Nonconvulsive status epilepticus with frontal features: quantitating severity of subclinical epileptiform discharges provides a marker for treatment efficacy, recurrence and outcome.

Nonconvulsive status epilepticus (NCSE) is difficult to diagnose but is an important cause of cognitive impairment. Electroencephalogram (EEG) monitoring is required for diagnosis and treatment. Little is known regarding the stability of subclinical epileptiform discharges (SEDs) preceding NCSE nor what strategies may optimize patient outcomes. We report extended follow-up of patients with recurrent frontal SEDs, integrating EEG and cognitive findings before and following treatment of NCSE, and show that quantitating SED severity provides an objective marker of treatment efficacy and recurrence.

Seizure-induced neuronal necrosis: implications for programmed cell death mechanisms.

PURPOSE: To determine definitively the morphology of neuronal death from lithium-pilocarpine (LPC)-and kainic acid (KA)-induced status epilepticus (SE), and to correlate this with markers of DNA fragmentation that have been associated with cellular apoptosis. Endogenous glutamate release is probably responsible for neuronal death in both seizure models, because neuronal death in both is N-methyl-D-aspartate receptor-mediated. METHODS: SE was induced for 3 hours in adult male Wistar rats with either LPC or KA, and 24 or 72 hours later the rats were killed. One group of rats had brain sections, stained with hematoxylin and eosin and the terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) technique, examined by light microscopy and by electron microscopy. A separate group of rats had DNA extracted from the same brain regions examined by electron microscopy in the first group. The extracted DNA was electrophoresed on an agarose gel with ethidium bromide and was examined for the presence or absence of internucleosomal DNA cleavage (DNA "laddering"). RESULTS: Twenty-four and 72 hours after 3 hours of LPC- or KA-induced SE, neuronal death in the hippocampus, amygdala, and piriform, entorhinal, and frontal cortices was morphologically necrotic, in spite of DNA laddering in these regions 24 and 72 hours after SE and positive TUNEL staining in some of the regions 72 hours after SE. Ultrastructurally, necrotic neurons were dark and shrunken, with cytoplasmic vacuoles and pyknotic nuclei with small, irregular, dispersed chromatin clumps. CONCLUSIONS: Our results, together with those of other reports, suggest that programmed cell death-promoting mechanisms are activated by SE in neurons that become necrotic rather than apoptotic and point to the possibility that such mechanisms may contribute to SE-induced neuronal necrosis.

Status epilepticus-induced neuronal loss in humans without systemic complications or epilepsy.

PURPOSE: To determine the regional distribution of neuronal damage caused strictly by status epilepticus (SE) without systemic complications, underlying brain pathology, or a history of preexisting epilepsy. METHODS: The medical records and electroencephalograms (EEGs) of three deceased patients who developed SE in the hospital were reviewed. Their brains were formalin-fixed, and 17 brain regions were selected, embedded in paraffin, and sectioned. Alternate sections were stained with either hematoxylin and eosin and cresyl violet to determine the extent of neuronal loss and gliosis or glial fibrillary astrocytic protein to confirm the extent of astrocytic proliferation. RESULTS: The three patients died 11 to 27 days after the onset of focal motor SE; none had hypotension, hypoxemia, hypoglycemia, or significant hyperthermia. Two patients had no prior seizures and no underlying brain pathology. The third patient, who had leptomeningeal carcinomatosis, had one seizure 2 months before the onset of SE. The duration of SE was 8.8 hours to 3 days. EEGs showed unilateral temporal lobe sharp-wave discharges in one patient and independent temporal lobe sharp-wave discharges bilaterally in the other two patients. In addition to widespread neuronal loss and reactive gliosis in the hippocampus, amygdala, dorsomedial thalamic nucleus, and Purkinje cell layer of the cerebellum, we report for the first time periamygdaloid (piriform) and entorhinal cortical damage occurring acutely after SE in humans. CONCLUSIONS: In the absence of systemic complications or preexisting epilepsy, SE produces neuronal loss in a distribution similar to that from domoic acid-induced SE in humans and from kainic acid- and pilocarpine-induced SE in rats.

