1708 ing tremor was observed. Laboratory investigations were normal except for brucella agglutination titres of 1/80 and nonagglutinating antibodies titre of 1/1280 in the serum. CSF protein, glucose, cells were normal and cultures were negative. Brucella agglutination titres were only slightly positive (1/8). CT scan of the head disclosed no abnormality. Brucella melitensis was isolated from two blood cultures performed when the patient had no fever. After levodopa and bromocriptine withdrawal, treatment with doxycycline by mouth 100mg bid, streptomycin 1 g IM daily and rifampin 600mg daily was introduced. One month later, the patient was able to walk without assistance, but still had a mild hypokinesia and resting tremor. Doxycycline and rifampin were continued for 2 months, and the patient was discharged free of symptoms. According to Roger and Poursine's classification, our patient suffered a late onset neurobrucellosis presenting as Parkinsonism. Though considered rare, extrapyramidal forms of neurobrucellosis have been well recorded, especially in Italian publications4; amongst these were two cases of subacute hemiparkinsonism.4 I Unlike ours, both may be included in the early form of Roger and Poursine's, their symptoms having appeared during the course of brucellosis. In both cases, right hemiparkinsonism persisted because treatment was limited by streptomycin toxicity and the lack of new antibiotics at the time. The exact pathogenisis of Parkinsonism in brucellosis is still unknown; there is a lack of pathological studies. Gianelli4 suggest a focal encephalitis with circumscribed symptoms, related to concurrent treatment (there being normal CSF), probably due to an idiosyncratic immunological reaction. In our opinion, Parkinsonism unresponsive to levodopa and bromocriptine suggests a post-synaptic dopamine receptor lesion, such as corpus striatum vascular inflammation similar to that observed in syphilitic Parkinsonism.6 Whether this reaction is immunological and not due to direct invasion by the organism, cannot be determined only on the basis of a negative CSF culture and agglutination titre, because serine pre-treatment CSF cultures obviously are not performed. Besides, CSF agglutination titres in chronic cases are usually low, and non-agglutinating antibodies or complement-fixation titres need to be included in adequate serologic study.2 Finally, we believe that previous treatment is the reason for the lack of CSF abnormality, as found in modern or partially treated neurosyphilis,7 especially in chronic and Letters has up until now only been reported either in patients with bilateral temporal lesions or with a unilateral lesion affecting the middle third of the superior temporal gyrus of the left temporal lobe.1 2 We have recently seen a patient with an agnosia predominantly for spoken words who had a unilateral right hemisphere temporo-parietal lesion demonstrated by computed tomography (CT). The patient, a 61 year old man who lived alone, noticed one morning when he turned on his radio that he was no longer able to recognise sounds; he could not make sense of what voices were saying and could not recognise any musical tunes. He was a keen CB radio operator and reported that "it was as though the audio was quite correct and loud but I could not make out who was A MOLINS speaking. It was as though some of the freJ MONTALBAN quencies had been chopped out". Although A CODINA spoken conversation with the patient was The Department ofNeurology, almost impossible because of his problem Vall d'Hebr6n Hospital, interpreting speech, he was able to continue Barcelona, his work as a commercial artist without any Spain. apparent problems. A friend eventually persuaded him to seek medical advice. Address for correspondence: Albert Molins, Although there was no history of any preAv/Meridiana 30-32 Esc B 3' 1i 08018 Barcelona, vious illness, hypertension or cardiovascular Spain. disease, an electrocardiogram showed changes of a previous, apparently asymptomatic, myocardial infarction. On examReferences ination the blood pressure was 110/70 mm Hg and the abnormalities found were I Hughes ML. Mediterranean, Malta or to the nervous system. The confined Undulant Fever. London: McMillan and patient's spontaneous speech was fluent, Company Ed. 1897. 