and that there was no real point in defining a critical threshold for block duration; we therefore set it at 1 hour. During the meeting, it was also suggested that a 10% difference in the proportion of successful blocks would be acceptable, but no effort was made to adapt the sample size to this end.

In rare cases in which complete block was not obtained by the end of the observation period (30 minutes), a value of 45 minutes was imputed.


One hundred patients were randomized to the D5W-LA (n = 44) and LA (n = 56) groups. Data from 95 patients could be analyzed: 43 in the D5W-LA group and 52 in the LA group (Fig. 2). Patient characteristics and surgical procedures were similar in the 2 groups (Table 1).

Figure 2
Figure 2:
CONSORT study flow diagram. LA = local anesthetic; D5W-LA = dextrose 5% in water–LA.
Table 1
Table 1:
Patient Characteristics and Surgical Procedures

Pain at injection site, paresthesia, and venous punctures were observed for 4 and 2, 5 and 1, and 1 and 0 patients, respectively, in the D5W-LA and LA groups (not significant). No intraneural injection was observed in the D5W-LA group, but there were 2 in the LA group.

Onset of sensory block (to cold and light touch testing) for the index finger and thenar eminence and onset of motor block were noninferior in the D5W group given the consensually agreed-upon critical limit of 7 minutes (P < 0.05), with the upper limit of the 95% confidence interval [CI] ranging from 1.2 to 5.8 minutes (Table 2).

Although a critical margin had not been formally specified, on a basis of a 10% difference in proportion, noninferiority was established for thenar cold and touch, and index cold. Because of limitations in sample size, noninferiority could not be established for index touch and motor function (Table 3). No patient, even among those excluded from analysis, required additional LA infiltration, and no major neurological complications were noted during outpatient follow-up.

Table 3
Table 3:
Percentages of Patients with a Complete Sensory or Motor Median Nerve Block at 30 Minutes


Our prospective double-blind randomized study shows that an ultrasound circumferential hydrodissection with D5W does not influence the onset time to complete light touch block nor the success rate of a median nerve block performed at the elbow. Our study did not find consequences of the dilution of LA solution by the D5W. Our results are consistent with those from previous studies performed during an axillary approach with a nerve stimulation technique, which reported that the dilution, i.e., the volume variation for a constant dose of LA drug, had no influence on block success rate and on the onset time of sensory and motor block.9,10 The high success rate in both groups suggests that lidocaine injected with needle redirections—in the LA group—or without needle redirection—in the D5W-LA group—surrounds the median nerve, allowing complete impregnation of the nerve fibers. One hypothesis is that mechanical separation of the nerve from its surrounding anatomical structures during hydrodissection with D5W contributes to the spread of LA injected secondarily. Nevertheless dynamic data are necessary to confirm this hypothesis. We also noted that the absence of needle redirections during injection of the LA in the D5W pocket did not hamper the quality of nerve block.

Several methodological points of our study should be stressed. Particular attention was given to excluding patients with neurological abnormalities that could modify block assessment and conclusions. We chose to perform a circumferential spread around the nerve, and our results would probably have been different if we had chosen a different endpoint spread pattern. However, obtaining a “doughnut” seems to be an effective endpoint, especially when a very low volume is used.1113 The volume of LA solution injected into the anatomical D5W spread space is questionable and could easily be reduced, according to recent studies using very low volume.1113 At present we routinely use 4 mL, reduced from 6 mL, to completely surround the median nerve, as recently described.14 We should also mention the choice of fluid to achieve hydrodissection. In contrast to saline, D5W has the advantage of not compromising functional motor response when a nerve stimulator is used with ultrasound, to confirm the observed image. Furthermore, the perineural injection of D5W does not lead to long-term neurological sequelae in humans or animals.15,16 The 7-minute duration threshold defining noninferiority of D5W-LA is arbitrary and based on clinical experience. However, our exploratory analyses showed that noninferiority would still be pronounced with a threshold of 6 minutes to obtain a complete light touch and of 4 minutes for the cold block of the index finger. The anesthesiologist performing the block was blinded only for the contents of the first syringe injection because it was considered impractical to set up a design blinded for the second syringe; however, the assessor remained blinded to the patients' treatment groups. Finally, analysis of block duration should be used with caution in view of the low accuracy of telephone interviews.

Our description of hydrodissection (isolating the nerve by D5W) is different from the hydrolocalization reported by Bloc et al.,17 which consists of repetitive injections of a small volume of fluid to increase the control of needletip progression towards the nerve during an out-of-plane procedure.

Because our study was not designed to show evidence of a reduction of systemic toxicity, the theoretical advantages on adverse consequences of the hydrodissection technique must be interpreted with caution. The injection of LA into D5W (both being hypoechoic solutes) is theoretically difficult to visualize, despite a real-time volume increase of the perineural space. The increased pressure in the perineural space produced by injection of 12 mL of fluid, and the consequent risk of nerve ischemia, must be considered. The risk of vascular injury should be considered during D5W hydrodissection because it requires per se needle redirections. However, it may be considered advantageous to avoid intraneural injection of LA at the cost of risking intraneural injection of a very small volume of D5W during the hydrodissection procedure.


