Abstract

A 63-year-old man developed acute transverse myelitis (ATM) with a rapid progression of sensory and motor deficits and autonomic dysfunction 2 days after chest surgery. Thoracic epidural anesthesia/analgesia (TEA) had been administered in this case. Since the temporal and spatial relationships between TEA and ATM are so close, one may easily mistake the TEA as the cause. Therefore, we discuss here the differential diagnoses for cord damage after TEA and the characteristics of ATM, and suggest that it is unlikely that TEA is the cause of ATM in this case.

Keywords

postoperative complications; thoracic epidural anesthesia; transverse myelitis;

1. Introduction

Thoracic epidural anesthesia/analgesia (TEA) has been a popular choice and one of the most effective means of providing perioperative analgesia after thoracic surgery. However, complications of TEA can result in severe spinal cord injury, including direct cord damage, epidural hematoma/abscess and neurotoxicity resulting from local anesthetics. Any neurological deficits in the postoperative period should be immediately evaluated for early management. In these cases, any neuraxial technique used, with close temporal and spatial relationships can easily be mistaken as the cause. However, this is not always the case. We present here a case of acute transverse myelitis (ATM) which manifested 2 days after thoracic surgery. In this case, TEA was adopted perioperatively.

2. Case report

A 63-year-old man, with American Society of Anesthesiologists physical classification II, was scheduled for thoracoscopic right upper lobectomy because of adenocarcinoma. His medical history was unremarkable except for being a heavy smoker for decades.

Non-intubated general anesthesia combined with TEA was adopted in this patient, as stated previously.¹ Briefly, a sensory blockade between T2 and T9 was achieved by thoracic epidural infusion of 2% lidocaine, and 1% propofol was intravenously infused to achieve a bispectral index value of between 50 and 70.¹

With 75 μg fentanyl intravenous injection and subcutaneous infiltration of local anesthetic, a 19-gauze polyamide epidural catheter was inserted via a 17-gauge Tuohy needle to the T6/T7 intervertebral space. The patient was fully aseptically prepared with chlorohexidine and surgically draped. The paramedian approach was adopted with the loss-of-resistance technique to confirm the epidural space. The epidural catheter was smoothly positioned in one attempt without unintentional wet or bloody tapping. The catheter was threaded 5 cm cephalad in the epidural space and then fixed. A test dose of 3 mL 2% lidocaine with 15 μg epinephrine was injected via the epidural catheter to exclude intravascular or intrathecal misplacement. The patient was then given intravenous propofol to achieve the targeted anesthetic depth and top-up titration of epidural 2% lidocaine injection to achieve adequate thoracic sensory blockade (T2–T9). The surgery was uneventful. There were neither episodes of hypotension (systolic arterial pressure was maintained at >90 mmHg), nor hypoxemia (SpO₂ was maintained at >94% via a facemask with oxygen) during surgery. Patient-controlled epidural analgesia (PCEA) was commenced at the end of surgery, using 0.1% bupivacaine and fentanyl (1.25 μg/mL) to maintain the visual analogue pain score below 3. The patient was comfortable with PCEA. There was no dysesthesia or weakness of his lower extremities.

He had low-grade fever (tympanic temperature: 38.3°C) which began in the morning of postoperative day (POD) 1. Although septic work-up was performed and a broad-spectrum antibiotic was upgraded accordingly, there were no infectious sources found, except that chest X-ray showed pulmonary atelectasis. However, the patient complained of numbness and hypoesthesia over the bilateral lower legs on the POD 2. Sensation to pain and temperature rapidly deteriorated with complete loss below T5 and partial loss below T1. The muscle power was also decreased to 0–1/5 of bilateral lower extremities with urinary incontinence. The PCEA was turned off after the advent of numbness and hypoesthesia and the epidural catheter was removed immediately. An emergent magnetic resonance imaging (MRI) scan of the thoracic spine showed a non-gadolinium-enhanced, ill-defined hyperintensity lesion within the cord at the level between T2 and T5 on the T2-weighted image (Fig. 1). Somatosensory evoked potentials (SSEPs) showed that the central conduction time from both median nerves was normal, but SSEPs from both tibial nerves were absent. Motor evoked potential (MEP) study by cortical magnetic stimulation was normal in both abductor digiti minimi, but was absent in the bilateral tibial nerves. Both SSEP and MEP studies were compatible with the diagnosis of thoracic myelopathy. In addition, hematological indices of inflammatory response showed leukocytosis (1.063 × 10⁹/L with segmented neutrophils: 86.4%, eosinophils: 0.2%, basophils: 0.1%, monocytes: 6.1%, lymphocytes: 7.2%) and an elevated level of C-reactive protein (18.9 mg/dL). Cerebrospinal fluid (CSF) analysis showed elevated white blood cells (84 cells/mm³, lymphocytes: 51%, neutrophils: 49%), mild elevated protein (65 mg/dL), glucose (78 mg/dL) and immunoglobulin G (IgG) index (0.928). The results of CSF for cytology and bacteria culture were negative. Immunological profiles, including antinuclear antibody, rheumatoid factor, C3 and C4 titers were all within normal limits.

Download full-size image

Fig. 1. Magnetic resonance imaging of thoracic spine in sagittal T2-weighted image showing an ill-defined hyper-intensity lesion within the cord at the level between T2 and T5 (between the arrows).

