Single, double or multiple-injection techniques for non-ultrasound guided axillary brachial plexus block in adults undergoing surgery of the lower arm

  • Review
  • Intervention

Authors


Abstract

Background

Regional anaesthesia comprising axillary block of the brachial plexus is a common anaesthetic technique for distal upper limb surgery. This is an update of a review first published in 2006 and updated in 2011.

Objectives

To compare the relative effects (benefits and harms) of three injection techniques (single, double and multiple) of axillary block of the brachial plexus for distal upper extremity surgery. We considered these effects primarily in terms of anaesthetic effectiveness; the complication rate (neurological and vascular); and pain and discomfort caused by performance of the block.

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library), MEDLINE, EMBASE and reference lists of trials. We contacted trial authors. The date of the last search was March 2013 (updated from March 2011).

Selection criteria

We included randomized controlled trials that compared double with single-injection techniques, multiple with single-injection techniques, or multiple with double-injection techniques for axillary block in adults undergoing surgery of the distal upper limb. We excluded trials using ultrasound-guided techniques.

Data collection and analysis

Independent study selection, risk of bias assessment and data extraction were performed by at least two investigators. We undertook meta-analysis.

Main results

The 21 included trials involved a total of 2148 participants who received regional anaesthesia for hand, wrist, forearm or elbow surgery. Risk of bias assessment indicated that trial design and conduct were generally adequate; the most common areas of weakness were in blinding and allocation concealment.

Eight trials comparing double versus single injections showed a statistically significant decrease in primary anaesthesia failure (risk ratio (RR 0.51), 95% confidence interval (CI) 0.30 to 0.85). Subgroup analysis by method of nerve location showed that the effect size was greater when neurostimulation was used rather than the transarterial technique.

Eight trials comparing multiple with single injections showed a statistically significant decrease in primary anaesthesia failure (RR 0.25, 95% CI 0.14 to 0.44) and of incomplete motor block (RR 0.61, 95% CI 0.39 to 0.96) in the multiple injection group.

Eleven trials comparing multiple with double injections showed a statistically significant decrease in primary anaesthesia failure (RR 0.28, 95% CI 0.20 to 0.40) and of incomplete motor block (RR 0.55, 95% CI 0.36 to 0.85) in the multiple injection group.

Tourniquet pain was significantly reduced with multiple injections compared with double injections (RR 0.53, 95% CI 0.33 to 0.84). Otherwise there were no statistically significant differences between groups in any of the three comparisons on secondary analgesia failure, complications and patient discomfort. The time for block performance was significantly shorter for single and double injections compared with multiple injections.

Authors' conclusions

This review provides evidence that multiple-injection techniques using nerve stimulation for axillary plexus block produce more effective anaesthesia than either double or single-injection techniques. However, there was insufficient evidence for a significant difference in other outcomes, including safety.

Plain language summary

Anaesthesia for hand and forearm surgery via single, double or multiple injections placed close to nerves in the armpit

A common method of regional anaesthesia for hand, wrist or forearm surgery is to inject local anaesthetic into the tissues surrounding nerves in the armpit. This is because in the armpit (axilla) the key nerves for the lower part of the arm are close together and are easier to locate. This type of anaesthesia is called axillary brachial plexus block. Successful blocking of the nerves produces a numb and limp arm that enables pain-free surgery. This review compared the effects of single, double and multiple (three or four) injections of local anaesthetic.

We searched the literature up until March 2013 and identified 21 randomized controlled trials for inclusion in the review. These trials involved a total of 2148 participants who were given regional anaesthesia for hand, wrist, forearm or elbow surgery. The trials used methods that were generally adequate and did not affect the validity of the findings. Eight trials compared double versus single injections. These found that fewer people in the double injection group required additional anaesthesia. However, the effect was more certain in the four trials where the nerves were located using the precise technique of neurostimulation. In the eight trials comparing multiple with single injections, and the 11 trials comparing multiple with double injections, there were significantly fewer people needing extra anaesthesia in the multiple injection groups. In addition, fewer patients in the multiple-injection group experienced tourniquet pain compared to the double-injection group. There were no other statistically significant differences in complications or patient discomfort between the two groups for any of the three comparisons. Single and double injections took less time to perform than multiple injections, but this did not reduce the total time required for adequate surgical anaesthesia to be established.

Overall, the evidence from these trials showed that injections of anaesthetic close to three or four nerves in the armpit provide more complete anaesthesia for hand and forearm surgery than one or two injections. There was, however, not enough evidence to determine if there was a significant difference in the other outcomes, including safety.

Summary of findings(Explanation)

Summary of findings for the main comparison. Double versus single-injection technique
  1. 1. Complete failure of block is defined as the need for general anaesthesia or a new plexus block to provide surgical anaesthesia.

    2. Time to readiness for surgery is defined as the time required to perform the block plus the time from completion of the block to development of surgical anaesthesia.

    3. Adverse effects lasting more than 24 hours refers mainly to neurological symptoms or deficits in the arm that was blocked.

Double injection versus single-injection technique of axillary brachial plexus block for hand, wrist or forearm surgery in adults

Patient or population: Adult patients undergoing hand, wrist or forearm surgery

Settings: Hospital

Intervention: Double-injection technique of axillary brachial plexus block

Comparison: Single-injection technique of axillary brachial plexus block

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Single injectionDouble injection
Primary anaesthesia failure38 per 10019 per 100
(11 to 32)

RR 0.51

(0.30 to 0.85)

497
(8 studies)
+++O
moderate
 
Secondary analgesia failure: Intraoperative sedation required27 per 10017 per 100
(8 to 35)

RR 0.64

(0.31 to 1.31)

129
(2 studies)
++OO
low
 
Secondary analgesia failure : Tourniquet pain16 per 1009 per 100
(4 to 25)

RR 0.58

(0.22 to 1.52)

104
(2 studies)
++OO
low
 
Complete failure of block116 per 100021 per 1000
(5 to 80)

RR 1.29

(0.33 to 5.01)

338
(6 studies)
+OOO
very low
There were no events in 4 out of 6 studies.
Time to readiness for surgery2 (minutes)See commentSee comment See comment+OOO
very low
None of the included studies assessed this outcome.
Intravascular injection55 per 1000322 per 1000
(14 to 7571)

RR 5.86

(0.25 to 137.66)

60
(1 study)
+OOO
very low
Only 1 event occurred in the study.
Adverse effects lasting more than 24 hours313 per 100016 per 1000
(4 to 77)

RR 1.25

(0.27 to 5.89)

119
(2 studies)
+OOO
very low
There were no events in 1 of the 2 studies.
*The assumed risk for the 'control' group is based on the mean value of the results for all single injection groups in the included trials reporting the outcome. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk Ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 2 Multiple versus single-injection technique

Summary of findings 2. Multiple versus single-injection technique
  1. 1. Complete failure of block is defined as the need for general anaesthesia or a new plexus block to provide surgical anaesthesia.

    2. Time to readiness for surgery is defined as the time required to perform the block plus the time from completion of the block to development of surgical anaesthesia.

    3. Adverse effects lasting more than 24 hours refers mainly to neurological symptoms or deficits in the arm that was blocked.

    4. Fanelli 1999 observed a 1% risk of transient neurological deficit in their study of 1650 patients receiving multiple-injection brachial plexus blocks.

Multiple-injection versus single-injection technique of axillary brachial plexus block for hand, wrist or forearm surgery in adults

Patient or population: Adult patients undergoing hand, wrist or forearm surgery

Settings: Hospital

Intervention: Multiple-injection technique of axillary brachial plexus block

Comparison: Single-injection technique of axillary brachial plexus block

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Single injectionMultiple injection
Primary anaesthesia failure41 per 10010 per 100
(6 to 18)

RR 0.25

(0.14 to 0.44)

682
(8 studies)
+++O
moderate
 
Secondary analgesia failure: Intraoperative sedation required27 per 10019 per 100
(11 to 32)

RR 0.70

(0.41 to 1.19)

482
(5 studies)
++OO
low
 
Secondary analgesia failure : Tourniquet pain16 per 10014 per 100
(4 to 44)

RR 0.97

(0.30 to 3.11)

379
(4 studies)
++OO
low
 
Complete failure of block124 per 10007 per 1000
(1 to 65)

RR 0.31

(0.03 to 2.71)

454
(6 studies)
+OOO
very low
There were no events in 3 out of 6 studies
Time to readiness for surgery2 (minutes)The mean block performance time ranged across control groups from 14.3 to 38.5 minutesThe mean block performance time ranged across intervention group from 21.1 to 25.0 minutes 206
(2 studies)
++OO
low
 
Intravascular injection55 per 100048 per 1000
(5 to 464)

RR 0.87

(0.09 to 8.44)

322
(3 studies)
+OOO
very low
 
Adverse effects lasting more than 24 hours313 per 10003 per 1000
(0 to 34)
RR 0.25 (0.02 to 2.59)244
(3 studies)
+OOO
very low
There were no events in 2 of the 3 studies.4
*The assumed risk for the 'control' group is based on the mean value of the results for all single-injection groups in the included trials reporting the outcome. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk Ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 3 Multiple versus double-injection technique

Summary of findings 3. Multiple versus double-injection technique
  1. 1. Complete failure of block is defined as the need for general anaesthesia or a new plexus block to provide surgical anaesthesia.

    2. Time to readiness for surgery is defined as the time required to perform the block plus the time from completion of the block to development of surgical anaesthesia.

    3. Adverse effects lasting more than 24 hours refers mainly to neurological symptoms or deficits in the arm that was blocked.

    4. Fanelli 1999 observed a 1% risk of transient neurological deficit in their study of 1650 patients receiving multiple-injection brachial plexus blocks.

Multiple-injection versus double-injection technique of axillary brachial plexus block for hand, wrist or forearm surgery in adults

Patient or population: Adult patients undergoing hand, wrist or forearm surgery

Settings: Hospital

Intervention: Multiple-injection technique of axillary brachial plexus block

Comparison: Double-injection technique of axillary brachial plexus block

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Double injectionMultiple injection
Primary anaesthesia failure26 per 1007 per 100
(5 to 10)

RR 0.28

(0.20 to 0.40)

936
(11 studies)
+++O
moderate
 
Secondary analgesia failure: Intraoperative sedation required19 per 10015 per 100
(11 to 20)

RR 0.75

(0.55 to 1.03)

716
(7 studies)
++OO
low
 
Secondary analgesia failure : Tourniquet pain13 per 1007 per 100
(4 to 11)

RR 0.53

(0.33 to 0.84)

719
(7 studies)
+++O
moderate
 
Complete failure of block123 per 10006 per 1000
(1 to 32)

RR 0.24

(0.04 to 1.41)

600
(8 studies)
+OOO
very low
There were no events in 6 out of 8 studies.
Time to readiness for surgery2 (minutes)The mean block performance time ranged across control groups from 8.8 to 38.0 minutesThe mean block performance time ranged across intervention group from 10.2 to 30 minutes 524
(5 studies)
++OO
low
 
Intravascular injection66 per 100036 per 1000
(15 to 87)

RR 0.55

(0.23 to 1.32)

476
(4 studies)
+OOO
very low
 
Adverse effects lasting more than 24 hours319 per 10004 per 1000
(0 to 31)
RR 0.20 (0.02 to 1.64)510
(6 studies)
+OOO
very low
There were no events in 5 of the 6 studies.4
*The assumed risk for the 'control' group is based on the mean value of the results for all double-injection groups in the included trials reporting the outcome. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; RR: Risk Ratio
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

An increase in the use of ambulatory hand surgery has generated the need for a method of regional anaesthesia that is comparable to general anaesthesia. Most anaesthesiologists agree that the regional technique has to satisfy four criteria for inclusion in their clinical practice. It should be effective, fast, safe and cause the patient either no, or only minimal, pain.

The three main nerves of the upper extremity (median, ulnar and radial) are enclosed in the axilla by the fascial neurovascular sheath. This limits the spread of fluid. Burnham 1958 discovered that filling this sheath with local anaesthetic could simplify the blocking procedure to a single axillary injection. The fourth main nerve of the upper extremity, the musculocutaneous nerve, usually leaves the brachial plexus more proximally, at the cord level in the infraclavicular area. De Jong 1961, using the mathematical formula for a cylinder and assuming equal proximal and distal spread from the injection site, calculated that 42 ml of local anaesthetic was sufficient to reach this area and thus anaesthetize the whole arm in the average adult. Thompson and Rorie (Thompson 1983) were the first to show (by computed tomograms) that the median, ulnar and radial nerves lie in separate fascial compartments within the neurovascular sheath. Small septae divide the neurovascular sheath and limit the circumferential spread of local anaesthetic. This provided a rational explanation for incomplete blocks. The anatomical study by Lassale and Ang, based on plaster injection into the axillary perivascular space, did not confirm the existence of a true neurovascular sheath (Lassale 1984). In a similar study, Vester-Andersen et al did not find the fascial septae separating the nerves but noticed that in all dissections only the median and ulnar nerves were engulfed by injected gelatine (Vester-Andersen 1986a). The musculocutaneous and radial nerves had either a partial contact or none at all. Partridge et al found interneural septae which were easily broken by injection of dyed latex (Partridge 1987). They therefore concluded that the septae did not limit fluid spread and had no clinical significance for anaesthesia. All these reports were based on either personal experience in a small number of patients or on cadaver studies, and none of them were controlled.

Before the 1960s, the prevalent block techniques were double or multiple axillary injections. After De Jong's report in 1961, the single-injection technique, being the simplest, became standard. Entry into the fascial neurovascular sheath was signalled either by a fascial 'click' or elicitation of paraesthesiae in the arm. The proximal spread of local anaesthetic was considered mandatory for success. The incomplete blocks were explained by insufficient volume of local anaesthetic. in the 1980s, however, Vester-Andersen et al repeatedly showed that, in spite of increased local anaesthetic volumes or concentrations, the rate of incomplete axillary blocks was high (Vester-Andersen 1984a; Vester-Andersen 1984b; Vester-Andersen 1986a). In the early 1990s, the double-injection transarterial technique using high doses of local anaesthetic gained popularity in the USA (Stan 1995; Urban 1994). In this technique, transfixion of the axillary artery was deliberately achieved to confirm entry into the neurovascular sheath; local anaesthetic was then injected behind (posterior to) as well as in front of (anterior to) the artery, in anticipation that this would increase the chance of spread to all components of the plexus.

At approximately the same time, technical development of peripheral nerve stimulators and insulated blunt needles allowed electrolocation of the individual plexus nerves. While electrolocation (also known as neurostimulation) may be applied to single and double-injection techniques, its greatest advantage is that it allows targeted injection around three or more of the main nerves to the arm. This is known as the multiple-injection technique. Lavoie et al and Koscielniak-Nielsen et al reported that this technique was superior to the single-injection method (K-Nielsen 1997a; Lavoie 1992a), and Koscielniak-Nielsen et al reported its superiority over the transarterial technique (K-Nielsen 1998a; K-Nielsen 1999c). Coventry et al and Sia et al drew similar conclusions when comparing triple injection with double injection (Coventry 2001a; Sia 2001a).

Why it is important to do this review

The first version of our review (Handoll 2006) reported that no recommendations were available as to which of these techniques (single, double or multiple injection) were preferable, and that the choice is left to the personal preferences of the anaesthesiologist. The findings of the systematic review were in favour of multiple-injection techniques using nerve stimulation for axillary plexus block in terms of providing more effective anaesthesia than either double or single-injection techniques. It emphasized, however, that there was insufficient evidence for other outcomes, especially longer-term outcomes and safety.

We updated this review in 2011 (Chin 2011) with data from eight newly-identified studies, which confirmed our previous findings. At this time we also clarified the scope of this review. We explicitly excluded ultrasound-guided axillary block techniques as these are clinically very different from the landmark and neurostimulation-guided techniques described above. Our review (Chin 2011) noted that ultrasound-guided axillary block was "a multiple injection technique where each of the four individual nerves are identified and targeted under direct vision". While there has been investigation since then comparing single, double, and multiple-injection ultrasound-guided techniques of axillary block (Bernucci 2012; Huynh 2012; Tran 2012), this remains a separate question from that addressed in this review.

Despite the fact that scientific research into peripheral nerve blockade is now focused primarily on ultrasound-guided techniques, this review remains clinically relevant as there are areas in the world, particularly developing countries, where ultrasound technology is not readily available to anaesthesia practitioners. This updated review therefore sought to determine which of the non-ultrasound guided techniques (single, double or multiple injection) are preferable in performing axillary block of the brachial plexus.

Objectives

To compare the relative effects (benefits and harms) of three injection techniques (single, double and multiple) of axillary block of the brachial plexus for distal upper extremity surgery. We considered these effects primarily in terms of anaesthetic effectiveness; the complication rate (neurological and vascular); and pain and discomfort caused by performance of the block.

Methods

Criteria for considering studies for this review

Types of studies

We included randomized controlled trials (RCTs) that compared single with double or multiple-injection techniques, or double with multiple-injection techniques, for axillary block.

Types of participants

We included adults (generally over 18 years of age) undergoing surgery of the distal upper extremity: the hand, wrist, forearm, elbow, or some combination thereof. We excluded trials that focused on children only.

Types of interventions

  1. Single injection in the axilla (including injection through a catheter)

  2. Double injection in the axilla (transarterial, elicitation of two paraesthesiae, electrolocation of two nerves, insertion of two needles)

  3. Multiple-injection techniques in the axilla (three or more paraesthesiae or electrolocations) regardless of the local anaesthetic, pH adjustment or additives

In this review, multiple-injection techniques, in particular nerve stimulator guided multiple-injection techniques, were the 'experimental' intervention. Single-injection and double-injection techniques represented 'standard' interventions. For comparisons of single versus double-injection techniques, the single injection was the 'standard'.

We distinguished between 'guided' (neurostimulation as the endpoint for nerve location) and 'blind' (fascial clicks, paraesthesia, or arterial puncture as the endpoints for nerve location) injection techniques.

Exclusions: ultrasound-guided techniques of nerve location; and routine supplementary analgesia (local anaesthetic infiltration of the surgical site; general anaesthetics and systemic opioids), with the exception of systemic opioids when used as a component of sedation (for example, small doses of opioids used in combination with benzodiazepines).

Types of outcome measures

Primary outcomes

Primary analgesia or anaesthesia failure. This was represented by the use of any additional anaesthetic or surgical intervention to ensure adequate surgical anaesthesia. This outcome can be measured or defined in various ways. It can be: a) incomplete overall sensory block or analgesia; or b) incomplete or inadequate sensory block or analgesia for the specific surgery undertaken at an appropriate (generally 30 minutes) time interval after completion of the blocking procedure.

Failure is also indicated by one or more of the following: use of supplementary analgesia either to ensure a) complete overall analgesia or b) analgesia for the surgical site; change in anaesthesia method, such as recourse to general anaesthesia; and the curtailment or modification of the planned surgical procedure due to inadequate anaesthesia. We also reported incomplete motor block.

Secondary outcomes
  1. Secondary analgesia failure, such as surgical site pain, tourniquet pain or need for intraoperative sedation.

