Secondary Logo

Journal Logo

Aetiology of convulsions due to stellate ganglion block: a review and report of two cases

Mahli, A.; Coskun, D.; Akcali, D. T.

European Journal of Anaesthesiology: May 2002 - Volume 19 - Issue 5 - p 376-380
Clinical Letter

Stellate ganglion block is a selective sympathetic block that affects the ipsilateral head, neck, upper extremity and upper part of the thorax. Convulsions are a recognized complication of intra-arterial injection during stellate ganglion block. As central nervous system toxicity depends ultimately on the concentration of the local anaesthetics presented to the brain, the likely causative factors are discussed as well as the types of toxic symptoms and their onset times. The paper considers the aetiological factors of such convulsions resulting from stellate ganglion block in two patients.

Gazi University, Department of Anesthesiology, Faculty of Medicine, Ankara, Turkey

Correspondence to: Ahmet Mahli, Gazi Universitesi Tip Fakülresi, Anesteziyoloji ve Reanimasyon AD, Besevler 06500, Ankara, Turkey. E-mail:; Tel: +00 312 2141080/5319; Fax: +00 0312-2124647

Accepted for publication May 2001 EJA 551

Back to Top | Article Outline


Stellate ganglion block is a common procedure. It is an important treatment modality for vascular insufficiency, reflex sympathetic dystrophia, causalgia and pain-induced herpes zoster infection [1-3]. Many approaches have been described for stellate ganglion block. However, despite all precautions, complications such as convulsions may occur even in the hands of an experienced physician [4,5].

Local anaesthetic toxicity is influenced by the speed and site of injection and by the type of local anaesthetic used [6]. Convulsing doses of intra-arterial local anaesthetics vary according to the location of the artery. In some patients, occasionally a very small dose of local anaesthetic can cause convulsions following vertebral or carotid artery injection. Therefore, before starting stellate ganglion block, all resuscitation drugs and equipment must be ready for use in any emergency. The aim of the paper is to emphasize the aetiological factors underlying convulsions due to stellate ganglion block as well as the advantages of the anterolateral approach over the anterior approach.

Of the many approaches described for stellate ganglion block, we are most familiar with the anterolateral and anterior approaches used in our institution. Before the procedure, patients are monitored by pulse oximetry, electrocardiography and non-invasive blood pressure. An i.v. cannula is inserted and resuscitation equipment prepared for use in case of potential side-effects or complications.

Back to Top | Article Outline


First patient

A 28-yr-old female, weight 75 kg, was admitted complaining of a cold sensation and ache in both hands. Her medical history was otherwise normal. Both hands were cold and cyanosis was prominent, especially of the distal phalanges. Raynaud's phenomenon was diagnosed. Treatment with stellate ganglion block by the anterolateral approach was planned, a technique first used by Leriche and later modified by Dosch [7]. In this technique, while the patient lies supine, the head is extended as far as possible in the opposite direction. Then, depending on the length of the patient's neck, three fingertips of the left hand are placed in such a way that the caudal finger lie on the upper edge of the sternoclavicular joint. The fingers carefully press the carotid artery and the jugular vein out of the way of the needle, and the distended apex of the pleura is also forced down completely. The needle is inserted a short distance immediately above the cranial finger and is then guided deeper to make contact with bone (Fig. 1, needle position 4), which is felt practically subcutaneously. After bone contact, the needle is withdrawn 2-3 mm to clear the muscles over the transverse process and to avoid the chance of subperiosteal injection, and the syringe is aspirated in two planes by turning the bevel 180°. So long as no blood, air or liquor appears in the syringe (a negative aspiration), local anaesthetic is then injected very slowly. In this patient, who presented for her fifth treatment session by stellate ganglion block, lidocaine 1% 5 mL was prepared for injection through a 20-G needle using the anterolateral approach. An aspiration test was negative in two planes and lidocaine 1% 1 mL was injected over 2-3 s. Since the second aspiration test was positive for blood, the needle was pulled back a little, and when the second attempt at injection was made, the patient immediately had severe generalized tonic-clonic convulsions. The injection was stopped and the needle removed. Although there were no apnoea or respiratory distress, cyanosis and nystagmus owing to the severe convulsions and contractions were observed. The patient's head was extended and the airway maintained by chin elevation. Thirty seconds after the injection, spontaneous respiration was normal and the patient opened her eyes. Arterial pressure was 140/80 mmHg with a pulse rate 100 beats min−1. The patient was completely conscious after 2 min.

