Intraoperative nerve action and compound motor action potential recordings in patients with obstetric brachial plexus lesions

Clinical article

Full access

Object

A typical finding in supraclavicular exploration of infants with severe obstetric brachial plexus lesions (OBPLs) is a neuroma-in-continuity with the superior trunk and/or a root avulsion at C-5, C-6, or C-7. The operative strategy in these cases is determined by the intraoperative assessment of the severity of the lesion. Intraoperative nerve action potential (NAP) and evoked compound motor action potential (CMAP) recordings have been shown to be helpful diagnostic tools in adults, whereas their value in the intraoperative assessment of infants with OBPLs remains to be determined.

Methods

Intraoperative NAPs and CMAPs were systematically recorded from damaged and normal nerves of the upper brachial plexus in a consecutive series of 95 infants (mean age 175 days) with OBPLs. A total of 599 intraoperative NAP and 836 CMAP recordings were analyzed. The severity of the nerve lesions was graded as normal, axonotmesis, neurotmesis, or root avulsion, based on surgical, clinical, histological, and radiographic criteria.

Results

The correlation of NAP and CMAP recordings with the severity of the lesion was assessed. The specificity of an absent NAP or CMAP to predict a severe lesion (neurotmesis or avulsion) was > 0.9. However, the sensitivity of an absent NAP or CMAP for predicting a severe lesion was low (typically < 0.3). The severity of the nerve lesion was related to CMAP and NAP amplitudes. Cutoff points useful for intraoperative decision making could not be found to differentiate between lesion types in individual patients.

Conclusions

Intraoperative NAP and CMAP recordings do not assist in decision making in the surgical treatment of infants with OBPLs. The authors' findings in infants cannot be generalized to adults.

Abbreviations used in this paper: ADST = anterior division of the superior trunk; CMAP = compound motor action potential; EMG = electromyography; NAP = nerve action potential; OBPL = obstetric brachial plexus lesion; PDST = posterior division of the superior trunk; ROC = receiver operator curve.

Abstract

Object

A typical finding in supraclavicular exploration of infants with severe obstetric brachial plexus lesions (OBPLs) is a neuroma-in-continuity with the superior trunk and/or a root avulsion at C-5, C-6, or C-7. The operative strategy in these cases is determined by the intraoperative assessment of the severity of the lesion. Intraoperative nerve action potential (NAP) and evoked compound motor action potential (CMAP) recordings have been shown to be helpful diagnostic tools in adults, whereas their value in the intraoperative assessment of infants with OBPLs remains to be determined.

Methods

Intraoperative NAPs and CMAPs were systematically recorded from damaged and normal nerves of the upper brachial plexus in a consecutive series of 95 infants (mean age 175 days) with OBPLs. A total of 599 intraoperative NAP and 836 CMAP recordings were analyzed. The severity of the nerve lesions was graded as normal, axonotmesis, neurotmesis, or root avulsion, based on surgical, clinical, histological, and radiographic criteria.

Results

The correlation of NAP and CMAP recordings with the severity of the lesion was assessed. The specificity of an absent NAP or CMAP to predict a severe lesion (neurotmesis or avulsion) was > 0.9. However, the sensitivity of an absent NAP or CMAP for predicting a severe lesion was low (typically < 0.3). The severity of the nerve lesion was related to CMAP and NAP amplitudes. Cutoff points useful for intraoperative decision making could not be found to differentiate between lesion types in individual patients.

Conclusions

Intraoperative NAP and CMAP recordings do not assist in decision making in the surgical treatment of infants with OBPLs. The authors' findings in infants cannot be generalized to adults.

Obstetric brachial plexus lesions are caused by traction during delivery.16 The reported incidence varies from 1.6 to 2.9 per 1000 births in prospective studies.1,6 The upper part of the brachial plexus, which includes the spinal nerves C-5 and C-6 and the superior trunk, is always involved. In more severe lesions, C-7 may be involved as well, and in ~ 15% of patients the spinal nerves at C-8 and T-1 are also damaged, resulting in impaired hand function.1,7,23

Depending on the severity of the traction injury, spontaneous recovery may occur. Neurapraxia and axonotmesis can eventually lead to complete recovery, but neurotmesis and root avulsion will result in permanent loss of function. Fortunately most children show good spontaneous recovery, but deficits are residual in 20–30% of children.19

At present, surgical exploration is performed when clinical examination does not show improvement in the associated muscles or movements at the age of 3,8 6,26 or 9 months,5 depending on the surgeon's approach.

The necessity for surgical nerve repair in cases of root avulsion or clear rupture is undisputed. The typical intraoperative finding in infants with OBPLs, however, is a neuroma-in-continuity of the superior trunk. A neuroma-in-continuity results from a traction injury in which the intraneural architecture is severely damaged, but the nerve does not completely rupture into 2 separate parts. Such lesions represent an intermediate position between axonotmesis and neurotmesis. When a significant part of the basal lamina tubes are spared, the potential for spontaneous axonal recovery remains and the lesion-incontinuity can be labeled as mainly axonotmetic. Resection followed by nerve grafting is not indicated in this case. When the lesion-in-continuity is mainly neurotmetic, however, it should be resected and grafting should be performed.

Differentiating between axonotmesis and neurotmesis relies on intraoperative appraisal by the surgeon, taking into account CT myelography findings. It depends on the amount of fibrosis, increase in cross-sectional diameter, loss of fascicular structure, and muscle contraction response to direct nerve stimulation. Such a differentiation may be difficult and depends on the surgeon's level of experience. Even in experienced hands misjudgment can be made in this situation, leading to unsatisfactory outcomes. Objective parameters for intraoperative assessment are therefore necessary to assess the severity of a neuroma-in-continuity.