  2. Timing, primarily time to achieve readiness for surgery.

  3. Complications and adverse effects: these included vascular complications such as haematoma; accidental intravascular injection and its sequelae (tachycardia, dizziness, loss of consciousness, seizures); and neurological complications, including residual neurapraxias not related to surgical site, that were present for more than 24 hours.

  4. Pain and discomfort during block performance.

Search methods for identification of studies

In the first version of this review (Handoll 2006), one author (Zbigniew J Koscielniak-Nielsen (ZK-N)) undertook the search (to August 2004) and Karen Hovhannisyan (KH) as Trials Search Co-ordinator, Cochrane Anaesthesia Review Group (CARG) supplemented this search (to March 2005).

In the second version of this review (Chin 2011), KH updated the database search strategies that had been used in the first version of this review and ran these (March 2011) for the following databases: Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library), MEDLINE and EMBASE.

For this update, we reviewed search downloads from CENTRAL, MEDLINE and EMBASE obtained using the same search strategy (see Appendix 1). The search dates for these searches were:

  • CENTRAL (2013, Issue 3) in The Cochrane Library;

  • Ovid MEDLINE (1956 to week 5 March 2013);

  • EMBASE (1980 to Week 14 2013).

As before, we applied no language restrictions.

The description of the search methods used in the first version of the review is given in Appendix 2.

Data collection and analysis

Selection of studies

In the first version of the review, one review author (ZK-N) compiled a set of reports of controlled trials testing various aspects of axillary brachial plexus neural blockade for surgery of the distal upper limb, using the author-performed search strategy, supplemented by his personal reference collection. ZK-N provided the second author, Helen Handoll (HH), with copies of the first pages (or more as required) of each report. Both authors independently selected a set of potentially eligible trials and then, based on full-text versions, independently selected trials that met the review inclusion criteria. All disagreements were resolved by discussion.

HH checked through the supplementary search results (March 2005) from three databases and put forward eligible trials for selection and future consideration.

In the second version of the review, both of the review authors (Ki Jinn Chin (K-JC) and HH) independently selected potentially eligible trials from the search downloads of CENTRAL, MEDLINE and EMBASE that were provided by the CARG Trials Search Co-ordinator (KH). Then, upon discussion and clarification of the inclusion criteria (see Differences between protocol and review), we independently selected trials from full-text versions.

In this updated version of the review, two of the review authors (K-JC and Husni Alakkad (HA)) independently selected potentially eligible trials from the search downloads of CENTRAL, MEDLINE and EMBASE that were provided by the CARG Trials Search Co-ordinator (KH). Full-text versions of these potentially eligible studies were obtained and trials were independently selected based on the inclusion and exclusion criteria.

Data extraction and management

In the first version of the review, three people (the two review authors and one other, Saúl Rugeles) performed data collection. For all versions, two people independently extracted trial information and results using a piloted data extraction form. Where available, we collected information on the following: trial methods (including methods of randomization and outcome assessment); details of the injection technique; the local anaesthetic agent; drugs used for sedation; baseline characteristics of the trial population (including sex, age, mental status and surgical interventions); and outcome measures such as pain and complications of the blocking procedure, as listed above. We resolved any differences by discussion via email correspondence. We contacted trial authors for further details of their trials.

In the first version of the review, because ZK-N was the lead investigator of four included trials, the other review authors undertook independent data entry into Review Manager (RevMan 4.2) and performed the presentation and interpretation of these four trials. However we took note of feedback, particularly corrections, from ZK-N.

In the second version of the review, both of the review authors (K-JC and HH) independently extracted trial information and results using a piloted data extraction form as described above. We resolved any differences by discussion via email correspondence. We contacted trial authors for further details of their trials. Both authors undertook independent data entry into RevMan 5.1.

In the current version of the review, two of the review authors (K-JC and Javier Cubillos (JC)) independently extracted trial information and results. Any differences were resolved by discussion, via email correspondence or video-conference. We contacted trial authors for further details of their trials. Data from newly-included trials were independently entered into RevMan 5.2 by two of the authors (K-JC and HA).

Assessment of risk of bias in included studies

For the first version of the review, two people independently assessed adequacy of study design using an adaptation of the eight-item scoring scheme (see Appendix 3) formerly developed by CARG. We assessed the following items: allocation concealment; description of study inclusion and exclusion criteria; intention-to-treat analysis (description of withdrawals); description of baseline characteristics of the trial population (in particular age, sex, mental status and type of surgery); comparability of care programmes other than the trial interventions (including anaesthetist experience with technique); outcome assessor blinding; and timing of outcome measurement (minimum 24 hours). As ZK-N was the lead investigator of four of the included trials in the first version of the review, these trials were reviewed independently of him. We resolved any differences by discussion.

In the second version of the review, as well as this update, we assessed risk of bias using the tool outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). This tool incorporates assessment of randomization (sequence generation and allocation concealment), blinding (of participants, treatment providers and outcome assessors), completeness of outcome data, selection of the outcomes reported and other sources of bias. We considered all outcomes in our assessment of blinding and completeness of outcome data. We assessed two additional sources of bias: selection bias resulting from major imbalances in key baseline characteristics (age, sex, type of surgery, mental status); and performance bias, such as that resulting from a lack of comparability in the experience of the anaesthetist with the interventions being compared.

In the second version of the review, one author (HH) assessed risk of bias of the already included trials, drawing on the previous assessments. Both authors independently assessed the newly included trials. We resolved any differences by discussion.

In the current version of the review, all three authors (KJ-C, JC, and HA) independently assessed the newly included trials. We resolved any differences by discussion.

Measures of treatment effect

We calculated risk ratios and 95% confidence intervals for dichotomous outcomes, and mean differences and 95% confidence intervals for continuous outcomes.

Assessment of heterogeneity

Heterogeneity was assessed by visual inspection of the forest plot (the analysis) along with consideration of the Chi² test for heterogeneity and the I² statistic (Higgins 2003).

Data synthesis

We reviewed the data from the included studies qualitatively and then, where possible and appropriate, presented data in the analysis and combined the data quantitatively. We pooled results of comparable groups of trials using the fixed-effect model and 95% confidence intervals. Where there was significant and unexplained heterogeneity among studies (P < 0.10 using Q statistics), we applied the random-effects model.

Subgroup analysis and investigation of heterogeneity

We planned subgroup analyses on the method of nerve location (paraesthesia, transarterial, nerve stimulation) and broad location of the surgery (hand, wrist, forearm and elbow). To test whether the subgroups were statistically significantly different from one another, we tested the interaction using the technique outlined by Altman and Bland (Altman 2003).

Sensitivity analysis

Where possible, we planned or undertook sensitivity analyses examining various aspects of trial and review methodology, including the effects of missing data and study quality (specifically allocation concealment and outcome assessor blinding).

Results

Description of studies

Results of the search

In the first version of this review (Handoll 2006), 73 studies were initially identified all of which involved investigation of some aspect of brachial plexus blockade for surgery of the distal upper limb. We rejected 55 studies at the first screening. The majority of rejected studies compared different types or doses of anaesthetic; the others investigated various physical aspects such as arm position, the use of digital pressure, and different techniques and approaches. We included 12 of the 18 remaining studies; the other six were excluded for reasons given in the Characteristics of excluded studies table.

In the second version of this review, we identified 18 articles that were related to new studies for potential inclusion, of which we excluded 10 after reviewing the full-text reports or after some reconsideration or clarification of the inclusion criteria of the review. We included six new articles related to eight trials (Hickey 1993; Imbelloni 2005; Rodriguez 2005; Rodriguez 2008; Sia 2010a; Sia 2010b; Sia 2010c; Turkan 2002). Three trials (Sia 2010a; Sia 2010b; Sia 2010c) were run concurrently and reported in the same article (see Figure 1). Of the remaining two articles, one was a letter (Geier 2006) commenting on Imbelloni 2005 and was hence excluded. The other was not included as it was awaiting translation and classification (Ramirez-Gomez 2010).

Figure 1.

Search flow diagram for Chin 2011.

For the current update, all three authors independently screened the search results from three databases: CENTRAL (28 references); EMBASE (356 references) and MEDLINE (272 references). We identified three articles related to new studies for potential inclusion (Bernucci 2012; Huynh 2012; Tran 2012) but excluded all three after reviewing the full-text reports as they involved ultrasound-guided axillary block techniques and therefore fell outside the scope of the review. We included the study that had been awaiting classification in the previous version (Ramirez-Gomez 2010).

Included studies

We included a total of 21 trials in this update, one of which was new (Ramirez-Gomez 2010). Details of individual trials are provided in the Characteristics of included studies table. All 21 included trials were reported in full. We obtained a translation for the two trials (Ramirez-Gomez 2010; Serradell Catalan 2001) not reported in the English language.

Setting

Each of the 21 trials took place in one of 10 countries (Brazil: 1; Canada: 1; Denmark: 4; Finland: 2; Italy: 4; Mexico: 1; Spain: 3; Turkey: 1; UK: 2; USA: 2). All four Danish trials had the same lead investigator (Koscielniak-Nielsen) and shared many trial characteristics. All four Italian trials also had the same lead investigator (Sia); three of these trials (Sia 2010a; Sia 2010b; Sia 2010c) were run concurrently and were published together.

Participants

The 21 trials included a total of 2148 participants; the number of participants in individual trials ranged from 50 (Inberg 1999; Pere 1993) to 138 (Sia 2010a; Sia 2010b; Sia 2010c). Fourteen patients were excluded after randomization because of the inability to locate the desired nerves, in the three concurrent trials conducted by Sia (Sia 2010a; Sia 2010b; Sia 2010c); the distribution of these 14 patients between the three trials is not known. The percentage of male participants ranged from 2% to 75% in the 17 trials providing this information. The mean ages of trial participants, reported by 19 trials, ranged between 37 and 58 years; the inclusion of exclusively adult participants was confirmed in 10 trials providing age-range data, or from their inclusion criteria. Eighteen trials reported the requirement for informed consent.

Five trials (Baranowski 1990; Goldberg 1987; Inberg 1999; Lavoie 1992; Ramirez-Gomez 2010) gave no exclusion criteria relating to anaesthesia. Of the other 16 trials, nine trials excluded people with an American Society of Anaesthesiologists (ASA) score greater than two (Imbelloni 2005; K-Nielsen 1999a; K-Nielsen 1999b; Pere 1993; Sia 2001; Sia 2010a; Sia 2010b; Sia 2010c; Turkan 2002) and seven trials excluded people with an ASA score greater than three.

The description of the types of surgery, including location or site and whether elective or acute, was generally limited in the trial reports but it was usually enhanced on receipt of further information from trialists. Details of the types of or indications for surgery were given for 11 trials (Coventry 2001; Goldberg 1987; Imbelloni 2005; K-Nielsen 1997; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b; Ramirez-Gomez 2010; Sia 2010a; Sia 2010b; Sia 2010c) but were quantified in full in only three trials (Coventry 2001; Goldberg 1987; Ramirez-Gomez 2010) and split by treatment group in only one trial (Coventry 2001). Surgery was explicitly restricted to the hand or wrist, or both, in five trials (Goldberg 1987; K-Nielsen 1997; Sia 2010a; Sia 2010b; Sia 2010c) and was probably limited to the same locations in Baranowski 1990. Eight trials also included forearm and elbow surgery (Inberg 1999; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b; Lavoie 1992; Pere 1993; Ramirez-Gomez 2010; Rodriguez 2008). While including forearm surgery, elbow surgery was not mentioned for Serradell Catalan 2001, Sia 2001, or Imbelloni 2005. There was no indication of location in Coventry 2001, although specific hand and wrist operations were listed. Surgery was referred to as 'elective' in three trials (Coventry 2001; K-Nielsen 1997; Sia 2001); 'scheduled' in another three trials (Baranowski 1990; Goldberg 1987; Inberg 1999); and 'post-traumatic' in Serradell Catalan 2001. Mixed elective and acute surgery were undertaken in three trials (K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b) and, probably, also in Lavoie 1992 and Ramirez-Gomez 2010. There was no information on the urgency of the operation in nine trials (Hickey 1993; Imbelloni 2005; Pere 1993; Rodriguez 2005; Rodriguez 2008; Sia 2010a; Sia 2010b; Sia 2010c; Turkan 2002).

Interventions

Fourteen trials (Coventry 2001; Imbelloni 2005; Inberg 1999; K-Nielsen 1997; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b; Pere 1993; Ramirez-Gomez 2010; Rodriguez 2008; Sia 2001; Sia 2010a; Sia 2010b; Sia 2010c) had two intervention groups. Four trials (Baranowski 1990; Goldberg 1987; Hickey 1993; Turkan 2002) had three intervention groups. Rodriguez 2005 had four intervention groups. The remaining two trials (Lavoie 1992; Serradell Catalan 2001) had five intervention groups. The trials made the following comparisons according to the aims of this review.

Double versus single-injection technique

Eight trials (Goldberg 1987; Hickey 1993; Inberg 1999; Lavoie 1992; Pere 1993; Rodriguez 2005; Serradell Catalan 2001; Turkan 2002) made this comparison.

Multiple versus single-injection technique

Eight trials (Baranowski 1990; K-Nielsen 1997; K-Nielsen 1999b; Lavoie 1992; Ramirez-Gomez 2010; Rodriguez 2005; Serradell Catalan 2001; Sia 2010a) made this comparison.

Multiple versus double-injection technique

Eleven trials (Coventry 2001; Imbelloni 2005; K-Nielsen 1998; K-Nielsen 1999a; Lavoie 1992; Rodriguez 2005; Rodriguez 2008; Serradell Catalan 2001; Sia 2001; Sia 2010b; Sia 2010c) made this comparison.

The method of nerve location varied among the studies (see Table 1) and can be broadly grouped into the following four methods: (1) transarterial (seven trials); (2) Winnie's perivascular (two trials); (3) paraesthesia (two trials); and (4) neurostimulation (18 trials).

Table 1. Methods of nerve location
  1. LA = local anaesthetic; mA = milliamperes

Method of nerve locationNumber of injectionsTrials

Transarterial

The axillary artery is palpated and deliberately transfixed with a needle. The needle is then either withdrawn to inject LA anterior (superficial) to the artery, or inserted deeper to inject LA posterior to the artery, or both.

Single - anterior Hickey 1993
Single - posterior Hickey 1993; K-Nielsen 1999b
Double Goldberg 1987; Hickey 1993; Imbelloni 2005; K-Nielsen 1998; K-Nielsen 1999a; Pere 1993

Winnie's perivascular technique

A needle is inserted adjacent to the axillary artery until a fascial click is felt, signifying penetration of the neurovascular fascial sheath. A catheter may be also inserted proximally within the sheath. LA is then injected, usually as a single bolus, while applying distal pressure to promote proximal spread of the LA.

Single Baranowski 1990; Turkan 2002

Paraesthesia

A needle is inserted adjacent to the axillary artery and manipulated to elicit paraesthesia in the distribution of one or more of the four terminal nerves. LA is then injected at these locations.

Single Goldberg 1987
Multiple Baranowski 1990

Neurostimulation (electrolocation)

A needle is inserted adjacent to the axillary artery and manipulated until it comes into close proximity to one or more of the four terminal nerves. An electric current is passed through the needle and needle-nerve proximity is signalled by an appropriate movement (motor response) of the forearm or hand, usually at currents of ≤ 0.5 mA. LA is injected at these locations.

Single Inberg 1999; K-Nielsen 1997; Pere 1993; Ramirez-Gomez 2010; Rodriguez 2005; Serradell Catalan 2001; Sia 2010a
Double Coventry 2001; Inberg 1999; Lavoie 1992; Rodriguez 2005; Rodriguez 2008; Serradell Catalan 2001; Sia 2001; Sia 2010b; Sia 2010c
Multiple Baranowski 1990; Coventry 2001; Imbelloni 2005; K-Nielsen 1997; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b; Lavoie 1992; Ramirez-Gomez 2010; Rodriguez 2005; Rodriguez 2008; Serradell Catalan 2001; Sia 2001; Sia 2010a; Sia 2010b; Sia 2010c
Outcomes

We have documented the length of follow-up and the types of outcomes assessed in individual trials in the Characteristics of included studies table. Further details of the methods used to assess and define sensory and motor blockade are presented in Appendix 4.

With the exception of six trials (Lavoie 1992; Ramirez-Gomez 2010; Sia 2010a; Sia 2010b; Sia 2010c; Turkan 2002), the included trials provided separate data on anaesthesia outcomes (for example sensory blockade) for named individual nerves. We have not presented these data in this review because our focus is on overall measures of incomplete or inadequate anaesthesia.

Monitoring of longer-term effects (24 hours or over), particularly adverse effects, was conducted in 11 trials.

Excluded studies

Nineteen studies were excluded for reasons given in the Characteristics of excluded studies table; six of these were identified in the first version of this review and 10 in the second version. Three new trials were identified and excluded in this update because they involved ultrasound-guided techniques (Bernucci 2012; Huynh 2012; Tran 2012).

Risk of bias in included studies

The risk of bias judgements on seven items for the individual trials are summarized in Figure 2 and Figure 3, and described in the risk of bias tables in Characteristics of included studies. We judged items as having a low, high, or 'unclear' risk of bias. An 'unclear' verdict often reflected a lack of information upon which to judge the item. Successful contact with trial investigators usually resulted in an improved assessment of one or more items. Lack of information on blinding, primarily assessor blinding, was always taken to imply that there was no blinding and was rated as high risk of bias. A high risk of bias rating was given for single items in six trials; this related to a lack of assessor blinding in four of these. A high risk of bias was given for two items in Ramirez-Gomez 2010.

Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

There was a general lack of detail on the method of randomization and measures taken to conceal treatment allocation in the included trials. Only four trials (K-Nielsen 1997; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b) were judged at low risk of bias, resulting from adequate sequence generation and allocation concealment. Allocation was judged as concealed in Coventry 2001 but there were insufficient details on the shuffling of the envelopes to confirm the generation of an adequate randomization sequence. The use of an open randomization list by Serradell Catalan 2001 meant this trial was judged at high risk of selection bias.

Blinding

Assessor blinding for the primary outcome was absent in Ramirez-Gomez 2010 and was not mentioned in four trials (Baranowski 1990; Hickey 1993; Imbelloni 2005; Pere 1993), which were thus judged at high risk of bias for this item. Blinding was incomplete for Turkan 2002, which was judged as 'unclear' for this item. While safeguards were rarely described, the risk of bias was considered low for those trials that reported blinding.

Incomplete outcome data

The short follow-up in most of these trials prevented loss of follow-up for the primary outcome being a serious issue and we judged that all trials performed intention-to-treat analysis in that there was no cross over. 'Unclear' ratings generally resulted from post-randomization exclusions but we noted also that none of the trials that followed up people after surgery explicitly reported that all trial participants attended their surgical follow-up. Unaddressed reporting inconsistencies in Ramirez-Gomez 2010 and Rodriguez 2005 were the reason behind the high risk of bias judgement for this item in these trials.