Figure 1

Figure 1

Back to Top | Article Outline

Second patient

A 31-yr-old diabetic male, weight 72 kg, who had a below-knee amputation 4 years previously, was diagnosed with Buerger's disease. He complained of cold sensation, pain, cyanosis and oedema of both hands, more on the right than on the left, the onset of which was 9 months previously. A stellate ganglion block by the anterior (paratracheal) technique was planned as the method of treatment. The technique was first used by Herget and Kirschner [8]. In this technique, while the patient lies supine and the head is maximally extended, the needle is inserted at the medial edge of the stenocleidomastoid muscle just below the cricoid cartilage at the level of the transverse process of C6 (3 cm above the clavicle). The non-operative hand retracts this muscle together with the carotid sheath before insertion of the needle. The needle is advanced to the transverse process and withdrawn 2-3 mm. An aspiration test is carried out in two planes before a 1 mL test dose excludes unintentional intravascular or subarachnoid injection into a dural sleeve.

The patient was about to undergo his third treatment session by stellate ganglion block. Lidocaine 1% 5 mL was prepared for injection through a 20-G 3.5 cm-long needle for the anterior approach. The patient was in a semisitting position. An aspiration test was negative in two planes and lidocaine was injected at 1 mL every 2-3 s. Suddenly convulsions, unconsciousness, mild cyanosis, and a rapid and shallow respiratory pattern were observed after injection of lidocaine 1 mL. Diazepam 10 mg i.v. was given while the airway was maintained. Arterial pressure was 160/100 mmHg with a pulse rate 110 beats min−1. The patient was conscious and respiration returned to normal with no further need for airway manipulations after 2 min.

Neither patient with convulsions had either early or late signs of Horner's syndrome on these occasions. The same specialist performed all the blocks. The patients received 10 treatments of stellate ganglion block, using an 18-G needle, for the remaining sessions without any complication.

Back to Top | Article Outline


Stellate ganglion block can be performed for diagnostic and therapeutic purposes and 16 possible approaches have been described [8], although only five (Fig. 1) are commonly practised clinically [9].

In the cases presented, the probable causes of the reactions observed during the application of the stellate ganglion block and the factors underlying these reactions were investigated and described. Various kinds of toxic reaction have been reported in humans in association with the use of local anaesthetic agents. The adverse reactions observed include systemic toxicity involving primarily the central nervous system (CNS) and the cardiovascular system, localized neural and skeletal muscle irritation, and specific side-effects such as methaemoglobinaemia and allergy. Moreover, most untoward effects are due to the inappropriate use of local anaesthetics, such as accidental intravascular or intrathecal injections or administration of an excessive dose [10]. In addition, it should be further emphasized that most toxic reactions to local anaesthetics in humans involve the CNS. Local anaesthetic-induced cardiovascular depression occurs less frequently but tends to be more serious and more difficult to manage. In convulsant dosage, local anaesthetics increase the blood pressure, heart rate and cardiac output. This represents a stimulatory effect to autonomic control centres in the brainstem. When supraconvulsant doses are administered, direct inhibition of cardiac contractility and conduction occur [6,10].

The systolic blood pressures of the two patients aged 28 and 31 yr were 140 and 160 mmHg while their heart rates were 100 and 110 beats min−1 respectively. This increase supports the above view that the dose (lidocaine 10 mg) accidentally injected into the vertebral artery stimulated the autonomic control centres in the brainstem. Thus, even this dose may be considered as convulsant. Allergic reactions to local anaesthetics have also been observed, but they are extremely rare (2%). They present as localized, cutaneous hypersensitivity or anaphylactic reactions involving widespread mediator release [7,11,12]. Local anaesthetics are not directly antigenic but act as haptens in producing immune-mediated responses. Cross-sensitivity to products containing para-aminobenzoic acid represents the most common allergic reaction to the amino-ester local anaesthetics. True allergic reactions to the amino-amide class are extremely rare. Preservatives in local anaesthetic solutions, such as parabens, may also induce allergic responses [6,13]. Of local anaesthetic reactions, 99% are not anaphylactic or allergic, but may be allergic if they simultaneously involve skin reactions [6]. Sensitivity reactions to injected drug are practically zero and so rare that they can be neglected and are thought to be toxic reactions because of intravascular injection. These reactions caused by a small dose of lidocaine are also not real allergic reactions. On the contrary, they are due to high drug plasma concentrations, like all reactions listed as sensitivity [10].

