Aseptic Meningitis Due to Trimethoprim-Sulfamethoxazole.

normal but electroencephalography revealed non-specific widespread slow-wave activity without focal features or triphasic complexes. The first CSF examination was acellular with an elevated protein of 0 93 g/l. A traumatic repeat lumbar puncture two days later, revealed 13,000 red blood cells, 280 white cells (90% polymorphs) and a protein of 2 7 g/l., the CSF/serum glucose ratio remaining normal. Bacteriological and viral cultures from a variety of sites were unhelpful and examination of paired serum and CSF revealed no rise in complement fixating antibodies to HSV1. He continued to deteriorate and developed right sided focal seizures. A further electroencephalogram 12 days after admission revealed sharp activity in the left temporal lobe while a CT scan showed a large area of low attenuation in the left frontal and temporal lobes with considerable mass effect. He was started on intravenous dexamethasone and acyclovir but died the same day. At necropsy, the heart was unremarkable, the coronary and intracranial vessels were patent and histological examination of the sinus node, the AV node and bundle of His showed no abnormality. The brain was swollen with evidence of softening of both temporal lobes and superficial petechial haemorrhages were seen both over the inferior aspect of the temporal lobes and the diencephalon. Histological examination revealed neuronal loss, perivascular lymphocytic cuffing, necrosis and the presence of macrophages and reactive astrocytes. These changes, typical of a viral encephalitis, were most marked in the temporal lobes but were also present in the diencephalon, hypothalamus and to a lesser extent in the brainstem. Positive immunoperoxidase stains for HSV I were obtained from both the temporal lobes and the pons, the causative agent being confirmed when the virus was isolated from the left temporal lobe. This patient with proven herpes simplex encephalitis ultimately developed typical radiological and pathological evidence of temporal lobe involvement. The case is of interest, however, because of the unusual presentation and progression of the illness. The presence of coma with intact respiration and brain stem reflexes but small reactive pupils strongly suggests that the infection initially affected the diencephalon and midbrain only. This may also be the explanation for the low serum sodium concentration since damage to the hypothalamus is a cause of inappropriate secretion of anti-diuretic hormone. Disturbance of cardiovascular control is a well recognised feature of diencephalic syndromes4 and there has been a recent report of herpetic brainstem encephalitis with recovery' in which labile hypertension, marked tachycardia, profuse sweating and mild hypothermia were attributed to diencephalic involvement. No primary cardiac cause for the arrhythmia could be demonstrated and in the absence of early temporal or brainstem involvement it