Selective reporting

The lack of protocols or trial registration entries hampered the assessment of risk of bias from selective reporting. However, we judged that selective reporting bias was avoided by virtue of the consistent approach taken in the planning of two series of trials headed by Koscielniak-Nielsen (K-Nielsen 1997; K-Nielsen 1998; K-Nielsen 1999a; K-Nielsen 1999b) and Sia (Sia 2010a; Sia 2010b; Sia 2010c) and the provision of additional data on request.

Other potential sources of bias

Bias resulting from major imbalances in baseline characteristics was judged as low in five trials and 'unclear' in the remainder. Generally the lack of information on the distribution in the types of surgery undertaken (and implicated nerves) in the intervention groups was the reason for uncertainty. The risk of performance bias, primarily based on an assessment of reported operator experience and comparability of this between intervention groups, was judged as low in 11 trials and 'unclear' in the rest. While we also based our judgement on an interpretation of individual trial procedures, we did not think the lack of reporting by trials on comparability of care programmes impacted on trial validity.

Effects of interventions

See: Summary of findings for the main comparison Double versus single-injection technique; Summary of findings 2 Multiple versus single-injection technique; Summary of findings 3 Multiple versus double-injection technique

The 21 included trials involved a total of 2148 participants who received regional anaesthesia for hand, wrist, forearm or elbow surgery.

Where data were available, we summed the results of the two or three intervention groups that fell into the same category (for example single injection) for the seven trials (Baranowski 1990; Goldberg 1987; Hickey 1993; Lavoie 1992; Rodriguez 2005; Serradell Catalan 2001; Turkan 2002) with more than two intervention groups. As stated a priori, we performed subgroup analysis according to the method of nerve location. We limited this to the outcome of primary analgesia failure and subgrouped according to whether nerves were located by nerve stimulation (or, more rarely, paraesthesia) or not, as in the transarterial method. Due to lack of data, we were unable to perform subgroup analyses according to the site of surgery. We were also unable to undertake sensitivity analyses to test aspects of trial methodology.

For primary analgesia or anaesthesia failure, we also presented data subgrouped according to whether this outcome was defined as incomplete overall sensory block, as determined by the individual trials, or incomplete sensory block as indicated by the need for supplementation at the surgical site.

Double versus single-injection technique

Eight trials (Goldberg 1987; Hickey 1993; Inberg 1999; Lavoie 1992; Pere 1993; Rodriguez 2005; Serradell Catalan 2001; Turkan 2002) made this comparison in a total of 498 participants. One person was excluded from Hickey 1993 following an aborted axillary block in which tachycardia and lightheadedness occurred during injection. The three incomplete procedures that occurred in the double-injection group of Rodriguez 2005 were included in an intention-to-treat analysis.

Primary analgesia or anaesthesia failure

The pooled results, using the random-effects model because of significant (P = 0.02) and substantial (I2 = 58%) heterogeneity, showed a statistically significant decrease in primary analgesia or anaesthesia failure (incomplete sensory block) in the double-injection group (see Figure 4) (RR 0.51, 95% CI 0.30 to 0.85). Figure 4 also presents the results for the trials subgrouped according to the technique used for double injection (transarterial versus neurostimulation). The results of the four trials (Goldberg 1987; Hickey 1993; Pere 1993; Turkan 2002) using transarterial injection showed no statistically significant difference between the double and single-injection groups (failure: RR 0.72, 95% CI 0.33 to 1.58), whereas a double injection was superior in those trials where location was by neurostimulation in both groups (failure: RR 0.40, 95% 0.22 to 0.73). A test of interaction based on fixed-effect model risk ratios showed that statistically the results of the two subgroups were significantly different from each other (two-tailed z-test = 0.0261). However, this was not the case for the random-effects model results (two-tailed z-test = 0.243) and the results in the two subgroups were also heterogeneous, hence, the differences in the method of nerve location do not appear to explain fully the heterogeneity of the overall result.

Figure 4.

Forest plot of comparison: 1 Double versus single-injection technique, outcome: 1.1 Primary anaesthesia failure (incomplete sensory block).

Analysis 1.2 presented the results subgrouped according to the definition of primary analgesia failure: incomplete overall sensory block (RR 0.43, 95% CI 0.24 to 0.76) or supplemental blocks required for surgical site (RR 0.43, 95% CI 0.17 to 1.11). The results for Inberg 1999 illustrated the difference in these two definitions. In the first, complete anaesthesia (sensory block) was sought, and supplemental blocks were performed if necessary to achieve this. In the second, only anaesthesia of the anticipated surgical site was sought and, as a result, the extent of supplementation was generally less.

The plexus block failed totally in seven people, six of whom had general anaesthesia and one (in Inberg 1999) had a new plexus block; there was no difference between the two groups in this outcome (see Analysis 1.3) (RR 1.29, 95% CI 0.33 to 5.01). There was no statistically significant difference between the two injection groups in the numbers of participants with incomplete motor block (see Analysis 1.4) (RR 0.78, 95% CI 0.0.58 to 1.03).

Secondary analgesia failure, timing, complications and other outcomes

None of the pooled differences between the two injection groups for secondary analgesia failure (surgical site pain, tourniquet pain or intraoperative sedation) were statistically significant (see Analysis 1.5). The only trial (Serradell Catalan 2001) reporting the time to perform the nerve block found that the double nerve block took significantly more time to perform (mean difference (MD) 1.65 minutes, 95% CI 0.72 to 2.58 minutes). None of the other differences in duration of operation, duration of tourniquet use and duration of block were statistically significant between the two groups (see Analysis 1.6). Four cases of venous puncture and six of paraesthesia occurred during nerve block in Serradell Catalan 2001; and one case of tachycardia and lightheadedness (signifying probable intravascular injection) in Hickey 1993. None of the differences between the two groups were statistically significant (see Analysis 1.7). The seven adverse effects, all lasting 24 hours, were all persistent paraesthesias in Serradell Catalan 2001 (see Analysis 1.8). The only persistent adverse effect, recorded at three months in Serradell Catalan 2001, that was noted in the 20 included trials was described as neurological dysfunction. This occurred in one participant of one of the two double-injection groups. Serradell Catalan 2001 found no statistically significant difference between the double and single-injection groups in patient discomfort or their dissatisfaction with the anaesthetic method (see Analysis 1.9).

Multiple versus single-injection technique

Eight trials (Baranowski 1990; K-Nielsen 1997; K-Nielsen 1999b; Lavoie 1992; Ramirez-Gomez 2010; Rodriguez 2005; Serradell Catalan 2001; Sia 2010a) made this comparison in a total of 684 participants. Two participants were excluded from K-Nielsen 1999b; one because of lack of comprehension of trial procedures and the other because of chest pain resulting in cancelled surgery. The one incomplete procedure that occurred in the multiple-injection group of Rodriguez 2005 was included in an intention-to-treat analysis.

Primary analgesia or anaesthesia failure

The pooled results, using the random-effects model because of significant (P = 0.04) and substantial heterogeneity (I² = 52%), showed a statistically significant decrease in primary analgesia or anaesthesia failure (incomplete sensory block) in the multiple-injection group (see Figure 5) (RR 0.25, 95% CI 0.14 to 0.44). Figure 5 also presents the trials subgrouped according to the technique used for single injection (neurostimulation versus no neurostimulation). The results of both groups of trials showed that multiple injections, all located via nerve stimulation, provided more complete sensory block than single injections located with (failure: RR 0.19, 95% CI 0.09 to 0.41) or without (failure: RR 0.40, 95% CI 0.25 to 0.65) the use of a nerve stimulator. A test of interaction showed that statistically the results of the two subgroups were not significantly different from each other (two-tailed z-test = 0.190).

Figure 5.

Forest plot of comparison: 2 Multiple versus single-injection technique, outcome: 2.1 Primary anaesthesia failure (incomplete sensory block).

Analysis 2.2 shows the results subgrouped according to the definition of primary analgesia failure: incomplete overall sensory block (RR 0.28, 95% CI 0.12 to 0.64), or supplemental blocks required for surgical site (RR 0.23, 95% CI 0.10 to 0.52). It should be noted that K-Nielsen 1997 was placed in the second category on the basis that it stipulated that supplementation of the musculocutaneous nerve was done only if necessary for surgery.

The plexus block failed totally in six people, all of whom then received general anaesthesia (see Analysis 2.3) (RR 0.31, 95% CI 0.03 to 2.71). The pooled results for incomplete motor block, using the random-effects model because of significant (P = 0.03) and substantial heterogeneity (I² = 66%), showed a statistically significant increase in incomplete motor block in the single-injection group (see Analysis 2.4) (RR 0.61, 95% CI 0.39 to 0.96).

Secondary analgesia failure, timing, complications and other outcomes

None of the pooled differences between the two injection groups for secondary analgesia failure (surgical site pain, tourniquet pain or intraoperative sedation) were statistically significant (see Analysis 2.5). Pooled analysis (using the random-effects model because of significant heterogeneity) of the three trials (K-Nielsen 1997; Serradell Catalan 2001; Sia 2010a) reporting the time to perform the nerve block found that the multiple nerve block took significantly more time to perform (see Analysis 2.6) (mean difference (MD) 3.34 minutes, 95% CI 2.66 to 4.03 minutes). There were conflicting findings in the two trials that measured the time from the start of the block until readiness for surgery (see Analysis 2.6). K-Nielsen 1997 found that this time period was significantly shorter in the multiple-injection group (MD -13.50 minutes, 95% CI -16.36 to -10.64 minutes) whereas Sia 2010a found it to be significantly longer in the multiple-injection group (MD 6.80 minutes, 95% CI 4.53 to 9.07 minutes). None of the differences in duration of tourniquet use, duration of the block or length of surgery were statistically significant between the two groups (see Analysis 2.6). Using the random-effects model because of significant (P = 0.01) and substantial heterogeneity (I² = 72% and 69% respectively) for the pooled results for paraesthesia and tachycardia, Analysis 2.7 showed that none of the differences between the two groups in the six listed complications occurring during nerve block were statistically significant. However, the statistically significant excess of paraesthesia and tachycardia as well as the two serious episodes of local anaesthetic toxicity in the single-injection group of K-Nielsen 1999b should not be disregarded given that these may reflect the method used for performing the single injection in this group (that is, transarterial). There appeared to be a trend for more arterial and venous punctures in the multiple-injection group (see Analysis 2.7). The three adverse effects, all lasting 24 hours, were all persistent paraesthesias in Serradell Catalan 2001 (see Analysis 2.8). Serradell Catalan 2001 found no statistically significant difference between the multiple and single-injection groups in patient discomfort. Pooled data from Serradell Catalan 2001 and Sia 2010a showed no statistically significant difference between the two groups in dissatisfaction with the anaesthetic method (see Analysis 2.9). K-Nielsen 1999b found no difference between the two groups in the pain experienced by the trial participants during performance of the block.

Multiple versus double-injection technique

Eleven trials (Coventry 2001; Imbelloni 2005; K-Nielsen 1998; K-Nielsen 1999a; Lavoie 1992; Rodriguez 2005; Rodriguez 2008; Serradell Catalan 2001; Sia 2001; Sia 2010b; Sia 2010c) made this comparison in a total of 937 participants. One participant of the multiple-injection group of K-Nielsen 1999a (who was taking cardiovascular medication) was excluded due to a severe reaction including loss of consciousness.

Primary analgesia or anaesthesia failure

The pooled results, using the fixed-effect model, showed a statistically significant decrease in primary analgesia or anaesthesia failure (incomplete sensory block) in the multiple-injection group (see Figure 6) (RR 0.28, 95% CI 0.20 to 0.40). Figure 6 also presents the trials subgrouped according to the technique used for double injection (transarterial versus neurostimulation). The clearly similar results of both groups of trials showed that multiple injections, all located via neurostimulation, provided more complete sensory block than double injections located with (failure: RR 0.28, 95% CI 0.18 to 0.44) or without (failure: RR 0.27, 95% CI 0.15 to 0.49) the use of a nerve stimulator.

Figure 6.

Forest plot of comparison: 3 Multiple versus double-injection technique, outcome: 3.1 Primary anaesthesia failure (incomplete sensory block).

Analysis 3.2 showed the results subgrouped according to the definition of primary analgesia failure (incomplete overall sensory block or supplemental blocks required for surgical site). While the results for both groups were in favour of multiple injections (incomplete overall sensory block: RR 0.24, 95% CI 0.15 to 0.37; supplemental blocks required for surgical site: RR 0.40, 95% CI 0.24 to 0.66), it was noteworthy that there were proportionately fewer participants in the double-injection group with primary anaesthesia failure when this outcome was defined according to the need for supplemental blocks for the surgical area rather than incomplete overall sensory blockade.

Six people required general anaesthesia for block failure (see Analysis 3.3) (RR 0.24, 95% CI 0.04 to 1.41). The pooled results for incomplete motor block, using the random-effects model because of significant (P = 0.02) and substantial heterogeneity (I² = 62%), showed a statistically significant decrease in incomplete motor block in the multiple-injection group (see Analysis 3.4) (RR 0.55, 95% CI 0.36 to 0.85).

Secondary analgesia failure, timing, complications and other outcomes

There was a statistically significant decrease in tourniquet pain in the multiple-injection group (RR 0.53, 95% CI 0.33 to 0.84) but not in the other outcomes of secondary analgesia failure (surgical site pain, and intraoperative sedation), although both favoured the multiple-injection group (see Analysis 3.5). Pooled results (using the random-effects model due to highly significant heterogeneity) from five trials (K-Nielsen 1998; Serradell Catalan 2001; Sia 2001; Sia 2010b; Sia 2010c) reporting the time to perform the nerve block found that the multiple-injection block took significantly more time to perform (see Analysis 3.6) (MD 1.74 minutes, 95% CI 1.04 to 2.45 minutes). In contrast, the time from the start of the block until readiness for surgery was similar between the multiple-injection and double-injection groups (MD -0.06 minutes, 95% CI -2.87 to 2.75 minutes) (see Analysis 3.6). Analysis 3.6 showed no statistically significant differences between the two groups for duration of tourniquet use, length of surgery or duration of block. Using the random-effects model because of significant (P = 0.01) and substantial heterogeneity (I² = 63%) in the pooled results for paraesthesia, Analysis 3.7 showed that there were no statistically significant differences between the two injection groups in the eight listed complications occurring during nerve block. It should be noted though that the more frequent occurrence of tachycardia (resulting from intravascular injection) and axillary haematoma, when the results of K-Nielsen 1998 and K-Nielsen 1999a were pooled, were consistent with the method of double injection used (transarterial without neurostimulation). The six adverse effects, all lasting 24 hours, were all persistent paraesthesias in Serradell Catalan 2001 (see Analysis 3.8). The only persistent adverse effect, recorded at three months in Serradell Catalan 2001, was neurological dysfunction that occurred in one participant in one of the two double-injection groups. There was no statistically significant difference between the multiple and double-injection groups in patient discomfort or their dissatisfaction with the anaesthetic method (see Analysis 3.9).

Discussion

Summary of main results

The ideal regional anaesthetic technique should meet four criteria; it should be effective, fast, safe and cause the patient little or no pain. While all 21 included trials reported on anaesthetic effectiveness (primary anaesthesia), the reporting of timing (block performance time, onset time, time to readiness for surgery), safety (early and late complications), and pain during block performance was incomplete. Though the latter three criteria are described as secondary outcomes in this review, they are as important as the primary outcome of anaesthetic effectiveness when considering the choice of anaesthetic technique. We summarize the findings of the three comparisons in turn and then provide some overall comments.

Double versus single-injection technique

Primary anaesthesia failure was much less likely in the double-injection group than the single-injection group (RR 0.51, 95% CI 0.30 to 0.85). This was true regardless of whether failure was defined as incomplete sensory block of all nerves or incomplete anaesthesia of the surgical site. However, when the data were subgrouped according to the technique used for double injection, double injections were significantly more effective than single injections only when neurostimulation was used in both intervention groups, and not when the transarterial technique was used. It should be noted that in the original review a test of interaction showed the results of these two subgroups to be significantly different from each other. This was not the case in the updated review (based on random-effects model risk ratios) and, given that the method of nerve location by itself does not explain the heterogeneity within the two subgroups, we have therefore pooled the data and reported the summary statistic for all trials. There were no statistically significant differences between the double and single-injection groups in the other reported outcomes (incomplete motor block, secondary analgesia failure, timings, complications and patient discomfort).

Multiple versus single-injection technique

Primary anaesthesia failure was much less likely in the multiple-injection group than the single-injection group (RR 0.25, 95% CI 0.14 to 0.44) and this held true across all subgroup analyses. Pooled data from four trials also showed a statistically significant decrease in incomplete motor block in the multiple-injection group. It took 3.3 minutes longer on average to perform the block in the multiple-injection group. However, it is unclear if this has any impact on the time to readiness for surgery as the two trials that reported this outcome had conflicting results. There were no statistically significant overall differences in the other reported outcomes (secondary analgesia failure, other timings, complications and patient discomfort). In one study (K-Nielsen 1999b) there was a statistically significant excess of paraesthesia and tachycardia and two serious episodes of local anaesthetic toxicity in the single-injection group, which can be attributed to the transarterial technique used.

Multiple versus double-injection technique

Primary anaesthesia failure was much less likely in the multiple-injection group than the double-injection group (RR 0.28, 95% CI 0.20 to 0.40); again, this held true across all subgroup analyses. In particular, it was irrespective of whether the double injections involved the transarterial injection technique or neurostimulation. Incomplete motor block and tourniquet pain were also significantly less likely in the multiple-injection group compared to the double-injection group. It took 1.7 minutes longer on average to perform the block in the multiple-injection group but the pooled data from five trials showed no overall difference in the time to readiness for surgery. There were no other statistically significant differences between the multiple and double-injection groups in the pooled results of other reported outcomes (secondary analgesia failure, other timings, complications and patient discomfort). The more frequent occurrence of tachycardia (resulting from intravascular injections) and axillary haematoma when the results of K-Nielsen 1998 and K-Nielsen 1999a were pooled are likely to reflect the method of double injection used (transarterial without neurostimulation) in these trials.

Overview

The results of this update confirm the original review’s conclusion that a multiple-injection technique (using neurostimulation) of non-ultrasound guided axillary block provides more effective anaesthesia than either a double or a single-injection technique. The question of whether three or four injections should be performed, or which nerves should be targeted in the multiple-injection technique, is not addressed in this review. The multiple-injection technique also appears to have other advantages, including more complete motor block and a reduced risk of tourniquet pain. Its primary disadvantage is that locating and injecting around three or more nerves in the axillary brachial plexus is much more complex, as reflected in the longer time required for performance of the multiple-injection technique compared to the single and double-injection techniques. Interestingly, this did not appear to significantly increase the time to readiness for surgery, although this is not necessarily conclusive given the limited data. The most likely explanation is that the increased anaesthetic efficacy of the multiple-injection technique offsets the longer block performance time.