Evidence of CNS dysfunction includes perioral or lingual numbness, sedation, lightheadedness, tinnitus, vertigo, visual disturbance, slurred speech, skeletal muscle twitching, disorientation, unconsciousness, generalized convulsions, coma and apnoea. The initial clinical presentation represents a state of central excitation that occurs before generalized CNS depression [6,10]. The mechanism by which local anaesthetic agents produce an initial state of CNS excitation involves the selective blockade of inhibitory pathways in the cerebral cortex. The initial inhibition of inhibitory pathways by local anaesthetic agents would allow facilitatory neurons to function in an unopposed fashion, which would result in an increase in excitatory activity, leading to convulsions [10]. Convulsions originate in areas of the brain such as the amygdala or hippocampus, and may result from an inhibition of γ-aminobutyric acid release [6].

Stellate ganglion block has some rare life-threatening complications because of the anatomical proximity of intravascular, subarachnoid and epidural spaces [2,14]. The anterolateral and anterior approach are reported to be easy, reliable and common approaches [7,9], but anterior stellate ganglion block especially is not always sufficient for upper extremity sympathetic nerve block [15] and has serious complications such as i.v. or intra-arterial injection at the time of, or just after, injection [16]. Bone contact and negative aspiration do not always guarantee avoidance of these complications [14].

CNS toxicity is a result of increased local anaesthetic concentration in the cerebral blood supply. Since the head and neck region is very vascular and richly perfused, under normal conditions local anaesthetic injected into the neck reaches toxic levels (dose related) in 5-30 min [12]. Several other studies maintain lidocaine plasma concentrations in the range 6-8 μg kg−1 at 20 min after injection [10]. Nevertheless, high local anaesthetic doses and/or blood-rich tissues are not always involved in CNS toxicity [12,17].

Accidental, direct intravascular injection during a regional anaesthetic block provides a common mechanism for unanticipated toxic effects. When a bolus dose of local anaesthetic enters the circulation, peak concentrations are higher and occur earlier than if the same dose were to undergo normal systemic absorption [18]. Intravascular injection is suspected when CNS toxicity signs and symptoms are seen during local anaesthetic injection [19]. Toxic signs can be observed by doses one-half or two-thirds of known maximum reliable doses. Usually, rapid injection is responsible for toxic venous plasma concentration of local anaesthetics [20]. For instance, when lidocaine 7 mg kg−1 is injected rapidly via the brachial or femoral vein, convulsions are seen in 30 s [12].

The convulsion, in these cases after lidocaine 1% 1 mL injection, was not related to sensitivity reactions since there were no skin reactions associated with high doses or i.v. injection. No adverse reactions were reported during the previous treatment sessions. The absence of motor or sensory loss and the briefness of the period of respiratory distress suggest that there was no injection into the subarachnoid or epidural spaces. The literature on convulsion aetiology reveals that intra-arterial injection is the cause of immediate convulsions, even at minimal doses.

In human beings convulsion incidence due to local anaesthetics is generally 0.017-0.12% and the convulsing dose (CD50) for lidocaine is 6.4 mg kg−1[2,21]. Plasma concentrations of lidocaine >5 μg mL−1 usually cause symptoms of systemic toxicity [11]. Arterial plasma concentrations of lidocaine after constant rate i.v. infusion in man has been studied. These results show a threefold increase in peak arterial lidocaine concentration when the drug is infused over 3 min (9.3 μg mL−1) compared with 19 and 31 min (3.2 and 2.5 μg mL−1) [18]. The estimated minimum toxic dose of lidocaine resulting from cerebral arterial injection is 15% of the minimum toxic i.v. dose divided by 4: in adults, (6.4 mg kg−1 × 70 kg × 0.15)/4 = 16.8 mg lidocaine [22]. The cerebral blood volume at any one time is 2.4-4.25 mL, which is approximately 3 mL 100 g−1 blood in the whole adult human brain at any given moment [23].

