Intra-operative neurophysiological prediction of upper trunk recovery in obstetric brachial plexus palsy with neuroma in continuity

Transcription

1 CHILDREN S ORTHOPAEDICS Intra-operative neurophysiological prediction of upper trunk recovery in obstetric brachial plexus palsy with neuroma in continuity K. F. Chin, V. P. Misra, G. M. Sicuri, M. Fox, M. Sinisi From Royal National Orthopaedic Hospital, Stanmore, United Kingdom We investigated the predictive value of intra-operative neurophysiological investigations in obstetric brachial plexus injuries. Between January 2005 and June 2011 a total of 32 infants of 206 referred to our unit underwent exploration of the plexus, including neurolysis. The findings from intra-operative electromyography, sensory evoked potentials across the lesion and gross muscular response to stimulation were evaluated. A total of 22 infants underwent neurolysis alone and ten had microsurgical reconstruction. Of the former, one was lost to follow-up, one had glenoplasty and three had subsequent nerve reconstructions. Of the remaining 17 infants with neurolysis, 13 (76%) achieved a modified Mallet score > 13 at a mean age of 3.5 years (0.75 to 6.25). Subluxation or dislocation of the shoulder is a major confounding factor. The positive predictive value and sensitivity of the intra-operative EMG for C5 were 100% and 85.7%, respectively, in infants without concurrent shoulder pathology. The positive and negative predictive values, sensitivity and specificity of the three investigations combined were 77%, 100%, 100% and 57%, respectively. In all, 20 infants underwent neurolysis alone for C6 and three had reconstruction. All of the former and one of the latter achieved biceps function of Raimondi grade 5. The positive and negative predictive values, sensitivity and specificity of electromyography for C6 were 65%, 71%, 87% and 42%, respectively. Our method is effective in evaluating the prognosis of C5 lesion. Neurolysis is preferred for C6 lesions. Cite this article: Bone Joint J 2013;95-B: K. F. Chin, MBChB, BMedSc, MRCS, Specialist Registrar G. M. Sicuri, MD, Visiting Fellow M. Fox, FRCS(Tr & Orth), Consultant Surgeon M. Sinisi, MD, Consultant Surgeon Royal National Orthopaedic Hospital, Peripheral Nerve Unit, Brockley Hill, Stanmore, Middlesex HA7 4LP, UK. V. P. Misra, MD, FRCP, Consultant Neurophysiologist Hammersmith Hospital, Peripheral Nerve Unit, Du Cane Road, London W12 0HS, UK. Correspondence should be sent to Mr M. Sinisi; marcosinisi@aol.it 2013 The British Editorial Society of Bone & Joint Surgery doi: / x.95b $2.00 Bone Joint J 2013;95-B: Received 19 September 2012; Accepted after revision 11 January 2013 Infants with obstetric brachial plexus injury (OBPP) who show no clinical recovery within the first few months may have a permanent deformity. 1-4 Neurophysiological investigations, such as mixed nerve action potential recording and electromyography (EMG), can help to determine the outcome 5 and those with favourable results can be managed conservatively. The accuracy of neurophysiological investigations in predicting recovery has been reported by Bisinella et al. 5 Their prediction of recovery of a C6 lesion was accurate in 92% of infants compared with 78% for C5, owing to the inability to record nerve action potentials for C5 and the high incidence of secondary shoulder pathology. 5 However, they concentrated on infants with favourable neurophysiological results who had been managed conservatively. Those with unfavourable results and failure of clinical recovery were operated upon but the accuracy of EMG in this group has not been reported. Also, the favourable natural history of recovery of biceps function 1 may have positively biased the accuracy of neurophysiological prediction reported by Bisinella et al. 5 In our practice, in contrast to the growing numbers of reports advocating early nerve grafting irrespective of conductivity in the neuroma, 6 we prefer combining clinical assessment and neurophysiological investigations to improve the accuracy of the evaluation of the neural lesion in neuroma-in-continuity, and thereby help in surgical decision making. Infants showing other signs of favourable recovery intra-operatively are managed with neurolysis alone, whereas those with unfavourable signs undergo either microsurgical repair or reconstruction. In this study we investigated the prediction of C5 and C6 recovery in a group of infants with OBPP having poor recovery and unfavourable neurophysiology results who underwent exploration with the intention of microsurgical intervention. Patients and Methods Infants with OBPP injury referred to the Peripheral Nerve Injury Unit of the Royal National Orthopaedic Hospital, London, are graded using the Narakas classification 7 : group 1, lesion of C5 and C6; group 2, lesion VOL. 95-B, No. 5, MAY