The method of nerve location used in the single or double-injection techniques appears to influence the effectiveness and safety of anaesthesia. Double injections are more effective than single injections when neurostimulation is used in both interventions, but not when double injection is performed using the transarterial method and single injection is performed using neurostimulation. There was also some evidence of a greater risk of short-term complications related to vascular puncture, such as intravascular injection and axillary haematoma, when the transarterial method was used. Taken together, this suggests that neurostimulation should be the method of choice when performing a double-injection axillary block.

While there were no significant differences observed in many of the other outcomes related to secondary anaesthesia, complications, and patient pain and discomfort, this cannot be regarded as conclusive due to the limited data. In particular, the safety of multiple-injection methods remains an important unresolved issue given the low complication event rates reported in this review. The inevitable increase in needle passes while searching for other nerves after the first or second injection carries an increased risk of vascular puncture and trauma to nerves that is difficult to quantify. However, one large, multicentre prospective study of multiple-injection techniques for upper and lower limb blockade found generally reassuring evidence for axillary brachial plexus block (Fanelli 1999). Although 17% (278/1650) of these multiple-injection axillary blocks elicited unintentional paraesthesiae prompting needle withdrawal, all 17 (1%) people sustaining transient neurological dysfunction recovered fully at an average of six weeks. Fanelli et al (Fanelli 1999) also found some evidence that high tourniquet pressure rather than multiple injections was associated with neurological dysfunction.

Overall completeness and applicability of evidence

In this update we included one additional trial that had been awaiting classification, bringing the total to 21 trials. No new trials were identified that met the inclusion and exclusion criteria. There are over 2100 participants in the updated review, although the numbers of participants for each of the three comparisons are obviously fewer (ranging from 497 to 937). The distinction between no evidence of an effect and evidence of no effect still needs to be considered where there are apparently comparable findings. Application of trial results to clinical practice is hampered where there is an inadequate description of trial inclusion and exclusion criteria (six trials) and the types of surgery undertaken (10 trials). Another common shortcoming (10 trials) was the failure to monitor longer-term effects, particularly adverse effects.

Quality of the evidence

The quality of the evidence, appraised using the risk of bias assessment tool recommended by The Cochrane Collaboration (Higgins 2011), varied in the 21 trials but showed that the included trials were generally well conducted and either at low or unclear risk of bias for the seven aspects rated in our assessment (see Figure 3). Only seven trials were rated at high risk of bias and this in one domain only for six of the seven trials. We consider that the findings of this review are therefore likely to be valid.

Potential biases in the review process

Publication bias

We may have missed trials that are not indexed in MEDLINE or EMBASE. In particular, we may have missed trials that remain unpublished in journals by not searching conference proceedings and other 'grey literature'. We did, however, approach trialists and contacts in the industry for information on existing trials. While the possibility of publication bias cannot be ruled out, we consider that a well-conducted trial on this topic would have stood a good chance of being published in specialist journals irrespective of its results. Our trial selection procedure was systematic and, after an initial filtering of the results from the electronic searches, each author carried out independent selection.

Pooling and heterogeneity

We chose to pool data from trials testing the same comparisons; however, no two trials were identical. There were notable differences in the interventions, such as in the method of location of nerves and selection of specific nerves (see Table 1), study populations, and definitions of outcomes (see Appendix 4). We performed subgroup analyses of the outcome of primary anaesthesia failure according to the method of nerve location and the definition of adequate sensory blockade; however, the data were insufficient to examine the effects of the other methodological differences.

Authors' conclusions

Implications for practice

This review provides evidence that multiple-injection techniques using neurostimulation for axillary brachial plexus block provide more effective anaesthesia than either double or single-injection techniques. There is insufficient evidence to determine the relative effects of single, double and multiple-injection techniques on the complication rate, secondary analgesia failure, patient discomfort and pain during the procedure. There is some evidence suggesting a greater risk of complications and less satisfactory anaesthesia with methods using the transarterial approach rather than neurostimulation.

Implications for research

Since the original review was published, the use of ultrasound to guide peripheral nerve blockade has become widespread and has largely supplanted neurostimulation techniques, particularly in developed countries with access to the technology. Hence while the maintenance of this review in the light of any new evidence from randomized trials is required, we do not consider that conducting further randomized trials on this subject is a priority. We however suggest that the systematic surveillance of people undergoing these injections to ascertain adverse effects, in particular serious and permanent neurological injuries, should be ongoing.

Acknowledgements

For this review update, we continue to be grateful for the help received for previous versions from the people listed in the previous acknowledgements (Chin 2011; Handoll 2006). We would like to acknowledge the significant contributions to the previous versions made by Zbigniew J Koscielniak-Nielsen and Helen Handoll, which we have drawn upon in this version.

We are very grateful to Karen Hovhannisyan for developing and running the search strategies for this update, and to Jane Cracknell for her continued support and patience. We thank Dr Melba Ramirez-Gomez for responding to our request for further details on her trial.

Data and analyses

Download statistical data

Comparison 1. Double versus single-injection technique
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Primary anaesthesia failure (incomplete sensory block)8497Risk Ratio (M-H, Random, 95% CI)0.51 [0.30, 0.85]
1.1 Transarterial injection (for double injection)4237Risk Ratio (M-H, Random, 95% CI)0.72 [0.33, 1.58]
1.2 Location by neurostimulation (for double injection)4260Risk Ratio (M-H, Random, 95% CI)0.40 [0.22, 0.73]
2 Primary anaesthesia failure - subgrouped by outcome definition8 Risk Ratio (M-H, Random, 95% CI)Subtotals only
2.1 Incomplete overall sensory block4238Risk Ratio (M-H, Random, 95% CI)0.43 [0.24, 0.76]
2.2 Supplemental blocks for surgical area5309Risk Ratio (M-H, Random, 95% CI)0.43 [0.17, 1.11]
3 Complete failure of block: general anaesthesia or new plexus block6338Risk Ratio (M-H, Fixed, 95% CI)1.29 [0.33, 5.01]
4 Incomplete motor block4229Risk Ratio (M-H, Fixed, 95% CI)0.78 [0.58, 1.03]
5 Secondary analgesia failure4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
5.1 Pain in surgical site/operative field3160Risk Ratio (M-H, Fixed, 95% CI)0.56 [0.25, 1.25]
5.2 Tourniquet pain2104Risk Ratio (M-H, Fixed, 95% CI)0.58 [0.22, 1.52]
5.3 Intra-operative sedatives2129Risk Ratio (M-H, Fixed, 95% CI)0.64 [0.31, 1.31]
6 Timing (in minutes)4 Mean Difference (IV, Fixed, 95% CI)Subtotals only
6.1 Time for block160Mean Difference (IV, Fixed, 95% CI)1.65 [0.72, 2.58]
6.2 Duration of operation150Mean Difference (IV, Fixed, 95% CI)9.0 [-8.19, 26.19]
6.3 Duration of tourniquet3154Mean Difference (IV, Fixed, 95% CI)2.44 [-5.24, 10.13]
6.4 Duration of block2129Mean Difference (IV, Fixed, 95% CI)11.98 [-6.73, 30.68]
7 Complications during nerve block3 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
7.1 Arterial puncture2110Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
7.2 Venous puncture2110Risk Ratio (M-H, Fixed, 95% CI)1.5 [0.17, 13.52]
7.3 Paraesthesia2110Risk Ratio (M-H, Fixed, 95% CI)2.5 [0.31, 19.99]
7.4 Tachycardia (intra-vascular injections)160Risk Ratio (M-H, Fixed, 95% CI)5.86 [0.25, 137.66]
8 Adverse effects (> 24 hours)2 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
9 Patient discomfort and dissatisfaction with method1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
9.1 Patient uncomfortable1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
9.2 Patient would not have method again1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
Analysis 1.1.

Comparison 1 Double versus single-injection technique, Outcome 1 Primary anaesthesia failure (incomplete sensory block).

Analysis 1.2.

Comparison 1 Double versus single-injection technique, Outcome 2 Primary anaesthesia failure - subgrouped by outcome definition.

Analysis 1.3.

Comparison 1 Double versus single-injection technique, Outcome 3 Complete failure of block: general anaesthesia or new plexus block.

Analysis 1.4.

Comparison 1 Double versus single-injection technique, Outcome 4 Incomplete motor block.

Analysis 1.5.

Comparison 1 Double versus single-injection technique, Outcome 5 Secondary analgesia failure.

Analysis 1.6.

Comparison 1 Double versus single-injection technique, Outcome 6 Timing (in minutes).

Analysis 1.7.

Comparison 1 Double versus single-injection technique, Outcome 7 Complications during nerve block.

Analysis 1.8.

Comparison 1 Double versus single-injection technique, Outcome 8 Adverse effects (> 24 hours).

Analysis 1.9.

Comparison 1 Double versus single-injection technique, Outcome 9 Patient discomfort and dissatisfaction with method.

Comparison 2. Multiple versus single-injection technique
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Primary anaesthesia failure (incomplete sensory block)8682Risk Ratio (M-H, Random, 95% CI)0.25 [0.14, 0.44]
1.1 No use of nerve stimulator (for single injection)2204Risk Ratio (M-H, Random, 95% CI)0.40 [0.25, 0.65]
1.2 Location by neurostimulation (for single injection)6478Risk Ratio (M-H, Random, 95% CI)0.19 [0.09, 0.41]
2 Primary anaesthesia failure - subgrouped by outcome definition8 Risk Ratio (M-H, Random, 95% CI)Subtotals only
2.1 Incomplete overall sensory block3264Risk Ratio (M-H, Random, 95% CI)0.28 [0.12, 0.64]
2.2 Supplemental blocks for surgical area5418Risk Ratio (M-H, Random, 95% CI)0.23 [0.10, 0.52]
3 Complete failure of block: general anaesthesia or new plexus block6454Risk Ratio (M-H, Random, 95% CI)0.31 [0.03, 2.71]
4 Incomplete motor block4304Risk Ratio (M-H, Random, 95% CI)0.61 [0.39, 0.96]
5 Secondary analgesia failure5 Risk Ratio (M-H, Random, 95% CI)Subtotals only
5.1 Pain in surgical site/operative field3244Risk Ratio (M-H, Random, 95% CI)0.53 [0.05, 5.37]
5.2 Tourniquet pain4379Risk Ratio (M-H, Random, 95% CI)0.97 [0.30, 3.11]
5.3 Intra-operative sedatives5482Risk Ratio (M-H, Random, 95% CI)0.70 [0.41, 1.19]
6 Timing (in minutes)5 Mean Difference (IV, Random, 95% CI)Subtotals only
6.1 Time for block3278Mean Difference (IV, Random, 95% CI)3.34 [2.66, 4.03]
6.2 Time for readiness for surgery2206Mean Difference (IV, Random, 95% CI)-3.33 [-23.23, 16.56]
6.3 Duration of tourniquet4379Mean Difference (IV, Random, 95% CI)2.30 [-2.22, 6.82]
6.4 Duration of block2107Mean Difference (IV, Random, 95% CI)-7.22 [-22.91, 8.47]
6.5 Length of surgery1138Mean Difference (IV, Random, 95% CI)2.0 [-3.53, 7.53]
7 Complications during nerve block4 Risk Ratio (M-H, Random, 95% CI)Subtotals only
7.1 Arterial puncture3278Risk Ratio (M-H, Random, 95% CI)1.90 [0.64, 5.66]
7.2 Venous puncture3278Risk Ratio (M-H, Random, 95% CI)2.58 [0.89, 7.48]
7.3 Paraesthesia4382Risk Ratio (M-H, Random, 95% CI)0.75 [0.20, 2.79]
7.4 Tachycardia (intra-vascular injections)3322Risk Ratio (M-H, Random, 95% CI)0.87 [0.09, 8.44]
7.5 Local anaesthesia toxicity (intra-arterial injections)1104Risk Ratio (M-H, Random, 95% CI)0.20 [0.01, 4.07]
7.6 Subcutaneous/axillary haematoma2184Risk Ratio (M-H, Random, 95% CI)0.33 [0.01, 7.95]
8 Adverse effects > 24 hours3 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
9 Patient discomfort and dissatisfaction with method2 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
9.1 Patient uncomfortable160Risk Ratio (M-H, Fixed, 95% CI)2.0 [0.77, 5.20]
9.2 Patient would not have method again2192Risk Ratio (M-H, Fixed, 95% CI)1.09 [0.43, 2.77]
Analysis 2.1.

Comparison 2 Multiple versus single-injection technique, Outcome 1 Primary anaesthesia failure (incomplete sensory block).

Analysis 2.2.

Comparison 2 Multiple versus single-injection technique, Outcome 2 Primary anaesthesia failure - subgrouped by outcome definition.

Analysis 2.3.

Comparison 2 Multiple versus single-injection technique, Outcome 3 Complete failure of block: general anaesthesia or new plexus block.

Analysis 2.4.

Comparison 2 Multiple versus single-injection technique, Outcome 4 Incomplete motor block.

Analysis 2.5.

Comparison 2 Multiple versus single-injection technique, Outcome 5 Secondary analgesia failure.

Analysis 2.6.

Comparison 2 Multiple versus single-injection technique, Outcome 6 Timing (in minutes).

Analysis 2.7.

Comparison 2 Multiple versus single-injection technique, Outcome 7 Complications during nerve block.

Analysis 2.8.

Comparison 2 Multiple versus single-injection technique, Outcome 8 Adverse effects > 24 hours.

Analysis 2.9.

Comparison 2 Multiple versus single-injection technique, Outcome 9 Patient discomfort and dissatisfaction with method.

Comparison 3. Multiple versus double-injection technique
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Primary anaesthesia failure (incomplete sensory block)11936Risk Ratio (M-H, Fixed, 95% CI)0.28 [0.20, 0.40]
1.1 Transarterial injection (for double injection)3270Risk Ratio (M-H, Fixed, 95% CI)0.27 [0.15, 0.49]
1.2 Location by neurostimulation (for double injection)8666Risk Ratio (M-H, Fixed, 95% CI)0.28 [0.18, 0.44]
2 Primary anaesthesia failure - subgrouped by outcome definition11 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
2.1 Incomplete overall sensory block7570Risk Ratio (M-H, Fixed, 95% CI)0.24 [0.15, 0.37]
2.2 Supplemental blocks for surgical area7586Risk Ratio (M-H, Fixed, 95% CI)0.40 [0.24, 0.66]
3 Complete failure of block: general anaesthesia or new plexus block8600Risk Ratio (M-H, Fixed, 95% CI)0.24 [0.04, 1.41]
4 Incomplete motor block6470Risk Ratio (M-H, Random, 95% CI)0.55 [0.36, 0.85]
5 Secondary analgesia failure8 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
5.1 Pain in surgical site/operative field5450Risk Ratio (M-H, Fixed, 95% CI)0.33 [0.04, 3.14]
5.2 Tourniquet pain7719Risk Ratio (M-H, Fixed, 95% CI)0.53 [0.33, 0.84]
5.3 Intra-operative sedatives7716Risk Ratio (M-H, Fixed, 95% CI)0.75 [0.55, 1.03]
6 Timing (in minutes)6 Mean Difference (IV, Random, 95% CI)Subtotals only
6.1 Time for block5556Mean Difference (IV, Random, 95% CI)1.74 [1.04, 2.45]
6.2 Time for readiness for surgery5524Mean Difference (IV, Random, 95% CI)-0.08 [-2.92, 2.77]
6.3 Duration of tourniquet5549Mean Difference (IV, Random, 95% CI)2.99 [-1.03, 7.01]
6.4 Duration of surgery3376Mean Difference (IV, Random, 95% CI)0.63 [-4.97, 6.24]
6.5 Duration of block2150Mean Difference (IV, Random, 95% CI)0.89 [-27.95, 29.73]
7 Complications during nerve block8 Risk Ratio (M-H, Random, 95% CI)Subtotals only
7.1 Arterial puncture6616Risk Ratio (M-H, Random, 95% CI)1.37 [0.66, 2.84]
7.2 Venous puncture6616Risk Ratio (M-H, Random, 95% CI)1.28 [0.75, 2.17]
7.3 Paraesthesia7716Risk Ratio (M-H, Random, 95% CI)0.71 [0.31, 1.62]
7.4 Tachycardia (intra-vascular injections)4476Risk Ratio (M-H, Random, 95% CI)0.55 [0.23, 1.32]
7.5 Local anaesthesia toxicity (intra-arterial injections)2170Risk Ratio (M-H, Random, 95% CI)1.0 [0.15, 6.82]
7.6 Axillary haematoma/bruises3260Risk Ratio (M-H, Random, 95% CI)0.30 [0.09, 1.06]
7.7 Accidental intravascular injection2200Risk Ratio (M-H, Random, 95% CI)0.81 [0.20, 3.26]
7.8 Transient bradycardia1100Risk Ratio (M-H, Random, 95% CI)0.33 [0.01, 7.99]
8 Adverse effects > 24 hours6 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
9 Patient discomfort and dissatisfaction with method4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
9.1 Patient uncomfortable180Risk Ratio (M-H, Fixed, 95% CI)1.33 [0.73, 2.45]
9.2 Patient would not have method again3356Risk Ratio (M-H, Fixed, 95% CI)1.13 [0.59, 2.13]
9.3 Patient dissatisfied170Risk Ratio (M-H, Fixed, 95% CI)0.25 [0.01, 5.98]
Analysis 3.1.

Comparison 3 Multiple versus double-injection technique, Outcome 1 Primary anaesthesia failure (incomplete sensory block).

Analysis 3.2.

Comparison 3 Multiple versus double-injection technique, Outcome 2 Primary anaesthesia failure - subgrouped by outcome definition.

Analysis 3.3.

Comparison 3 Multiple versus double-injection technique, Outcome 3 Complete failure of block: general anaesthesia or new plexus block.

Analysis 3.4.

Comparison 3 Multiple versus double-injection technique, Outcome 4 Incomplete motor block.

Analysis 3.5.

Comparison 3 Multiple versus double-injection technique, Outcome 5 Secondary analgesia failure.

Analysis 3.6.

Comparison 3 Multiple versus double-injection technique, Outcome 6 Timing (in minutes).

Analysis 3.7.

Comparison 3 Multiple versus double-injection technique, Outcome 7 Complications during nerve block.

Analysis 3.8.

Comparison 3 Multiple versus double-injection technique, Outcome 8 Adverse effects > 24 hours.

Analysis 3.9.

Comparison 3 Multiple versus double-injection technique, Outcome 9 Patient discomfort and dissatisfaction with method.