Approximately 1 mL lidocaine 1% (10 mg) was injected. Assuming that it is diluted in 30 mL blood, the total concentration is 10 000/30 = 333 μg mL−1. If there were no reverse flow into the carotid system, then, as the flow through the vertebral artery represents approximately 20% of the cerebral blood flow, the concentration of lidocaine delivered to the brain in the above patients may have been 1665 μg mL−1. As CNS toxicity depends ultimately on the concentration of local anaesthetics presented to the brain, the focal symptoms in the present patients resulted from a highly localized toxic dose.

Two problems arise from the proximity of the stellate ganglion and the arteries perfusing the brain. First, the tight anatomical relationship between ganglion and arteries increases the risk of intra-arterial injection; second, cerebral toxicity can result even when very low doses of local anaesthetics are injected into the arteries that perfuse the brain [24]. Two types of toxic reaction - excitatory and depressive - to cerebral intra-arterial injections of local anaesthetics are described. Excitatory reactions are more common. The lack of excitatory reactions and the presence of only depressive symptoms suggest direct carotid artery injection, whereas only excitatory reactions suggest direct vertebral artery injection of local anaesthetic while the stellate block is performed [10,23,25]. Both vertebral arteries, after forming the basilar artery, and the carotid arteries supply the circle of Willis, thus a low-dose local anaesthetic injected into either vertebral artery causes generalized convulsions [26]. Local anaesthetic injection into the vertebral artery is a rare complication. It causes unconsciousness, convulsions and apnoea. Sometimes, a locked-in syndrome characterized by transient eye fixation, an inability to move and respiratory distress without loss of consciousness may constitute a complication [25].

In both patients reported here, brief excitatory signs without depressant signs caused by small doses of local anaesthetic injected directly to the proximal cerebral circulation proved vertebral artery injection. The injections were immediately stopped. Rapid elimination of local anaesthetic from the brain due to the high cerebral blood flow was the reason for the brief excitatory symptoms. On the other hand, if the local anaesthetic dose was more than the calculated intra-arterial toxic dose, potential excitatory symptoms would have lasted longer and treatment would have been more difficult. The minimum dose of lidocaine toxicity for cerebral arterial injection has been calculated as 16.8 mg. However, it should be noted that accidental injections into either the vertebral or the carotid arteries even with half this dose, partly also depending on the rate of injection, may produce toxic reactions (excitatory or depressive) ranging from the mildest to the most severe.

It is essential that resuscitative equipment is available before the performance of any block. Patients should also be warned not to cough, swallow or take deep breaths while the stellate ganglion block is performed to reduce the risk of serious complications. In addition, the injection should be made using a well-lubricated glass syringe with an easy movement and fitted with a large-diameter needle with the bevel directed cranially or caudally, and only low concentrations of local anaesthetic should be injected. If toxicity symptoms are observed, the injection should be stopped immediately and the needle withdrawn. The anterolateral approach for stellate ganglion block should be the method of choice as small and large vessels are more distant than encountered in the anterior approach. Moreover, regardless of the technique used, we recommend using less than the calculated minimum arterial toxic dose as the initial test dose, and any subsequent doses (each lidocaine 1% 0.5 mL) after aspiration in two planes, to minimize and allow the observation of any potentially disastrous side-effects of an intra-arterial injection, thus giving more time to intervene.

