RESUMO
BACKGROUND: Several techniques are available to identify, among patients with mild cognitive impairment (MCI), those at risk of conversion to Alzheimer dementia (CAD). However, simple cost-effective methods to assess the risk are not available yet. METHODS: This retrospective study included 143 MCI outpatients (76.6±5.2 years, 46.8% women). Baseline variables were common neuropsychological tests (including Mini Mental State Examination-MMSE and Montreal Cognitive Assessment-MoCA), brain CT and 18F-fluorodeoxyglucose (FDG)-PET. Outcome variable was CAD after 1 year. RESULTS: At follow-up, 31 (21.7%) patients had CAD. In multivariable analysis (OR, 95% CI), female sex (4.7, 1.6-14.0), MoCA-executive component <3 (6.3, 2.1-19.2), left medial temporal atrophy (MTA) ≥3 (5.4, 1.9-15.7) and FDG-PET suggesting CAD (5.4, 1.9-15.7) were associated with CAD (area under ROC curve 0.873). Without FDG-PET, MMSE score <28 remained associated with CAD (6.0, 2.2-16.9). As first step (before FDG-PET execution), we counted 1 point for MMSE <28, executive MoCA <3 and left MTA ≥3. With 2-3 points CAD probability was high (75%) and with 0 points it was low (6.5%). Thus, FDG-PET (second step) might be performed only in patients with 1 point (probability 19.7%, 42.7% of patients). Among them, 35% had a positive FDG-PET, suggesting high risk. Overall, 28.0% of patients were considered at high risk (specificity 83.9%, sensitivity 71.0%, accuracy 81.1%). CONCLUSION: With a 2-step procedure, less than half of MCI patients might undergo FDG-PET and nearly a quarter of our patients was found to be at high CAD risk, including almost three quarters of future CADs.