2 700 K. F. CHIN, V. P. MISRA, G. M. SICURI, M. FOX, M. SINISI Table I. The modified Mallet score 8 for evaluation of C5 and the Gilbert and Raimondi score 9 for evaluation of C6 Score Action Mallet Abduction < to 90 > 90 External rotation 0 < 20 > 20 Hand to head Impossible Difficult Easy Hand to back Impossible To S1 To T12 Hand to mouth Trumpet sign Partial trumpet sign < 40 abduction Gilbert & Raimondi Flexion No contraction Partial contraction Full contraction Extension None Weak Good Lack of extension > to 50 0 to 30 Table II. The motor response to electrical stimulation. Type A and Type B (favourable) findings were interpreted as favourable responses Type A Electrophysiological findings Amplitude > 1 mv and non-dispersed (50% drop of amplitude between distal and proximal stimulation, if present, is defined as conduction block) Type B Favourable Unfavourable Amplitude of > 500 μv or > 250 μv and a non-dispersed response Amplitude < 250 μv or < 500 μv but dispersed Type C Absent or minimal response (< 100 μv) of C5, C6 and C7; group 3, total lesions without Horner s syndrome; and group 4, total palsy and Horner s syndrome. The outcome of the treatment for C5 lesions were evaluated using the modified Mallet score 8 and the outcome of C6 lesions were evaluated using the Gilbert and Raimondi score. 9 The modified Mallet score assesses the ability of the infant to abduct and externally rotate the shoulder, whereas the Gilbert and Raimondi score assesses active flexion and extension of the elbow (Table I). Those with poor recovery undergo mixed nerve action potential (NAP) recording and needle EMG. 10 Those with unfavourable results are re-assessed after two months with, if necessary, further neurophysiological studies. If the adverse clinical features and neurophysiological investigations persist, the brachial plexus is explored. At two to three months we also explore babies born by the breech delivery, or those with Narakas Group 4 lesions, many of whom have a complete rupture or avulsion of spinal nerves. 11 In 2005, we started to use intra-operative motor nerve stimulation and EMG recording as an aid to decision making. 12 A total of 206 infants with OBPP were referred between January 2005 and June Of these, 32 underwent exploration of the brachial plexus. There were ten male and 22 female infants, with a mean age at operation of six months (3 to 15). Their mean age at follow-up was 40 months (8 to 75). During the exploration, we first relocated the shoulder if there was subluxation or dislocation and carried out neurolysis of the brachial plexus. The infants did not receive intra-operative muscle relaxants. Needle EMG recordings looking for spontaneous activity such as fibrillations and motor unit action potentials from direct motor nerve stimulation were made, together with sensory evoked potential (SEP) at C2. The nerve action potential for C5 and gross muscular response to stimulation were also determined. Following neurolysis, the response to motor stimulation was recorded with a concentric needle from the deltoid and supraspinatus for C5 and from biceps for C6. The action potential from each muscle was recorded at two sites and the best response taken. EMG sampling was undertaken using a concentric needle electrode inserted into each muscle at three different sites. The overall motor unit morphology was assessed qualitatively by passive movement and electrical nerve stimulation. For example, absence of spontaneous activity and normal morphology of motor units would suggest conduction block, which would be favourable. Profuse spontaneous activity and few motor units with polyphasic morphology would suggest unfavourable axonal damage with limited reinnervation. The final decision on whether the neurophysiology was favourable overall was made on the basis of the electrical nerve stimulation and EMG findings together (Table II). Infants with favourable features underwent neurolysis alone, and those with unfavourable features had nerve reconstruction/grafting. The intra-operative recordings were analysed by a neurophysiologist (VPM) who was blinded to the outcome. The THE BONE & JOINT JOURNAL

