were 4.54 in AD, 7.37 in MCI, and 7.79 in normal controls. The distribution of scores in each group was as follows: 3/2/1/0: 3(12.5%)/0(0%)/0(0%)/ 27(87.5%) in AD, 11(40.7%)/1(3.7%)/9(33.3%)/6(22.2%) in MCI, 11(57.9%)/1(5.3%)/7(36.8%)/0(0%)) in normal controls. A total of 25 pa- tients were scored 3. The mean duration of their performance was 17.66 s in AD, 13.70 s in MCI, 14.70 s in normal controls. The most common error was omission of task, especially "edge" in AD whereas sequential error and addition of wrong motion were commonly observed in MCI. Conclusions: MCI patients showed more common errors than normal. Most of AD pa- tients were not able to do FEP. Our study suggests that FEP can be used in bed-side screening tool in the dementia clinic. P1-204 PERFORMANCE OF THE 16-ITEM INFORMANT QUESTIONNAIRE ON COGNITIVE DECLINE FOR THE ELDERLY IN AN ARABIC-SPEAKING OLDER POPULATION WITH LOW EDUCATION Kieu T. Phung 1 , Monique Chaaya 2 , Khalil El Asmar 2 , Samir Atweh 3 , Hussam Ghusn 3 , Rose Mary Khoury 4 , Martin Prince 5 , Gunhild Waldemar 1 , 1 Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark; 2 American University of Beirut, Beirut, Lebanon; 3 American University of Beirut Medical Center, Beirut, Lebanon; 4 American University of Beirut, Beirut, Lebanon; 5 King’s College London, London, United Kingdom. Contact e-mail: pttk2003@yahoo.com Background: The Middle East has high illiteracy rates among older people, making direct cognitive testing a great challenge. The 16-item Informant Questionnaire on Cognitive Decline for the Elderly (IQCODE) is an indirect way to screen for dementia through an informant interview. The aim of this study was to validate the Arabic version of the IQCODE. Methods: 236 par- ticipants older than 65 years, 93 with dementia and 143 with normal cogni- tion were recruited from social organizations and clinics. Half of the participants had no formal education. Physicians diagnosed dementia ac- cording to DSM-V criteria. The Geriatric Mental State was used to diagnose depression. Trained interviewers blind to the cognitive status of the partic- ipants administered the IQCODE to the informants. The ability of IQCODE to screen for dementia was evaluated against clinical diagnoses. Results: IQCODE scores differed significantly between normal cognition, mild, and moderate dementia and were not affected by gender, age, education, type of informants (lay versus health professional), or type of relationship (spouse, child, other relative, friend/social worker, health professional). In- formants of depressed participants tended to rate higher score (p¼0.07). IQ- CODE had excellent overall predictability (area under the ROC ¼ 0.96). The cut-off point of 3.34 yielded the best sensitivity and specificity, 92.5% and 94.4% respectively. There was no statistically significant differ- ence in the psychometric properties of IQOCDE through sensitivity analysis stratified by type of informants and type of residence (community versus nursing homes). Conclusions: The 16-item IQCODE is particularly useful as a brief screening tool for dementia among Arabic speaking older adults with no formal education living in the community or in the nursing homes. P1-205 PREDICTORS OF COGNITIVE DECLINE AMONG CHINESE ALZHEIMER’S DISEASE PATIENTS TREATED WITH ALZHEIMER’S MEDICATIONS: A THREE-YEAR PROSPECTIVE STUDY Leung-wing Chu, The University of Hong Kong, Hong Kong, Hong Kong. Contact e-mail: lwchu@hkucc.hku.hk Background: Alzheimer’s disease (AD) is the most common type of de- mentia in older adults. The existing treatment