Appendices

Appendix 1. Search strategies for current update

CENTRAL, The Cochrane Library

#1    MeSH descriptor Anesthesia, Conduction, this term only
#2    MeSH descriptor Anesthesia, Local, this term only
#3    MeSH descriptor Nerve Block, this term only
#4    ((analg* or an?esth*) near (local* or regional)):ti,ab
#5    (par?esthes* or dys?esthes* or h?ematom* or seizur*):ti,ab
#6    (pain near (per?operativ* or post?operativ*)):ti,ab
#7    (#1 OR #2 OR #3 OR #4 OR #5 OR #6)
#8    (surg* near (hand* or wrist* or forearm* or elbow*))
#9    (#7 AND #8)
#10   (((brachial or axillary) near (block* or an?esthesia)) or midhumer*):ti,ab
#11   (#9 OR #10)

EMBASE (OvidSP)

1     regional anesthesia/ or local anesthesia/ or nerve block/ or ((exp brachial plexus/ or exp axilla/) and block*.mp.) or ((analg* or an?esth*) adj3 (local* or regional)).ti,ab. or (par?esthes* or dys?esthes* or h?ematom* or seizur*).ti,ab. or (pain adj3 (per?operativ* or post?operativ*)).ti,ab.
2     exp hand surgery/ or (surg* adj3 (hand* or wrist* or forearm* or elbow*)).ti,ab.
3     1 and 2
4     (((brachial or axillary) adj3 (block* or an?esthesia)) or midhumer*).ti.
5     3 or 4

Ovid MEDLINE(R)

1     Anesthesia-Conduction/ or Anesthesia-Local/ or Nerve Block/ or ((exp Brachial-Plexus/ or exp Axilla-/) and block*.mp.) or ((analg* or an?esth*) adj3 (local* or regional)).ti,ab. or Postoperative-Complications/ or Pain-Postoperative/ or (par?esthes* or dys?esthes* or h?ematom* or seizur*).ti,ab. or (pain adj3 (per?operativ* or post?operativ*)).ti,ab.
2     (exp Surgery/ and exp Hand/) or (surg* adj3 (hand* or wrist* or forearm* or elbow*)).ti,ab.
3     1 and 2
4     (((brachial or axillary) adj3 (block* or an?esthesia)) or midhumer*).ti.
5     3 or 4

Appendix 2. Search strategies in first version of review (Handoll 2006)

We searched the following databases: the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and EMBASE.

Author-led literature search

One author (ZK-N) performed literature searches up to August 2004 and identified RCTs using the following strategy.

  1. Searching MEDLINE (Ovid-Web) from 1966 to August 2004 using a series of free-text and MESH terms (see below). The results from each term were inspected in turn.

  2. Using similar search terms (free text and MESH) for EMBASE (Ovid-Web) from 1988 to August 2004, and the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 3, 2004).

  3. Checking reference lists of RCTs identified through the electronic searches.

  4. Contacting trial authors and the medical industry.

  5. Scruntiny of article titles of the following anaesthesia journals for mention of axillary or midhumeral block:

  • Acta Anaesthesiologica Scandinavica (1980 to 2004),

  • Anaesthesia (1980 to 2004),

  • Anaesthesia and Intensive Care (1980 to 2004),

  • Anesthesia and Analgesia (1980 to 2004),

  • Anesthesiology (1980 to 2004),

  • British Journal of Anaesthesia (1980 to 2004),

  • Canadian Journal of Anaesthesia (1980 to 2004),

  • European Journal of Anaesthesiology (1990 to 2004),

  • Regional Anesthesia/Regional Anesthesia and Pain Medicine (1985 to 2004).

Supplementary search

Karen Hovhannisyan ((KH) Trials Search Co-ordinator, Cochrane Anaesthesia Review Group (CARG)) supplemented these searches up to March 2005 on CENTRAL, MEDLINE and EMBASE.

  • CENTRAL (Issue 1, 2005).

  • SilverPlatter MEDLINE (WebSPIRS) (up to April Week 3 2005/04).

  • SilverPlatter EMBASE (WebSPIRS) (up to 2005/03).

KH combined the subject-specific terms for MEDLINE and EMBASE with optimal search strategies for RCTs for these databases.

We applied no language restrictions.

MEDLINE (Ovid-Web) search terms

Search numberSearch term
#1 (Free terms)Search axillary or midhumeral block
Field: All Fields, Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#2 (Free terms)Search anesthesia and axillary or midhumeral block
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#3 (Free terms)Search plexus anesthesia and axillary or midhumeral
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#4 (Free terms)Search anesthesia and brachial plexus and surgery
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#5 (Free terms)Search anesthesia and brachial plexus and injection
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#6 (Free terms)Search nerve block and brachial plexus and injection technique
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#7 (Free terms)Search axillary or midhumeral block and injection technique
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#8 (Mesh)Search anesthesia,conduction and brachial plexus
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#9 (Mesh)Search anesthesia,conduction and brachial plexus and axilla
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#10 (Mesh)Search anesthesia,conduction and surgery,hand
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#11 (Mesh)Search nerve block and surgery,hand
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#12 (Mesh)Search nerve block and brachial plexus
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#13 (Mesh)Search nerve block and axilla
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans
#14 (Mesh)Search nerve block and axilla and surgery
Limits: All Adult: 19+ years, Publication Date from 1966 to 2004, only items with abstracts, Randomized Controlled Trial, Humans

 

Appendix 3. Former methodological quality assessment tool

ItemScore
1. Was the assigned treatment adequately concealed prior to allocation?3 = allocation was concealed (e.g. sequentially numbered, sealed, opaque envelopes)
2 = small but possible chance of disclosure of assignment
1 = states random but no description
0 = quasi-randomized or open list/tables
2. Were the inclusion and exclusion criteria for entry clearly defined?1 = clearly defined (including contra-indications)
0 = inadequately or not defined
3. Were the outcomes of patients who withdrew described and included in the analysis (intention-to-treat)?1 = Outcomes of patients who withdrew or were excluded after allocation were EITHER detailed separately OR included in an intention-to-treat analysis OR the text stated there were no withdrawals
0 = Outcomes of patients who withdrew or were excluded after allocation were NEITHER detailed separately NOR included in an intention-to-treat analysis
4. Were important baseline characteristics reported?1 = Intervention groups were adequately described at entry. A minimum of 3 admission details were described: age, sex, type of surgery, mental status.
0 = Intervention groups were NOT adequately described at entry
5. Were care programmes, other than the trial options, identical?
(Example of a clinically important difference is anaesthetist experience)
1 = The text stated that the care programmes other than trial options were identical (or clear from the text)
0 = The text stated that the care programmes other than trial options were NOT identical
6. Were the outcome measures used clearly defined?1 = Outcome measures were clearly defined in the text
0 = Outcome measures were NOT clearly defined in the text
7. Were the outcome assessors blinded to treatment status?1 = Outcome assessors were blind to the allocation of patients
0 = Not mentioned or outcome assessors were NOT blind to the allocation of patients
8. Was the timing (e.g. duration of surveillance) clinically appropriate?1 = The timing of the measurement of the outcomes was appropriate (e.g. at least 24 hours)
0= The timing of the measurement of the outcomes was NOT appropriate

Appendix 4. Measurement of sensory and motor blockade

Study IDSensory: methodSensory: timingSensory: ratingNerve areas testedNerves: block Y/N?Motor: ratingNerves tested
Baranowski 1990Use of blunt end of a 27 gauge dental needleEvery 5 minutes for 30 minutes0 = no sensory loss
1 = loss of pinprick
2 = loss of touch
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = 3 or 4 of the following 4 nerves were blocked to sensory loss score of 2 (loss of touch) at 30 minutes:
Median
Musculocutaneous
Radial
Ulnar
Not reportedNot reported
Coventry 2001Use of a short-bevelled 27 gauge needleEvery 10 minutes for 30 minutesComplete sensory lossMedial cutaneous nerve of arm
Medial cutaneous nerve of fore arm
Median
Musculocutaneous
Radial
Ulnar
Success = sensory blockade of 6 nerves:
Median
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Musculocutaneous
Radial
Ulnar
Inability to move relevant muscle groups against gravityBlockade of 4 nerves reported:
Median
Musculocutaneous
Radial
Ulnar
Goldberg 1987Skin pinched with Allis clampNot statedNo pain on pinching the skinMedian
Musculocutaneous
Radial
Ulnar
Success = sensory blockade (no pain) for all 4 nerves:
Median
Musculocutaneous
Radial
Ulnar
Not reportedNot reported
Hickey 1993Pinprick2, 5, 10, 15, 20, 25 and 30 minutes following injection0 = no loss of sensation to pinprick
1 = analgesia (patient felt touch but not sharp)
2 = anaesthesia (patient did not feel touch)
Axillary
MusculocutaneousMedian
Radial
Ulnar
Medial brachial cutaneous
Medial antebrachial cutaneous
Intercostobrachial
Overall block success was not strictly defined. Instead they looked primarily at anaesthesia and analgesia in individual nerve territoriesHand grip
0 = no weakness
1 = paresis
2 = paralysis
Not reported
Imbelloni 2005Skin clamp and "observing patients’ pain manifestations"Not statedNot stated. The term “analgesia” is used.Musculocutaneous
Median
Ulnar
Radial
“Blockade was considered complete if all nerves were blocked”.
Incomplete = “need for additional injection”.
Failure = “need for general anesthesia”.
Not reportedNot reported
Inberg 1999Skin pinched40 minutes0 = normal sensation
1 = hypalgesia
2 = analgesia
3 = anaesthesia
Axillary
Lateral cutaneous
Medial cutaneous
Median
Musculocutaneous
Radial
Ulnar
Success = sensory blockade (no pain: score 2 or 3) and motor blockade (little or no power: score 2 or 3) for all 4 nerves:
Median
Musculocutaneous
Radial
Ulnar
0 = normal muscular function
1 = slight depression of power
2 = weak function without power
3 = no muscular function
Median
Musculocutaneous
Radial
Ulnar
K-Nielsen 1997Painful pinch with a plastic clampEvery 10 minutes until ready for surgery; supplementation from 20 minutes0 = no analgesia/anaesthesia
1 = loss of pain
2 = loss of sensation
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = no need to supplement any of 4 nerves:
Median
Musculocutaneous (only if necessary for surgery)
Radial
Ulnar
Poor = no obvious relaxation
Satisfactory = minor movement of digits
Good = completely limp hand
Not reported
K-Nielsen 1998Painful pinch with a plastic clampEvery 10 minutes until ready for surgery; supplementation from 30 minutes0 = no analgesia/anaesthesia
1 = loss of pain
2 = loss of sensation
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = no need to supplement. No pain felt in any area below elbow.
Incompletely blocked nerves were:
Axillary
Median
Musculocutaneous
Radial
Ulnar
Poor = no obvious relaxation
Satisfactory = minor movement of digits
Good = completely limp hand
Not reported
K-Nielsen 1999aPainful pinch with a plastic clampEvery 10 minutes until ready for surgery; supplementation from 30 minutes0 = no analgesia/anaesthesia
1 = loss of pain
2 = loss of sensation
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = no need to supplement. No pain felt in any area below elbow.
Incompletely blocked nerves were:
Axillary
Medial cutaneous nerve of arm
Median
Musculocutaneous
Radial
Ulnar
Poor = no obvious relaxation
Satisfactory = minor movement of digits
Good = completely limp hand
Not reported
K-Nielsen 1999bPainful pinch with a plastic clampEvery 10 minutes until ready for surgery; supplementation from 30 minutes0 = no analgesia/anaesthesia
1 = loss of pain
2 = loss of sensation
Axillary
Median
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Musculocutaneous
Radial
Ulnar
Success = no need to supplement. No pain felt in any area below elbow.
Incompletely blocked nerves were:
Axillary
Medial cutaneous nerve of arm
Median
Musculocutaneous
Radial
Ulnar
Poor = no obvious relaxation
Satisfactory = minor movement of digits
Good = completely limp hand
Not reported
Lavoie 1992Use of Wartenberg pinwheelEvery 5 minutes up to 30 minutesNeedles of pinwheel no longer felt"Each dermatome of the upper limb". Nerves not listed but would be:
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = "dermatomes of the nerves implicated in the surgical site were anaesthetized". (All nerves at surgical site: skin, muscles and bones)0% = flexion/extension movements in hand and arm against resistance
33% = flexion/extension movements in hand and arm against gravity but not against resistance
66% = flexion/extension movements in hand only
100% = no movement of upper limb against gravity
Not reported
Pere 1993Pinprick5, 10, 20 and 30 minutes and 3 hours; supplementation from 20 minutesPainful pinprick /
pinprick analgesia
Axillary
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
(Supraclavicular)
Ulnar
Success = no supplementation of nerves at site of planned surgery
Incompletely blocked nerves were:
Axillary
Median
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Musculocutaneous
Radial
Supraclavicular
Ulnar
Strength of extensors and flexors:
No reduction in strength / reduced strength / no muscular movement
Grip strength (kg / cm2): 0 / 0.1-0.4 / > 0.4
Not reported
Ramirez-Gomez 2010Pinprick with a sterile needleEvery 10 minutes until 30 minutes after block completion.

0 = normal sensation of pinprick

1 = reduced sensation of pinprick

2 = sensation to touch only

3 = absence of sensation

Musculocutaneous
Radial
Median
Ulnar
Criteria for successful block not clearly defined.

0 = normal muscle power

1 = slight decrease in muscle power

2 = very weak muscle power

3 = absence of muscle power

Musculocutaneous
Radial
Median
Ulnar
Rodriguez 2005Pinprick with 18G long bevel needle5, 20 minutes after block completion

0 = painful
1 = analgesia to pinprick
2 = anaesthesia to pinprick (no perception)

Global quality scale = sum of scores for all 6 nerves (0-12)

Musculocutaneous
Radial
Median
Ulnar
Medial brachial cutaneous (arm?)
Medial antebrachial cutaneous (forearm?)
Not specifically stated. Blocks were supplemented pre-operatively if there was absence of complete anaesthesia in surgical sites

Elbow flexion / extension
Wrist flexion / extension
Fingers flexion / extension
Thumb adduction

0 = no paresis
1 = paresis
2 = complete paralysis

Global quality scale = sum of scores for all 7 areas (0 to 14)

Not reported
Rodriguez 2008Pinprick with 18G long bevel
needle
10, 20, and 30 minutes after injection of the total dose of LA (time zero)0 points = pinprick
perceived as painful
1 point = analgesia to pinprick (tactile sensation)
2 points = anaesthesia to pinprick (no perception)
Musculocutaneous
Radial
Median
Ulnar
Medial cutaneous

Global quality scores for both sensory block (minimum, 0 point; maximum, 12 points) and motor block (minimum, 0 point; maximum, 14 points) were based on the sum of the individual scores obtained at 10, 20, and at
30 minutes in each cutaneous nerve distribution or joint movement. 

“Blocks were supplemented preoperatively with additional peripheral nerve blocks when the cutaneous nerve distributions corresponding to the operative area did not have complete anesthesia (i.e., score < 2) before the operation”

Motor block was assessed for flexion and extension of the elbow, flexion and extension of the wrist, flexion and extension of the fingers, and adduction of the thumb.

0 points = no paresis
1 point = paresis
2 points = complete paralysis

Not reported
Serradell Catalan 2001PinprickEvery 10 minutes up to 40 minutesNone: normal sensation
Partial: analgesia
Total: anaesthesia
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = sensory blockade for all 5 nerves:
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
None: normal movements
Partial: reduced movements
Total: flaccid hand and forearm
Median
Musculocutaneous
Radial
Ulnar
Sia 2001Used 22 gauge needleEvery 10 minutes up to 30 minutesAnalgesia: loss of pinprick
Anaesthesia: loss of touch
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Success = complete block of all 6 nerves:
Medial cutaneous nerve of arm
Medial cutaneous nerve of forearm
Median
Musculocutaneous
Radial
Ulnar
Absent: no block
Satisfactory: minor movements of digits possible
Complete: no movements against gravity
Not reported
Sia 2010aCold test5, 10, 15, 20, 25 and 30 minutes after end of procedureYes = "I feel cold"
No = "I do not feel cold"
Musculocutaneous
Median
Radial
Ulnar
Medial cutaneous
Loss of cold sensation at 30 minutes sufficient for surgery. Unblocked nerves implicated in the surgical site were blocked by the anaesthesiologist.
Intraoperatively, if the patient complained of pain at the surgical field, supplementation with LA was done by the surgeon.

Motor block was assessed for wrist extension, forearm flexion, index flexion, little finger flexion.

Grade 1 = no loss of force
Grade 2 = reduced force compared with contralateral arm
Grade 3 = complete motor block

Not reported
Sia 2010bCold test5, 10, 15, 20, 25 and 30 minutes after end of procedureYes = "I feel cold"
No = "I do not feel cold"
Musculocutaneous
Median
Radial
Ulnar
Medial cutaneous
Loss of cold sensation at 30 minutes sufficient for surgery. Unblocked nerves implicated in the surgical site were blocked by the anaesthesiologist.
Intraoperatively, if the patient complained of pain at the surgical field, supplementation with LA was done by the surgeon.

Motor block was assessed for wrist extension, forearm flexion, index flexion, little finger flexion.

Grade 1 = no loss of force
Grade 2 = reduced force compared with contralateral arm
Grade 3 = complete motor block

Not reported
Sia 2010cCold test5, 10, 15, 20, 25 and 30 minutes after end of procedureYes = "I feel cold"
No = "I do not feel cold"
Musculocutaneous
Median
Radial
Ulnar
Medial cutaneous
Loss of cold sensation at 30 minutes sufficient for surgery. Unblocked nerves implicated in the surgical site were blocked by the anaesthesiologist.
Intraoperatively, if the patient complained of pain at the surgical field, supplementation with LA was done by the surgeon.

Motor block was assessed for wrist extension, forearm flexion, index flexion, little finger flexion.

Grade 1 = no loss of force
Grade 2 = reduced force compared with contralateral arm
Grade 3 = complete motor block

Not reported
Turkan 2002

Pinprick

Testing with an Allis clamp by surgeon also mentioned

Sensory testing every 3 minutes following injection. No other details.

Quality of analgesia
1 = no pain
2 = discomfort
3 = pain

 

Musculocutaneous
Median
Radial
Ulnar

 

Overall block success was when the patient felt no pain in all four nerve distributions when tested by a surgeon (with an Allis clamp).Not reportedNot reported

What's new

Last assessed as up-to-date: 5 March 2013.

DateEventDescription
7 August 2013New citation required but conclusions have not changedThis review is an update of the previous Cochrane systematic review (Chin 2011). There is a change in authorship: Helen Handoll has been replaced by Husni Alakkad and Javier Cubillos.
7 August 2013New search has been performed

We updated the review as follows.

1. The title was amended to make it explicit that the scope of the review is restricted to non-ultrasound guided techniques of axillary brachial plexus blockade.

2. We updated the literature search from March 2011 to March 2013.

3. We included one study that was identified and awaiting classification in the previous review (Ramirez-Gomez 2010). We excluded three newly-identified studies on the basis of the use of ultrasound guidance.

History

Protocol first published: Issue 4, 2002
Review first published: Issue 1, 2006

DateEventDescription
14 March 2011New citation required but conclusions have not changedThis review is an update of the previous Cochrane systematic review (Handoll 2006). There is a change in authorship, including the lead author and the contact author: Zbigniew J Koscielniak-Nielsen has been replaced by Ki Jinn Chin.
14 March 2011New search has been performed

We updated the review as follows.

1. The title was changed to make it explicit that the scope is restricted to adults.

2. The inclusion criteria were revised to exclude children and trials using ultrasound-guided techniques of nerve location.