2 Pedro Pons A, Foz M, Codina A. Les Neu- appropriate and comprehensible, though robrucelloses. Cahiers de Medicine (Europa with occasional paraphasic errors. He was Medica) 1972;13:855-62. able to give a reasonable account of his ill3 Sahs AL. Brucellosis. In: Handbook of Clinical ness. In contrast, his comprehension of the Neurology 1978;33:305-26. spoken word was restricted to simple direct 4 Gianelli A. Contribution to the study of questions about his age, name and Parainfective hemineurobrucellosis. though comprehension parkinsonism with chronic development. Riv occupation, improved if he was aware of the context of Pat Nerv Ment 1959;80: 1135-52. 5 D'Amore. Emisindrome parkinsoniana nel the question. He sometimes understood corso della febbre di Malta. Morgagni other direct questions such as "how are you?", but he could not reliably obey simple 1933;75: 1208. 6 Guillain G. Les syndromes Parkinsoniens commands such as "sit down" or "stand d'origine syphilitique. Progres Med up". Repetition of spoken material was 1 935;23:978. poor and limited to three digits and two sim7 Hotson JR. Modern neurosyphilis: a partially ple words. He could understand the written treated chronic meningitis. Medical word normally, responding rapidly and progress. West J Med 1981;135:191-200. 8 Sanz J, Puras A, Alguacil A, Alonso F, Franch accurately to complex written commands. 0, Damiano A. Neurobrucelosis. Rev Clin Object naming and reading out loud were slightly impaired. He had occasional Esp 1985;176:349-53. difficulty recognising non-verbal sounds Accepted 13 July 1987 such as rattling keys and tearing paper. He reported that he was no longer able to appreciate music or recognise familiar melodies. Visuospatial and constructional abiliPure word deafness and unilateral right ties were normal, he was able to perform the temporo-parietal lesion: a case report block design test of the WAIS without difficulty and perhaps more importantly, his Sir: The syndrome of pure word deafness employers had not noticed any impairment encephalitic forms. All these considerations lead us to recommend serine blood cultures when suspecting brucellosis in patients with unusual neurological manifestations and normal or inconvincing CSF changes, even if fever or other general symptoms are absent. Treatment with two, three or even four drugs, including those with high CSF penetration (rifampin, trimethoprimsulfamethoxazole), seems to be desirable8 but there are neither comparative studies nor large series to verify the best drug combination. We conclude by emphasising the importance of this sometimes underestimated disease, with a marked CNS affinity, which is still far from being eradicated. Letters 1709 alised epileptic discharges are evoked by intermittent photic stimulation, constituting a photoconvulsive response.3 Sodium valproate is the drug of choice for photosensitive epilepsy. Three patients are reported who developed photosensitivity related to sodium valproate withdrawal. The first case was a 47 year old man who had had four generalised tonic-clonic convulsions, nocturnal or on awakening, from 18 years to 24 years of age. He was seizurefree until 36 years of age, off phenobarbitone therapy since aged 29. Sodium valproate treatment (15 mg/kg/die, plasma level 73-3 ug/ml) was started at 36 years of age because of a relapse with a morning generalised convulsion. EEG initially showed generalised spike and wave discharges at rest and during hyperventilation and then normalised. Intermittent photic stimulation always had no effect. There was gradual sodium valproate reduction at 45 years of age. Six months later, on 5 mg/kg/die of sodium valproate (plasma level 18 5 ug/ml) for 3 weeks, he suffered from two tonicclonic convulsions when watching television. EEG showed generalised epileptic discharges, evoked by intermittent photic stimulation. Sodium valproate dosage was increased to 10 mg/kg/die (plasma level 44 0 pg/ml). He was seizure-free for the following year and photoconvulsive response disappeared. A 32 year old man had two morning generalised convulsive seizures at 20 and 22 years of age. EEG showed generalised discharges during hyperventilation but no paroxysmal response to intermittent photic stimulation (fig). Sodium valproate therapy and left hemisphere lesions. In: Mountcastle (13 mg/kg/die, plasma level 62-7 pg/ml), VB (ed), Interhemispheric Relations and started after the second fit, controlled the Cerebral Dominance. Baltimore, Johns Hopseizures and normalised the EEG. Sodium kins 1962. 