Our study demonstrated that creating a circumferential perineural area of spread by hydrodissection with D5W into which LA solution is injected is feasible and leaves the onset time for sensory block unchanged. Whether hydrodissection modifies the spread of LA around the nerve is not known. Further studies should investigate whether a lower volume of D5W and LA solution would influence nerve block effectiveness and whether our results could be extrapolated to other LA solutions or to other peripheral nerves. The assumption that this procedure could reduce the risk of intravascular injection and systemic toxicity remains to be demonstrated.


Name: Eric Dufour, MD.

Contribution: This author helped design the study, conduct the study, and prepare the manuscript.

Name: Nicolas Donat, MD.

Contribution: This author helped conduct the study.

Name: Souhail Jaziri, MD.

Contribution: This author helped conduct the study.

Name: Okba Kurdi, MD.

Contribution: This author helped conduct the study.

Name: Christian Couturier, MD.

Contribution: This author helped conduct the study.

Name: Jean-François Dreyfus, MD, PhD.

Contribution: This author helped design the study, analyze the data, and prepare the manuscript.

Name: Marc Fischler, MD.

Contribution: This author helped design the study, analyze the data, and prepare the manuscript.

This manuscript was handled by: Terese T. Horlocker, MD.


1. Sites BD, Chan VW, Neal JM, Weller R, Grau T, Koscielniak-Nielsen ZJ, Ivani G. The American Society of Regional Anesthesia and Pain Medicine and the European Society of Regional Anaesthesia and Pain Therapy Joint Committee recommendations for education and training in ultrasound-guided regional anesthesia. Reg Anesth Pain Med 2009;34:40–6
2. Casati A, Danelli G, Baciarello M, Corradi M, Leone S, Di Cianni S, Fanelli G. A prospective, randomized comparison between ultrasound and nerve stimulation guidance for multiple injection axillary brachial plexus block. Anesthesiology 2007;106:992–6
3. Loubert C, Williams SR, Helie F, Arcand G. Complication during ultrasound-guided regional block: accidental intravascular injection of local anesthetic. Anesthesiology 2008;108:759–60
4. Zetlaoui PJ, Labbe JP, Benhamou D. Ultrasound guidance for axillary plexus block does not prevent intravascular injection. Anesthesiology 2008;108:761
5. Gray AT, Schafhalter-Zoppoth I. A concerning direction. Anesthesiology 2004;100:1325
6. Brull R, Perlas A, Cheng PH, Chan VW. Minimizing the risk of intravascular injection during ultrasound-guided peripheral nerve blockade. Anesthesiology 2008;109:1142
7. Dufour E, Cymerman A, Nourry G, Balland N, Couturier C, Liu N, Dreyfus JF, Fischler M. An ultrasonographic assessment of nerve stimulation–guided median nerve block at the elbow: a local anesthetic spread, nerve size, and clinical efficacy study. Anesth Analg 2010;111:561–7
8. Rimbaud L. Précis de Neurologie. Paris: G. Doin, 1939:668
9. Krenn H, Deusch E, Balogh B, Jellinek H, Oczenski W, Plainer-Zochling E, Fitzgerald RD. Increasing the injection volume by dilution improves the onset of motor blockade, but not sensory blockade of ropivacaine for brachial plexus block. Eur J Anaesthesiol 2003;20:21–5
10. Bertini L, Palmisani S, Mancini S, Martini O, Ioculano R, Arcioni R. Does local anesthetic dilution influence the clinical effectiveness of multiple-injection axillary brachial plexus block? A prospective, double-blind, randomized clinical trial in patients undergoing upper limb surgery. Reg Anesth Pain Med 2009;34:408–13
11. Eichenberger U, Stockli S, Marhofer P, Huber G, Willimann P, Kettner SC, Pleiner J, Curatolo M, Kapral S. Minimal local anesthetic volume for peripheral nerve block: a new ultrasound-guided, nerve dimension-based method. Reg Anesth Pain Med 2009;34:242–6
12. O'Donnell BD, Iohom G. An estimation of the minimum effective anesthetic volume of 2% lidocaine in ultrasound-guided axillary brachial plexus block. Anesthesiology 2009;111:25–9
13. Marhofer P, Eichenberger U, Stockli S, Huber G, Kapral S, Curatolo M, Kettner S. Ultrasonographic guided axillary plexus blocks with low volumes of local anaesthetics: a crossover volunteer study. Anaesthesia 2010;65:266–71
14. Harper GK, Stafford MA, Hill DA. Minimum volume of local anaesthetic required to surround each of the constituent nerves of the axillary brachial plexus, using ultrasound guidance: a pilot study. Br J Anaesth 2010;104:633–6
15. Tsui BC, Wagner A, Finucane B. Electrophysiologic effect of injectates on peripheral nerve stimulation. Reg Anesth Pain Med 2004;29:189–93
16. Tsui BC, Kropelin B. The electrophysiological effect of dextrose 5% in water on single-shot peripheral nerve stimulation. Anesth Analg 2005;100:1837–9
17. Bloc S, Ecoffey C, Dhonneur G. Controlling needle tip progression during ultrasound-guided regional anesthesia using the hydrolocalization technique. Reg Anesth Pain Med 2008;33:382–3
© 2012 International Anesthesia Research Society