Detecting prodromal Alzheimer's disease in mild cognitive impairment: utility of the CAMCOG and other neuropsychological predictors

BACKGROUND: The Cambridge cognitive examination (CAMCOG) is a mini neuropsychological battery which is well established and widely used. The utility of the CAMCOG in detecting prodromal Alzheimer's disease (AD) in patients with mild cognitive impairment (MCI) has not been determined. The objectives of this study are: to establish which subtests of cognitive domains contained within the CAMCOG are predictive of conversion to AD, to compare these with an extended version of the delayed word recall (DWR) test and to establish optimal cut points for all measures used. METHODS: 182 patients with MCI were identified from consecutive referrals to a memory clinic. Logistic regression, cox regression and receiver operating characteristic curve (ROC) analyses were conducted. RESULTS: The DWR displayed the best sensitivity (77%) and specificity (76%). The composite memory score contained within the CAMCOG achieved similar sensitivity (78%) and specificity (74%). The recognition component of the extended DWR demonstrated good specificity (85%) but poor sensitivity (57%). The optimal predictive model combined category fluency with the DWR and achieved predictive accuracy of 83%. CONCLUSION: The DWR, which is a test specifically designed to have high predictive accuracy for AD, performed best. The composite measure of memory contained within the CAMCOG performed similarly well. The DWR has the advantage of being brief, easy to administer and suitable for use in non-specialist settings. The CAMCOG takes longer to administer but provides information regarding additional cognitive domains and is sensitive to change over time. Category fluency may be usefully combined with the DWR to improve predictive accuracy.