3. We now assess risk of bias; this replaced the previous methodological quality assessment.

4. We updated our literature search from March 2005 (date of last search in the previous review) to March 2011.

5. We included eight new trials (Hickey 1993; Imbelloni 2005; Rodriguez 2005; Rodriguez 2008; Sia 2010a; Sia 2010b; Sia 2010c; Turkan 2002). We excluded a further seven newly identified studies (Carre 2000; Liu 2005; Sites 2006; Tuominen 1987; Youssef 1988; Yu 2007). One trial (Ramirez-Gomez 2010a) is currently awaiting translation and classification.

6. We added summary of findings tables for the three comparisons.

14 March 2011AmendedThe review was amended to the format of RevMan 5.1.
31 July 2008AmendedConverted to new review format.

Contributions of authors

Updated review

Ki Jinn Chin (KJC), Husni Alakkad (HA), Javier E Cubillos (JEC)

Conceiving the review: KJC
Co-ordinating the review: KJC
Screening search results: KJC, HA
Organizing retrieval of papers: KJC, HA
Screening retrieved papers against inclusion criteria: KJC, HA
Appraising quality of papers: KJC, JEC
Abstracting data from papers: KJC, JEC
Writing to authors of papers for additional information: JEC

Data management for the review: KJC, JEC
Entering data into Review Manager (RevMan 5.2): KJC HA
RevMan statistical data: KJC
Interpretation of data: KJC, JEC
Writing the review: KJC, HA, JEC

Performing previous work that was the foundation of the present study: KJC
Guarantor for the review (one author): KJC
Person responsible for reading and checking review before submission: KJC

Previous update

See Chin 2011

Dr KJ Chin and Dr H Handoll updated the review.

Original review

See Handoll 2006

Dr ZJ Koscielnak-Nielsen conceived the idea for the review and wrote the protocol.
Dr H Handoll and Dr ZJ Koscielnak-Nielsen wrote the review.

Declarations of interest

Ki Jinn Chin: none known

Husni Alakkad: none known

Javier E Cubillos: none known

Sources of support

Internal sources

  • University of Teesside, Middlesbrough, UK.

  • Department of Anaesthesia and Operative Services, HOC, Rigshospital, Copenhagen, Denmark.

  • University of Toronto, Toronto, Canada.

External sources

  • No sources of support supplied

Differences between protocol and review

Changes made for the second update of the review

The title of the review was changed to make it explicit that the review is restricted to non-ultrasound guided techniques of nerve location.

Changes made for the first update of the review

Changes made to the inclusion criteria and methods before processing the included trials were as follows.

  1. We clarified that we would exclude trials focusing on children only ('Types of participants'). The restriction to adults was made explicit in the title of the review.

  2. We excluded trials that used ultrasound-guided techniques of nerve location.

Risk of bias assessment replaced the eight-item methodological quality assessment scoring scheme.

Changes made for the first version of the review

Important changes made to the protocol before processing the included trials were as follows.

  1. The midhumeral approach was no longer specified as included ('Types of studies').

  2. The exclusion of trials involving supplementary anaesthesia was moderated to allow for trials using systemic opioids as a component of sedation ('Types of interventions').

  3. The addition of a third primary outcome, failed anaesthesia ('Types of outcome measures').

  4. Adjustments to the methods to accommodate the change in review authorship ('Methods').

  5. The expansion of the quality assessment of the included trials to include all eight items suggested in the generic scoring scheme of CARG ('Methods').

  6. The prior specification of sensitivity and subgroup analyses.

Before publication of the review (Handoll 2006), the name was changed from that in the protocol: 'Single, double or multiple injection techniques for axillary brachial plexus block for surgery of the distal upper extremity'.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Baranowski 1990

MethodsMethod of randomization: not stated
No blinding indicated
No loss to follow-up
ParticipantsLondon, UK
Period of study: not stated
100 people scheduled for outpatient hand surgery. Informed consent
Male: not stated; mean age: 49 years
Excluded: no details
Interventions

Multiple (neurostimulation method) versus multiple (paraesthesia method) versus single injection.

All received local anaesthetic (LA solution): up to 40 ml lignocaine 1.5% with adrenaline 200 µg.

No premedication. No sedation or IV analgesia. All blocks performed or supervised by 1 of the 2 trial authors.

1. Multiple injection using neurostimulator: unsheathed block needle. Attempts made to locate 3 to 4 main branches of brachial plexus; nerve specific muscle twitches. Incremental LA injections.
2. Multiple injection using paraesthesia: 22 gauge regional block needle. Attempts made to locate 3 to 4 main peripheral nerves. Incremental LA injections. Distal pressure applied.
3. Single injection via catheter on its insertion (introduction with 18 gauge needle) in brachial plexus sheath. Fascial 'click' and easy insertion used to identify sheath. Distal pressure applied.

OutcomesLength of follow-up: 30 minutes
Sensory blockade
Motor blockade (no data)
Anaesthesia failure (less than 3 nerves fully blocked, general anaesthesia, failure to penetrate brachial plexus sheath)
NotesRequest for details of methods, types of surgery and results, including adverse effects, sent to Dr Baranowski on 02/12/04.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“One hundred patients ... were randomly allocated”. No details of method
Allocation concealment (selection bias)Unclear riskNo details of method
Blinding (performance bias and detection bias)
All outcomes
High riskNo mention of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up seemed likely.
Selective reporting (reporting bias)Unclear riskPossible but no protocol available
Balance in baseline characteristics?Unclear riskIncomplete information to judge: no gender or type of surgery. Single injection group was 6 years younger (not statistically significant)
Free from performance bias?Low risk"All of the blocks were performed or supervised by one of the two authors.” There was a register and a consultant. Other treatment (none) seemed comparable

Coventry 2001

MethodsMethod of randomization: "Technique written on card and placed in envelope. Envelopes sealed, shuffled and numbered 1-60." "Envelope opened immediately pre-op by 'regional anaesthetist'".
Double-blind: anaesthetist carrying out assessments and surgeon were blind to injection technique
Blinded outcome assessor
No loss to follow-up
ParticipantsDundee, UK
Period of study: 1995
60 people undergoing elective upper limb surgery: Dupuytren's, carpal tunnel, tendon surgery, arthrodesis/arthroplasty, wrist arthroscopy and miscellaneous
Male: 45%; age range: 20 to 85 years
Excluded: patients refusing a local anaesthetic technique, dementia; age <17 years; peripheral neuropathy, sensitivity to amide local anaesthetics; ASA physical status > 3 (see notes).
Interventions

Multiple versus double injection (both groups using the neurostimulation method).

All received LA solution: 30 ml lidocaine 15 mg/ml with epinephrine 5 ug/ml. Nerve blockade facilitated using 22G insulated short-bevelled needle and peripheral nerve stimulator. All blocks carried out by one operator.

Initial sedation with midazolam. Skin anaesthetized with 1 to 2 ml plain lidocaine 10 mg/ml. Musculocutaneous nerve was first located and 5 ml LA solution injected.
1. Multiple injection: 15 ml LA to median nerve followed by 10 ml LA to radial nerve.
2. Double injection: single injection of 25 ml LA to median nerve.

OutcomesLength of follow-up: 30 minutes (and duration of surgery)
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia)
Duration of surgery
Tourniquet discomfort
Problems (all nerves located; no problems indicated)
Notes

Reply to request for details of methods and adverse effects received from Dr Coventry on 13/10/04.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“randomly allocated” “Technique written on card and placed in envelope. Envelopes sealed, shuffled and numbered 1-60.” Unclear how well shuffled
Allocation concealment (selection bias)Low riskSealed "Envelope opened immediately pre-op by 'regional anaesthetist'".
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind. Blinded investigator anaesthetist then carried out all assessments. This investigator was totally blind as was the surgeon. No assessment was carried out by the regional anaesthetists thus ensuring blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up
Selective reporting (reporting bias)Unclear riskLikely but no protocol available and took 6 years to publish
Balance in baseline characteristics?Low riskBaseline characteristics including types of surgery detailed and appeared balanced
Free from performance bias?Low riskAll blocks carried out by one operator. No cause for concern

Goldberg 1987

MethodsMethod of randomization: not stated
Blinded outcome assessor: operating surgeon
No loss to follow-up
ParticipantsPhiladelphia, USA
Period of study: not stated
59 people scheduled for upper extremity surgery, wrist or more distal, amenable to brachial plexus block outpatient hand surgery: carpal tunnel repair/median nerve release; Dupuytren's contracture release; arthroplasty of interphalangeal joint; ganglion excision; distal radial and/or ulnar plating; foreign body excision; miscellaneous finger operations. Consenting.
Male: not stated; mean age: 50 years, age 18+ years
Excluded: no details
Interventions

Double versus single (neurostimulation method) versus single (paraesthesia method) injection.

All received local anaesthetic (LA) solution: 40 ml/70 kg mepivacaine 1.5%.

No mention of premedication, sedation or IV analgesia. All blocks performed by first or second year anaesthesiology residents supervised by staff anaesthesiologist (usually first author).

1. Double transarterial injection: 22 gauge short bevel needle inserted transarterial fixation - half of LA volume administered posterior to axillary artery and half anterior to artery.
2. Single injection using nerve stimulator: 23 gauge insulated needle connected to stimulator. Whole volume of LA injected when maximum stimulation no longer produced muscle activity.
3. Single injection using paraesthesia: 22 gauge short bevel needle inserted until hand paraesthesia elicited - whole volume of LA injected.

OutcomesLength of follow-up: hand clinic (timing not specified)
Sensory blockade
Anaesthesia failure (non-blocked nerves)
Adverse effects (none at hand clinic)
NotesRequest for details of methods, types of surgery and results, including adverse effects, sent to Prof Goldberg on 15/12/04.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“Three methods of blockade were randomly selected.” No details of method
Allocation concealment (selection bias)Unclear riskNo details of method
Blinding (performance bias and detection bias)
All outcomes
Low riskPain was tested “by a surgeon (with an Allis clamp) unaware of the method utilized.”
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of loss to follow-up
Selective reporting (reporting bias)Unclear riskPossible but no protocol available
Balance in baseline characteristics?Unclear riskIncomplete information to judge: no gender or type of surgery. Single injection groups were 6 and 11 years younger than transarterial groups
Free from performance bias?Unclear riskNo difference in “the level of training of the residents who performed the blocks, which ranged from 1-19 months”. No other problems detected - blinded surgeon

Hickey 1993

MethodsMethod of randomization: not specified
No blinding of patient, care-giver or outcome assessor described
No loss to follow-up
ParticipantsTexas, USA
Period of study: not stated
60 adults scheduled for surgery of the upper extremity (not otherwise specified)
Male: 1.7%; mean age 56 years
Excluded: ASA physical status > 3 (see notes)
Interventions

Double versus single posterior versus single anterior injection (transarterial method in all groups).

All received local anaesthetic (LA) solution: 1.5% mepivacaine with 5 mcg/ml epinephrine; total volume 50 ml. Transarterial technique in all cases with a 22G short-bevel block needle. All blocks performed by residents supervised by one staff anaesthesiologist.

Sedative premedication with IV midazolam: up to 3 mg

1. Double injection: injection of half of LA volume (25 ml) anterior to axillary artery, injection of other half (25 ml) posterior to axillary artery.
2. Single posterior injection: injection of all of LA volume (50 ml) posterior to axillary artery.
3. Single anterior injection: injection of all of LA volume (50 ml) anterior to axillary artery.

Subcutaneous injection of 3 ml of LA over the axillary artery to block the intercostobrachial nerve.

OutcomesLength of follow-up: 30 minutes
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; general anaesthesia)
Onset of analgesia and anaesthesia in individual nerve territories
Complications during block injection and operation
Notes

One double injection patient excluded from the analyses because of an aborted block.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"Randomized study". No further details
Allocation concealment (selection bias)Unclear riskNo details on allocation concealment
Blinding (performance bias and detection bias)
All outcomes
High riskNo blinding stated (seems unlikely given the supervisory aspect)
Incomplete outcome data (attrition bias)
All outcomes
Low riskAll participants accounted for
Selective reporting (reporting bias)Unclear riskNo protocol. Insufficient information to judge this
Balance in baseline characteristics?Unclear riskInsufficient information to judge this in terms of lack of details on surgical interventions. Double injection group on average 8 years older
Free from performance bias?Low riskAll blocks performed by residents directly supervised by the same member of staff

Imbelloni 2005

MethodsMethod of randomization: not stated
No blinding of patient, care-giver or outcome assessor described
No loss to follow-up
ParticipantsSao Jose do Rio Preto, Brazil
Period of study: not stated
70 adults scheduled for orthopaedic surgery of the forearm and hand. Informed consent
Male: 56%; mean age 38 years
Exclusions: ASA physical status > 2 (see notes), Age < 20 years or > 60 years
Interventions

Multiple (neurostimulation method) versus double injection (transarterial method).
All received local anaesthetic (LA) solution: 1.6% lidocaine with 5 mcg/ml epinephrine; total volume 50 ml. The identity and experience level of the operator performing the blocks was not stated.

No premedication given.

1. Multiple injection: neurostimulation-guided, injection of 20 ml to ulnar or distal radial nerve response, 20 ml to median nerve, 10 ml to musculocutaneous nerve.
2. Double injection: transarterial technique, injection of 30 ml posterior to axillary artery, injection of 20 ml anterior to artery.

Incomplete blocks were supplemented but the definition of incomplete blocks and the timing of supplementation were not specified.

OutcomesLength of follow-up: sensory and motor block follow-up duration was not specified. Patients were followed up for 48 hours postoperatively to assess for complications
Sensory blockade
Analgesic failure (use of supplementary anaesthesia; general anaesthesia, pain at surgical site)
Time to readiness for surgery
Tourniquet discomfort and pain
Complications during block injection and operation, and up to 48 hours after
Block duration
Patient satisfaction
Notes

Block outcomes were vaguely defined, and assessment timing not specified.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk“prospective study”;  “who were randomly distributed in two groups (group MNS= 40 patients and group TA = 30 patients) according to the technique.” Unexplained imbalance
Allocation concealment (selection bias)Unclear riskNo details of method
Blinding (performance bias and detection bias)
All outcomes
High riskNo mention
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskLikely that all participants accounted for, but not stated explicitly
Selective reporting (reporting bias)Unclear riskNo prior protocol and vague in the definition of outcomes
Balance in baseline characteristics?Unclear riskBalanced for sex, age, weight and height but no information on surgery
Free from performance bias?Unclear riskInsufficient information to judge this

Inberg 1999

MethodsMethod of randomization: computer based, organized by the statistical department
Patient blinded
Blinded outcome assessor
No loss to follow-up
ParticipantsTampere, Finland
Period of study: 1996 to 1997
50 adults scheduled for upper limb surgery under axillary block anaesthesia. Informed consent
Male: 74%; mean age: 44.5 years
Excluded: weight < 50 kg or > 100 kg; surgery proximal from the elbow joint
Interventions

Double versus single injection (both groups using neurostimulation method).

All received local anaesthetic (LA) solution: Prilocaine 1% and bupivacaine 0.5% in 1:1 ratio; total volume 0.7 ml/kg body weight. Nerve blockade facilitated using 22G insulated needle and nerve stimulator. All blocks carried out by one operator.

Initial sedation with diazepam. Initial subcutaneous injection of 2 to 3 ml LA.
1. Double injection: injection (half volume) LA to median nerve followed by injection (rest of volume) LA to radial nerve (14 cases) or ulnar (11 cases).
2. Single injection: single injection of LA to median nerve (23 cases); radial nerve (1 case) or ulnar nerve (1 case).

Subcutaneous injection of 5 ml LA to block medial cutaneous nerves of the arm.

OutcomesLength of follow-up: 40 minutes (and duration of surgery)
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; new plexus block, general anaesthesia; use of opioids for tourniquet pain or pain in operation area)
Duration of surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during operation)
NotesReply to request for details of methods, type of surgery and details of participants receiving general and another full plexus block received from Dr Annila on 21/12/04. Original patient papers are now missing.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"computer-based randomization", organized by the statistical department
Allocation concealment (selection bias)Unclear riskNo information on concealment
Blinding (performance bias and detection bias)
All outcomes
Low risk“The evaluation of the sensory and motor blocks was blinded, and the patient was unaware of the method used, which makes the study double blind.”
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up. Information supplied by trial author on the particular circumstances of two patients, who received respectively a new plexus block and general anaesthesia
Selective reporting (reporting bias)Unclear riskPossible but no protocol available
Balance in baseline characteristics?Unclear riskNo information on type of surgery or mental status. No statistically significant differences between the two groups in age, gender, weight or height
Free from performance bias?Low riskSeemingly so: same experienced operator and anaesthetic procedures. "All blocks were performed by the first author, who is experienced in axillary block."

K-Nielsen 1997

MethodsMethod of randomization: computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room.
Blinded outcome assessor
No loss to follow-up
ParticipantsCopenhagen, Denmark
Period of study: 1995 to 1996
80 people undergoing elective hand surgery: arthrodeses, arthroplasties, osteosyntheses (K-wire, Herbert screw etc), nerve sutures, finger amputations, neuroma or ganglion removals, hardware removals, wrist arthroscopies. Informed consent.
Male: 66%; mean age: 46 years, range: 18 to 75 years
Excluded: ASA physical status > 3 (see notes). Allergy to amide type LA, pregnancy, inability to communicate, neurological disorders affecting peripheral nerves and resulting sensory loss and/or motor weakness (e.g. advanced neuropathies - uraemic or diabetic)
Interventions

Multiple versus single injection (both groups using neurostimulation method).

All received local anaesthetic (LA) solution: 1% mepivacaine with adrenaline 5 µg/ml.

Nerve blockade facilitated using 22G insulated short-bevelled needle and nerve stimulator. All blocks carried out by one operator.

Initial sedation with diazepam. Initial subcutaneous injection of 5 ml LA to anaesthetize medial cutaneous nerves of arm or forearm.
1. Multiple injection: injection of 10 ml LA cephalad to artery, then injection below artery - 20 ml at point of maximum stimulation OR if 2 nerves located: 15 ml close to each nerve. If just one located then final 10 ml into coracobrachial muscle.
2. Single injection: injection of 40 ml LA to one nerve: median nerve (32), ulnar (6), radial (2); then 20 ml to radial nerve.

OutcomesLength of follow-up: surgical follow-up (not stated) for adverse neurological outcomes; onset of sensory block assessed every 10 minutes (and duration of surgery)
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; general anaesthesia; use of opioids for tourniquet pain in operation area)
Time to be ready for surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during injection and operation)
Long-term complications (none)
Notes

Reply to request for details of types of surgery, exclusion criteria and so on received from Dr Koscielniak-Nielsen on 11/11/2004.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand ...."
Allocation concealment (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room."
Blinding (performance bias and detection bias)
All outcomes
Low risk“All the blocks were performed by the first author [who left the room] and assessed by the others. .. The [randomization] envelope was then resealed and attached to the assessment form.”
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various elective hand operations
Free from performance bias?Low riskAll the blocks were performed by the first author [an experienced anaesthetist]. Other care seemed comparable

K-Nielsen 1998

MethodsMethod of randomization: computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room.
Blinded outcome assessor
No loss to follow-up
ParticipantsCopenhagen, Denmark
Period of study: 1996 to 1997
100 people undergoing acute (nerve and/or tendon sutures, K-wire or Hoffmann osteosynthesis) or elective hand, wrist or forearm surgery. Informed consent
Male: 64%; mean age: 47 years, range: 18 to 80 years
Excluded: ASA physical status > 3 (see notes). Allergy to amide type LA, pregnancy, inability to co-operate, diseases affecting sensory or motor function of the upper extremity
Interventions

Multiple (neurostimulation method) versus double injection (transarterial method).