In adults, recording of intraoperative NAPs and evoked CMAPs is advocated to distinguish objectively between axonotmetic and neurotmetic lesions.10,24 It has been shown that the presence of an NAP across the lesion site requires at least 3000–4000 nerve fibers with a diameter > 5 μm. The presence of these fibers in a recovering nerve indicates that spontaneous functional recovery will take place, and that therefore resection and grafting are not indicated.11,12 The presence of an NAP in a damaged nerve will precede the reinnervation of the muscle and consequently be detectable weeks to months before there is EMG evidence of reinnervation and even longer before clinical recovery.

In contrast to adults, little is known about intraoperative neurophysiological assessment in infants, and the use of this technique is not widespread. In a study of 10 patients, the lesion-in-continuity was not resected in 5 patients because the NAP recordings were considered indicative of neural continuity across the neuroma.13 These patients, however, did not recover well, so the NAP predictions were considered too optimistic.13 Other authors have only mentioned the use of intraoperative neurophysiological evaluation as a potential assessment tool,14,21,22 but their methods and results were not published in detail.

The present study was undertaken to explore the use of intraoperative NAP and CMAP measurements in patients with OBPLs. Our objectives were to assess the validity of this method, the correlation of NAP and CMAP measurements with the severity of the lesion on both an individual and group basis, and the factors affecting the recordings.

Methods

Patient Population

The NAP and CMAP recordings were obtained in 98 consecutive OBPL surgical procedures by the senior neurosurgeon (M.J.A.M.) between 2001 and 2005. The infants' parents gave informed consent for the extensive EMG recordings.

The decision to perform surgery was based either on complete paralysis of the arm and hand at 3 months of age or on insufficient recovery of upper arm function at 4–6 months of age, as outlined in previous reports.17,18 In all patients, CT myelography and ultrasonography of diaphragm movements had been performed preoperatively.

Three of 98 patients were excluded from analysis because they were > 1 year of age at surgery because of late referral. Because surgery is not routinely performed in patients of that age, these 3 cases were outliers. The analyzed group consisted of 48 boys and 47 girls. In 7 infants the lesion had occurred during a breech delivery, and the remaining infants had a cephalic presentation. The mean age at first clinical assessment at our outpatient clinic was 74 days (range 2–283 days). The lesion was strictly limited to the upper part of the brachial plexus (C5–6) in 20 patients, and in 55 patients C-7 was also involved. In some of these 55 patients, the muscles innervated by C-7 (the extensors of the arm, wrist, and fingers) had recovered partially or totally by the time of surgery. In 14 patients the hand was also partially affected, and 6 flail arms were seen. The right arm was affected in 52 and the left in 41 patients. Two patients had bilateral lesions, and in both cases the left side recovered spontaneously and the right side required an operation. The mean age at surgery was 175 days (range 90–299 days).

Surgical Procedure and Assessment of Lesion Severity

The operative procedure, performed with the patient in a state of general anesthesia without the use of muscle blocking agents, consisted of exposing the supraclavicular part of the brachial plexus in the lateral neck triangle through a straight incision parallel to the clavicle. In all cases the spinal nerves at C-5 and C-6 and the superior trunk were exposed. The C-7 spinal nerve was exposed when affected or used as a control for NAP and CMAP recordings when exposing this nerve would not result in additional surgical complications. Depending on the extent of injury, the infraclavicular part of the brachial plexus was also exposed. Direct electrical stimulation was performed with a bipolar nerve stimulator (GN 15; Aesculap) with stimuli of 5–10 mA.

The severity of the lesion of each exposed spinal nerve was classified as avulsion, partial avulsion, neurotmesis, intraforaminal neurotmesis, axonotmesis, or normal, based on CT myelography results, intraoperative morphological characteristics, direct nerve stimulation, and frozen-section examination. The surgeon was blinded to the results of NAP and CMAP recordings during this evaluation.

Avulsed Nerves

A spinal nerve root was considered avulsed if the nerve at the intraforaminal and juxtaforaminal level exhibited root filaments, the dorsal root ganglion was visible, neuroma formation was absent, and there were no muscle contractions in response to direct stimulation. These intraoperative findings corresponded in general with root filament discontinuity noted on CT myelography. Partial avulsion was diagnosed when the nerve appeared normal at the intraforaminal level, but neuroma formation was absent and direct stimulation resulted in only minimal muscle contractions. Computed tomography myelography demonstrated a clear asymmetry compared with the healthy side and/or a pseudomeningocele.

Neurotmetic Nerves

A spinal nerve was considered neurotmetic when the following features were present: a normal appearance at the intraforaminal level, a clear increase of the cross-sectional diameter at the juxtaforaminal level, abundant epineurial fibrosis, loss of fascicular continuity, increased consistency, and increased length of the nerve elements with concomitant distal displacement of the trunk divisions. On direct electrical stimulation of the spinal nerve proximal to the neuroma, contractions of related muscles were not strong enough to move the limb. Neurotmetic nerve elements were resected stepwise until the proximal stump showed normal fascicular architecture, which was confirmed by frozen-section examination.15 If frozen-section examination of the most proximal section of the nerve root in the foramen showed abundant neuroma formation, the nerve lesion was considered an intraforaminal neurotmesis. In these cases the proximal stump was not considered a suitable outlet for nerve grafting.

