The case

A 61-year-old Libyan man was admitted under a medical team with a 3-month history of lethargy and mild generalised weakness. He reported minimal right ear hearing loss and impaired balance. His family reported mild irritability during this period.

He had recently started a statin for hypercholesterolaemia but had no other significant medical history. He was a self-employed businessman who had moved to the UK 15 years earlier.

On initial examination, cognition was normal. Cranial nerve examination identified minimal right hearing loss. He had a broad-based gait and could not perform tandem gait. Romberg's test was negative. He had minimal (Medical Research Council scale for muscle power (MRC) 5-/5) weakness in the small muscles of his right hand and a very mild asymmetric pyramidal weakness in his legs, with the right leg slightly weaker. Reflexes were preserved throughout with the exception of absent ankle jerks. Sensation remained intact in all four limbs. Neurological examination was otherwise normal.

Initial investigations

He had a mild normocytic anaemia, normal liver function tests, urea and electrolytes, vitamin B₁₂ and folate and negative antinuclear antibody, anti-neutrophil cytoplasmic antibody and Venereal Disease Research Laboratory (VDRL). Inflammatory markers including erythrocyte sedimentation rate and C reactive protein were both normal. MR scan of brain without contrast was normal. MR scan of the whole spine without contrast showed minor cervical spondylosis without cord compression or signal change.

He was discharged with a request for an outpatient neurological opinion.

Four weeks following discharge, he was readmitted with deteriorating mobility and increased weakness. He had more significant weakness in the small muscles of his right hand with mildly reduced upper limb reflexes. He had a waddling gait. Both quadriceps muscles had become wasted and there were fasciculations in the right quadriceps. Leg tone was flaccid. Power had decreased dramatically in both legs, in a pyramidal distribution. Leg reflexes were absent with initially flexor plantar responses, though these responses later became unresponsive. Sensation remained unaffected.

Further investigations

Repeat MR scan of whole spine was normal. Nerve conduction studies were normal. Needle electromyography showed chronic partial denervation changes in both lower limbs (right more than left) with positive sharp waves and fasciculation potentials. There were similar but less severe changes in both arms. The findings suggested a subacute proximal neurogenic process, most likely affecting the anterior horn cells or motor roots. Table 1 gives his cerebrospinal fluid (CSF) results.

We gave him empirical intravenous ceftriaxone and aciclovir. More detailed systemic enquiry identified a several month history of low-grade headaches and intermittent chills. These first appeared following a trip to Libya. While an inpatient he had an intermittent low-grade fever but blood cultures were repeatedly negative. HIV testing was negative. There was no history of contact with either tuberculosis or polio.

Table 2 gives the result of his second lumbar puncture 1 week later.

His weakness increased and he lost anti-gravity power in his leg muscles, other than knee extension and ankle plantar flexion. Leg reflexes became absent. Plantar responses were mute.

CSF India ink and acid-fast bacilli staining, herpes simplex virus/varicella zoster virus/cytomegalovirus/Epstein-Barr virus/enterovirus PCR, cryptococcal antigen and tuberculosis PCR were all negative. A Heaf test was strongly positive (grade IV necrotic reaction). We started him on empirical anti-tuberculosis quadruple therapy (rifampicin, isoniazid, pyrazinamide and ethambutol) with prednisolone 80 mg daily (tapered slowly over 6 weeks) and pyridoxine. His weakness plateaued but did not improve.

Given the CSF cytology findings, we looked for an occult malignancy. Tumour markers, whole body CT-positron emission tomography and bone marrow examination were normal, as was repeat CSF cytology.

Further microbiological and virological tests including Borrelia and Brucella serology (complement fixation test) and micro-agglutination assay were performed at the National Reference Laboratory. Human T-lymphotropic virus and schistosomiasis serology, enterovirus PCR (stool sample), atypical pneumonia screen (Q fever, chlamydia and mycoplasma) and anti-neuronal antibodies, all returned negative. Prolonged CSF tuberculosis and fungal cultures were both negative. CSF ACE was elevated at 2.84 U/L (normal <1.2). MR scan of whole spine with contrast showed diffuse enhancement of some nerve roots within the cauda equina of uncertain significance but was considered at the time to be related to multiple lumbar punctures.

Further lumbar punctures showed a rising CSF white cell count (peaking at 217 monocytes), an absence of atypical cells and unmatched CSF oligoclonal bands. CSF West Nile virus PCR was negative. CSF immunotyping found no monoclonal expansion of lymphocytes.

Given the ongoing suspicion of neurological infection, a further opinion was sought from HM. During this consultation, the patient was further questioned about his trips to Libya. He had always stayed in Tripoli but had drunk unpasteurised milk purchased from a small dairy shop, of the type commonly found in Libya. The onset of his presenting symptoms were closely related to his most recent trip to Libya. As such, despite the recent negative serological testing, HM strongly suspected neurobrucellosis, which remains endemic in Libya. Further blood and CSF samples were sent to the National Reference Laboratory for retesting. Both CSF and serum ELISA returned strongly positive for Brucella IgG, providing serological confirmation of neurobrucellosis. Subsequently, the patient's original blood sample was reanalysed and found to be strongly positive. An investigation disclosed that there had been a simple administrative error, resulting in an incorrect result being released for the initial sample (table 3).

Following receipt of these results, we started him on oral doxycycline 100 mg twice daily and continued the rifampicin 300 mg twice daily. We stopped all his other anti-tuberculous treatment. Within 1 week his headache and intermittent low-grade fever resolved and his leg muscle strength improved (table 4).

With additional prolonged neurorehabilitation, he went on to make an excellent recovery without residual weakness. Six months after adding doxycycline, he had mild sensory ataxia, normal limb muscle bulk and power, bilateral extensor plantar responses and minor proprioceptive loss in both big toes. Hearing and cognition were normal. His serum vitamin B₁₂ level was at the lower limit of normal and so he was loaded with intramuscular vitamin B₁₂. Spinal MRI remained normal. Repeat electromyography showed mild-to-moderate chronic partial denervation with re-innervation in leg muscles. He completed a total of 6 months of combined doxycycline and rifampicin, preceded by 3 months of anti-tuberculous quadruple therapy, including rifampicin (table 5).