WE report a case of permanent phrenic nerve paralysis with hemidiaphragmatic palsy after interscalene brachial plexus block.
A 60-yr-old man, 160 cm tall, weighing 75 kg, with American Society of Anesthesiologists physical status class II, was admitted for elective right shoulder surgery. His medical history was unremarkable except for recent mild diabetes with no related neuropathy, controlled by diet and glimepiride. Physical examination results were unremarkable, and the results of laboratory studies were all within normal limits, including preoperative glycemia and chest x-ray. He agreed to undergo a combination of regional and general anesthesia.
Hydroxyzine, 100 mg, was administered 2 h preoperatively. After application of routine monitors, intravenous access was secured. He was positioned supine with the head turned to the contralateral side, and the right side of the neck was prepared as a sterile field. The elbow was flexed, with the forearm lying on the patient’s abdomen. Thereafter, interscalene brachial plexus block was performed as described by Winnie but using a nerve stimulator to ascertain that the needle’s tip was in the brachial plexus. The plexus was located with a nerve stimulator (Stimuplex HNS 11; B/Braun, Melsungen, Germany) and an insulated needle, 25 mm long with a short 30° bevel (Stimuplex, B/Braun). Three attempts at needle insertion were required to achieve an appropriate motor response: the brachial plexus was first located using a high current intensity (2 mA; 0.1 ms and 1 Hz), and then it was decreased to 0.5 mA to refine the approach. After obtaining a motor response of the deltoid muscle, a mixture of 30 ml ropivacaine, 0.75%, and 75 μg clonidine was injected. No blood could be aspirated, and the patient reported neither pain nor paresthesia during the procedure, although phrenic nerve stimulation was transiently observed.
After 20 min, profound surgical anesthesia was established on C5–C7 dermatomes. Then, general anesthesia was induced with 2 mg midazolam, 100 μg fentanyl, 200 mg propofol, and 30 mg atracurium to facilitate tracheal intubation. General anesthesia was maintained with 1–2% sevoflurane and 50% nitrous oxide, and the patient underwent a right rotator cuff repair via a deltopectoral approach. He was placed in a “beach chair” position with his head turned the opposite direction. Vital signs and standard parameters remained stable throughout the 2-h procedure.
At the end of surgery, the trachea was extubated, and the patient was observed for 1 h in the postanesthesia care unit. He did not report any pain. Vital signs and postoperative glycemia were normal. The interscalene brachial plexus blockade was still effective. Postoperative analgesia consisted of regular administration of a combination of propacetamol and nefopam intravenously. The patient was discharged to the ward. Postoperative follow-up was unremarkable.
Ten days later, the patient was readmitted to the hospital because of increasing shortness of breath. A chest roentgenogram revealed marked elevation of the right hemidiaphragm when compared with the preoperative chest film. No signs of infection or other disorders were shown on the film. This pattern was suggestive of acquired phrenic nerve palsy.
Because the moderate difficulty in breathing persisted despite physiotherapy, a complete checkup was made 3 months after the block. A new chest x-ray confirmed that the elevation of the right hemidiaphragm was unchanged and revealed atelectasis limited to the lower part of the right lung field, probably related to the right ventilatory deficit. No movement of the hemidiaphragm was observed during fluoroscopy, and paradoxical motion was shown by sniffing maneuver. Pulmonary function tests showed mild restrictive lung disease: vital capacity, forced expiratory volume in 1 s, forced vital capacity, and total lung capacity were respectively reduced to 89, 79, 88, and 76% of predicted values. By contrast, peak expiratory flow rate, arterial oxygen tension (Pao2), and arterial carbon dioxide tension (Paco2) were in the normal range. Computed tomography and nuclear magnetic resonance scans of the neck and thorax were also normal.
A definitive diagnosis of phrenic nerve dysfunction as the cause of hemidiaphragm paralysis was obtained by electromyography using phrenic nerve stimulation in the neck and the measurements of phrenic nerve latencies and conduction velocities. Stimulating electrodes were placed over the phrenic nerve in the supraclavicular fossa. The compound action potential of the hemidiaphragm was recorded using surface electrodes placed on the anterolateral aspect of the chest in the seventh intercostal space in the anterior axillary line. Results showed the absence of a right phrenic nerve compound action potential, whereas the left phrenic nerve conduction velocity was normal, suggesting that the right phrenic nerve was completely interrupted or significantly demyelinated. Although this examination failed to identify the mechanism or the precise location of the lesion, it was useful in confirming the lack of electromyographic pattern of diffuse neuropathy. One year after surgery, the patient still reported exertional dyspnea with no functional improvement.
In 1985, Bashein et al. 1reported a case of permanent hemidiaphragmatic paralysis after interscalene block performed using a paresthesia technique as described by Winnie. These authors suggested that phrenic nerve paralysis was related to a direct needle trauma. In the current situation, we describe right phrenic nerve paralysis after interscalene brachial plexus block despite the use of a nerve stimulator and B-bevel needle.