Axonotmetic Nerves

A spinal nerve was considered axonotmetic when neurolysis revealed no substantial increase of the cross-sectional diameter, only limited epineurial fibrosis, and intact fascicular continuity. Furthermore, muscle contractions induced limb movements (Medical Research Council score ≥ 3) on stimulation of related nerve elements. Axonotmetic nerves were left in situ because spontaneous nerve regeneration was apparently in process, although not yet clinically detectable. The infants with axonotmetic lesions were followed up for at least 2 years in the outpatient clinic. At 6-month intervals the neurological recovery of axonotmetic nerve elements was scored as follows: abduction for the PDST (range of motion), elbow flexion for the ADST, and extension of the elbow, wrist, and fingers for C-7 (Medical Research Council scale).

Normal Nerves

A spinal nerve was considered normal when preoperative neurological examination showed good function, in addition to a normal appearance during surgical exploration and strong contractions of related muscles after direct electrical stimulation.

Nerve Action Potential and CMAP Recording

Recordings were obtained in the supraclavicular and retroclavicular areas (Fig. 1). Care was taken to manipulate the nerve elements as little as possible prior to recording. The nerve elements were exposed as far proximally and distally as possible to provide maximum recording distance. Proximally, the lateral anterior scalene muscle fibers are routinely resected to allow sufficient exposure of the spinal nerves up to the intraforaminal level. Distally, dissection of the divisions of the trunks is routinely performed up to the infraclavicular level. This is executed by pulling the clavicle upwards. In this fashion recording distance was usually 3–4 cm. The nerve elements under study were electrically isolated by dissecting them 360° free, and isolating them from the surrounding tissue with a dry surgical patty. Electrical stimuli were administered and recorded with a J-shaped isolated bipolar forceps. Care was taken to ensure that both arms of the forceps touched only the nerve under analysis, and not the surrounding tissues or nerve elements.

Fig. 1.
Fig. 1.

Schematic drawing of the superior and middle parts of the brachial plexus. The different shades of gray represent the predominant innervation. The NAPs were stimulated (^) on C-5, C-6, and C-7, and recorded (^) on the suprascapular nerve (ss), the anterior division of the superior trunk (adst), the posterior division of the superior trunk (pdst), and the middle trunk (mt). Biceps and triceps CMAPs were stimulated in C-5, C-6, C-7, ADST, and PDST, and recorded in the biceps and triceps muscles. Additional nerve elements in the drawing: superior trunk (st), lateral cord (lc), posterior cord (pc), lateral part of the median nerve (lpmn), axillary nerve (ax), radial nerve (rad), and musculocutaneous nerve (mc).

Nerve action potential recordings were performed first. The proximal spinal nerve was stimulated with increasing amperage (stimulus 0.05 msec, 2 Hz) until a maximum amplitude was obtained at the most distally exposed position of the nerve element. The amperage of the supramaximal stimulus ranged from 2.5 to 9.8 mA. Nerve action potentials were recorded for several “trajectories,” which here refers to a specific combination of stimulation site and recording site. The NAP trajectories were recorded in the following order (stimulation site/recording site): C-5/suprascapular nerve, C-5/PDST, C-5/ADST, C-6/suprascapular nerve, C-6/PDST, C-6/ADST, and C-7/middle trunk. Usually ~ 8 stimuli were delivered to obtain a reproducible NAP recording.

Subsequently CMAP recordings of specific nerve and muscle combinations (trajectories) were performed. Different nerve elements were consecutively stimulated (stimulus 0.05 msec, 1 Hz) with an isolated bipolar forceps. The CMAPs were recorded with standard Ag/AgCl electroencephalography surface electrodes on the belly of the muscle, with a reference electrode on the tendon of the muscle. The following CMAP trajectories were recorded (stimulation site/ recording site): C-5/biceps, C-5/triceps, C-6/biceps, C-6/triceps, C-7/biceps, C-7/triceps, ADST/biceps, ADST/triceps, PDST/biceps, and PDST/triceps. Stimulation continued until 4 reproducible recordings were obtained.

The recordings were acquired using a Viking NT apparatus (v 5.0, Nicolet Biomedical, Inc.). The recordings were amplified and displayed (NAP: timesweep 2 msec/ division, 20 μV/division, filter 150 Hz–1.5k Hz; CMAP: timesweep 5 msec/div, 200–500 μV/div depending on the amplitude, filter 30 Hz–1.5 kHz). The amplitude of the first negative wave was measured after identification of the baseline and the top of the wave (Fig. 2). When a waveform was absent, the amplitude was marked as zero.

Fig. 2.
Fig. 2.

Example of NAP (A) and CMAP (B) recordings. Note the different grid scales of both curves and the different sizes of the stimulation artifacts. The beginning of the NAP (+) overlaps partly with the stimulation artifact. – = top of the amplitude; | = end of the amplitude measurement.

Statistical Analysis

Commercially available software was used for statistical analysis (SPSS 11.0.1; SPSS, Inc.). A probability value < 0.05 was considered statistically significant. The paired Wilcoxon rank-sum test was used to compare NAP amplitudes of trajectory pairs and CMAP trajectory pairs in the same infant. These pairs were analyzed to test whether standard brachial plexus anatomy could be ascertained based on this data.

A qualitative analysis was then performed in which the presence or absence of an NAP or CMAP was considered a dichotomous variable. In this qualitative analysis the sensitivity and specificity of absent NAPs or CMAPs were calculated to predict whether the lesion had an unfavorable (neurotmesis and avulsion) or favorable prognosis (normal and axonotmesis).

Subsequently, a quantitative amplitude analysis was performed in which absent NAPs or CMAPs were assigned the value zero. The nonparametric Mann–Whitney U-test and the nonparametric Kruskal–Wallis test were used to analyze NAP and CMAP values in relation to the severity of the lesion. When only a limited number of measurements were available per diagnosis, statistical analysis was not considered relevant. Receiver operator curves were computed and the sensitivity and specificity were calculated to predict an unfavorable (neurotmesis and avulsion) or a favorable diagnosis (normal and axonotmesis) depending on the cutoff point of the amplitude of NAPs or CMAPs.

