Permanent Loss of Cervical Spinal Cord Function Associated with Interscalene Block Performed under General Anesthesia

PERFORMING an interscalene block (ISB) during general anesthesia (GA) for postoperative shoulder surgery analgesia is a common procedure. Although epidural, 1–3subarachnoid, 4–6and motor nerve root 7injections have been reported during the performance of ISB, there have been no cases reported in the peer-reviewed literature of injection of local anesthetic solution into the substance of the cervical spinal cord. Although injection of local anesthetic solution into the cervical spinal cord would be unlikely in awake patients, performing the block in anesthetized patients may increase the risk .This report describes four cases in which the performance of ISB during GA was followed by total spinal anesthesia and extensive permanent loss of bilateral cervical spinal cord function, and in which magnetic resonance imaging (MRI) showed syrinx or cavity formation in the cervical spinal cord. Each of the patients in the cases was involved in a separate medical malpractice law suit. In the first three cases the author was an expert witness, and in the fourth case the medical data was taken from the defense-expert timeline of clinical events and findings (obtained by interrogatory questioning of patient 2); in each case the relevant MRI was supplied by the plaintiff's attorney.Case 1The patient was a 39-yr-old woman, 172 cm, 109 kg, who presented for right shoulder arthroscopic repair of a rotator cuff tear. A cervical spine MRI 16 months before surgery was normal. Medical and surgical histories were unremarkable. After sedation with use of midazolam and methohexital administered in divided doses over 6 min to eliminate head and neck movement, the patient was unconscious but spontaneously breathing. The right interscalene groove was palpated and right upper extremity (RUE) muscle twitches were obtained via a 2-in (5.0 cm) 22-gauge Stimuplex (B. Braun Medical, Bethlehem, PA) needle at 0.81 mA. Alternating negative aspirations with 5-ml injections, a total of 40 ml of a solution containing 0.75% mepivacaine and 0.25% bupivacaine in 1:200,000 epinephrine was injected. Less than 30 s after the injection was completed, the patient became apneic, oxygen saturation as measured by pulse oximetry (Spo2) decreased from 96% to 60%, blood pressure decreased by 25%, and heart rate was unchanged. The patient easily underwent ventilation via mask, was tracheally intubated, and stabilized within 2 min. The 45-min surgery was performed uneventfully while the patient underwent ventilation with positive pressure with 0.4% isoflurane in oxygen. Two and one half hours post-ISB, the patient opened her eyes to command and began spontaneous ventilation; 3 h post-ISB the patient was extubated. After extubation, the patient had complete loss of motor and sensory function of the RUE and allodynia and burning pain of the left upper extremity; the fully evolved neurologic deficit also included bilateral lower extremity weakness, in the right much more than in the left. Cervical spine MRI performed 3 weeks and 9 months post-ISB showed a syrinx or cavity in the central portion of the right half of the C4–C7 spinal cord (fig. 1). The patient was a healthy 52-yr-old woman, 155 cm, 86 kg, who underwent arthroscopic repair of a right rotator cuff tear during GA. At the end of the operation, and with the patient breathing spontaneously during GA, a right ISB was performed using a 2-in (5.0 cm) 22-gauge Stimuplex needle, 1.0 mA stimulation, and 20 ml bupivacaine, 0.25% (alternating negative aspirations with injections). After the injection, a decrease in blood pressure of 25% was responsive to 10 mg intravenous ephedrine, but the patient became apneic, was unresponsive to noxious stimuli, and had dilated pupils. One hour post-ISB, the patient awakened, and 2 hours post-ISB the patient was extubated. However, the patient had no motor or sensory function in the left upper extremity and had significant weakness of the RUE. MRI of the cervical spine 20 h post-ISB showed myelomalacia extending from C1–T1 and a 7 × 7 × 10 mm intradural, but extramedullary, right anterolateral mass at the T2 level, which slightly compressed the spinal cord. Spinal tap on postoperative day 3 showed intensely xanthrochromic cerebral spinal fluid. Cervical spine MRI 25 months post-ISB showed one continuous C2–T1 intramedullary cervical spinal cord lesion and slight enlargement of the T2 extramedullary mass (fig. 2). The fully evolved neurologic deficit at 25 months consisted of weakness and atrophy of the upper extremities, in the left more than in the right, loss of pain and temperature in both upper extremities, preservation of touch, vibration and proprioception in both upper extremities, and preservation of motor and sensory function in both lower extremities. The T2 extramedullary mass was resected at 28 months post-ISB and was diagnoses as a meningioma; the patient's neurologic status was unchanged. A healthy 52-yr-old man, 167 cm, 93 kg, underwent an open repair of a 2-yr-old right rotator cuff tear during GA. After the completion of surgery, all anesthetics were discontinued. While the patient was still unconscious, a right ISB with a 1.5-in (3.75 cm) 25-gauge needle was thought to have been performed by "walking" the needle off the posterior aspect of the C6 transverse process and injecting 15 ml of a solution containing 1% lidocaine and 0.5% bupivacaine without epinephrine by alternating three negative aspirations with three 5-ml injections. During the ISB, patient coughing, straining, and neck movement with the needle in situ necessitated administration of 75 mg intravenous sodium thiopental and resumption of 1.0% isoflurane. After ISB, isoflurane was discontinued and the patient underwent suctioning and extubated. However, after 5–10 min of ventilation via mask, the patient remained apneic, unconscious, and unresponsive to noxious stimuli and the pupils were dilated fully. The patient was reintubated and taken to the intensive care unit. Approximately 1 h post-ISB the patient started to breathe spontaneously and open his eyes to command, and, 3 h post-ISB, the patient was extubated and awake and alert. However, he had no motor or sensory function in his RUE, whereas the left upper extremity had a moderate decrease in finger and wrist strength. MRI of the cervical spine on postoperative day 7 showed