3 INTRA-OPERATIVE NEUROPHYSIOLOGICAL PREDICTION OF UPPER TRUNK RECOVERY IN OBSTETRIC BRACHIAL PLEXUS PALSY 701 Table III. The proposed correlation between the severity of C5 and C6 lesions and electromyography, with predicted outcomes, using Mallet 8 and Gilbert and Raimondi 9 scores Expected outcome Lesion type C5 C6 Type A Normal or nearly normal shoulder with abduction Complete and powerful elbow flexion (Gilbert and Raimondi 5) and forward flexion > 150 (Gilbert 5+, Mallet 15) Type B favourable Abduction and forward flexion > 90 (Gilbert 3+ or Complete and powerful elbow flexion (Gilbert and Raimondi 5) better, Mallet 13) Type B unfavourable Abduction and forward flexion > 45 (Gilbert 2+; Weak elbow flexion or deficit of supination (Gilbert and Raimondi 4) Mallet 11 and 12) Type C Mallet < 11 and Gilbert < 2 Gilbert and Raimondi 3 Table IV. Electromyography (EMG) prediction and its accuracy for C5 in each Narakas group. Figures in parentheses are the number of infants whose EMG matched the predicted outcome Intra-operative C5 EMG prediction Narakas group B favourable B unfavourable C Total 2 5 (4) 3 (0) 1 (1) (3) 2 (0) (1) Match 8/11 0/5 1/1 9/17 = 53% modified Mallet scores 8 and outcomes were also compared with the intra-operative neurophysiological findings using the expectation reported by Bisinella et al 5 (Table III). Our decision to perform neurolysis alone depended on the presence of a combination of favourable prognostic factors. However, we followed these infants closely and were prepared for reconstruction if subsequent recovery was unsatisfactory. Statistical analysis. The results were analysed using the IBM SPSS v21 (SPSS Inc., Chicago, Illinois). Categorical comparisons were carried out using chi-squared (groups achieving Mallet score 13 and < 13) and Wilcoxon s signed rank test (difference between pre- and intra-operative EMG predictions). Correlation assessments were carried out with Spearman s rho test (Mallet score and prognostic factors). The threshold for statistical significance was set at p < Sensitivity, specificity, positive predictive and negative predictive values of the EMG test for each nerve root were also carried out. Results Out of 32 infants, 22 had neurolysis alone as the first operation. One infant (Narakas group 4) had recurrent dislocation of the shoulder and underwent glenoplasty. As the outcome cannot yet be confirmed, this infant was excluded from the analysis of C5 lesions. One infant was lost to follow-up and excluded. In ten infants neurotisation of the spinal accessory to the suprascapular nerve was undertaken; two of these also had nerve grafts from C6 to the upper trunk and one had neurotisation of the spinal accessory nerve to the anterior division of the upper trunk. Of the 20 children in the neurolysis group, 13 achieved a Mallet score 13 compared with three in the reconstruction group. Although the difference is significant (chi-squared test, p = 0.048), this result does not indicate that neurolysis is superior because C5 nerves with more severe and unfavourable injuries would receive primary reconstruction. Of the 32 infants, 13 needed relocation of the shoulder as part of the first operation, of whom ten had neurolysis alone. Two had recurrent subluxation while a further three who had neurolysis but did not have subluxation at the primary operation developed late subluxation and dislocation. None of the infants from the reconstruction group developed late shoulder pathology (chi-squared test, p = 0.001). Three infants with subluxation or dislocation of the shoulder who had neurolysis as primary operation needed subsequent spinal accessory to supraclavicular nerve transfer. One was Narakas type 3 and two were type 4 and, according to pre-operative EMG, two were type B unfavourable and one was type C. However, the intra-operative EMG recruitment pattern improved and post-operatively one EMG was type B unfavourable and the other two were type B favourable, with a good motor response to nerve stimulation for all three. The SEP was diminished in one infant and normal in the other two. We felt that subluxation or dislocation of the shoulder impeded recovery because there was no stable fulcrum for abduction. Therefore, we followed up these three infants and undertook neurotisation of the spinal accessory to suprascapular nerve, as recovery was still not detected after two to three months. It is unclear whether this observation period was too brief, especially as full recovery of C6 nerve and biceps VOL. 95-B, No. 5, MAY 2013