All received local anaesthetic (LA) solution: 1% mepivacaine with adrenaline 5 µg/ml.

Nerve blockade facilitated using 22G or 24G insulated short-bevelled needle with nerve stimulator in place. All blocks carried out by one operator.

Initial sedation with diazepam. Initial subcutaneous injection of 5 ml LA to anaesthetize medial cutaneous nerves of arm or forearm.
1. Multiple injection: injection of 10 ml LA to 4 terminal motor nerves (musculocutaneous, median, radial and ulnar). Connected to nerve stimulator, current = 1.5 mA.
2. Double injection "transarterial": injection of 20 ml LA deep to and 20 ml superficial to axillary artery. Connected to nerve stimulator but current = 0 mA.

OutcomesLength of follow-up: surgical follow-up (not stated) for adverse neurological outcomes; onset of sensory block assessed every 10 minutes (and duration of surgery)
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Time to be ready for surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during injection and operation)
Long-term complications (none)
Notes

Reply to request for details of types of surgery received from Dr Koscielniak-Nielsen on 15/11/2004.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand ...."
Allocation concealment (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room."
Blinding (performance bias and detection bias)
All outcomes
Low risk“All the blocks were performed by the first author [who left the room] and assessed by the others, who were unaware of the applied nerve block."
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various operations, which were elective or acute hand, wrist or forearm surgery
Free from performance bias?Low riskAll the blocks were performed by the first author [an experienced anaesthetist]. Other care seemed comparable

K-Nielsen 1999a

MethodsMethod of randomization: computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room.
Blinded outcome assessor(s)
No loss to follow-up but 1 excluded due to serious adverse effect
ParticipantsCopenhagen, Denmark
Period of study: 1998
101 people undergoing acute (nerve and/or tendon sutures, fracture osteosyntheses, amputations, wound revisions etc) or elective (Dupuytrens, arthroplasties, arthrodeses, ligament reconstructions, Hunter I or II, caput ulnae resections, scaphoideum osteosyntheses, neuroma or ganglion removals, carpal tunnel surgery etc) hand, wrist or forearm surgery. Informed consent.
Male: 64%; mean age: 49.5 years, range: 18 to 80 years
Excluded: ASA physical status > 2 (see notes). Allergy to amide type LA, pregnancy, inability to co-operate, diseases affecting sensory or motor function of the upper extremity
Interventions

Multiple (neurostimulation method) versus double injection (transarterial method).

All received local anaesthetic (LA) solution: 2% mepivacaine with adrenaline 5 µg/ml.

Nerve blockade facilitated using 22G or 24G insulated short-bevelled needle with nerve stimulator in place. Blocks carried out by one operator or under his supervision.

Initial sedation with diazepam. Initial subcutaneous injection of 5 ml LA (1% mepivacaine) to anaesthetize medial cutaneous nerves of arm or forearm.
1. Multiple injection: injection of 10 ml LA to 4 terminal motor nerves (Musculocutaneous, median, radial and ulnar). Connected to nerve stimulator, current = 1.5 mA.
2. Double injection "transarterial": injection of 20 ml LA deep to and 20 ml superficial to axillary artery. Connected to nerve stimulator but current = 0 mA.

OutcomesLength of follow-up: surgical follow up (5 to 10 days, then 3 to 4 weeks after) for adverse neurological outcomes; onset of sensory block assessed every 10 minutes
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Time to be ready for surgery
Pain during block
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during injection and operation)
Long-term complications (none)
Notes

Excluded patient was a cardiovascularly medicated participant of the multiple injection group who developed hypertension, atrial fibrillation, became agitated and lost consciousness 12 minutes after block performance. After intravenous administration of labetol, metoprolol and midazolam his condition improved and he woke up 15 minutes later.

Details of types of surgery, study period received from Dr Koscielniak-Nielsen on 02/12/2004.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand ...."
Allocation concealment (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room."
Blinding (performance bias and detection bias)
All outcomes
Low risk“All blocks were assessed by the anaesthetists, who were unaware of the applied nerve block technique."
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up but 1 excluded due to serious adverse effect
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various operations, which were elective or acute hand, wrist or forearm surgery
Free from performance bias?Unclear riskWhile 38% versus 32% of blocks were done by residents and other staff members (rather than the first author [an experienced anaesthetist], the first author supervised all blocks. Other care seemed comparable

K-Nielsen 1999b

MethodsMethod of randomization: computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room.
Blinded outcome assessor
No loss to follow-up
ParticipantsCopenhagen, Denmark
Period of study: 1998
106 people undergoing acute (nerve and/or tendon sutures, fracture osteosyntheses, amputations, wound revisions etc) or elective (Dupuytrens, arthroplasties, arthrodeses, ligament reconstructions, Hunter I or II, caput ulnae resections, scaphoideum osteosyntheses, neuroma or ganglion removals, carpal tunnel surgery etc) hand, wrist or forearm surgery. Informed consent
Male: 57%; mean age: 45.5 years, range: 18 to 80 years
Excluded: ASA physical status > 2 (see notes). Allergy to amide type local anaesthetic (LA), pregnancy, inability to co-operate, diseases affecting sensory or motor function of the upper extremity
Interventions

Multiple (neurostimulation method) versus single injection (transarterial/paraesthesia method).

Blocks carried out by first author - consultant anaesthetist - other staff members or supervised residents.

Initial sedation with diazepam to apprehensive patients. Initial subcutaneous injection of 5 ml mepivacaine 1% with adrenaline 5 µg/ml to anaesthetize medial cutaneous nerves of arm or forearm.
1. Multiple injection: injection using 24 gauge, 25 mm long insulated short-bevelled cannula of 5 ml mepivacaine 1% with adrenaline 5 µg/ml to 4 terminal motor nerves (Musculocutaneous, median, radial and ulnar). Connected to nerve stimulator, current = 1.5 mA.
2. Single injection: injection with 25 gauge, 35 mm long hypodermic needle of 80 ml mepivacaine 1% with adrenaline 2.5 µg/ml LA behind transfixed axillary artery (beforehand if hand paraesthesia elicited). Connected to nerve stimulator but current = 0 mA.

OutcomesLength of follow-up: surgical follow up (5 to 10 days, then 3 to 4 weeks after) for adverse neurological outcomes; onset of sensory block assessed every 10 minutes
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area; use of propofol for restlessness caused by tourniquet)
Time to be ready for surgery
Pain during block
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during injection and operation)
Long-term complications (none)
Notes

One of the two excluded trial participants was an Inuit who didn't understand trial procedures. The other participant, who already had coronary artery disease, developed chest pain - the surgery was cancelled.
Details of types of surgery, study period and other clarification received from Dr Koscielniak-Nielsen on 14/12/2004.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand ...."
Allocation concealment (selection bias)Low risk"A computer-generated random allocation assuring an equal number of patients in both groups was obtained beforehand and sealed in numbered envelopes. An envelope containing the random assignment was opened after the patient's arrival at the anaesthesia room."
Blinding (performance bias and detection bias)
All outcomes
Low risk“The anaesthetist performing the block ... restarted the stopwatch and left the room. The blocks were ...assessed by an anaesthesiologist unaware of the applied technique."
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up although 2 were excluded;  because of language and heart problems respectively
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various operations, which were elective or acute hand, wrist or forearm surgery
Free from performance bias?Low riskBlocks carried out by first author - consultant anaesthetist - other staff members or supervised residents. Other care seemed comparable

Lavoie 1992

MethodsMethod of randomization: use of random number table. "The 75 patients were blocked in 15 groups of 5 patients each (group 1, group 2 etc allocated in a random way into each one of the 15 groups)."
Double-blind: anaesthetist carrying out assessments and patients were blind to injection technique
No loss to follow-up
ParticipantsSherbrooke, Quebec, Canada
Period of study: 1991
75 people undergoing upper limb surgery including the elbow down to the hand: fractures, soft tissues. Informed consent
Male: 55%; mean age: 41 years
Excluded: no details
Interventions

Multiple versus double versus single (radial nerve) versus single (median nerve) versus single (ulnar nerve) injection.

All received local anaesthetic (LA) solution: 30 ml/square metre body surface (approximately 50 ml/70 kg body weight) lidocaine 1% with adrenaline 5 µg/ml. A tourniquet was used in all cases.

No premedication, sedation or IV analgesia mentioned. Anaesthetist performing the blocks was aware of the surgical site. After locating, by palpation, the axillary nerve in the axilla, 2 ml of 2% lidocaine injected subcutaneously to block medial cutaneous nerves of the arm. A 22-gauge insulated needle connected to peripheral nerve stimulator used to locate the nerves (0.5 mA current).

1. Multiple injection: to musculocutaneous, radial, median and ulnar nerves. LA volume equally divided between the 4 injections.
2. Double injection: to musculocutaneous and one of radial, median or ulnar nerves directly related to surgical site. LA volume equally divided between the 2 injections.
3. Single injection: to radial nerve of full volume of LA.
4. Single injection: to median nerve of full volume of LA.
5. Single injection: to ulnar nerve of full volume of LA.

OutcomesLength of follow-up: 30 minutes
Sensory blockade
Motor blockade
Anaesthesia failure (incomplete sensory blockade, general anaesthesia: none)
Adverse effects (none recalled by contact trialist)
NotesReply to request for details of methods, types of surgery and results, including adverse effects, received from Dr Martin on 05/01/05.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"We used a random number table. The 75 patients were blocked in 15 groups of 5 patients each (group 1, group 2 etc allocated in a random way into each one of the 15 groups)."
Allocation concealment (selection bias)Unclear riskNo information, but some predictability may have occurred at the end of each block
Blinding (performance bias and detection bias)
All outcomes
Low risk“The anaesthetist performing the block was aware of the surgery but another anaesthetist unaware of the patients’ group evaluated the sensory and motor blocks ..” ".. the patient did not know what aspect of his axillary block was studied specifically .."
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up
Selective reporting (reporting bias)Unclear risk Possible but no protocol available
Balance in baseline characteristics?Low riskFine. Although no details of surgery, there was balance in the implicated nerves
Free from performance bias?Low riskSeems likely and the author suggested that the "technique of randomization by blocks allows that the learning of the technique is uniform on the 5 groups of patients".

Pere 1993

MethodsMethod of randomization: not stated
No blinding indicated
No loss to follow-up (assumed for 3 hours follow-up)
ParticipantsHelsinki, Finland
Period of study: not stated
50 people undergoing hand, forearm or elbow surgery. Informed consent
Male: not stated; mean age: 37 years
Excluded: ASA physical status > 2 (see notes)
Interventions

Double (transarterial method) versus single injection (neurostimulation method).

All received LA solution: mepivacaine 1% with adrenaline 5 µg/ml.

Premedication with diazepam and oxycodone.
1. Double injection "transarterial": injection using 0.7 x 50 mm needle advanced through the artery. Injection of half of 45 ml LA deep to and half superficial to axillary artery.
2. Single "perivascular" injection: injection using 0.7 x 50 mm needle and facilitated by nerve stimulator of 45 ml LA after location of axillary brachial plexus.

Neurovasular sheath compressed during the injection and for 3 minutes afterwards.

OutcomesLength of follow-up: 3 hours
Sensory blockade - 8 nerves
Motor blockade
Analgesic failure (use of supplementary blocks; general anaesthesia; use of opioids)
Duration of tourniquet use
Notes

Subsidiary radiological study of 16 people also performed. It was not clear if the people were randomized to the same comparison as the clinical trial. Need for supplementary blocks (2/8 versus 2/8) and more than 1 dose of opoid (1 versus 1) was the same in both groups.

Response from Dr Pere, received 10/02/05, indicated, with regret, that there was now no more information available on this trial.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"Patients were allocated randomly to two groups."
Allocation concealment (selection bias)Unclear riskNo details of method
Blinding (performance bias and detection bias)
All outcomes
High riskNo mention of blinding
Incomplete outcome data (attrition bias)
All outcomes
Low riskSeemed fine
Selective reporting (reporting bias)Unclear riskProtocol not available. No mention of complications
Balance in baseline characteristics?Unclear risk"There were no differences between the groups in patient age, height or weight.." However, no details of type of surgery or gender.
Free from performance bias?Unclear riskNo details of who administered the anaesthesia

Ramirez-Gomez 2010

Methods

Method of randomization: not stated

No blinding of outcome assessor

No loss to follow-up; excluded 3 patients who received general anaesthesia from certain outcome assessments based on assumed block failure

Participants

Tamaulipas, Mexico

Period of study: not stated

50 adult patients undergoing surgery of the elbow, forearm, wrist and hand

Male:female ratio 27:23. ASA 1-3. Age not stated

Interventions

Multiple versus single injection, both neurostimulation guided, motor response sought at 0.4mA or less.

Multiple injection: radial nerve + one or more of the following: median, ulnar, musculocutaneous.

Single injection: any motor response.

All patients received 30 ml of local anaesthetic mixture: 20 ml of 1.5% lidocaine + 10 ml 0.25% bupivacaine.

All patients received pre-operative ketorolac 1 mg/kg and block sedation of midazolam 0.1 mg/kg and nalbuphine 0.1 mg/kg.

Patients received local anaesthetic supplementation if sensory and motor block was inadequate after 30 minutes, and general anaesthesia if there was intraoperative pain.

Outcomes

Need for local anaesthetic supplementation

Need for general anaesthesia

Duration of motor and sensory block

Surgical site pain score at 8 hours and 24 hours postoperatively

NotesStudy was published in Spanish, and translated by co-author (JC). The methods state that patients were undergoing trauma surgery of the lower arm, however the data supplied indicate that the majority underwent surgery for non-traumatic conditions.There were inconsistencies in the paper between data reported in the text and in the tables. The authors were contacted for clarification of methodology and data; a response was received which provided some additional information but did not answer all queries.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot described
Allocation concealment (selection bias)Unclear riskNot described
Blinding (performance bias and detection bias)
All outcomes
High riskThere was no blinding (personal communication from lead author)
Incomplete outcome data (attrition bias)
All outcomes
High riskThe reported denominators for each group vary between the text and the tables. The authors clarification also contains inconsistencies in the numbers. It is unclear if this is simple error or reflects incomplete outcome data. Communication with the authors was unhelpful in clarifying this.
Selective reporting (reporting bias)Unclear riskThey appear to have collected data which was not reported, e.g. onset of block in the various dermatomes. Complication data was not reported. Communication with the authors was unhelpful in clarifying this.
Balance in baseline characteristics?Unclear riskThere was a high proportion of minor surgery, e.g. carpal tunnel release. The distribution of this surgery between groups was not stated.
Free from performance bias?Unclear riskMultiple-injection group: the proportion of patients receiving 2, 3, or 4 injections was not stated; which could have an influence on the results. It was not clear how skilled the investigators were at performing the more complex multiple-injection technique.

Rodriguez 2005

MethodsMethod of randomization: computer-generated randomization list
Blinded outcome assessor
No loss to follow-up. Four incomplete procedures included in intention-to-treat analysis
ParticipantsSantiago, Spain
Period of study: not stated
120 people undergoing surgery of the upper limb (not otherwise specified). Informed consent
Male: 27%; mean age: 51 years
Excluded: ASA physical status > 3 (see notes)
Interventions

Multiple versus double versus single (median nerve) versus single (radial nerve) injection.

All received local anaesthetic (LA) solution: 1.5% mepivacaine; total volume 40 ml. All blocks were neurostimulation-guided with a 22G insulated block needle. All blocks were performed by one of two senior anaesthesiologists.

Sedative premedication with 1 to 3 mg of IV midazolam according to clinical judgement.

1. Multiple injection: injection of 15 ml to radial nerve, 15 ml to median nerve, 10 ml to musculocutaneous nerve.
2. Double injection: injection of 35 ml on radial nerve, injection of 5 ml on musculocutaneous nerve.
3. Single injection (median): injection of 40 ml on median nerve.
4. Single injection (radial): injection of 40 ml on radial nerve.

Blocks were supplemented preoperatively if the operative nerve distributions did not have complete sensory block before operation; timing of this was not specified.

Intraoperative pain was treated with infiltration of local anaesthetic at the site, or with injection of 50 to 100 mcg of fentanyl. General anaesthesia was used if pain was persistent.

OutcomesLength of follow-up: sensory and motor block assessed at 5 and 20 minutes. No follow-up detailed beyond that
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; general anaesthesia; pain in operative field)
Block performance time
Notes

Request for additional information on method of randomization, blinding, and results sent to Dr Rodriguez on 13/07/2010; no reply received.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk“Assignment was performed by means of a computer-generated randomization list.”
Allocation concealment (selection bias)Unclear riskNo information on allocation concealment
Blinding (performance bias and detection bias)
All outcomes
Low riskParticipants - not blinded. Caregivers - not blinded. Outcome assessors - blinded. No mention of safeguards, but plausible. Blinding is within study limitations and unlikely to influence outcome
Incomplete outcome data (attrition bias)
All outcomes
High riskIntention-to-treat analysis was done and there were none lost to follow-up. But the length of follow-up was only 20 minutes, yet authors state “we had the clinical impression that many incomplete blocks progressed until 30 minutes.” It is unclear if they supplemented the blocks after 20 minutes, or later, yet this is reported as an outcome.
Selective reporting (reporting bias)Unclear riskProtocol not available. No reporting of complications, which is unusual for this type of study
Balance in baseline characteristics?Unclear riskNo details of type of surgery performed
Free from performance bias?Low riskBlocks were administered by one of two senior anaesthesiologists

Rodriguez 2008

MethodsMethod of randomization: computer-generated randomization list
Blinded outcome assessor
Loss for follow-up: one patient in the double group excluded after randomization and block performance as assessment was not possible
ParticipantsSantiago, Spain
Period of study: not stated
60 people undergoing surgery of the hand (49), forearm (3), elbow (8). Informed consent
Male: 48.3%; mean age: 58 years
Excluded: ASA physical status > 3 (see notes)
Interventions

Multiple versus double injection (both groups using neurostimulation method).

All received local anaesthetic (LA) solution: 2% mepivacaine of volume 30 ml, and 1% mepivacaine of volume 5 ml (to musculocutaneous nerve); total volume 35 ml. All blocks were neurostimulation-guided with a 22G insulated block needle. Identity and experience level of operators performing block were not specified.

Sedative premedication with 1 to 3 mg of IV midazolam.

1. Multiple injection: injection of 15 ml to radial nerve, 15 ml to median nerve, 5 ml to musculocutaneous nerve.
2. Double injection: injection of 30 ml on radial nerve, injection of 5 ml on musculocutaneous nerve.

