1,2) indicates petechial blood in the region of infarction and paren- chymal haematoma (PH 1,2) indicates blood clot with mass effect. Results: Imaging data was complete (pre and post tPA/placebo PWI and DWI) in 96 of 100 patients. HT occurred in 46 of 96 patients. Parenchymal haematomas (PH 1,2) occurred in 14 (4 symptomatic), asymptomatic haemorrhagic infarcts (HI 1,2) in 32. All VLCBV thresholds had similar results - the <2.5 th percentile is reported here. Median volume of VLCBV (mL) was significantly higher in HT com- pared with non-HT (4.04mL Vs 0.43mL p < 0.001) and in PH com- pared with non-PH patients (7.47mL Vs 1.22mL p = 0.003). There was a trend towards higher VLCBV volume in PH compared to HI (7.47 Vs 3.29 p = 0.11). Using VLCBV volume in ROC analysis gave an AUC of 0.74 for HT Vs no HT and 0.75 for PH Vs no PH indicating good performance as a predictive test. In subgroup analysis VLCBV was a better predictor of HT Vs no HT in reperfused patients (AUC 0.80) compared to non-reperfused patients (AUC 0.65) and in tPA treated patients (AUC 0.82) compared to untreated patients (AUC 0.68). A cut-point at 1 mL VLCBV had sensitivity 91%, specificity 73%, PPV 74%, NPV 90% for any HT and 100% NPV for PH. DWI volume correlated with VLCBV (R = 0.78) but was less accurate in predicting any HT (AUC 0.69) or PH (AUC 0.70). Univariate ordinal logistic regression using 5 HT categories as dependent variable was signifi- cant for VLCBV (coeff -1.27, p < 0.001) and DWI (coeff -0.14, p = 0.016). In multivariate analysis VLCBV (coeff -2.44, p = 0.001) was strongly predictive but DWI (coeff +0.23, p 0.06) did not im- prove prediction. Conclusions: This study has found that VLCBV is a better predic- tor of HT following thrombolysis than DWI volume in a large patient cohort. The effect holds true at a range of CBV thresholds and predic- tion is better in patients who reperfuse. Specificity may be improved by incorporating other predictive factors. The addition of VLCBV to pre-thrombolysis decision making could reduce the incidence of HT. doi:10.1016/j.jocn.2009.07.006 Encephalitis Aetiology in Hunter New England, New South Wales, 1998-2007 Gawarikar Yash 1 , Huppatz Clare 2,3 , Levi Chris 1 , Kelly Paul 3 , Williams David 1 , Dalton Craig 2 , Durrheim David 2 1 Department of Neurology, John Hunter Hospital and Centre for Brain and Mental Health Research HMRI, Newcastle, Australia 2 Hunter New England Population Health Unit, Newcastle, Australia 3 National Centre for Epidemiology and Population Health, College of Medicine, Biology and Environment, Australian National University, Canberra, Australia Introduction: Encephalitis is an uncommon yet important syn- drome from both the clinical and public health perspectives. The aetiology of up to 70% of cases of suspected encephaltis remains un- known. From a public health perspective, this means that preventa- tive measures cannot be implemented in the majority of cases. We retrospectively ascertained and reviewed all identified cases of encephalitis admitted to three major hospitals in the Hunter New England area, NSW over a ten year period, and examined the diag- nostic work-up in each case. Methods: All cases admitted to the John Hunter Hospital, the Maitland Hospital and Tamworth Base Hospital between July 1998 and December 2007, with an ICD discharge diagnosis consistent with encephalitis, were included in the study. Case notes were reviewed using a specially designed audit tool to collect information concern- ing clinical features, history, investigations and outcome. Findings are described as a case series. Results: A total of 74 cases were included in the case series. The mean age was 50.3 years (range 18.4-85.1 years) and 41 (55.4%) were male. The presenting symptoms and signs included fever (77.0%), headache (62.1%), altered consciousness (63.5%), lethargy (32.4%), sei- zures (25.7%), focal neurological deficits (31.1%) irritability (12.1%) and abnormal behaviour (12.1%). In searching for a causative organ- ism, the most common laboratory test performed was cerebrospinal fluid (CSF) analysis (n = 67, 91%). Herpes virus polymerase chain reac- tion (PCR) (n = 53, 71.6%) and cryptococcal antigen (n = 46, 62.2%) were the most common tests performed on CSF. Neuroradiological procedures employed were computerised tomographic brain scan- ning (68/74 cases) and magnetic resonance imaging of the brain (35/74 cases) but only 10 cases demonstrated focal imaging abnor- malities. Thirty-five patients (47.3%) had electroencephalograms (EEG) but only four had patterns reported as consistent with enceph- alitis. The treating clinicians suspected a causative organism in 14/74 cases (18.9%), of which nine (12.1%) were confirmed by laboratory testing. A total of 70 (94.6%) patients survived and 4 (5.4%) died. Of the survivors, 40 (54.1%) did not have any residual deficit, whereas 25 (33.7%) were left with neurological or psychological impairment. Conclusion: Although a causative organism was found for only a few patients, history taking and laboratory tests were not optimised to identify a pathogen in all cases. This study suggests that there is little uniformity in the diagnostic workup of patients with suspected encephalitis. We suggest that an algorithm and guidelines for the workup of encephalitis cases would help to optimise laboratory test- ing so that clinical care can be best tailored to the pathogen, and appropriate Public Health measures implemented. doi:10.1016/j.jocn.2009.07.007 Predicting Response to Intravenous Immunoglobulin in Lower Motor Neuron Syndromes Burrell James R 1,2 , Yiannikas Con 1 , Rowe Dominic 3 , Kiernan Matthew C 2 1 Department of Neurology, Concord Repatriation General Hospital 2 Prince of Wales Medical Research Institute and Prince of Wales Clinical School, University of NSW 3 Department of Neurology, Royal North Shore Hospital Objective: In patients presenting with purely lower motor neuron (LMN) syndromes, the distinction between motor neuropa- thy and an anterior horn cell disorders may be difficult to appreciate initially. As such, the selection of patients for treatment with intra- venous immunoglobulin (IVIg) is often a clinical dilemma, particu- larly as IVIg is an expensive, limited resource, ideally reserved for treatable patients. Consequently, the present study was undertaken to identify clinical and neurophysiological features that may predict a response to IVIg in patients presenting with pure LMN syndromes. Methods: Patients with LMN syndromes, who had been treated with at least 3 months of IVIg, were recruited from three specialised neuromuscular clinics. Patients with upper motor neurone features, bulbar involvement or objective sensory disturbance were excluded. Demographic data, symptom development, pattern of weakness, ganglioside (anti-GM1) antibody status and clinical follow-up were recorded for each patient. The presence of conduction block (definite and probable) and findings on electromyography were also assessed. Results: 42 patients (30 males, 12 females) aged 18 to 83 years (mean 48 years) were included in the study. After treatment with IVIg, 24 patients improved, 7 stabilised and 11 deteriorated. Compared to patients who deteriorated, patients who improved or stabilised were typically younger (46 compared to 56 years, Abstracts / Journal of Clinical Neuroscience 16 (2009) 1514–1546 1515