4 702 K. F. CHIN, V. P. MISRA, G. M. SICURI, M. FOX, M. SINISI Table V. The predictive value of EMG Case positive Case negative Prediction positive 8 3 Sensitivity: 73% Prediction negative 5 1 Specificity: 17% Positive prediction value 62% Negative prediction value 25% Table VI. The outcome in infants with and without dislocation of the shoulder by Mallet score 8 Mallet score < 13 Mallet > 13 Total Without shoulder dislocation With shoulder dislocation Total Table VII. Electromyography (EMG) prediction and its accuracy for C5 in each Narakas group 7 in infants without subluxation or dislocation of the shoulder. Figures in parentheses are the numbers of infants whose EMG matched the predicted outcome Intra-operative C5 EMG prediction without dislocation Narakas group B favourable B unfavourable or C Total 2 2 (2) 1 (0) (3) (1) 1 Match 6/6 0/1 85.7% (6/7) function was observed in two of the infants without neurotisation or grafting of C6. The other had damage to the biceps at birth and needed a subsequent free muscle transfer. Overall, we considered the decision to neurotise to be potentially a confounding rather than determining factor for the accuracy of the prediction. These three infants were therefore also excluded, leaving 17 infants for the analysis and C5 nerve root EMG prediction presented below. The mean Mallet score for these 17 infants was 12.5 (8 to 15). The mean abduction, forward flexion, external rotation and internal rotation were 98 (30 to 160 ), 100 (40 to 150 ), 43 (0 to 90 ) and 70 (0 to 90 ), respectively. The pre- and intra-operative EMGs were significantly different (Wilcoxon s signed ranks test, p = 0.035), but there was no significant correlation between the Narakas type and Mallet score (Spearman s rho, p = 0.714). Tables IV and V illustrate the intra-operative EMG prediction and its accuracy for C5 in each Narakas group. As there was only one patient with a type C prediction, we combined types B unfavourable and C. The positive predicted case is defined as a favourable lesion with a Mallet score 13. The prediction value of the EMG is shown in Table V. The intra-operative EMG correctly identified eight infants with a type B favourable outcome and 8/11 (73% sensitivity) achieved Mallet > 13. The intra-operative EMG identified a type B unfavourable/c lesion in six infants, but only one had Mallet < 13. The positive and negative predictive values were 62% and 25%, respectively. As also shown by Bisinella et al 5 we found that dislocation of the shoulder significantly influenced the outcome (Table VI). In the neurolysis group, all seven patients without dislocation had a Mallet outcome > 13, whereas only six of ten patients with dislocation achieved this (chisquared test, p = 0.031). As shoulder subluxation or dislocation is a major confounding factor, we analysed the prediction in the infants without it and found that in six of seven cases the prediction matched the outcome (Table VII). The positive predictive value (PPV) was 85.7% (6 of 7), indicating that a positive result predicted by the intra-operative EMG is reasonably accurate if shoulder subluxation/dislocation is absent. We assessed whether we had made the correct overall clinical decision for C5 lesions. Of 21 infants who had neurolysis, three needed revision nerve transfer/grafting, although this may have been expected as the other intraoperative findings, such as SEP, muscle response and macroscopic appearance of the nerve, were ambiguous Another infant needed subsequent relocation of the shoulder and glenoplasty. Of the remaining 17, four had a Mallet score < 13. The results of the remaining 13 (76%) were as anticipated. THE BONE & JOINT JOURNAL