Kainic acid-induced seizures produce necrotic, not apoptotic, neurons with internucleosomal DNA cleavage: implications for programmed cell death mechanisms.

Prolonged seizures (status epilepticus) induced by kainic acid activate programmed cell death mechanisms, and it is believed that kainic acid-induced status epilepticus induces neuronal apoptosis. In order to test this hypothesis, adult rats were subjected to 3-h kainic acid-induced seizures, with 24- or 72-h recovery periods. Neuronal death was assessed by light microscopy with the Hematoxylin and Eosin stain and with in situ terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL stain), by electron microscopy, and by agarose gel electrophoresis of DNA extracted from five vulnerable brain regions. Spontaneous and MK-801-induced apoptotic neurons from retrosplenial cortex of neonatal rats, evaluated by light and electron microscopy, were used as positive controls for apoptosis. Surprisingly, the large chromatin clumps of apoptotic neurons were TUNEL negative, whereas the cytoplasm showed light-to-moderate TUNEL staining, consistent with a lack of identifiable nuclear membranes ultrastructurally, and with intermingling of nuclear and cytoplasmic contents. Ultrastructurally, the acidophilic neurons produced by kainic acid-induced status epilepticus, identified with Hematoxylin and Eosin stain, were dark, shrunken and necrotic, with pyknotic nuclei containing small, dispersed chromatin clumps, and with cytoplasmic vacuoles, some of which were swollen, disrupted mitochondria. No apoptotic cells were seen. Acidophilic neurons were found in up to 20 of 23 brain regions examined and comprised 10-25% of the total number of neurons examined. A subset of these neurons (<10% of the total number of neurons in five of 23 regions) had TUNEL-positive nuclei 72h but not 24h after status epilepticus. Internucleosomal DNA cleavage (DNA "laddering") occurred in the four most damaged brain regions examined by electron microscopy 24h after SE and the three most damaged regions 72h after status epilepticus. Our results demonstrate that kainic acid-induced status epilepticus produces neuronal necrosis and not apoptosis in adult rats. The necrotic neurons show nuclear pyknosis, chromatin condensation and DNA laddering. Programmed cell death mechanisms activated by kainic acid-induced status epilepticus occur in neurons which become necrotic and could contribute to necrotic, as well as apoptotic, neuronal death.

Lithium-pilocarpine-induced status epilepticus produces necrotic neurons with internucleosomal DNA fragmentation in adult rats.

Prolonged and continuous epileptic seizures [status epilepticus (SE)] produce a widespread pattern of neuronal death, primarily in limbic brain regions. Because it has been suggested that seizure-induced neuronal death may be apoptotic in nature, we tested the hypothesis that lithium-pilocarpine-induced status epilepticus (LPCSE) produces apoptotic neurons. LPCSE lasting 3 h was induced in male Wistar rats which were allowed to recover for 24 or 72 h before perfusion-fixation. Neuronal death was assessed by light microscopy with the haematoxylin-and-eosin stain (H&E), with in situ DNA nick-end labelling (TUNEL stain), by electron microscopy, and by agarose gel electrophoresis of DNA extracted from vulnerable brain regions. Ultrastructurally, acidophilic neurons identified with H&E were dark, shrunken and necrotic in appearance, exhibiting pyknotic nuclei, irregular, dispersed chromatin clumps and cytoplasmic vacuolization. No cells with apoptotic features were seen. Acidophilic neurons were found in 21 out of 23 brain regions examined, and comprised 26-45% of the total number of neurons examined. A subset of these neurons (< 10% of the total number of neurons) were TUNEL-positive at 72 h, but not 24 h, after SE. Internucleosomal DNA cleavage (DNA 'laddering') was found in the six brain regions examined ultrastructurally 24 and 72 h after SE. These results indicate that, in adult rats, LPCSE produces neuronal injury with the appearance of necrosis rather than apoptosis. The necrotic neurons show nuclear pyknosis, chromatin condensation and internucleosomal DNA fragmentation, confirming the nonspecificity of these nuclear changes. Internucleosomal DNA cleavage and other programmed cell death mechanisms can be activated by SE in neurons which become necrotic.