Linear regression was used to correlate age and the amplitude of NAPs or CMAPs (split per diagnosis group).

Results

The diagnosis for each nerve is provided in Table 1. Histological examination confirmed the presence of abundant neuroma formation in all samples that had been classified as neurotmesis. Infants with nerve elements that had been classified as axonotmesis underwent clinical follow-up, and all showed good function of elbow flexion and shoulder abduction.

TABLE 1

Summary of diagnoses by nerve root*

No. of Lesions
DiagnosisC-5C-6C-7
partial avulsion227
avulsion31421
intraforaminal neurotmesis––5––
neurotmesis74639
axonotmesis161139
normal––––19

* –– = not applicable.

† Not used in statistical analysis.

In total, 599 NAPs and 836 CMAPs were recorded (Table 2). In 97 NAPs (16%) and 132 CMAPs (16%) a waveform was absent. The recordings did not show a normal binomial distribution: skewness and kurtosis were positive (data not shown).

TABLE 2

Summary of NAP and CMAP recordings*

TrajectoryNo.No. of 0 Values (%)MinMaxMeanMedian
NAP(μV)
 C-5/SSN8911 (12.4)0824.058.0823.81
 C-5/ADST8820 (22.7)0280.232.5215.86
 C-5/PDST8911 (12.4)0490.240.0423.21
 C-6/SSN8520 (23.5)0678.543.7620.48
 C-6/ADST8919 (21.3)0201.130.3520.11
 C-6/PDST8911 (12.4)0539.043.9920.86
 C-7/C-7705 (7.1)0853.9128.0554.59
CMAP(mV)
 PDST/biceps8421 (25.0)01.000.1620.079
 ADST/biceps856 (7.1)03.530.9000.808
 PDST/triceps776 (7.8)02.710.4420.237
 ADST/triceps7718 (23.4)01.480.1610.078
 C-5/biceps9411 (11.7)02.180.4820.417
 C-6/biceps9513 (13.7)02.020.4300.352
 C-5/triceps8313 (15.7)00.660.1500.104
 C-6/triceps8518 (21.2)00.640.1470.105
 C-7/biceps8118 (22.2)02.300.2390.128
 C-7/triceps758 (10.7)04.330.6260.306

*SSN = suprascapular nerve.

A subset of 8 patients with the least severe lesion type (axonotmesis of both C-5 and C-6) was selected to explore whether the amplitudes of CMAPs and NAPs corresponded to known neuroanatomical connections. In each of these 8 patients a given CMAP trajectory was compared with another CMAP trajectory in the same patient with the paired Wilcoxon rank-sum test. The combinations of CMAP trajectories were chosen such that either the stimulation site was the same or the recording site was the same.

We found that the amplitude of the CMAP in the C-6/ biceps trajectory was higher than the amplitude of the CMAP in the C-7/biceps trajectory in all 8 patients (p = 0.01). Conforming to expectation, the CMAP C-5/biceps was larger than the C-7/biceps in 7 of 8 patients (p = 0.02). Amplitudes of CMAPs measured at the C-5/biceps and the C-6/biceps were equivalent (p = 0.67), as were those of the CMAP C-5/triceps and C-6/triceps (p = 0.92). The CMAP ADST/biceps was larger than the CMAP PDST/biceps in 8 of 8 patients (p = 0.01), and the CMAP ADST/biceps was bigger than the ADST/triceps in 7 of 8 patients (p = 0.07). The NAP trajectories were explored in a similar fashion, but no statistical differences were found between NAP trajectory pairs (data not shown).

Qualitative Analysis

Absence of NAPs or CMAPs was not distributed evenly between diagnosis groups (Table 3). The percentage of absent NAPs or CMAPs was higher when the diagnosis was neurotmesis or avulsion than in normal or axonotmetic cases. The sensitivity and specificity of an absent NAP or CMAP to predict an unfavorable lesion (neurotmesis and avulsion) were calculated. The specificity of this test ranged from 0.9 to 1.0, but the sensitivity was < 0.30 with 1 exception (Table 3).

TABLE 3

Summary of absent recordings grouped by diagnosis*

TrajectoryNormal (%)Axonotmesis (%)Neurotmesis (%)Avulsion (%)SpecificitySensitivityPPVNPV
NAP (μV)
 C-5/SSN0/16 (0)8/68 (12)2/3 (67)1.0000.1411.0000.208
 C-5/PDST0/16 (0)9/68 (13)2/3 (67)1.0000.1551.0000.211
 C-5/ADST1/16 (6)16/67 (24)3/3 (100)0.9380.2710.9500.227
 C-6/SSN1/11 (9)9/53 (17)6/14 (43)0.9090.2240.9380.161
 C-6/PDST0/11 (0)4/57 (7)3/14 (21)1.0000.0991.0000.147
 C-6/ADST0/11 (0)7/58 (12)8/13 (62)1.0000.2111.0000.164
 C-7/C-70/18 (0)0/33 (0)0/2 (0)4/12 (33)1.0000.5331.0000.879
 total0/18 (0)2/114 (2)53/373 (14)28/62 (45)0.9850.1950.9770.270
CMAP (mV)
 C-5/biceps1/16 (6)8/73 (11)1/3 (33)0.9380.1410.9230.170
 C-6/biceps0/11 (0)5/63 (8)5/14 (36)1.0000.0911.0000.155
 ADST/biceps0/11 (0)3/54 (6)2/14 (14)1.0000.1181.0000.155
 C-7/triceps0/17 (0)1/32 (3)2/8 (25)5/12 (42)0.9800.1820.8000.727
 total0/17 (0)2/70 (3)18/198 (9)12/43 (28)0.9770.1240.9380.287

* Fractions represent the number of flat recordings/total number of recordings. Abbreviations: NPV = negative predictive value; PPV = positive predictive value.