5 INTRA-OPERATIVE NEUROPHYSIOLOGICAL PREDICTION OF UPPER TRUNK RECOVERY IN OBSTETRIC BRACHIAL PLEXUS PALSY 703 Table VIII. The number of infants achieving Mallet score 8 13 with the presence of a favourable score Mallet score < 13 Mallet 13 Total Total favourable factors score < Table IX. The accuracy of the scoring system Case positive Case negative Prediction positive 10 0 Sensitivity: 100% (10/10) Prediction negative 3 4 Specificity: 57% (4/7) Positive prediction value 77% (10/13) Negative prediction value 100% (4/4) Table X. Electromyography (EMG) prediction for C6 in each Narakas group 7 and its accuracy. Figures in parentheses are the number of infants whose EMG matched the predicted outcome Narakas group Intra-operative C6 EMG prediction B favourable or better B unfavourable or worse Total 2 7 (6) 5 (1) (5) 5 (3) (2) 2 (1) 4 Match 13/15 = 87% 5/12 = 42% 18/27 = 67% Our decision was made by analysing three intra-operative prognostic factors: the EMG, SEP and muscular response to stimulation. We also assessed the severity of the lesion but this generally matches the neurophysiological tests: for example, a rupture with severe scarring and diffuse involvement will also have a poor SEP across the lesion. In order to apply this analytical approach to a mathematical model, we assigned one point to favourable prognostic factors (favourable B EMG, good muscular response, good SEP), half a point to less favourable factors (unfavourable type B, sluggish muscle response and weak SEP) and no points for unfavourable factors (type C, absence of muscle response and SEP). We considered that dislocation of the shoulder is a significant prognostic factor and assigned one negative point if it was present. The points were added and correlated with the Mallet score. There was a significant positive correlation between the points scored and the Mallet grade (Spearman s rho, correlation coefficient and p = 0.001, two-tailed). There were also significantly more infants with two or more favourable prognostic factors achieving a Mallet score > 13 (chi-squared test, p = 0.001) (Table VIII). Although this approach is sensitive and will pick up all those infants with a favourable outcome (Table IX; 100% sensitivity), its positive predictive value (PPV) is only 77%. It is not, therefore, very specific and not all infants with poor prognostic factors will have a poor outcome. With regard to the accuracy of C6 prediction, of the 32 infants, three underwent nerve reconstruction for C6 (two grafts and one neurotisation) and 29 had neurolysis. For one infant who underwent neurolysis the intra-operative EMG for C6 was measured but not recorded. Another from the neurolysis group needed pectoralis minor muscle transfer for a congenitally damaged biceps. These two infants were excluded, leaving 27 in the neurolysis group. The result of nerve reconstruction was unpredictable and only one infant regained full biceps function (Gilbert and Raimondi score 9 of 5). Conversely, 20 infants in the neurolysis group achieved Raimondi 5. Of note, 19 infants had no biceps movement clinically and four had only a flicker of movement pre-operatively at a mean age of 5.5 months (3 to 15). Only four infants had anti-gravity power pre-operatively. Using the same method as Bisinella et al 5 the accuracy of the intra-operative C6 prediction is shown in Tables X and XI. A type A or B favourable lesion equates to complete and powerful flexion of the elbow (Gilbert and Raimondi 5), whereas a type B unfavourable or C lesion indicates weak flexion or supination (Gilbert and Raimondi 4). Type B unfavourable and type C were combined for analysis. The prediction matched 67% (18 of 27) of the outcomes (Table X). The specificity and sensitivity were 42% and 87%, respectively, and the negative and the positive predictive values were 71% and 65%, respectively (Table XI). VOL. 95-B, No. 5, MAY 2013

6 704 K. F. CHIN, V. P. MISRA, G. M. SICURI, M. FOX, M. SINISI Table XI. The predictive value for C6 Case negative (poor outcome) Case positive (good outcome) Prediction positive (good result) 2 13 Sensitivity: 87% (13/15) Prediction negative (poor result) 5 7 Specificity: 42% (5/12) Positive prediction value (predict good result) 65% (13/20) Negative prediction value (predict poor result) 71% (5/7) Discussion Birch et al 1 reported the strategy for identifying infants with OBPP for surgical exploration. There are two outstanding issues, namely the limited ability to predict the outcome of C5 lesions, and when and which operation is needed. Ideally, the reliability of prediction of the intra-operative EMG should be compared with the long-term clinical outcome in a cohort of patients who have had no surgical intervention. However, as operative treatment is warranted if other clinical