Electrical stimulation of the nervous system with the aim of treating chronic pain is performed at different levels of the neuraxis, e.g. the peripheral nerve, trigeminal ganglion and/or root, spinal cord, thalamus and motor cortex. Shealey and colleagues first described the use of spinal cord stimulation in 1967 . We report on a case of subacute epidural abscess 1 yr after spinal cord stimulator implantation and explain the clinical features in order to highlight the significance of the radiological diagnosis.
A 59-yr-old male had a fall and was admitted to hospital because of unrelieved sciatic pain. Magnetic resonance imaging (MRI) revealed an L5-S1 disc herniation as the source of pain and lumbosacral spinal surgery was performed. Surgery consisted of an L5-S1 decompression (lumbar discectomy). A few weeks later, he developed intense low back pain with right episodic sciatica. A computed tomographic scan, MRI and myelography showed periradicular fibrosis and arachnoiditis with right-sided predominance.
Two years later, the patient consulted the pain clinic complaining of chronic low back pain resulting from 'failed back surgery syndrome'. The pain was controlled by medical treatment (analgesics, opioids, adjuncts), epidural sessions (local anaesthetics, corticosteroids) and transcutaneous electrical nerve stimulation (TENS).
Another 2 yr later, the low back pain became uncontrollable and it was decided by anaesthesiologists and neurosurgeons to perform an L4-L5-S1 stabilization procedure. The immediate postoperative course was good. The low back pain improved, but still the episodic sciatica and radicular pain resulting from paraesthesia in the right foot and numbness in the heel persisted. The pain was successfully controlled with TENS and medical treatment. A few months later his situation worsened with invalidating low back pain again, dysesthesia in the left leg and gemellus neurological amyotrophy. Later, and with clear evidence of failed back surgery syndrome, it was decided to implant a T10-T11 spinal cord stimulator. Before the operation, the patient underwent a baseline quantitative evaluation of functional capacities by a psychologist. The system was implanted through a standard laminectomy, and placing a temporary electrode under fluoroscopic control for a routine 7 day trial. The patient reported an estimated 80% relief of pain (medication was maintained) and his physical activity improved. In a second operation, the definitive implant (Medtronic-Itrel 3®; Medtronic ibérica, Barcelona, Spain) was introduced through the laminectomy. The procedures were performed by an experienced consultant under aseptic conditions and with antibiotic prophylaxis (cefotaxime 1 g intravenously (i.v.)). The implant was placed with particular care in order to prevent paraparesis and paralysis (Fig. 1).
One month later, the spinal cord stimulation was ineffective owing to tolerance and it was necessary to increase the medication. The patient described severe pain in the site of the pulse generator with hyperalgesia and allodynia and without inflammatory signs, and he wanted the implant to be removed. On removal, the site was surgically explored and wound infection was observed, and septic material sent for cultures was negative. After placing a Penrose-type drainage system in the wound, ciprofloxacine 750 mg i.v. was administered every 8 h for 1 week. Given that the patient had responded well to spinal cord stimulation, and with the agreement of the patient and accepting the risk of infection, it was decided to remove the generator and the wire and leave the epidural electrode in place.
Six months later the patient consulted again for progressive leg weakness (paraparesis), functional impotence of the right psoas muscle, dysesthesia in the right leg and physical disability. He had apyrexia and no sphincteral dysfunction. He was unable to flex his right hip in the physical examination. The cardiopulmonary and abdomen examinations were unremarkable. The laboratory studies showed a white blood cell count of 10.8 × 109L−1 and a C-reactive protein of 4.8 mg dL−1. Chest and abdominal radiographs were normal. The computed tomographic scan showed a density increase in the epidural posterior space (T12) extending into the paravertebral space (Fig. 2). MRI showed an inflammatory process at the right psoas with muscular atrophy and did not detect an intraspinal space-occupying lesion. Electromyography showed L5-S1 irritation and denervation. Electromyography showed L5-S1 irritation and denervation confirming the existence of myeloradicular compression. As an empyema was suspected, a decompressive T11-T12 laminectomy was performed, and when the electrodes were removed, pus and fibrotic material were revealed. Staphylococcus aureus resistant to methicillin and Pseudomonas aeruginosa were found after being isolated from the tissue and cultured. The therapy was changed to ciprofloxacine 750 mg i.v. 12 hourly. The patient slowly recovered from surgery, showed improvement in the clinical symptoms and was discharged after 15 days of antibiotic therapy with a residual paresis of the right psoas muscle.