OutcomesLength of follow-up: sensory and motor block assessed at 10, 20 and 30 minutes
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; general anaesthesia; pain in operative field)
Block performance time (reported as median and ranges)
Acute complications during block procedure
Notes

Request for additional information on method of randomization, blinding, and results sent to Dr Rodriguez on 13/07/2010; no reply received.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk“Assignment was performed by computer-generated randomization list.”
Allocation concealment (selection bias)Unclear riskNo information provided
Blinding (performance bias and detection bias)
All outcomes
Low riskParticipants - not blinded. Caregivers - not blinded. Outcome assessors - blinded. No mention of safeguards, but plausible. Blinding is within study limitations and unlikely to influence outcome
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskMostly: four incomplete procedures were included -  but there is slight concern over the one excluded patient; the percentages in Table 4 imply patient was included in the analysis
Selective reporting (reporting bias)Unclear riskNo protocol and side effects not reported
Balance in baseline characteristics?Unclear riskBaseline characteristics balanced but no details of distribution of types of surgery
Free from performance bias?Unclear riskThe identity and experience of care providers was not stated

Serradell Catalan 2001

MethodsMethod of randomization: use of a computer-generated table of random numbers to generate a randomization list
Blinded outcome assessor
No loss to follow-up
ParticipantsBarcelona, Spain
Period of study: 1999 to 2000
100 adults (> 18 years) undergoing upper limb (forearm, wrist or hand) post-traumatic orthopaedic surgery. ASA physical status 1-3 (see notes). Informed consent
Male: 56%; mean age: 55 years
Excluded: usual contraindications for axillary nerve blockade and regional anaesthesia. Motor or sensory disease of limb involved in surgical procedure. Non-palpable axillary artery pulse
Interventions

Multiple (musculocutaneous, radial, median and ulnar nerves) versus multiple (triple: musculocutaneous + two of radial/ median/ ulnar nerves) versus double (two of radial/ median/ ulnar nerves) versus
double (musculocutaneous + radial/ median/ ulnar nerve) versus single (radial/ median/ ulnar nerve) injection.

All received local anaesthetic (LA) solution: 40 ml mepivacaine 1%.

Oral premedication with lorazepam 1 mg and sedation with IV midazolam. One anaesthetist performed all the blocks. After locating the axillary artery in the axilla, lidocaine 1% injected subcutaneously over the arterial pulse. A 22 gauge 50 mm long insulated needle connected to nerve stimulator was used to locate the nerves.

1. Multiple injection: to musculocutaneous, radial, median and ulnar nerves.
2. Multiple injection: to musculocutaneous nerve (10 ml LA) and either to the radial and median nerves or the radial and ulnar nerves or the median and ulnar nerves.
3. Double injection: to the radial and median nerves or the radial and ulnar nerves or the median and ulnar nerves.
4. Double injection: to musculocutaneous nerve (10 ml LA) and either to the radial, median or ulnar nerve.
5. Single injection: to the radial or median or ulnar nerve.

OutcomesLength of follow-up: 3 months surgical follow-up (also 24 hours) for adverse neurological outcomes; onset of sensory block assessed every 10 minutes, 40 minutes; also duration of block
Sensory blockade
Motor blockade
Anaesthesia failure (incomplete sensory blockade, incomplete motor blockade)
Time for block
Duration of sensory blockade
Duration of tourniquet use
Tourniquet discomfort
Problems (during injection)
Long term neurological complication
Participant opinion of technique
Notes

Part translation from Spanish provided by co-author (SR).
Response to request for details of methods, trial setting and use of intra-operative opioids and sedatives received from Dr Serradell Catalan on 03/02/05.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"A computer-generated table of random numbers.”
Allocation concealment (selection bias)High risk"read off the allocation from a list."
Blinding (performance bias and detection bias)
All outcomes
Low risk"Single blind" Blinded doctor for motor and sensory block evaluation
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss to follow-up
Selective reporting (reporting bias)Unclear riskPossible but no protocol available. There was a sample size calculation
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various operations, which were upper limb (forearm, wrist or hand) post-traumatic orthopaedic surgery
Free from performance bias?Low riskAll blocks were performed by the same doctor. Other care seemed comparable

Sia 2001

MethodsMethod of randomization: not stated
Double-blind
Blinded outcome assessor
No loss to follow-up (assumed for 30 days follow-up)
ParticipantsFlorence, Italy
Period of study: 2000?
100 people undergoing elective upper limb surgery in hand, wrist or forearm. Informed consent
Male: 55%; mean age: 41.5 years
Excluded: ASA physical status > 2 (see notes)
Interventions

Multiple versus double injection (both groups using neurostimulation method).

All received local anaesthetic (LA) solution: lidocaine 2% and bupivacaine 0.5% in 1:1 ratio. Nerve blockade facilitated using 22G insulated short-bevelled needle and nerve stimulator. All blocks carried out by one operator.

All received IV midazolam (sedation) and fentanyl 5 minutes before block. Initial subcutaneous injection of 4 ml LA to anaesthetize medial cutaneous nerves of arm or forearm.
1. Multiple (triple) injection: injection of 10 ml LA to musculocutaneous nerve; then 10 ml to median nerve and 20 ml to radial nerve.
2. Double injection: injection of 20 ml LA to median nerve; then 20 ml to radial nerve.

OutcomesLength of follow-up: nerve injury at 48 hours, neurological sequelae at 10 and 30 days; 30 minutes or until sensory block (and duration of surgery)
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Duration of anaesthesia
Duration of surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Problems (during injection and operation)
Long-term neurological complication (none)
Notes

Request for details of method of randomization, types of surgery and some of the results sent to Dr Sia on 09/11/04.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"One hundred patients were randomly allocated to 2 groups." No details of method
Allocation concealment (selection bias)Unclear riskNo information
Blinding (performance bias and detection bias)
All outcomes
Low risk"double-blind study"; "All blocks .... were assessed by an investigator unaware of group assignment."
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo loss of follow-up apparent for block performance
Selective reporting (reporting bias)Unclear riskPossible but no protocol available. The selection of primary outcome for sample size calculation of the blocking of the musculocutaneous nerve is unusual
Balance in baseline characteristics?Unclear riskYes, aside from the distribution of the various operations, which were elective forearm, wrist or hand surgery
Free from performance bias?Low risk"All blocks were performed or supervised by the first author." Other care seemed comparable

Sia 2010a

MethodsMethod of randomization: computer-generated randomization list
Blinded outcome assessor
No loss to follow-up (14 patients - distributed between the three trials Sia 2010 a,b+c - were excluded after randomization because of inability to locate the desired nerves)
ParticipantsFlorence, Italy
Period of study: 2005 to 2008
138 people undergoing surgery on the fifth finger (fractures, neoformations, tendon injuries) and on the fifth metacarpal bone. Informed consent
Male: 56.5%; mean age: 44 years
Excluded: ASA physical status > 2 (see notes)
Interventions

Multiple (triple) versus single (ulnar) injection (both groups using neurostimulation method).

All received local anaesthetic (LA) solution: lidocaine 2% and bupivacaine 0.5% in 1:1 ratio; total volume 40 ml. Nerve blockade facilitated using 22G insulated short-bevelled needle and nerve stimulator. All blocks carried out by one experienced operator.

All received IV midazolam 20 µg/kg and fentanyl 1 µg/kg 5 minutes before block. Initial subcutaneous injection of 4 ml LA over the axillary artery to anaesthetize medial cutaneous nerves of arm and forearm.
1. Multiple (triple) injection: injection of 10 ml LA to median nerve; 6 ml to musculocutaneous nerve and 20 ml to radial nerve.
2. Single injection: injection of 36 ml LA to ulnar nerve.

OutcomesLength of follow-up: 30 minutes for sensory and motor block; 48 hours for nerve injury; neurological sequelae at 10 and 30 days
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Block performance time
Block onset time
Time to readiness for surgery
Duration of surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Need for intraoperative sedation
Problems (during injection and operation)
Long-term neurological complication (none)
Notes

Request for clarification on patient enrolment and additional data on complications sent to Dr Salvatore Sia on 14/10/2010; reply received 06/11/2010.

Although the Results section for the 3 trials states that 6 patients were excluded in the triple injections groups (TNS) and 8 patients in single injection groups (SEL), Dr Sia clarified in a personal communication that these were treated as "pre-operative" dropouts and were replaced by other patients.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"Patients undergoing each type of surgery were randomly assigned by a computer-generated list ..."
Allocation concealment (selection bias)Unclear riskAllocation concealment was not stated
Blinding (performance bias and detection bias)
All outcomes
Low riskPatients - not blinded. Caregiver - unblinded. Assessors - blinded. “All the blocks were... assessed by a blinded investigator.” Blinding is within study limitations
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk“Six patients in group TNS and 8 in group SEL were excluded from the study because all the prearranged nerves were not located by nerve stimulation.” Personal communication from the author indicates that these were post-randomization exclusions.
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Low riskBalanced
Free from performance bias?Unclear riskAll the blocks were performed or supervised by the first author, however the number and experienced level of supervised care providers is not stated.

Sia 2010b

MethodsMethod of randomization: computer-generated randomization list
Blinded outcome assessor
No loss to follow-up (14 patients - distributed between the three trials Sia 2010 a,b+c were excluded after randomization because of inability to locate the desired nerves)
ParticipantsFlorence, Italy
Period of study: 2005 to 2008
138 people undergoing superficial surgery (without bone involvement) on the palm (e.g., Dupuytren contracture, tendons or nerve injuries, neoformations) or on the dorsum of the hand (e.g., cysts, neoformations, pathologies of extensor tendons). Informed consent
Male: 45%; mean age: 49.5 years
Excluded: ASA physical status > 2 (see notes)
Interventions

Multiple (triple) versus double (median and ulnar) injection.

All received local anaesthetic (LA) solution: lidocaine 2% and bupivacaine 0.5% in 1:1 ratio; total volume 40 ml. Nerve blockade facilitated using 22G insulated short-bevelled needle and nerve stimulator. All blocks carried out by one experienced operator.

All received IV midazolam 20 mcg/kg and fentanyl 1 mcg/kg 5 minutes before block. Initial subcutaneous injection of 4 ml LA over the axillary artery to anaesthetize medial cutaneous nerves of arm and forearm.
1. Multiple (triple) injection: injection of 10 ml LA to median nerve; 6 ml to musculocutaneous nerve and 20 ml to radial nerve.
2. Double injection: injection of 18 ml LA to ulnar nerve, injection of 18 ml LA to median nerve.

Outcomes

Length of follow-up: 30 minutes for sensory and motor block; 48 hours for nerve injury; neurological sequelae at 10 and 30 days
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Block performance time

Block onset time

Time to readiness for surgery
Duration of surgery
Duration of tourniquet use
Tourniquet discomfort and pain

Need for intraoperative sedation
Problems (during injection and operation)
Long-term neurological complication (none)

Notes

Request for clarification on patient enrolment and additional data on complications sent to Dr Salvatore Sia on 14/10/2010; reply received 06/11/2010.

Although the Results section for the 3 trials states that 6 patients were excluded in the triple injections groups (TNS) and 8 patients in single injection groups (SEL), Dr Sia clarified in a personal communication that these were treated as "pre-operative" dropouts and were replaced by other patients.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"Patients undergoing each type of surgery were randomly assigned by a computer-generated list ..."
Allocation concealment (selection bias)Unclear riskAllocation concealment was not stated
Blinding (performance bias and detection bias)
All outcomes
Low riskPatients - not blinded. Caregiver - unblinded. Assessors - blinded. “All the blocks were... assessed by a blinded investigator.” Blinding is within study limitations
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk“Six patients in group TNS and 8 in group SEL were excluded from the study because all the prearranged nerves were not located by nerve stimulation.” Personal communication from the author indicates that these were post-randomization exclusions.
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Low riskBalanced
Free from performance bias?Unclear riskAll the blocks were performed or supervised by the first author, however the number and experienced level of supervised care providers is not stated.

Sia 2010c

MethodsMethod of randomization: computer-generated randomization list
Blinded outcome assessor
No loss to follow-up (14 patients - distributed between the three trials Sia 2010 a,b+c - were excluded after randomization because of inability to locate the desired nerves)
ParticipantsFlorence, Italy
Period of study: 2005 to 2008
138 people undergoing any surgery on the first three fingers in which only 1 or 2 nerves were involved. Informed consent
Male: 52%; mean age: 45.5 years
Exclued: ASA physical status > 2 (see notes)
Interventions

Multiple (triple) versus double (median and radial) injection.

All received local anaesthetic (LA) solution: lidocaine 2% and bupivacaine 0.5% in 1:1 ratio; total volume 40 ml. Nerve blockade facilitated using 22G insulated short-bevelled needle and nerve stimulator. All blocks carried out by one experienced operator.

All received IV midazolam 20 µg/kg and fentanyl 1 µg/kg 5 minutes before block. Initial subcutaneous injection of 4 ml LA over the axillary artery to anaesthetize medial cutaneous nerves of arm and forearm.
1. Multiple (triple) injection: injection of 10 ml LA to median nerve; 6 ml to musculocutaneous nerve and 20 ml to radial nerve.
2. Double injection: injection of 18 ml LA to median nerve, and injection of 18 ml LA to radial nerve.

OutcomesLength of follow-up: 30 minutes for sensory and motor block; 48 hours for nerve injury; neurological sequelae at 10 and 30 days
Sensory blockade
Motor blockade
Analgesic failure (use of supplementary anaesthesia; use of opioids for tourniquet pain in operation area)
Block performance time
Block onset time
Time to readiness for surgery
Duration of surgery
Duration of tourniquet use
Tourniquet discomfort and pain
Need for intraoperative sedation
Problems (during injection and operation)
Long-term neurological complication (none)
Notes

Request for clarification on patient enrolment and additional data on complications sent to Dr Salvatore Sia on 14/10/2010; reply received 06/11/2010.

Although the Results section for the 3 trials states that 6 patients were excluded in the triple injections groups (TNS) and 8 patients in single injection groups (SEL), Dr Sia clarified in a personal communication that these were treated as "pre-operative" dropouts and were replaced by other patients.

The ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk"Patients undergoing each type of surgery were randomly assigned by a computer-generated list ..."
Allocation concealment (selection bias)Unclear riskAllocation concealment was not stated
Blinding (performance bias and detection bias)
All outcomes
Low riskPatients - not blinded. Caregiver - unblinded. Assessors - blinded. “All the blocks were... assessed by a blinded investigator.” Blinding is within study limitations
Incomplete outcome data (attrition bias)
All outcomes
Unclear risk“Six patients in group TNS and 8 in group SEL were excluded from the study because all the prearranged nerves were not located by nerve stimulation.” Personal communication from the author indicates that these were post-randomization exclusions
Selective reporting (reporting bias)Low riskNo protocol but same plan for series of trials. Clearly reported primary outcomes. Ethics committee acceptance reported
Balance in baseline characteristics?Low riskBalanced
Free from performance bias?Unclear riskAll the blocks were performed or supervised by the first author, however the number and experienced level of supervised care providers is not stated.

Turkan 2002

  1. a

    ASA = American Society of Anaesthesiologists
    LA = local anaesthetic
    IV = intravenous

MethodsMethod of randomization: not stated
Blinded outcome assessor
No loss to follow-up apparent
ParticipantsAnkara, Turkey
Period of study: not stated
69 people undergoing orthopaedic or trauma surgery of the upper extremity (not otherwise specified). Informed consent
Male: 75%; mean age: 49 years
Excluded: ASA physical status > 2 (see notes)
Interventions

Double versus single (Winnie's technique) versus single (transarterial) injection.

All received local anaesthetic (LA) solution: 2% prilocaine and 0.5% bupivacaine in 1:1 ratio; total volume 40 ml. Identity of operators performing block were not specified; but they were described as experienced.

Sedative premedication with 0.15 mg/kg of IM midazolam.

1. Double injection: injection of 20 ml using Winnie's technique (endpoint of fascial click and paraesthesia in hand or forearm), and injection of 20 ml using transarterial technique posterior to the axillary artery.
2. Single injection: injection of 40 ml using Winnie's technique (endpoint of fascial click and paraesthesia in hand or forearm).
3. Single injection: injection of 40 ml using transarterial technique posterior to the axillary artery.

When patient in extreme anxiety or block was incomplete, propofol (≤ 3 mg/kg) and/or fentanyl (≤ 1µg/kg) was administered intraoperatively.

OutcomesLength of follow-up: sensory and motor block assessed at 10, 20 and 30 minutes
Sensory blockade
Analgesic failure (use of supplementary anaesthesia)
Tourniquet discomfort and pain
NotesThe ASA (American Society of Anesthesiologists) physical status classification is a system for assessing the fitness of patients before surgery. It has five categories (1-5): 1 = healthy, 2 = mild systemic disease, 3 = severe systemic disease, 4 = severe disease that is a constant threat to life, 5 = moribund.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear risk"The patients were divided randomly into three groups..". No further details given
Allocation concealment (selection bias)Unclear risk Allocation concealment not stated
Blinding (performance bias and detection bias)
All outcomes
Unclear riskNot stated for participants, operators. It is mentioned that part of sensory testing was done by a blinded surgeon, but no detail given for other outcomes
Incomplete outcome data (attrition bias)
All outcomes
Low riskNot evident
Selective reporting (reporting bias)Unclear riskA bit vague in definition of outcomes and some of the P values seem excessive
Balance in baseline characteristics?Unclear riskBalanced for sex, age, weight and height but no information on surgery
Free from performance bias?Unclear riskInsufficient detail given regarding operator experience. “Experienced hands” implied in Discussion but no information to judge this

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Bernucci 2012Not in scope of review: use of ultrasound-guided method
Bloc 2010Not in scope of review: use of ultrasound-guided method
Bouaziz 1997Not in scope of review: comparison of two approaches: midhumeral versus axillary
Carre 2000Not in scope of review: children only
Gianesello 2010Not in scope of review: review of full text revealed that the study compared two different multiple-injection methods
Huynh 2012Not in scope of review: use of ultrasound-guided method
Imasogie 2010Not in scope of review: use of ultrasound-guided method
K-Nielsen 2000Not in scope of review: comparison of two approaches: subcoracoid versus axillary
Kjelstrup 2006Non-randomized study
Liu 2005Not in scope of review: use of ultrasound-guided method
Sia 2001bNot in scope of review: both interventions tested belonged to the multiple-injection group
Singelyn 1992Not in scope of review: single injection in both groups
Sites 2006Not in scope of review: use of ultrasound-guided method
Tran 2012Not in scope of review: use of ultrasound-guided method
Tuominen 1987Not in scope of review: review of the full text revealed that these were both single-injection techniques
Vester-Andersen 1984Not in scope of review: single injection into the same site via indwelling catheter
Vester-Andersen 1986Not in scope of review: single injection into the same site via indwelling catheter
Youssef 1988Not in scope of review: review of the full text revealed that these were both single-injection techniques
Yu 2007Not in scope of review: use of ultrasound-guided method

Ancillary