Introduction and Aims: An important prognostic factor in neonatal brachial plexus injury is the timing of biceps recovery. Although the natural history is not clear, biceps recovery after three months has been used to predict universally poor long-term shoulder function. The absence of biceps function at three months has been adopted as an indication for early microsurgery, in an attempt to improve upper limb function. e

Method: Between 1980 and 1992, 170 patients with neonatal brachial plexus injury were entered into a prospective study, which recorded details of the birth and serial examinations. Patients were grouped according to the level of injury and timing of biceps return. Twenty-nine patients were observed to have absent biceps at three months. Twenty-eight of 29 patients were available for long-term review at an average of 11 years and one month (range 5yrs 7mths–16 years). At follow-up, patients answered a questionnaire, underwent muscle strength testing, sensory evaluation, and had their shoulder function graded according to Gilbert’s modification of the Mallet score. The level of injury and time of biceps return were compared to the final outcomes.

Results: The level of injury was C5/6 in 13/28 (46%), C5/6/7 in 5/28 (18%) and pan-plexus in 10/28 (36%), with a Horner’s syndrome in four patients. Of the 28 patients with absent biceps function at three months, biceps contraction was observed in 20 patients (71%) by six months. 13/13 C-5,6 patients, 3/5 C-5,6,7 patients, and 4/9 pan-plexus patients (0/4 of those with a Horner’s syndrome) regained biceps by six months. Twenty-two patients did not have plexus surgery, nine of these had subsequent shoulder surgery. At follow-up, 27 of 28 patients had at least anti-gravity biceps function. Patients who regained biceps function between three and six month had better scores for abduction (p=0.04), hand to neck (p=0.05) and hand to back (p< 0.001) than patients who regained biceps after six months. Patients with C-5, 6 lesions had better scores for external rotation (p=0.04), hand to neck (p=0.05), hand to mouth (p< 0.01) and hand to back (p< 0.001) than patients with pan-plexus lesions. Twelve of the 22 (55%) patients who did not have plexus surgery had Mallet class IV shoulders (good function). Of the C-5, 6 patients who did not have plexus surgery, 8/12 had class IV shoulders.

Conclusion: This study supports the associations between prolonged neurological recovery, more extensive level of injury and worse long-term shoulder function. In patients with C-5, 6 injuries and absent biceps function at three months, good shoulder function is common in the long-term without plexus surgery. This finding is important in considering the indications for and the results of early microsurgical intervention.