|
(Continued from page 139)
were given a Mini-Mental State (MMS) examination and a spatial ability test,4 and had complete blood cell count and blood chemistry profiles reviewed. Each person was either excluded on the basis of these criteria, or assigned to the AD or control group.
The control group was examined and determined not to have dementia. Before admission to the nursing home, the patients with AD had their AD diagnosed at outside AD centers by use of the AD center's standard criteria. The patients' AD was diagnosed antemortem, without the definitive method of diagnosis, brain biopsy. Diagnosis or non-diagnosis of AD was reconfirmed on the basis of medical and p~ycho1ogic history; an MMS score of less than 23 for the AD group and greater than 23 for the control group; and performance on the spatial test.
The subjects in the AD group met the criteria of the Diagnostic and Statistical Manual of Mental Disorders III for primary degenerative dementia, had AD diagnosed at least 6 months before the beginning of this pilot study, had AD of insidious onset, and a progressive, deteriorating course. We believe that the experimental group reasonably represents AD patients.
The study was approved by the Institutional Review Board of the Chicago College of Osteopathic Medicine. Informed consent was obtained from the competent patient or the legal guardian. A 30-mL fasting blood sample was drawn without adding anticoagulant but not allowed to clot. The blood was placed on ice for transport, centrifuged at 7000g for 10 minutes at 2°C, and divided into aliquots in plastic tubes, which were immersed in liquid nitrogen before being stored at -- 80°C. For the separation and quantification, preliminary precautions were taken to avoid contamination. The plasma and standards were processed at 0°C, analyzed in triplicate, in a coded, blinded fashion, according to standard procedures,6~8 using the protocol for spectrophotometry as detailed in Methods in Enzymology,68 and other methods9'4 as outlined in Methods of Enzymatic Analysis, to determine the concentrations of aspartate, -y-aminobutyric acid, glutamate, glutamine, and a-ketoglutarate and activities of glutamate decarboxylase, glutamate dehydrogenase, glutaminase, and glutamine synthetase. The data were analyzed for significance by the Student's t-test and verified by the Bonferroni approach to multiple t-tests.
After the analysis, a discriminant analysis number was obtained by use of the four significantly different compounds. Linear discriminant analysis is based on a concept of distance. One starts by assuming that the (in this case two) groups have multivariate normal distributions in the variables measured. Then, the centers of the groups (their multivariate means) are found. We assume that the covariance structures for the two groups are also the same, having determined this common structure by pooling the covariance information. Finally, we classify points by seeing which of the two centers is closest. Because we have already assumed normality and we know the covariance structure, we can express the probability (Continued on page 141)
|
|
