Paraplegia after coeliac plexus block

Anaesthesia, 1989, Volume 44, pages 4 8 7 4 8 9 CASE REPORT

Paraplegia after coeliac plexus block

M . J. WOODHAM

AND

M . H. H A N N A

Summary A case is described in which paraplegia followed a coeliac plexus block performed using 90% alcohol under X ray screening. The likely cause was an ischaemic injury to the cord secondary to damage to the artery ofrldamkiewicz. This rare complication seems difficult to avoid.

Key words Anaesthetic techniques, regionat; coeliac plexus block. Complications; paraplegia.

Coeliac plexus block is a widely used technique for the control of upper abdominal visceral pain secondary to malignant disease.’ Classically it has been indicated for the treatment of pain secondary to carcinoma of the pancreas, although it has been used in patients with chronic pancreatitis. A hundred patients have been treated with coeliac plexus blockade in this centre over the last 4 years and the indications for this procedure are listed in Table 1. This table shows that a large number of our patients have either primary or secondary carcinoma of the liver, which reflects our pattern of patient referral. There had been no serious side effects until the case that is reported here. Case history Three months before referral to the pain relief unit the patient, a 62-year-old woman, was diagnosed as having disseminated carcinomatosis. She presented with complaints of abdominal pain, nausea, general malaise and occasional headaches. Her liver function tests were abnormal and a liver biopsy was performed; the histology showed ‘infiltrating adenocarcinoma with dense surrounding fibrous tissue consistent with metastatic tumour’. This was treated by inserting an hepatic artery portacath at laparotomy followed by a course of 5-fluorouracil. The patient was then discharged and readmitted 3 months later. She complained of worsening pain after re-admission and was referred to the pain relief unit. Her main complaint was of right upper-quadrant pain which radiated to her epigastrium. This was due to further enlargement of the liver metastases. She was taking regular paracetamol at that time but this was not controlling the pain and she

Table 1. Indications for coeliac plexus block in the last 100 patients in this unit. Liver metastases Hepatoma Chronic pancreatitis Carcinoma pancreas Others

= = = = =

40% 16% 13% 12% 19%

was unwilling to take stronger analgesics. It was decided, after discussion with the patient, to proceed with a coeliac plexus block; the procedure was explained and formal consent obtained. The technique used in this centre is similar to that described by Thompson.’ The patient was starved on the morning of the procedure; no premedication was given. On arrival in the X ray department she was placed prone, an intravenous cannula inserted and an infusion of compound sodium lactate solution started. Sedation was achieved with midazolam and alfentanil. Local anaesthetic was infiltrated into the skin 8 cm lateral to the first lumbar vertebra bilaterally, and the two needles were then advanced at an angle of 45” to the skin, towards the midline, to strike the first lumbar vertebra. This was verified by X ray screening. The needles were then withdrawn before they were advanced at a more acute angle to the skin, with a slight cephalad angulation, testing for ‘loss of resistance’ continuously. The positions of the two needles were then checked with anteroposterior (AP) and lateral screening, when they were seen to lie 1.5 cm anterior and slightly lateral to the cephalic end of the first lumbar vertebra. The positions of the needles were

M.J. Woodham, FFARCS, Research Fellow, M.H. Hanna, FFARCS, Consultant, Pain Relief Unit, King’s College Hospital, Denmark Hill, London SES. Correspondence should be addressed to Dr Hanna please. Accepted 6 December 1988. 0003-2409/89/060487

+ 03 $03.00/0

@ 1989 The Association of Anaesthetists of Gt Britain and Ireland

487

M.J. Woodharn and M.H. Hanna satisfactory and a 2-ml dose of radio-opaque dye (omnipaque 240) was injected down both needles consecutively.The spread on the left was satisfactory and conformed to the classical picture; however, on the right the spread of dye did not conform to the classical picture. It was decided to inject 90% alcohol 30 ml on the left and 0.25% bupivacaine 20 ml on the right. It is our normal practice to prevent hypotension prophylactically; therefore ephedrine 3 mg was given intravenously and a further 15 mg intramuscularly at the same time as the injection of alcohol. There were no significant changes in either arterial blood pressure or pulse rate during the procedure or after the blockade. The patient was transferred back to the ward, and at this point she was still drowsy but responding to commands. Three hours later, she complained of numbness and weakness of her legs. A full neurological assessement showed minimal power in the legs with a total flaccid paralysis, except for a flicker of movement in the left second toe. There was a sensory level to pain, vibration and temperature at TI,,,,, but there was partial preservation of sensation to light touch and a proportion of stimuli were sensed, in particular over the sacral dermatomes. The epigastric pain was relieved after the procedure. A myelogram performed the next day was normal. A presumptive diagnosis of ischaemia of the cord secondary to damage to the artery of Adamkiewicz was made. There was minimal change in the patient’s neurological condition in the 2 weeks after the procedure, and on her discharge she had no bladder control and minimal muscle power in her legs. She returned to the West Indies and died 2 months later.