extensive edema and hemorrhage (clotted blood) throughout the cervical spinal cord (fig. 3A). MRI at 15 months post-ISB showed a syrinx in the central aspect of the right spinal cord extending the entire length of the cervical spinal cord (fig. 3B). The fully evolved motor and sensory deficits were improved slightly compared with the initial examination and were consistent with the location of the syrinxes in the cervical spinal cord. A 38-yr-old woman, 160 cm, 71 kg (body mass index = 27.8 kg), required arthroscopic repair of a right rotator cuff tear. The patient was in good health and had an unremarkable medical and surgery history. GA (propofol bolus and infusion, vecuronium, nitrous oxide, isoflurane) and the 100-min surgery were uneventful. During GA, a right ISB was performed with a 2-in (5.0 cm) 22-gauge Stimuplex needle, 1.0 mA nerve stimulation, and 35 ml bupivacaine, 0.5%, with 1:200,000 epinephrine injected in divided doses (alternating with negative aspirations). For the first 2.5 h after ISB the patient was apneic, remained unconscious and unreactive to painful stimuli, had dilated nonreactive 7-mm pupils, and was hemodynamically stable. Five hours post-ISB, the patient was extubated awake and alert; however, the patient had extensive loss of RUE motor and sensory function and left upper extremity burning pain and allodynia. Cervical spinal cord MRI at 18 h post-ISB showed central intramedullary hemorrhage extending to C5–C7 and at 14 months post-ISB showed myelomalacia and syrinx formation in most of the cervical spinal cord.In each of these four patients ISB performed during GA was followed by the same three in-series events: spinal anesthesia, loss of cervical spinal cord function, and radiologic evidence of severe cervical spinal cord damage. Several possible mechanisms are responsible for ISB-induced cervical spinal cord damage in these patients. First, there is an experimental basis for proposing peripheral intraneural injection and retrograde intraneural dissection of local anesthetic into the cervical spinal cord theory of causation, 8,9including increased toxicity of undiluted (with cerebrospinal fluid) local anesthetic. 10In addition, during GA a patient would be unable to protest the pain of an intraneural injection. However, the principle practical arguments against such a cause consist of the absence of damage to structures outside the cervical spinal cord and the technical improbability of achieving such an injection. 11,12Second, in patient 2, the T2 meningioma could have caused cephalad pooling of local anesthetic if the local anesthetic had gained access to the cerebrospinal fluid space. The principle arguments against pooling of local anesthetic above the T2 meningioma consist of absence of a rim lesion, absence of a T2 lesion, the presence of asymmetrical central cord damage, and the nontoxicity of 0.25% bupivacaine topically applied to the spinal cord. 10Third, in all four patients, all clinical and radiologic data are explained well by injection of local anesthetic solution directly into the substance of the cervical spinal cord at C6. Indeed, this cause explains why the lesions in all four patients are centered on C6. Although the tip of a 22-gauge needle may be expected to cause some damage to the spinal cord, most of the damage caused by direct injection of fluid into the spinal cord probably is a result of the phenomena of pressure and separation of cells or neurons, and the extent of neurologic damage is likely to be proportional to the volume of local anesthetic injected into the spinal cord. 9–11The clinical symptomatology caused by the fluid injected into the substance of the cervical spinal cord depends on the path that the fluid follows away from the point of injection.In patients 1 and 2 the principal argument against "intracord" injection was that a 1.5- to 2.0-in (3.75- to 5.00-cm) needle could not reach the spinal cord by passing through the C6 intervertebral foramen. However, four considerations make it probable that the needle could pass into the cervical spinal cord. First, indentation of the skin line and compression of the deeper tissue by palpation of the interscalene groove can shorten the skin-line-to-spinal-cord distance considerably. In addition, the interscalene groove can be palpated with much more force in a patient during GA than in an awake patient. Second, previous reports of injection of local anesthetic solution into the motor nerve root and axial subarachnoid space 4–7show that the tip of the needle can get very close to the cervical spinal cord. Third, the brachial plexus is very close to halfway between the skin line and cervical spinal cord. 13–15Because the distance from the interscalene groove skin line to the brachial plexus in the majority of patients is less than 1 in (2.5 cm), and in healthy patients is approximately 0.5 in (1.25 cm), 12–15it follows that the cervical spinal cord will be within 2 in (5.0 cm) of the skin line in the majority of patients. Unfortunately, MRI in individual cases cannot be used to determine the IS-groove-to-spinal-cord distance because the lateral border of the neck at the C6 level is formed by the trapezius muscle, and the IS groove is well within (mesiad to) the trapezius muscle, or the IS groove cannot be identified by MRI. Fourth, in patient 3, sudden and unexpected forceful movement of the patient's neck may have caused the 1.5-in (3.75 cm) needle to forcefully go through the C6 intervertebral foramen and penetrate the cervical spinal cord.If the clinical and radiologic lesions in these four patients were caused by intracord injection of local anesthetic, there are several clinical lessons to be learned or relearned. The lessons include the following: GA can be considered as a relative contraindication for ISB, ISB needles greater than 1.0–1.25 in should not be inserted, ISB needles less than 1.5 in (< 3.75 cm) should be used, the physician should ensure the patient does not unexpectedly move, the ISB needle should have a definite caudad direction at the C6 level, and in obese patients landmarks may be obscure. With these lessons in mind, the risk of intracord injection during ISB should be minimized.The author thanks attorneys Edwin Krieger, Thomas Ryan, Michael McNally, and Mark Kamitomo for assistance in obtaining the relevant MRIs, and neuroradiologist doctors John Arrington, Colin Bramford, and Charles Kerber for accurate labeling of the MRIs.