signs and investigations indicate a poor prognosis for spontaneous recovery, this approach cannot be adopted. All infants in our study had poor pre-operative electrophysiological results and a lack of clinical signs of recovery. In the presence of large neuromas the minimum required was neurolysis. We suggest that neurolysis does not alter the natural history of recovery but allows the recovery to happen effectively by removing the external constriction to regenerating nerve fibres. However, once surgical intervention has been performed the natural history of nerve recovery and the prediction of the intra-operative EMG could not be confirmed, other than for those who had neurolysis. With regard to improving the pre-operative prediction of the outcome of C5 lesions, neurophysiological evaluation might be limited by the difficulty of performing an accurate examination in the newborn, and the inability to record nerve action potentials for C5. The use of intraoperative neurophysiology in OBPP is largely limited to the diagnosis of root avulsions with the accuracy of this technique in predicting prognosis being disappointing. 13,14 However, several technical and secondary confounding factors may have contributed to these results. König et al 13 reported poor outcomes in five infants undergoing neurolysis because of the presence of recordable NAPs and good muscular response to motor stimulation proximal to the lesion. However, the NAPs were recorded at the infraclavicular brachial plexus rather than our central motor action potential (CMAP) recording, and the mean age of ten months in their series may be too late for neurolysis to be effective. Pondaag et al 14 found that the accuracy of NAP and CMAP registration was matched by the severity of nerve lesions defined by clinical, surgical, radiological and histological criteria. However, although the specificity of absent NAPs and CMAP in predicting a severe lesion was > 90%, the sensitivity was only < 30%. In the study by Pondaag et al, 14 CMAP registration was percutaneous. The needle EMG technique alone in our study has a sensitivity of 73% but a specificity of 17%. We defined a positive outcome as a favourable lesion, in contrast to Pondaag et al, 14 and hence our specificity is equivalent to their sensitivity. However, our study also showed that subluxation or dislocation of the shoulder is a major confounding factor, and removing it improved the sensitivity for C5 to 100% and the PPV to 85.7%. Our approach of combining NAPs, CMAP and gross motor response for assessing C5 conferred a sensitivity of 100%, specificity of 71.4%, PPV of 83% and NPV of 100% in all infants, including those with subluxation or dislocation of the shoulder. The specificity of only 71.4% suggests that the intra-operative neurophysiological investigation was not as optimistic as once thought. Although epidemiological studies indicate that the current incidence of OBPP is similar to that of 40 years ago, 3 the number of infants with severe OBPP seems to be decreasing, as 37% of infants who underwent C5 repair in the study by Birch et al 1 had Narakas type 4 injury, compared with 18% in our study. The change might reflect an increase in the number of elective Caesarean sections. With the prospect of an increasing incidence of less severe paralysis we believe the decision to reconstruct the superior trunk based on clinical examination alone is no longer acceptable. Clinical evaluation is not sufficiently accurate 15 and infants with a conduction block lasting more than three months have been described. 10 There were significantly more children without shoulder joint pathology who achieved a Mallet score > 13 (p = 0.031) (Table VI). This highlights the need to treat the shoulder pathology more rigorously. In all, ten infants had primary subluxation or dislocation of the shoulder at the time of exploration, of which four were Narakas type 2 injuries, four type 3 and two type 4. One worrying and evolving feature is that secondary subluxation of the shoulder only occurred in the neurolysis group (five of 22). This finding will be closely monitored. In the management of C6 lesions the sensitivity of the EMG for C6 is 87% and the negative predictive value (NPV) 71%. The intra-operative EMG predicted a poor outcome in seven infants with a C6 lesion but good outcomes were achieved. Whereas 74% (20 of 27) of the infants who had neurolysis achieved Raimondi grade 5 biceps function, only one of three infants who had nerve reconstruction recovered to the same level. The remaining seven in the neurolysis group had either grade 3 or 4 function, whereas one of those who underwent nerve reconstruction has no biceps function. This confirms the earlier finding 1 that the prognosis for C6 is generally good, reconstruction less predictable, and the only bad result was seen in those who were repaired. THE BONE & JOINT JOURNAL