Discussion The vascular supply of the spinal cord was first studied in 1882 by Adamkiewicz, and it was he who first described the arteria radicularis magna of Adamkiewicz (the largest of the anterior medullary feeders for the supply of the lumbar spinal cord). This area has since been studied further by other workers, and the artery of Adamkiewicz is known to have a variable entry site to the spinal cord; thus it may occur at any level between T, and L, although it is usually found at T, to T , , . 2 , 3The artery of Adamkiewicz occurs on the left in 80% of cases. It is interesting, and may be relevant, that alcohol was only injected on the left in this patient. The artery of Adamkiewicz is generally held to be important, but in some animal studies it has been ligated without any subsequent neurological consequences. There is a degree of variation between patients as to the number of anterior feeder arteries. This may vary from two to 17, although there are rarely more than nine.3 Obviously the greater the number of other anterior medullary feeders the less critical will be the damage to the artery of Adamkiewicz. It may be that in older patients with a degree of arteriosclerosis the system of collaterals is compromised, thus increasing the importance of this artery. Direct injection of alcohol into the cerebrospinal fluid (CSF) was considered extremely unlikely, since lateral and AP X ray screening was used and there were no neurological signs above T , , after the procedure. Had

alcohol been injected directly into the CSF then some spread of alcohol cephalad to the site of injection would be expected. It is also unlikely that alcohol was injected directly into the artery of Adamkiewicz, since no blood was aspirated before injection, and on screening one would expect rapid disappearance of the dye if the needle was situated in the artery itself. The neurological damage in this case was probably due to spasm of the artery of Adamkiewicz in response to irritation caused by alcohol injected in close proximity to it, or mechanical damage by one of the needles, most probably the one sited on the left. There have been three previous case reports of neurological lesions after coeliac plexus Galizia and Lahiri in 19744 reported a case that involved the use of phenol which led to complete paraplegia, with a sensory level at T12. X ray screening in this case was not used and there was a rapid onset of leg weakness at the time of injection. This was postulated to be secondary to an ischaemic injury to the spinal cord, but the most likely cause was damage to the artery of Adamkiewicz. Thompson in 19775 reported on a series of 100 coeliac plexus blocks which included one case of partial leg paralysis. This procedure was performed without the use of X ray screening and in the lateral position. The most likely cause of this neurological lesion was believed to be direct irritation of the lumbar nerve roots by alcohol. The last and most recent case report that involved neurological damage after coeliac plexus block was by Cherry and Lamberty in 1984.6Thisagain resulted in paraplegia with a sensory level at T I , _ , , . The sensory loss was variable whereas the motor deficit was profound. X ray screening was used in this last case, and once again the postulated cause of the neurological lesion was ischaemia of the cord secondary to damage (or spasm) of the artery of Adamkewicz. There was a period in this patient of 2-3 hours after the procedure before the onset of weakness and numbness. The patient received a general anaesthetic and the procedure was performed with the patient in the right lateral position. It seems that there have been two previous case reports of ischaemic cord damage after injury to the artery of Adamkiewicz. It is difficult to say if the damage to the artery is secondary to injection of alcohol (or phenol) leading to spasm, or direct trauma by the needles used. X ray screening does not seem to reduce the chances of this occurring. The syndrome that is seen after injury to the artery of Adamkiewicz has certain characteristics, in particular dense motor weakness and variable sensory deficit, with a level around T, 2. This adverse complication of coeliac plexus block is fortunately rare and there have been few reported cases in the literature. It may happen in the hands of the most experienced of operators, and cannot be prevented by the use of X ray screening, or probably even the use of a CAT scanner. It could however be postulated that a coeliac plexus block performed using one needle on the right might be as effective as the bilateral approach and avoid this rare but serious complication. Acknowledgments The authors thank Sister J. Latham and Dr S. Peat for their help with the preparation of this manuscript.

Paraplegia after coeliac plexus block References 1. THOMPSON GE. Celiac plexus, intercostal and minor peripheral

blockade. In: COUSINSMJ, BRIDENBAUGH PO, eds. Neural blockade in clinical anaesthesia and management of pain. Philadelphia: Lippincott, 1980: 384404. 2. GUTTMANN L. Anatomical data on vertebral column and spinal cord. In: GUTMANN L. Spinal cord injuries: comprehensive management and research. Oxford: Blackwell Scientific, 1973: 47-70. 3. DOMMISSE GF. The arteries, arterioles, and capillaries of the

489

spinal cord: surgical guidelines in the prevention of postoperative paraplegia. Annals of the Royal College of Surgeons of England 1980; 62: 369-76. 4. GALIZIAEJ, LAHIRISK. Paraplegia following coeliac plexus block with phenol. Case report. British Journal of Anaesthesia. 1974; 46: 53940. 5. THOMPSON GE, MOOREDC, BRIDENBAUGH LD, ARTIN RY. Abdominal pain and alcohol celiac plexus block. Anesthesia and Analgesia 1977; 5 6 1-5. 6. CHERRY DA, LAMBERTY J. Paraplegia following coeliac plexus block, Anaesthesia and Intensive Care 1984; 1 2 59-61,