Computed Tomography, Electroencephalography, and Clinical Features in the Differential Diagnosis of Senile Dementia

phy, electroencephalography, and clinical features in the differential diagnosis of senile dementia was studied prospectively. Out of 50 demented patients, autopsy revealed 32 cases with either senile dementia of the Alzheimer's type (SDAT), multi-infarct dementia (MID), or a combination of both. Eighteen patients had dementia caused by other diseases. Based on a combination of computed tomography, electroencephalography, and clinical features, senile dementia of the Alzheimer's type was differentiated from all 50 patients, with a specificity of 83% and a sensitivity of 80%. Focusing on senile dementia of the Alzheimer's type, multi-infarct dementia, or a combination of both, specificity decreased to 65% and sensitivity to 47%. Comparing the different methods, multi-infarct processes were diagnosed with a higher sensitivity by the clinical features (73%) than by computed tomography (18%) or electroencephalography (18%). None of the methods validly differentiated multi-infarct dementia from a combination of multi-infarct dementia and senile dementia of the Alzheimer's type. (Arch Neurol. 1989;46:1217-1220) T^he accuracy in diagnosing different types of senile dementia has been the topic of many studies. Few re¬ searchers, however, have attempted to evaluate the clinical diagnosis by autopsy.1-5 In these studies, the diag¬ noses are based either on predictive clinical features (ischemie score) alone,2-4 on electroencephalography (EEG) alone,5 on EEG and cerebral blood flow (CBF),3 or on ischemie score in combination with EEG and com¬ puted tomography (CT).1 Two of these reports yield a sensitivity in the diag¬ nosis of Alzheimer's disease between 70%2 and 87%,' and a specificity be¬ tween 70%2 and 78%.' These rates consider the differentiation of Alzhei¬ mer's disease from all other causes of dementia. Specificity and sensitivity would be lower if senile dementia of the Alzheimer's type (SDAT), multiinfarct dementia (MID), and a combi¬ nation of the two (MIX) were taken into account alone. In clinical practice the differential diagnostic problems arise commonly in these three groups. The differentiation between MID and MIX remains extremely difficult.12-4 Other pathologic findings causing de¬ mentia, eg, intracerebral mass lesions and hemorrhages, are usually easy to recognize. Another problem deals with the age of the patients. Some of the clinical predictors may be variable with aging. The cited data refer to patients with presenile dementia,3 senile dementia,2-4