The indications for spinal cord stimulation include syndromes causing neuropathic pain in the extremities, axial pain related to mechanical factors such as spinal stenosis, mixed pain conditions (e.g. failed back surgery syndrome), intractable angina pectoris and peripheral vascular disease . The pain resulting from failed back surgery syndrome is difficult to treat. Although some patients develop fibrosis in the operated nerve root after disc herniation, there is not always an association between epidural fibrosis and sciatica in the lumbar postdiscectomy syndrome. Spinal cord stimulation can be considered if the pain persists  despite conservative treatment and there are no significant psychological antecedents and co-morbidity. Complications associated with column stimulator implants may be technical or functional. Many authors state that the technique does not share the risks of surgery; however, this procedure is not risk free and can occasionally be life threatening . Technical and functional complications after the implantation of spinal cord stimulation devices have been reported . Technical complications are defined as those occurring in the immediate postoperative period due to problems associated with the surgical procedure (e.g. transitory or severe pain, seromas or leakage of cerebrospinal fluid). These are mainly difficulties with the hardware such as a displaced electrode, treated by reoperation and repositioning, or hardware malfunction. Fortunately, the possibility of changing electrode combinations non-invasively to optimize the topography or stimulation reduces the number of displacements requiring further procedures. Other complications recorded are superficial infections (successfully treated by antibiotics or removal), subcutaneous haematomas, cerebrospinal fluid leakage (spontaneously resolved) and psoas abscess. Serious neurological injuries requiring surgical intervention have been reported but never as a subacute event. Previously reported functional complications are mostly based on stimulation failures.
The correct implantation of the spinal cord stimulation device is a prerequisite for a successful outcome. The implantation of the electrode should be performed percutaneously or through a small laminectomy under fluoroscopy in strictly aseptic conditions. Antibiotics should be given before, during and after operation to reduce the risk of wound infection and meningitis. The clinical symptoms of spinal epidural abscess are pain, fever and rapidly progressing weakness. Pain is usually present either along the spine or is radicular. The duration of pain before the appearance of symptoms is generally 2 weeks or less, but in some chronic cases, it may be several months or longer. Risk factors include impaired immune status (e.g. diabetes mellitus, renal failure, alcoholism, malignancy), intravenous drug abuse and infections of the skin or other tissues. Two-thirds of epidural infections are the result of haematogenous spread from the skin, soft tissue or deep viscera. One-third results from the direct extension of a local infection to the subdural space. Most cases are due to S. aureus, although Gram-negative bacilli, Streptococcus, anaerobes and fungi can also cause epidural abscesses.
In the present case, the patient had no risk factors and the epidural abscess developed as a subacute process 1 yr after spinal cord stimulator implantation. Perhaps the original operative cultures were falsely negative. The presence of a foreign body and the growth of S. aureus (resistant to methicillin) in purulent material cultures suggest that the contaminating agent was introduced by the catheter from the skin. It is not known whether the skin was the original focus of the infection, which would seem unlikely 1 yr after the last surgical procedure, or the result of a haematogenous dissemination from other tissue. Screening for methicillin-resistant S. aureus colonization should be considered in patients at high risk of this infection because the result could affect the choice of antimicrobial prophylaxis and empirical therapy. We took a calculated risk and decided not to remove the epidural electrode in the hope of avoiding another invasive procedure and possibly attaching a new generator.
In conclusion, even when a spinal cord stimulation lumbar catheter is implanted following all the appropriate steps, it is important to consider the possibility of subacute epidural abscess as a complication.
Hospital Universitari de Girona 'Doctor Josep Trueta'; Department of Anesthesia and Pain Therapy; Girona, Catalonia, Spain
1. Shealey CN, Mortimer JT, Reswick JB. Electrical inhibition of pain by stimulation of dorsal column: preliminary clinical report. Anesth Analg
2. Andersen C. Complications in spinal cord stimulation for treatment of angina pectoris. Differences in unipolar and multipolar percutaneous inserted electrodes. Acta Cardiol
3. Marwalder TM, Battaglia M. Failed back surgery syndrome. Part I: Analysis of the clinical presentation and results of testing procedures for instability of the lumbar spine in 171 patients. Acta Neurochir
4. Markwalder TM, Reulen H. Diagnostic approach in instability and irritative state of a 'lumbar motion segment' following disc surgery - failed back surgery syndrome. Acta Neurochir
5. Gutiérrez V, Solans R, Pigrau C, Pahissa. Psoas abscess due to an epidural spinal cord stimulation catheter. Scand J Infect Dis