The diagnosis of ATM was made. High dose intravenous methylprednisolone therapy (30 mg/kg for 15 minutes, 5.4 mg/kg/h for next 23 hours and 1000 mg/day for the next 3 days) was commenced. After steroid therapy, his muscle power of bilateral lower extremities was partially recovered (2/5 and 3/5 of right and left legs, respectively), despite persistent sensory loss below the T2 level and autonomic dysfunction, as demonstrated by studies on sympathetic skin response and R–R interval variation. Four months later, with aggressive rehabilitative exercises, his motor function of lower extremities improved to 3/5 (right leg) and 4/5 (left leg), and at this juncture he was capable of ambulatory walking with a walker. However, he still suffered from complete sensory loss below T2, mild hyperalgesia between T11 and L1, and had persistent neuropathic pain over the right scapular and surgical wound sites. Gabapentin was thus prescribed for neuropathic analgesia. His autonomic dysfunction was partially recovered without fecal incontinence but he still needed urinary catheterization.

3. Discussion

Acute paraplegia in the early postoperative period is a challenging and potentially devastating anesthetic complication, which can occur when neuraxial techniques are used. Several clinical differential diagnoses are possible. Emergent MRI in this patient excluded cord compression by lesions such as herniated intervertebral discs, dislocated vertebral body or bony material, tumor, epidural abscess or epidural hematoma. According to the topographical distribution of neurological deficits, a thoracic intramedullary hyperintensity lesion (between T2 and T5) on the T2-weighted MRI scan and elevated IgG index in the CSF analysis, the diagnosis of thoracic ATM was established,² which was supported by both SSEP and MEP results.

ATM is a rare clinical syndrome with an incidence of one to four patients per million, per year.³ It is characterized by acute inflammation affecting the white and gray matter in multiple spinal cord segments, and causes rapid development of neurological deficits in motor, sensory, and autonomic functions. Motor weakness typically occurs initially in the lower legs and ascends to the waist rapidly, although bilateral arms may be involved occasionally. Loss of sensation in all modalities below a fixed level usually can be documented. Autonomic dysfunction varies, however, including increased urinary urgency, bowel or bladder incontinence, difficulty or inability to void, incomplete evacuation, or bowel constipation.⁴ Mostly, these neurological deficits progress to their nadir within hours or days.² The most common level in adults is the mid-thoracic region, whereas children have a higher frequency of involvement of cervical levels.⁵ Based on expert opinion of the Transverse Myelitis Consortium Working Group, causes of ATM can be classified as either idiopathic or disease-associated ATM.² The latter is usually a part of manifestations of systemic autoimmune disease, such as systemic lupus erythematous, multiple sclerosis, Sjogren syndrome, or neurosarcoidosis.⁶ The remaining idiopathic ATM is difficult to find the definite causes but is generally thought as a parainfectious or autoimmune-based neuroinflammatory process.² Nearly half of the patients diagnosed with idiopathic ATM have a preceding febrile event such as respiratory, gastrointestinal, or systemic illness.7, 8

The temporal relationship and proximity of an inserted thoracic epidural catheter for TEA to the affected regions of spinal cord easily raises the concern that TEA or local anesthetic neurotoxicity may be the cause of the neurological deficits.⁹ Although sporadic cases of ATM have been described following spinal¹⁰ or epidural11, 12 anesthesia and even general anesthesia without neuraxial invasion,¹³ none of them suggested a direct causal relationship after cautiously ruling out vascular and compressive etiologies of cord injury. Direct cord trauma either by needle or catheter tip was hardly possible in our patient since the catheter was inserted while he was fully conscious. Once the cord was inadvertently penetrated or injected with local anesthetic, he should be aware of the injury and reported his suffering at the first moment. Importantly, the patient was doing well after surgery with freely movable extremities until acute deterioration of spinal cord function on POD 2. Excessive sensory and motor blockade from epidural local anesthetics was excluded since the sensory and motor deficits were not reversed after removal of epidural catheter. Therefore, the progression of sensory, motor, and autonomic deficits, along with development of characteristic inflammatory spinal cord lesion determined by an MRI scan, strongly support the diagnosis of ATM.2, 10, 11, 12, 13 The causes of ATM in this patient in the early postoperative period may be secondary to an occult infection or by an immune-mediated neuroinflammatory mechanism. Since the autoimmune profile was normal, negative bacterial cultures may be explained by the aggressive use of postoperative antibiotics. Although high dose intravenous corticosteroid treatment to this situation is usually anecdotal, it was reasonable in this patient who yielded partial improvement in motor function.⁶ In spite of failure of identification of an infectious source, the cord lesion was evidently neuroinflammatory in etiology, which explained the effect of high-dose steroid therapy.

In summary, this report describes the presentation and course of a patient with ATM in the early postoperative period after TEA with an epidural catheter in situ. Using MRI, SSEP/MEP and CSF analysis, ATM was established based on the neuroinflammatory origin of spinal cord injury. Therefore, we suggest that ATM and TEA coincidentally occurred in this patient. For early differential diagnoses and prompt management to improve outcomes of this rare and devastating complication, awareness of the potential complications of TEA, continuous evaluation of neurological function after surgery and early MRI examination, if indicated, are warranted.

Acknowledgments

Attributed to the Department of Anesthesiology, National Taiwan University Hospital and National Taiwan University College of Medicine. Support was provided solely from institutional and/or departmental sources.