options included two groups of symptomatic drugs - namely the cholinesterase inhibitors and N-methyl D-aspartate (NMDA) receptor antagonist. Long-term progression data on cognitive function in AD patients treated with anti-dementia drugs in the real world settings have been reported from Western but not Asian AD pop- ulations. Objectives: The primary objective of the present study was to investigate the predictors of cognitive decline among Chinese AD patients treated with currently approved symptomatic AD drugs, in a real-life or real-world clinical setting. The secondary objective was to investigate the rate of long-term cognitive decline. Methods: This was a 3-year prospective study conducted in the Memory Clinic of Queen Mary Hospital, The Uni- versity of Hong Kong. Chinese AD patients who fulfilled the NINCDS- ADRDS criteria were recruited. Cognitive assessment with the Chinese ver- sions of ADAS-cog, MMSE and Delayed 10-Word Recall Test (DWRT) were done at baseline and 3-year follow-up. Baseline data on demographic, co-morbid diseases and apolipoprotein E (APOE) genotype were deter- mined. Results: 135 Chinese AD patients who completed 3-year follow- up were recruited. 67.1% were females. The mean (SD) age and educational level were 78.7 (7.5) and 3.6 (4.6) years, respectively. The mean (SD) 3-year declines in ADAS-cog, MMSE and DWRT were 8.74 (10.94), 3.66 (4.04) and 0.16 (0.72) points, respectively; and the calculated annual declines in ADAS-cog, MMSE and DWRT were 2.91, 1.22 and 0.05 points, respec- tively. These rates of declines were less than those in a historical control group of untreated Chinese AD patients, who showed annual declines of 4.09, 1.26 and 0.11 points for ADAS-cog, MMSE and DWRT, respectively. In multivariate analyses with general linear models for predictors of decline in ADAS-cog, the only significant predictor was age. APOE4 genotype, gender, education and BMI were non-significant. Conclusions: In this 3- year prospective study on long-term progression in Chinese AD patients, the cognitive decline is attenuated by treatments of currently approved AD medications. The only significant for predictor of cognitive decline is age, while APOE4 genotype and gender are both BMI are not significant predictors. P1-206 CONCISE COGNITIVE EVALUATION BATTERY (CCEB): A RELIABLE COGNITIVE BATTERY FOR MEMORY CLINIC Linda Lay Hoon Lim 1 , Jude Russell Chander 1 , Nyu Mei Mei 1 , Tanya Marie Choong 2 , Eveline Silva 1 , Nagaendran Kandiah 3 , 1 National Neuroscience Institute, Singapore, Singapore; 2 National Neuroscience Institute, Singapore, Singapore; 3 Department of Neurology, National Discriminatory ability of IQCODE at different cut-off points Cut-off point (Case if IQCODE score was greater than) N¼236 Sensitivity Specificity Youden index PPV NPV Correctly identified >3.22 94.6 (87.3,98.0) 86.7 (79.8,91.6) 0.813 82.2 (73.4,88.7) 96.1 (90.7,98.6) 89.8 (85.3,93.1) >3.28 92.5 (84.6,96.7) 90.2 (83.8,94.3) 0.827 86.0 (77.3,91.9) 94.9 (89.3,97.7) 91.1 (86.8,94.1) >3.34 92.5 (84.6, 96.7) 94.4 (88.9, 97.4) 0.869 91.5 (83.4, 96.0) 95.1 (89.7, 97.8) 93.6 (89.8,96.1) >3.40 87.1 (78.2, 92.9) 95.1 (89.8,97.8) 0.822 92.0 (83.8,96.5) 91.9 (86.0,95.6) 91.9 (87.8,94.8) >3.47 83.9 (74.5,90.4) 95.8 (90.7,98.3) 0.797 92.9 (84.5,97.1) 90.1 (84.0,94.2) 91.1 (86.8,94.1) IQCODE ¼ Informant Questionnaire on Cognitive Decline for the Elderly PPV ¼ Positive Predictive Value, NPV ¼ Negative Predictive Value All values except Youden Index are percentages with 95% confidence interval in parentheses Poster Presentations: P1 P378