Quantitative Analysis of Recordings

In case of an absent NAP or CMAP, the amplitude was marked as zero. Nerve action potential and CMAP measurements were analyzed quantitatively in relation to diagnosis on group level. With the Kruskal–Wallis test, a statistically significant difference was found between the groups, as it was for most NAP and CMAP trajectories (Tables 4 and 5).

TABLE 4

Comparison of diagnosis groups in different NAP trajectories using the Kruskal–Wallis test for the hypothesis that the groups are different*

NAP TrajectoryDiagnosisdfp Value
C-5
 C-5/SSNneurotmesis & axonotmesis10.035
 C-5/ADSTneurotmesis & axonotmesis10.016
 C-5/PDSTneurotmesis & axonotmesis10.026
C-6
 C-6/SSNaxonotmesis, neurotmesis, & avulsion20.033
 C-6/ADSTaxonotmesis, neurotmesis, & avulsion20.001
 C-6/PDSTaxonotmesis, neurotmesis, & avulsion20.185
C-7
 C-7/C-7normal, axonotmesis, & avulsion20.002

* df = degrees of freedom.

TABLE 5

Comparison of diagnosis groups in different CMAP trajectories using the Kruskal–Wallis test for the hypothesis that the groups are different

CMAPDiagnosisdfp Value
C-5
 C-5/bicepsneurotmesis & axonotmesis10.040
 C-5/tricepsneurotmesis & axonotmesis10.861
 ADST/bicepsneurotmesis & axonotmesis10.077
C-6
 C-6/bicepsaxonotmesis, neurotmesis, & avulsion20.000
 C-6/tricepsaxonotmesis, neurotmesis, & avulsion20.021
 ADST/bicepsaxonotmesis, neurotmesis, & avulsion20.007
C-7
 C-7/bicepsaxonotmesis, normal, neurotmesis, & avulsion20.002
 C-7/tricepsaxonotmesis, normal, neurotmesis, & avulsion20.000

An absolute cutoff value to separate diagnosis groups on the basis of NAP or CMAP values could not be determined because of overlapping values. An example is provided to show the overlapping values of NAP trajectory C-6/ADST and CMAP trajectory C-6/biceps (Fig. 3). The ROC was calculated for different NAP and CMAP trajectories. The ROC for NAP trajectories was close to the diagonal reference line (the area under the curve was 0.60 for NAPs at C-5/PDST and 0.66 for NAPs at C-6/ADST, other data not shown). The ROCs for CMAPs proved statistically stronger. The ROC for the CMAP trajectory C-6/ biceps resulted in the biggest area under the curve (0.88), whereas NAPs at C-5/biceps resulted in an area under the curve of 0.72. The cut-off value with the highest likelihood ratio (5.73) for the CMAP trajectory C-6/biceps was determined at 0.70 mV, which resulted in a sensitivity of 0.82 and a specificity of 0.86. When the cutoff point was set at 0.75 mV, the sensitivity dropped to 0.55 (likelihood ratio 4.20).

Fig. 3.
Fig. 3.

Box plot graphs showing CMAP (upper) and NAP (lower) C-6/biceps recordings by C-6 root diagnosis. The whisker lines show 95% of values, the boxes show the 25th to the 75th percentile, and the bold lines show the medians.

Factors of Influence

A positive correlation was found between CMAP amplitude and the age at surgery. Linear regression showed that age had the largest effect on the CMAP trajectory C6-biceps. This effect was largest in cases of axonotmesis of C-6: the explained (p = 0.009) proportion of the variance was 55% and the slope of the regression line was 0.74 mV/day. A less strong positive correlation was found in case of avulsion of C-6, neurotmesis of C-6, and neurotmesis of C-5, and no correlation was found for the following diagnoses: axonotmesis of C-5, avulsion of C-7, axonotmesis of C-7, or normal C-7 (data not shown).

When calculating correlation for NAP and age at surgery, only in the diagnosis group neurotmesis of C-6 did the NAP trajectory C6-ADST reach a statistically significant level. However, the explained proportion of the variance was < 10% (data not shown).

Discussion

We analyzed the results of intraoperative NAP and CMAP recordings in 95 patients with OBPLs to assess the predictive values of these recordings in the diagnosis of axonotmesis, neurotmesis, avulsion, or normal spinal nerves. Theoretically, NAPs should be more informative than CMAPs during early surgical exploration because NAPs reflect the number of axons across the lesion, whereas CMAP amplitudes reflect the number of axons that have already reached the target muscle, which takes several weeks to months.12 We found that NAP and CMAP amplitudes both showed a wide numerical range. The NAPs had a larger range than the CMAPs, which is in accord with evoked recordings data in adults. Because of the larger and longer action potential of CMAPs compared with NAPs, they are theoretically less susceptible to disturbing factors.9

Statistically significant differences were found between diagnosis groups. For the individual patient, however, a clinically useful cutoff point for NAP and CMAP recordings for differentiating between avulsion, neurotmesis, axonotmesis, and normal nerve could not be found. The sensitivity for an absent NAP or CMAP was too low for clinical use. We conclude that intraoperative NAP and CMAP recordings do not add to the assessment of lesion severity of affected brachial plexus elements in infants with OBPLs.