7 INTRA-OPERATIVE NEUROPHYSIOLOGICAL PREDICTION OF UPPER TRUNK RECOVERY IN OBSTETRIC BRACHIAL PLEXUS PALSY 705 The timing of EMG investigation was not addressed in this study. As pointed out by Bisinella et al, 5 the relationship between EMG and outcome may not be linear. A type B favourable lesion may not necessarily equate to Mallet 13+ or Raimondi 4+. The recovery of an OBPP lesion is a dynamic process and the EMG represents only a snapshot of the state of recovery. It has been suggested that three months may not represent a stable state in OBPP injury, 16 and therefore a type B unfavourable lesion at operation may not necessarily equate with a poor outcome. Further work is needed to define the temporal relationship of EMG recording with the accuracy of prediction. In conclusion, the decision to limit the operation to neurolysis rather than grafting remains a viable option in infants showing signs of recovery intra-operatively where EMG is useful in predicting C5 nerve recovery. However, its accuracy in predicting the long-term outcome is confounded by secondary shoulder pathology. Nerve action potentials, motor action potentials and gross motor response to nerve stimulation can be used to improve the accuracy of intra-operative decision making. Greater emphasis needs to be placed on the treatment of secondary shoulder pathology in order to improve outcome. Neurolysis alone is preferred for C6 nerve lesions. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by D. Jones and first-proof edited by G. Scott. References 1. Birch R, Ahad N, Kono H, Smith S. Repair of obstetric brachial plexus palsy: results in 100 children. J Bone Joint Surg [Br] 2005;87-B: Gilbert A, Whitaker I. Obstetrical brachial plexus lesions. J Hand Surg Br 1991;16: Evans-Jones G, Kay SP, Weindling AM, et al. Congenital brachial palsy: incidence, causes, and outcome in the United Kingdom and Republic of Ireland. Arch Dis Child Fetal Neonatal Ed 2003;88:F185 F 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] 1999;81-A: Bisinella GL, Birch R, Smith SJ. Neurophysiological prediction of outcome in obstetric lesions of the brachial plexus. J Hand Surg [Br] 2003;28: Lin JC, Schwentker-Colizza A, Curtis CG, Clarke HM. Final results of grafting versus neurolysis in obstetrical brachial plexus palsy. Plast Reconstr Surg 2009;123: Narakas AO. Obstetric brachial plexus injuries. In: Lamb DW, ed. The paralysed hand. Edinburgh, Churchill Livingstone, 1987: Mallet J. Paralysie obstetricale du Plexus Brachial. Rev Chir Orthop Reparatrice Appar Mot 1972;58: (in French). 9. Haerle M, Gilbert A. Management of complete obstetric brachial plexus lesions. J Pediatr Orthop 2004;24: Smith SJ. The role of neurophysiological investigation in traumatic brachial plexus lesions in adults and children. J Hand Surg Br Birch R. Birth lesions of the brachial plexus. In: Birch R, Bonney G, Wynn Parry C, eds. Surgical disorders of the peripheral nerves. London: Churchill Livingstone, 1998: Chin KF, Di Mascio L, Holmes K, Misra VP, Sinisi MM. The value of preoperative and intraoperative electromyography in the management of obstetric brachial plexus injury. J Neurosurg Pediatr 2010;6: König RW, Antoniadis G, Börm W, Richter HP, Kretschmer T. Role of intraoperative neurophysiology in primary surgery for obstetrical brachial plexus palsy (OBPP). Childs Nerv Syst 2006;22: Pondaag W, van der Veken LP, van Someren PJ, van Dijk JG, Malessy MJ. Intraoperative nerve action and compound motor action potential recordings in patients with obstetric brachial plexus lesions. J Neurosurg 2008;109: Michelow BJ, Clarke HM, Curtis CG, et al. The natural history of obstetrical brachial plexus palsy. Plast Reconstr Surg 1994;93: Smith NC, Rowan P, Benson LJ, Ezaki M, Carter PR. Neonatal brachial plexus palsy: outcome of absent biceps function at three months of age. J Bone Joint Surg [Am] 2004;86-A: VOL. 95-B, No. 5, MAY 2013