An electroencephalographic study comparing maximum blink rates in schizophrenic and nonschizophrenic psychiatric patients and nonpsychiatric control subjects.

BACKGROUND: We did a retrospective electroencephalographic (EEG) analysis of blink rates in patients with psychiatric disorders and control subjects to determine whether maximum blink rates under different conditions were higher in patients with psychiatric disorders. METHODS: Maximum blink rates in those with schizophrenia (n = 23), those with nonschizophrenic psychiatric illnesses (n = 21), and nonpsychiatric control subjects (n = 35) were obtained from patients' EEGs and compared with one-way analysis of variance and post hoc tests. In addition, correlation analysis was performed to determine if patients' medications affected maximum blink rates. RESULTS: Patients with schizophrenic and nonschizophrenic psychiatric disorders had twofold higher maximum resting blink rates compared to controls (p < .05 respectively). No difference was found between those with schizophrenic and nonschizophrenic psychiatric disorders. The maximum blink rate during cognitive testing was also twofold higher in those with nonschizophrenic psychiatric disorders (n = 11) compared to controls (n = 16; p < .05). Within each group, maximum blink rates during quiet rest and cognitive testing did not differ, nor were there differences between groups in the duration of high-frequency blinking (greater than 40 blinks per minute) during quiet rest. In psychiatric patients, none of the medications taken at the time of EEG recording correlated with maximum blink rates. CONCLUSIONS: High maximum blink rates recorded by EEG may suggest the presence of a psychiatric disorder.

Effects of N-methyl-D-aspartate-receptor blockade on high-potassium-induced neuronal death and glutamate release.

Microdialysis probe delivery of an isotonic high-K+ solution to the rat amygdala for 60-70 min produces neuronal necrosis, edematous neuropil and increases in extracellular glutamate and aspartate. In this study we determined the effects of the competitive N-methyl-D-aspartate (NMDA)-receptor antagonist CGP 40116. Bilateral microdialysis probes were inserted through previously placed guide cannulae in the basolateral amygdaloid nuclei of adult Wistar rats. Following a 2 h baseline perfusion in freely-moving rats, the solution was switched bilaterally for 60-70 min to an isotonic 100 mM K+ solution; 12% of the 100 mM K+ is extracted by tissue. One side contained either 36 or 360 microM CGP 40116 during baseline and high-K+ perfusion, of which 42 and 48% were extracted respectively. Although neither concentration reduced tissue edema, 360 microM CGP 40116 was neuroprotective and prevented the high-K+-induced elevations of glutamate.

In vivo elevation of extracellular potassium in the rat amygdala increases extracellular glutamate and aspartate and damages neurons.