The most important methodological drawback of our study is that a gold standard for assessment of the severity of the nerve lesion in OBPLs does not exist. In other words, the diagnosis may not have been correct. However, histological examination confirmed abundant neuroma formation in all samples of assumed neurotmetic lesions. Obviously, we did not obtain histological samples from the cases of axonotmesis to serve as a control. These children underwent clinical follow-up, and all demonstrated good spontaneous recovery of elbow flexion and shoulder abduction. We therefore believe that the overall conclusions of our study cannot be strongly influenced by falsely diagnosed lesions.

Another consequence of the absence of a gold standard is that the key features that eventually determined the severity of the lesion in this study (the clinical picture and the CT myelography and surgical findings) are the same factors that define the pretest probability of a severe lesion in the normal clinical decision process. In this aspect, the present study differs from the clinical decision process in which neurological examination, CT myelography, and surgical findings are known, and further investigations are expected to provide additional information.

The optimal methodology for assessing NAP and CMAP recordings would be to execute these recordings in all infants with OBPLs at a certain age regardless of the extent of the lesion, for instance at the age of 3 months. After recording, surgical reconstruction should not be performed, and all nerve elements should remain untouched to establish the natural course of regeneration. Functional outcome should then be assessed in each individual at the minimum age of 3 years, or even at an adult age. Optimally, repeated inspection of the affected nerves and their histological examination should be performed to finally classify the extent of the lesion of each involved nerve element. These findings would then serve as the gold standard for assessing correlations with the intraoperative NAP and CMAP recordings obtained years before. Of course, such a study would be ethically unacceptable. The standard that we have applied in this study is, in our opinion, currently the best we could achieve, despite its shortcomings.

A minor drawback of this study is that the EMG technician was not formally blinded to the surgical findings. Although the surgical findings were not explicitly discussed during the recordings, unintentional verbal or nonverbal communication may have influenced measurements. Although the surgeon was blinded to the results of the NAP and CMAP recordings, stimulation may have elicited muscle contractions or limb movements revealing the severity of the lesion. The amperages of the stimuli for NAP and CMAP recordings were of the same magnitude as of the amperage of the direct nerve stimulation that is part of our routine assessment. The final assessment of the severity of the lesion was performed after the NAP and CMAP recordings were finished, and the data were linked months after surgery. It is therefore not likely that recordings were biased by the applied methodology.

Anatomical Connections

Our first step for appraisal of intraoperative recordings of NAPs and CMAPs was an analysis of paired recordings from each patient to compare NAP and CMAP trajectories to known neuroanatomical connections. Preferably, only noninjured nerves should be analyzed for this purpose instead of nerves subsequent to a traction injury. In the present study, the nerve elements with the least severe lesions (such as axonotmesis of both C-5 and C-6) were included because, because of the composition of this surgical group, the number of measurements of noninjured nerve elements was limited.

The analysis of paired CMAP trajectories within each patient correlated well with known neuroanatomical connections. For instance, in all infants in this subset a higher CMAP was found from the trajectory C-6/biceps compared with C-7/biceps, which corresponds well to the fact that the biceps muscle is mainly innervated by the C-6 spinal nerve and not C-7. Similarly, a higher CMAP was found for the C-7/triceps trajectory compared with C-7/biceps. No statistical differences were found between paired NAP measurements.

Some of our recordings are more difficult to explain. For instance, an anatomical connection from the anterior division of the superior trunk to the triceps muscle or from the posterior division of the superior trunk to the biceps muscle does not exist. In the present study, however, these recordings resulted in a measurable CMAP in 80% of recordings (Table 2). A number of explanations are possible. First, electrical stimulation of the PDST resulted in unwanted costimulation of the ADST (or vice versa) despite the extensive precautions that were applied in the surgical field to avoid such costimulation. Second, stimulation of the PDST resulted in retrograde stimulation of axons in the neuroma-in-continuity of the superior trunk. In the neuroma, as a result of cross-excitation (or “ephaptic stimulation”), axons to the ADST were stimulated as well, thereby resulting in the recording of a CMAP in the biceps.20 Third, branching in the superior trunk neuroma may have led to branches of the same sprouting axon in both the ADST and the PDST. Stimulation of 1 branch may then result in costimulation of the other branch. Fourth, coregistration may have occurred. This may have resulted from suboptimal positioning of the recording electrode in the surgical field or recording of a distant action potential in a nearby nerve or muscle.

Qualitative Analysis of NAP and CMAP Recordings

Absence of NAPs and CMAPs was not evenly distributed among different diagnosis groups, but showed a relation to severity of the nerve lesion. In adults the presence of NAPs is indicative of the presence of 3000–4000 myelinated axons, which will eventually lead to spontaneous recovery.10 The absence of NAPs predicts that spontaneous recovery will not occur, so the involved nerve element should be resected and repaired.10 In the present series of infants with OBPLs, absent NAPs were often found in cases of severe nerve lesion (neurotmesis or avulsion) just as they are in adults. However, present NAPs were recorded also in neurotmetic or avulsed roots. For instance, analysis of the trajectory C-6/ADST showed that in 51 of 58 cases of neurotmesis and 5 of 13 cases of root avulsion, NAPs were detected. In other words, the presence of an NAP or CMAP results in an overly optimistic judgment of the severity of a nerve lesion in infants with OBPLs. The absence of an NAP or CMAP is not a useful test in clinical practice because of its low sensitivity (and thus low negative predictive value).

The authors of only 1 paper have previously reported the outcome in patients in whom NAP recordings were performed.13 Their study was conducted in 10 nonconsecutive patients, and in 5 of these infants spontaneous recovery did not take place despite the presence of NAP recordings. Clarke et al.3–5 used direct nerve stimulation and observed muscle contractions. They found that neurolysis of conducting neuromas-in-continuity resulted in a less favorable outcome than resection and grafting of such lesions. In other words, other studies have confirmed that electrical conduction in a neuroma does not imply good spontaneous recovery in patients with OBPLs.