6 Milner B. Interhemispheric differences in the valproate was gradually stopped at 30 years localisation of psychological processes in of age. Three weeks after drug withdrawal he had a tonic-clonic convulsion while using man. Br Med Bull 1971;27:272. 7 Shankweiler DP. Performance of brain- a video-computer. EEG showed generalised damaged patients on two tests of sound epileptic discharges exclusively during interlocalisation J Comp Physiol Psychol mittent photic stimulation (fig). On pheno1961;54:375. 8 Spreen 0, Benton AL, Fincham W. Auditory barbitone treatment (15 mg/kg/die) there agnosia without aphasia. Arch Neurol was no seizure recurrence and no abnormal response to intermittent photic stimulation 1965;13:84-92. for the following 2 years. The third case was a 16 year old girl who Accepted 26 June 1987 otherwise neurologically normal had two nocturnal clonic seizures involving the right half of the face at 8 and 9 years. CT of the Photosensitivity related to valproate with- head was normal. EEG showed left centrotemporal sharp waves but no photodrawal convulsive response. EEGs of the parents Sir: Photosensitive epilepsy comprises and the older sister were normal. Sodium patients in whom seizures are provoked by valproate therapy (20 mg/kg/die), started environmental flicker stimulation. 1 2 Gener- following the first seizure, was gradually in his drawing skills. Handedness was right, presumably non-dominant, hemiassessed by informal questioning about sphere in the fronto-temporo-parietal areas. hand, foot and eye preference for various In contrast, our patient had only a single activities, direct observation and the results clinical episode, a more restricted neuroof the Handedness Inventory.3 All measures psychological deficit and a smaller lesion. used indicated strong right preference with Nonetheless, although the relationship of no evidence of left hand bias. There was no the lesion to the deficit appears much family history of left handedness. Left hemi- clearer, other possibilities must be considsphere language representation was indi- ered: it is possible that there was infarction cated. His IQ was estimated to be above of the contralateral hemisphere, not visible average. Pure tone audiometry was not per- on the CT scan; it is also possible that the formed. Bedside testing of hearing was nor- patient was right hemisphere dominant for mal. The CT scan, performed 3 months after language (or had bilateral language repreonset, showed an infarct which was predom- sentation), despite his strong right-handed inantly in the right parietal region, with preference. However, since all these alternasome extension inferiorly into the posterior tive explanations seem unlikely, we conclude temporal lobe (fig). There was no evidence that the right hemisphere infarct was the of any abnormality in the left hemisphere. cause of our patient's pure word deafness. We postulate that our patient's word MARK ROBERTS deafness was due to the right temporoPETER SANDERCOCK parietal lesion seen on CT. Patients with Fazakerly Hospital, Liverpool tumours of, or injury to, the right hemiERIC GHADIALI sphere may develop "crossed" aphasia, Walton Hospital, though such lesions may cause aphasia by Rice Lane, pressure or distortion of the left hemiLiverpool L9 IAE, UK sphere.4 Lesions in the non-dominant temporal lobe have been reported to produce an References agnosia for non-verbal sounds with pre- I Buchman AS, Garron DC, Trost-Cardamone served comprehension of the spoken word' JE, Wichter MD, Schwartz M. Word deafand impaired appreciation of music.6 ' ness: one hundred years later. J Neurol NeuSpreen8 described a man with an agnosia for rosurg Psychiatry 1986;49:489-99. non-verbal sounds, but intact language 2 Kirshner HS. Behavioural Neurology. A Practical Approach. New York; Churchill Livcomprehension. He had had a stroke causingstone 1986. ing a marked left hemiparesis and subGG, Nebes RD. Patterns of hand prefsequently developed emotional lability, sei- 3 Briggs erence in a student population. Cortex zures and bilateral pyramidal signs. 1975;11:2302.8. Handedness was not tested. At necropsy, the 4 Benson FD. Aphasia, Alexia and Agraphia. patient had bilateral subdural haematomas New York; Churchill Livingstone 1979:145. and a very extensive area of infarction in the 5 Hecaen H. Clinical symptomatology in right I. fe. I~~~~~~~~~~~~~~~~~~~~~~_K- te_Mpoo-aietl area f _ | m