It is well known that high potassium (K+) solutions introduced by microdialysis into normal brain increase the extracellular concentration of the excitatory amino acid glutamate, and in vitro studies suggest that a high exogenously applied glutamate concentration can produce excitotoxic neuronal death. However, only recently were in vivo studies undertaken to determine whether high-K+ exposure damages neurons. We implanted microdialysis probes into rat amygdalae bilaterally, and after a 2-h baseline period exposed one side to a modified Krebs-Ringer-bicarbonate solution containing 100 mmol/l KCl for 30,50 and 70 min, followed by a 2-h recovery period, and 70 min and 3 h without a recovery period. Of 100.9 +/- 2.0 mmol/l KCl, 12.0 +/- 1.0% was extracted by amygdalar tissue in vivo. Election of the extracellular K+ concentration in the amygdala for 70 min or longer without a recovery period produced extensive neuronal damage and edematous-appearing neuropil in the tissue dialysed, as well as loss of normal neurons. Histological evidence of edema subsided in the groups with a 2-h recovery period. Although the number of damaged neurons was not significantly higher in the group with a 70 min high-K+ exposure and 2-h recovery period, the number of normal neurons was reduced, suggesting cell loss. During 70-min high-K+ exposure, the extracellular glutamate concentration increased to 242-377% of baseline during the first 60 min, and extracellular aspartate rose to 162-213% during the first 50 min; extracellular taurine rose even higher, to 316-567% of baseline, and glutamine fell to 14-27% of baseline. Extracellular serine was decreased at 20, 50 and 70 min of high-K+ exposure; extracellular glycine was unchanged. The elevated extracellular glutamate and aspartate concentrations suggest that exposure of the amygdala to high extracellular K+ may produce cell death through an excitotoxic process, and point the way to future studies to define the specific mechanisms involved.

The temporal evolution of neuronal damage from pilocarpine-induced status epilepticus.

The temporal evolution of irreversible neuronal damage from pilocarpine-induced seizures was studied by light microscopy. Neuronal cell death was judged on a 0-3 scale by estimating the percentage of acidophilic neurons in each of 23 brain regions. In addition, in the dorsal dentate hilus (CA4), quantitative cell counts of normal and acidophilic neurons were also performed. A few dead neurons (grade 0.5 damage) appeared in ventral hippocampal CA1 and CA3 regions after 20-min status epilepticus (SE). Slight-to-mild damage (grades 0.5-1.5) occurred in 14 and 12 brain regions after 40-min and 1-h SE respectively, and slight-to-moderate damage (grades 0.5-2.0) was found in 15 regions after 3-h SE. Twenty-four h and 72 h after 3-h SE, there was slight-to-severe damage (grade 0.5-3.0) in 22 and 21 regions respectively. Three-h SE produced more severe damage to 7 brain regions compared to 1-h SE, and 16 regions had more pronounced neuronal injury 24 h after rather than 0-4 h after 3-h SE. Eight brain regions had less damage 72 h compared to 24 h after SE, probably because of progressive neuronal lysis and dropout, but in mediodorsal and lateroposterior thalamic nuclei damage worsened from 24 to 72 h after SE. Neuronal cell counting revealed 20% acidophilic neurons in dorsal dentate hilus after 40-min SE and no difference between the 1-h and 3-h seizure groups (31% vs. 43% acidophilic neurons respectively). Among the 3 groups of rats with 3-h SE and varying recovery periods, the 24-h and 72-h recovery groups had higher percentages of acidophilic neurons (65% and 54% respectively) than the 0-4-h group (43%). Finally, the hippocampal CA2 region and dentate granule cell layer and the caudate-putamen, considered resistant to seizure-induced cell injury, were all damaged from SE lasting 40 min or more.

Neuroprotective effect of ketamine administered after status epilepticus onset.

We investigated the neuroprotective effect of the noncompetitive N-methyl-D-asparatate (NMDA) antagonist ketamine when administered after onset of lithium-pilocarpine-induced status epilepticus (SE). Seizures were induced in Wistar rats with lithium chloride (3 mEq/kg) and pilocarpine (PC) (30-60 mg/kg intraperitoneally, i.p.). Fifteen minutes after SE onset, either ketamine 100 mg/kg or normal saline was injected i.p., and 3 h after SE onset atropine, diazepam (DZP), and phenobarbital (PB) were administered i.p. to terminate the seizures. Twenty-four hours later, rats underwent brain perfusion-fixation, with subsequent brain processing for light-microscopic examination. Rats adminstered saline (n = 5) had neuronal damage in 24 of 25 brain regions examined. Rats administered ketamine (n = 7) had significant neuroprotection in 22 of 24 damaged regions. Ketamine reduced the amplitude of seizure discharges, and in 3 rats EEG seizure activity ceased in 30 min; none of these rats had neuronal damage. In the other 4 rats, EEG seizure discharges persisted > 90 min; in these animals, 21 of 24 damaged regions were protected. In contrast, rats with 1-h high-dose PC-induced SE (400 mg/kg i.p. without lithium chloride preadministration) had 14 damaged regions, of which 7 were significantly different from the undamaged regions of the ketamine subgroup with persistent electrographic seizures. Thus, ketamine is remarkably neuroprotective when administered after onset of SE, whether or not seizure discharges are eliminated.