There are various factors that may affect NAP or CMAP recordings and are unique to OBPLs (compared with traumatic brachial plexus injury in adults), as we discussed in a review of the limited usefulness of EMG examination in infants with OBPL.25 These factors include differences in trauma mechanism, a different potential for nerve regeneration in the infant population, an overly pessimistic clinical examination, overestimation of EMG recruitment because of small muscle fibers, persistent fetal innervation, plasticity of the spinal cord, developmental apraxia, and misdirection, in which axons reach inappropriate muscles.25 Especially the occurrence of abnormal branching (misrouting) may be important also in the NAP and CMAP evaluation. The results of experiments in animals have shown that misrouting occurs more often in young than in adult animals.2 Another influential factor may be axonal cross-excitation. It could be hypothesized, that stimulation of nerve fibers from C-5 may lead to crossexcitation of nerve fibers from C-6 in the neuroma.20 The resulting amplitude in the ADST (or the resulting CMAP in the biceps) thus overestimates the number of axons in the C-5/ADST trajectory in such a case, as the recordings might be amplified because of axons in the C-6/ADST trajectory.

In general, NAPs and CMAPs may be influenced by a number of technical failures, which can be divided into factors in the surgical field (poor electrode contact, inadequate nerve isolation, temperature, too little space between electrodes, stimulation and recording electrodes distal to intended lesion), equipment factors (cracked wires), and factors related to the interpretation and recording (scale set too high or low, sweep set too fast or slow, stimulation voltage too low).24 In our setup, much care was taken to avoid such technical failures.

Another technical factor is the presence of a stimulation artifact in many NAP recordings. We could identify a point where the NAP deflects from the stimulation artifact in most recordings, but the stimulation artifact may still have influenced NAP amplitude measurements. As such, this may affect the quantitative analysis of amplitudes, but on a qualitative level (presence or absence) the stimulation artifact is not important. Especially the presence (or absence) of an NAP is claimed to be the main criterion to evaluate nerve lesions in continuity according to Kline and coworkers in experiments in adult patients and in animals.10–12,24

Quantitative Analysis of NAP and CMAP Recordings

In the quantitative analysis, the absent NAPs and CMAPs were analyzed with their amplitudes set at zero. Analysis of the diagnosis groups showed that CMAP measurements correlated with the diagnosis. A decreasing amplitude was observed in the following order: axonotmesis > neurotmesis > avulsion. In the C-7 spinal nerve, the trend was for the CMAP amplitude in normal cases to be greater than that in anoxotmesis, but this finding did not reach statistical significance (p = 0.07). These findings correspond to the pathophysiological assumption that a nerve element contains a decreasing number of fibers when the lesion is more severe. Nerve action potential measurements yielded similar findings, but were statistically less robust than the CMAP measurements.

In the individual patient, however, NAP or CMAP measurements did not lead to a definite diagnosis because absolute cutoff values of the CMAP and NAP data per diagnosis could not be determined. The ROC for the CMAP trajectory C-6/biceps resulted in a sensitivity of 0.82 and a specificity of 0.86. These were the best specifications of the NAP and CMAP recordings. Such a diagnostic test seems quite acceptable on the basis of these statistics. However, a few serious shortcomings should be mentioned.

First, the question should be posed whether a sensitivity of 0.8 is sufficient to perform an irreversible surgical intervention. In other words, a 20% chance of mistaken resection of a recovering nerve element remains. Second, as previously stated, the additional value of NAP/CMAP recording when the clinical picture, the CT myelography, and the surgical findings are known, cannot be determined because of the study setup. Third, when the cutoff point was changed from 0.70 to 0.75 mV, the sensitivity dropped to 0.55. Probably the lack of a normal data distribution caused such large differences in sensitivity with a relatively small change in cutoff value. Such a small difference of the cut-off point may be difficult to measure reliably in clinical practice, which limits the clinical usefulness of CMAP recordings.

Effect of Age

A positive correlation between age and CMAP amplitude was found, which explained up to 55% of the variance of the CMAP. The influence of age on the amplitude of the CMAP differs between the diagnosis groups. This influence is greatest in cases of axonotmesis, which can be explained by a higher number of outgrowing axons that will reach the target muscle in time because of spontaneous recovery. Another factor that might partially explain the age effect is that muscle fiber size increases with age, which in turn might lead to an increase of the amplitude of the measured CMAP.

Conclusions

We found a larger variance in the NAP than in CMAP recordings, and normal brachial plexus neuroanatomy correlated better with CMAPs than with NAPs.

Statistically significant differences were found between diagnosis groups. In the individual patient, however, neither NAP nor CMAP recordings could differentiate between the diagnosis of avulsion, axonotmesis, and neurotmesis. Therefore, intraoperative NAP and CMAP recordings do not add to the intraoperative assessment of the severity of the lesion and surgical decision making in the individual infant with an OBPL.

In view of the applied methodology, it seems unlikely that technical refinements may optimize NAP and CMAP recordings. Other solutions should be sought after to facilitate intraoperative diagnostics and improve treatment in this patient population. This can only happen with increased understanding of the complex pathophysiology of OBPLs.