Back to Top | Article Outline


1. Arden RL, Bahu SJ, Zuazu MA, Berguer R. Reflex sympathetic dystrophy of the face: current treatment recommendations. Laryngoscope 1998; 108: 437-442.
2. Wulf H, Maier C. Complications and side effects of stellate ganglion blockade. Results of a questionnaire survey. Anaesthesist 1992; 41: 146-151.
3. Sugimoto M, Shimaoka M, Taenaka N, Kiyono H, Yoshiya I. Lymphocyte activation is attenuated by stellate ganglion block. Reg Anesth Pain Med 1999; 24: 30-35.
4. Bruyns T, Devulder J, Vermeulen H, DeColvenaer L, Rolly G. Possible inadvertent subdural block following attempted stellate ganglion blockade. Anaesthesia 1991; 46: 747-749.
5. Mishio M, Matsumoto T, Okuda Y, Kitajima T. Delayed severe airway obstruction due to hematoma following stellate ganglion block. Reg Anesth Pain Med 1998; 23: 516-519.
6. Mitchell ME. Local anesthetic toxic effects. In: Atlee JL (ed.). Complications in Anesthesia. Philadelphia, USA: Saunders, 1999: 249-256.
7. Dosch P. Manual of Neural Therapy: According to Hunecke, 11th edn. Heidelberg, Germany: Karl F Haug, 1984: 375-386.
8. Nolte H. Blockade des sympathikus und Vagus. In: Hans Killian (ed.). Lokalanaesthesie und Lokalanaesthetika zu operutiven, diagnostischen und therapeutischen Zwecken, 2nd edn. Stuttgart, Germany: George Thieme, 1973: 460-482.
9. Alam S. Continuous stellate ganglion blockade. Anesth Analg 1996; 83: 1350-1351.
10. Covino BG, Wildsmith JAW. Clinical pharmacology of local anesthetic agents. In: Cousins MJ, Bridenbaugh PO (eds). Neural Blockade in Clinical Anesthesia and Management of Pain, 3rd edn. Philadelphia, USA: Lippincott-Raven, 1998: 97-128.
11. Covino BG. Systemic toxicity of local anesthetic agents. Anesth Analg 1978; 57: 387-388.
12. Aldrete JA, Romo-Salas F, Arora S, Wilson R, Rutherford R. Reverse arterial blood flow as a pathway for central nervous system toxic responses following injection of local anesthetics. Anesth Analg 1978; 57: 428-433.
13. DeJong RH, Ronfeld RA, DeRosa RA. Cardiovascular effects of convulsant and supraconvulsant doses of amide local anesthetics. Anesth Analg 1982; 61: 3-9.
14. Raj PP. Stellate ganglion block. In: Waldman SD, Winnie AP (eds). Interventional Pain Management. Philadelphia, USA: Saunders, 1996: 267-274.
15. Soriano C, Ortiz M, Fas MJ, et al. A possibility of central diffusion during stellate ganglion blockade: the sheath of the spinal rachidian nerve. Rev Esp Anestesiol Reanim 1999; 46: 123-125.
16. Larsen B, Macher F, Bolte M, Larsen R. Stellate ganglion block with transcutaneous electric nerve stimulation (TENS): a double-blind study with healthy probands. Anasthesiol Intensivmed Notfallmed Schmerzther 1995; 30: 155-162.
17. Yokoyama M, Mizobuchi S, Nakatsuka H, Hirakawa M. Comparison of plasma lidocaine concentrations after injection of a fixed small volume in the stellate ganglion, the lumbar epidural space, or a single intercostal nerve. Anesth Analg 1998; 87: 112-115.
18. Tucker GT, Boas RA. Pharmacokinetic aspects of intravenous regional anesthesia. Anesthesiology 1971; 34: 538-549.
19. Bourne JG. Deaths with dental anaesthetics. Anaesthesia 1970; 25: 473-481.
20. Scott DB, Thorburn JT. Evaluation of the toxicity of local anaesthetic agents in man. Br J Anaesth 1975; 47: 56-60.
21. Reiz S, Nath S. Cardiotoxicity of local anaesthetic agents. Br J Anaesth 1986; 38: 736-746.
22. Scott DL, Ghia JN, Teeple E. Aphasia and hemiparesis following stellate ganglion block. Anesth Analg 1983; 62: 1038-1040.
23. Szeinfeld M, Laurencio M, Pallares VS. Total reversible blindness following attempted stellate ganglion block: a case report. Anesth Analg 1981; 60: 689-690.
24. Korevaar WC, Burney RG, Moore PA. Convulsions during stellate ganglion block: a case report. Anesth Analg 1979; 58: 329-330.
25. Dukes RR, Alexander LA. Transient locked-in syndrome after vascular injection during stellate ganglion block. Reg Anaesth 1993; 18: 378-380.
26. Yasargil MG. Microneurosurgery. New York: Thieme, 1984: 128-129.

GANGLIA, SYMPATHETIC, stellate ganglion; SEIZURES, convulsions

© 2002 European Academy of Anaesthesiology