A curious case of convulsion

Abstract

In July, 2007, an afebrile 61-year-old man presented with a fi rst episode of a generalised convulsion. Medical history included previous smoking and haemoptysis in the previous few weeks. Laboratory blood tests, viral serology, and blood cultures were unremarkable, except for a C-reactive protein of 35 mg/L (normal <5) and leucocytes of 17×109/L (3-9×109/L). MRI of the brain showed multiple bihemispherical foci, ring-enhanced with perifocal oedema, without midline shift (fi gure A). An opacity was seen on the chest radiograph, so CT was done, which showed a mass in the upper lobe of the right lung that was centrally liquefi ed, and modestly enlarged mediastinal lymph nodes. Dexamethasone and levetiracetam were given to prevent brain swelling and convulsions, and antibiotic treatment was initiated. Our patient improved over 3 days. We suspected he had lung cancer with metastases. At the meeting of the interdisciplinary tumour board, the diff erential diagnosis of cerebral abscesses was discussed, but considered unlikely because of absence of clear indicators of infection in clinical, laboratory, and radiological examination. The subtle signs of infection were attributed to superinfection of pulmonary malignancy. Chemotherapy and cerebral radiotherapy were planned. Surprisingly, bronchoalveolar lavage (BAL) and transbronchial and CTguided transthoracic biopsies failed to confi rm malignancy. Cytologically, BAL fl uid yielded druses structures surrounded by neutrophils, suggestive of actinomyces (fi gure B). Nonetheless, lobectomy was done because the lobe was completely destroyed and non-functional and also histological tissue examination was necessary for defi nitive exclusion of malignancy, potentially hidden by infection. Histological examination of the resected tissue showed an abscessing infl ammation with caverns. Malignancy was ruled out. No microorganisms were detected in staining or cultures. Unexpectedly, panbacterial PCR showed Streptococcus intermedius, but no actinomyces. Restaining of preoperative BAL fl uid showed gram-positive cocci in colonies or chains within the actinomyces-like druses (fi gure C). No fi lamentous structures characteristic of actinomyces were seen. It became apparent that a S intermedius lobar pneumonia was the primary site of infection which had subsequently spread to the brain. Our patient was discharged in good condition 8 days after surgery under appropriate antibiotic and antiepileptic treatment. Antibiotics were continued for 1 year. At fi nal follow-up in March, 2010, the patient had fully recovered and the cerebral lesions had completely regressed aside from a few scarred residues. S intermedius, S constellatus, and S anginosus form the Streptococcus milleri group (SMG). Purulent infections, often abscessing, subacute presentation, and absence of bacteraemia, are characteristic of SMG, and each species has certain sites that it characteristically infects. Typically, S intermedius aff ects brain or liver, whereas thoracic infections, in particular lobar pneumonia, are rare. Identifi cation of SMG is challenging; confusingly, in our case S intermedius had formed structures reminiscent of actinomyces druses. Clinical presentation of brain abscesses includes headache, fever, vomiting, mental change, focal neurological defi cit, and, seizures; SMG species, primarily S intermedius, are the most common pathogens. Brain abscesses and SMG infection tend to be associated with immunodefi ciency, but SMG species are increasingly recognised as important pathogens in previously healthy individuals. Radiological diff erentiation from neoplastic brain masses is diffi cult and can be aided by magnetic resonance spectroscopy. Management by surgery and/or antibiotics depends on abscess number, size, location, and mass eff ects. Even with a seemingly clear working diagnosis of malignancy with brain involvement, a thorough work-up must carefully distinguish neoplastic from infectious disease.