Disclaimer

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

References

Article Information

Address correspondence to: Willem Pondaag, M.D., Department of Neurosurgery, Leiden University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands. email: w.pondaag@lumc.nl.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Schematic drawing of the superior and middle parts of the brachial plexus. The different shades of gray represent the predominant innervation. The NAPs were stimulated (^) on C-5, C-6, and C-7, and recorded (^) on the suprascapular nerve (ss), the anterior division of the superior trunk (adst), the posterior division of the superior trunk (pdst), and the middle trunk (mt). Biceps and triceps CMAPs were stimulated in C-5, C-6, C-7, ADST, and PDST, and recorded in the biceps and triceps muscles. Additional nerve elements in the drawing: superior trunk (st), lateral cord (lc), posterior cord (pc), lateral part of the median nerve (lpmn), axillary nerve (ax), radial nerve (rad), and musculocutaneous nerve (mc).

  • View in gallery

    Example of NAP (A) and CMAP (B) recordings. Note the different grid scales of both curves and the different sizes of the stimulation artifacts. The beginning of the NAP (+) overlaps partly with the stimulation artifact. – = top of the amplitude; | = end of the amplitude measurement.

  • View in gallery

    Box plot graphs showing CMAP (upper) and NAP (lower) C-6/biceps recordings by C-6 root diagnosis. The whisker lines show 95% of values, the boxes show the 25th to the 75th percentile, and the bold lines show the medians.

References

1

Bager B: Perinatally acquired brachial plexus palsy—a persisting challenge. Acta Paediatr 86:121412191997

2

Brown MCHardman VJ: A reassessment of the accuracy of reinnervation by motoneurons following crushing or freezing of the sciatic or lumbar spinal nerves of rats. Brain 110:6957051987

3

Capek LClarke HMCurtis CG: Neuroma-in-continuity resection: early outcome in obstetrical brachial plexus palsy. Plast Reconstr Surg 102:155515621998

4

Clarke HMAl Qattan MMCurtis CGZuker RM: Obstetrical brachial plexus palsy: results following neurolysis of conducting neuromas-in-continuity. Plast Reconstr Surg 97:974 9821996

5

Clarke HMCurtis CG: An approach to obstetrical brachial plexus injuries. Hand Clin 11:5635801995

6

Dawodu ASankaran-Kutty MRajan TV: Risk factors and prognosis for brachial plexus injury and clavicular fracture in neonates: a prospective analysis from the United Arab Emirates. Ann Trop Paediatr 17:1952001997

7

DiTaranto PCampagna LPrice AEGrossman JA: Outcome following nonoperative treatment of brachial plexus birth injuries. J Child Neurol 19:87902004

8

Gilbert ATassin JL: [Surgical repair of the brachial plexus in obstetric paralysis.]. Chirurgie 110:70751984. (Fr)

9

Kimura JMachida MIshida TYamada TRodnitzky RLKudo Y: Relation between size of compound sensory or muscle action potentials, and length of nerve segment. Neurology 36:6476521986

10

Kline DG: Nerve surgery as it is now and as it may be. Neurosurgery 46:128512932000

11

Kline DGDeJonge BR: Evoked potentials to evaluate peripheral nerve injuries. Surg Gynecol Obstet 127:123912481968

12

Kline DGHackett ERMay PR: Evaluation of nerve injuries by evoked potentials and electromyography. J Neurosurg 31:1281361969

13

König RWAntoniadis GBörm WRichter HPKretschmer T: Role of intraoperative neurophysiology in primary surgery for obstetrical brachial plexus palsy (OBPP). Childs Nerv Syst 22:7107152006

14

Laurent JPLee RShenaq SParke JTSolis ISKowalik L: Neurosurgical correction of upper brachial plexus birth injuries. J Neurosurg 79:1972031993

15

Malessy MJvan Duinen SGFeirabend HKThomeer RT: Correlation between histopathological findings in C-5 and C-6 nerve stumps and motor recovery following nerve grafting for repair of brachial plexus injury. J Neurosurg 91:6366441999

16

Metaizeau JPGayet CPlenat F: [Brachial plexus birth injuries. An experimental study.]. Chir Pediatr 20:1591631979. (Fr)

17

Pondaag Wde Boer Rvan Wijlen-Hempel MSHofstede-Buitenhuis SMMalessy MJ: External rotation as a result of suprascapular nerve neurotization in obstetric brachial plexus lesions. Neurosurgery 57:5305372005

18

Pondaag WMalessy MJ: Recovery of hand function following nerve grafting and transfer in obstetric brachial plexus lesions. J Neurosurg 105:1 Suppl33402006

19

Pondaag WMalessy MJvan Dijk JGThomeer RT: Natural history of obstetric brachial plexus palsy: a systematic review. Dev Med Child Neurol 46:1381442004

20

Seltzer ZDevor M: Ephaptic transmission in chronically damaged peripheral nerves. Neurology 29:106110641979

21

Shenaq SMKim JYArmenta AHNath RKCheng EJedrysiak A: The surgical treatment of obstetric brachial plexus palsy. Plast Reconstr Surg 113:54E67E2004

22

Sherburn EWKaplan SSKaufman BANoetzel MJPark TS: Outcome of surgically treated birth-related brachial plexus injuries in twenty cases. Pediatr Neurosurg 27:19271997

23

Sjoberg IErichs KBjerre I: Cause and effect of obstetric (neonatal) brachial plexus palsy. Acta Paediatr Scand 77:357 3641988

24

Tiel RLHappel LT JrKline DG: Nerve action potential recording method and equipment. Neurosurgery 39:1031081996

25

van Dijk JGPondaag WMalessy MJ: Obstetric lesions of the brachial plexus. Muscle Nerve 24:145114612001

26

Waters PM: Comparison of the natural history, the outcome of microsurgical repair, and the outcome of operative reconstruction in brachial plexus birth palsy. J Bone Joint Surg Am 81:6496591999

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 3 3 3
Full Text Views 53 53 18
PDF Downloads 59 59 34
EPUB Downloads 0 0 0

PubMed

Google Scholar