Phantom phenomena are common complications of limb amputations and may occasionally follow traumatic paraplegia and severe injuries of peripheral nerves. However, they have not been previously reported in patients with non-traumatic paraplegia. The following case history describes a patient with transverse myelitis resulting in complete paraplegia who experienced persistent movements and abnormal positions of her paralysed lower limbs. These findings suggest that disruption of the anatomical and functional integrity of the spinal cord may be the most important factor in the pathogenesis of phantom sensations.

A 61 year old woman presented with severe weakness of both legs, skin sensory loss and paraesthesia of the lower limbs, and bowel and bladder symptoms. She was well until 3 months earlier when she started to develop a tingling sensation and numbness over the outer side of her left leg. These symptoms gradually progressed and by the time she was admitted to hospital she had paraesthesia and sensory impairment of the whole of the left leg and in the distal half of the right leg. A month before admission she had become unsteady on her feet and developed urinary frequency, urgency of micturition, and constipation. There was also a rapidly progressive weakness of both legs, but no other symptoms.

Four years earlier the patient had had paraesthesia in both feet. This was thought to be due to peripheral neuropathy, but the diagnosis was not confirmed with neurophysiological tests. The symptoms resolved in a few weeks. The patient had a partial thyroidectomy for a nodular goitre 15 years ago. There was no other medical or family history of note. She was not taking any medication.

Physical examination confirmed the presence of complete flaccid paraplegia with skin sensory loss of all sensory modalities to the waist. The knee and ankle jerks were absent and both plantar responses were extensor. She had retention of urine and symptoms, signs, and radiological features of a paralytic ileus. The rest of the neurological and general physical examination was unremarkable. A full blood count, urea and electrolytes, and liver and thyroid function tests were within normal limits. An MRI of the cervical spine confirmed the presence of mild degenerative changes in the cervical spine at the level of C5-C7. There was no radiological evidence of an intrinsic or extrinsic cord compression or demyelination. However, the five distal segments of the thoracic cord appeared swollen and there was loss of the normal CSF rim ventral and dorsal to the cord on T1 weighted images. The T2 signal was prolonged and there was no contrast enhancement of the lesion. The appearances were considered consistent with oedema of the thoracic spinal cord. Brain MRI was normal. Visual evoked responses and brain stem auditory evoked potentials were within the normal limits. Somatosensory evoked potentials of the posterior tibial nerve could not be obtained because the patient developed severe myoclonic jerks of the entire leg at very low stimulus intensities. Her CFS protein concentration was 0.88 g/dl. No oligoclonal bands were detected on CSF protein electrophoresis. There were 2 lymphocytes/mm3 and four polymorphs/mm³. There was no bacterial growth on CFS culture.

Shortly after admission the patient started to experience phantom sensations in her lower limbs. At times she thought that her legs were crossed and on other occasions she felt that that she was standing on tiptoes. These symptoms were persistent and appeared to be spontaneous. Careful questioning did not disclose any specific stimuli. Their intensity remained unchanged until the patient was started on 200 mg carbamazepine three times a day. With this treatment the phantom sensations became less frequent and the images were less intense but they did not resolve completely. The paralytic ileus resolved with conservative treatment. However, the patient's neurological impairments remained unchanged until she was discharged from hospital 6 months later.

Non-painful phantom phenomena are continuous or intermittent sensations emanating from an amputated or deafferented part of the body. The missing or denervated part may be perceived in its premorbid shape, size, and other physical characteristics1 or in a distorted form.2 Patients often report normal functions associated with the absent organ—for example, penile erection, ejaculation, and orgasm after removal of the genitalia3 or voluntary or involuntary movements of an amputated limb. Often, sensations such as touch, pressure, and cold are experienced in the phantom organ. Phantom sensations often occur after limb amputations1 and have also been reported in about 15% of patients after a mastectomy.4 Sometimes they may follow spinal cord injury.5 However, their occurrence after transverse myelitis has not been previously reported.

Understanding the pathogenesis of phantom sensations is important for developing the appropriate treatment strategies. However, the mechanisms that underlie these phenomena are not fully understood at present. It has been suggested that they may be a manifestation of a psychological disorder or due to organic neurophysiological abnormalities.

Psychological factors such as denial or grief for the lost body part have been suggested as the cause of the postamputation phantom phenomena. However, this explanation is not supported by the currently available evidence. For example, the occurrence of phantom phenomena does not correlate with poor psychological adjustment or with the incidence of depressive symptoms in these patients.6Another hypothesis is that damaged peripheral somatosensory receptors fire spontaneously and give rise to the painful or abnormal experiences.7 However, phantom sensations have been reported by patients after spinal anaesthesia in the absence of damage to the peripheral nervous system.1 At present the neuromatrix theory8 offers the most plausible explanation for phantom sensations and pain.

According to this theory the symptoms associated with the phantom phenomena originate from genetically predetermined sensory images (or sensory engrams) that are stored in the cerebral cortex. It was postulated that the sensory images are triggered when neural impulses from the periphery are blocked. The patient reported here had complete “functional” transection of the spinal cord. The occurrence of phantom sensations in this patient was therefore independent of the neural input from the peripheral nervous system. This case provides further evidence that phantom phenomena are due to a central neurophysiological mechanism, probably triggered by impulses arising spontaneously from damaged spinal cord neurons. This is in accord with a previous report of structural and functional changes in the spinal cord in the acute stage after deafferentation.9

Ramachandran and Hirstein10 reviewed the studies of the topographical reorganisation of the cerebral cortex after limb amputations and concluded that the mechanism of phantom experiences is “remapping” of specific brain areas. The present study did not consider this question. However, the diversity of the illusionary experiences of movement reported by our patient suggests a more diffuse cortical reorganisation. This is more in keeping with the neuromatrix theory,9 and the presence of “diffuse neural matrix”.

The occurrence of phantom limb phenomena in patients with non-traumatic CNS lesions had also been previously described in a few patients with stroke. Halligan et al 11carried out a detailed study of a 65 year old man with severe left sided weakness, sensory loss, and hemianopia who, for several weeks, consistently reported a phantom (or supernumerary) third limb. Like our patient, he had good insight into his neurological deficits and his behaviour was completely rational, suggesting that the phantom experience was not a delusional belief but a direct result of organic brain damage.