The 100% incidence of ipsilateral hemidiaphragmatic paresis reported in patients undergoing interscalene brachial plexus block has been related to the spread of the local anesthetic solution either on C3–C5 roots or through the scalene anterior fascia. 2This is not prevented by digital pressure and occurs with a variety of local anesthetics and doses. The main mechanism of this hemidiaphragmatic paresis is transient phrenic nerve block. This hypothesis is supported by the time profile of the paralysis, which is usually correlated with the pharmacologic properties of the local anesthetics used. However, this hemidiaphragmatic paresis is not usually associated with adverse clinical symptoms in healthy patients.
The mechanism of phrenic nerve paralysis in the current case may be caused by ischemic, mechanical, or chemical factors, which may occur either alone or in combination. The neural toxicity of ropivacaine has been studied after intraneural injection. 3Results showed that ropivacaine seems to be devoid of toxicity. Moreover, considering the small diameter of a phrenic nerve (1.5 mm), an intraneural injection seems unlikely. However, needle trauma may easily explain the phrenic nerve injury observed in the current case and in the report of Bashein et al. 1Indeed, such trauma has already been observed in patients undergoing central venous catheterization. 4–6Etiologies usually reported in the literature are needle trauma or compression resulting from hematoma. 7Functional respiratory recovery seems to be less after trauma than after hematoma.
Other causes of hemidiaphragmatic paralysis were considered. These etiologies were not compatible with the clinical course, the lack of electromyographic evidence of diffuse neuropathy, and the normal postoperative computed tomographic scan of the chest. In the current case, the surgical procedure is unlikely to explain the phrenic nerve paralysis. Moreover, even though the patient’s head was rotated during surgery, this posture is unlikely to explain diaphragmatic paralysis. Indeed, such stretch mechanisms have been reported to be associated with cervical nerves roots injuries, but they were induced by severe trauma 8or cervical chiropractic manipulation 9and often lead to transient phrenic nerve palsy only. Conversely, the chronology of the events is in favor of a complication related to the interscalene block, and the symptoms can be attributed to a severe phrenic neuropathy, such as axonotmesis or neurotmesis. Moreover, taking into account both the distance between the diaphragm and the site of nerve injury, the lesion can be considered to be permanent beyond 12 months after the initial injury, according to the usual speed of regeneration (1 mm/day). Such a long follow-up study has been performed only in the case published by Bashein et al. 1and in the current one.
This report stresses that the use of a nerve stimulator or a B-bevel needle does not guarantee that complications will not occur. Preliminary data from France also suggest that nerve complications can occur despite the use of a nerve stimulator, even with an apparently uneventful block (in 4 of 9 patients with neurologic complications reported, no risk factor could be identified [S.O.S. Regional Anesthesia Hot Line, Paris, France, unpublished data, obtained August 1998–May 1999]). It is noteworthy that the patient was installed with his forearm resting on his abdomen. With the forearm positioned that way, an ipsilateral diaphragmatic contraction may have been misinterpreted as an elbow contraction. Therefore, we recommend that patients be placed supine, with their arms by their sides. The sudden occurrence of diaphragmatic contractions should alert and indicate that the needle should be gently moved posteriorly.
It is interesting to note that the majority of patients with phrenic nerve paralysis reported in the literature did not report during the procedure paresthesia or shoulder pain that lead to nerve injury. 4–6Nevertheless, because referred pain from phrenic nerve irritation is mainly located in the shoulder, it is conceivable for a patient to report right shoulder pain while the needle tip comes in contact with the phrenic nerve. This has been reported during central venous catheterization. 10Therefore, subjective paresthesia to the shoulder during interscalene block could also be misinterpreted as an appropriate location of the needle tip, which is actually anterior to the plexus.
Appendix: The S.O.S. Regional Anesthesia Hot Line Service
Yves Auroy, M.D., Laurent Bargues, M.D. (Staff Anesthesiologists, Département d’Anesthesie, Hôpital D’instruction des Armees, Percy, Clamart, France); Dan Benhamou, M.D. (Professor and Chairman, Département d’Anesthésie et de Réanimation, Hôpital de Bicètre, Université Paris Sud, France); Hervé Bouaziz, M.D., Ph.D. (Staff Anesthesiologist, Département d’Anesthésie et de Réanimation, Hôpital Central, Nancy, France); Claude Ecoffey, M.D. (Professor and Chairman, Département d’Anesthésie et de Réanimation 2, Hôpital Ponchaillou, Université de Rennes, France); Frédéric J. Mercier, M.D. (Staff Anesthesiologist, Département d’Anesthésie, Hôpital Béclère, Clamart, France); Kamran Samii, M.D. (Professor and Chairman, Coordination d’Anesthésie et de Réanimation, Université de Toulouse, France).