The competitive NMDA receptor antagonist CGP 40116 protects against status epilepticus-induced neuronal damage.

We studied the efficacy of the competitive NMDA receptor antagonist CGP 40116 in protecting against seizure-induced neuronal necrosis from lithium-pilocarpine-induced status epilepticus (SE). Rats were given CGP 40116 either before SE (12 mg/kg i.p.) or 15 min after the onset of SE (4, 12 and 24 mg/kg); controls received normal saline 15 min after SE began. Diazepam and phenobarbital were given i.p. after 3 h of SE to stop the seizures. Rats were killed 24 h later, and their brains were processed for light microscopic examination. Neuronal damage occurred in 24 of 25 brain regions examined in saline-injected animals. Protection was maximal in rats given 12 and 24 mg/kg CGP 40116 after SE onset: 19 and 21 of the 24 damaged regions were protected respectively, but the 24 mg/kg group had a mortality rate comparable to saline-injected controls. No necrotic neurons were found in posterior cingulate and retrosplenial neurons at the two highest CGP 40116 doses, suggesting that the transient cytoplasmic vacuolization induced by NMDA receptor antagonists does not progress to frank necrosis. In rats given CGP 40116 seizure discharges were not eliminated, but their amplitudes were significantly reduced 2 h after SE began. The periodic epileptiform discharge (PED) EEG pattern, probably a sign of widespread neuronal damage, developed in saline-injected controls after 2-2.5 h of SE but not in rats given 12 and 24 mg/kg of CGP 40116. CGP 40116 provided widespread protection against seizure-induced neuronal necrosis, suggesting that an essential step in its production is NMDA receptor activation by endogenous glutamate. The neuroprotection provided was not simply an antiepileptic effect, since electrographic seizures persisted despite NMDA receptor blockade. CGP 40116 and NMDA receptor antagonists in general could be useful as adjunctive neuroprotectants in patients with refractory SE.

Pathophysiological mechanisms of brain damage from status epilepticus.

Human status epilepticus (SE) is consistently associated with cognitive problems, and with widespread neuronal necrosis in hippocampus and other brain regions. In animal models, convulsive SE causes extensive neuronal necrosis. Nonconvulsive SE in adult animals also leads to widespread neuronal necrosis in vulnerable regions, although lesions develop more slowly than they would in the presence of convulsions or anoxia. In very young rats, nonconvulsive normoxic SE spares hippocampal pyramidal cells, but other types of neurons may not show the same resistance, and inhibition of brain growth, DNA and protein synthesis, and of myelin formation and of synaptogenesis may lead to altered brain development. Lesions induced by SE may be epileptogenic by leading to misdirected regeneration. In SE, glutamate, aspartate, and acetylcholine play major roles as excitatory neurotransmitters, and GABA is the dominant inhibitory neurotransmitter. GABA metabolism in substantia nigra (SN) plays a key role in seizure arrest. When seizures stop, a major increase in GABA synthesis is seen in SN postictally. GABA synthesis in SN may fail in SE. Extrasynaptic factors may also play an important role in seizure spread and in maintaining SE. Glial immaturity, increased electronic coupling, and SN immaturity facilitate SE development in the immature brain. Major increases in cerebral blood flow (CBF) protect the brain in early SE, but CBF falls in late SE as blood pressure falters. At the same time, large increases in cerebral metabolic rate for glucose and oxygen continue throughout SE. Adenosine triphosphate (ATP) depletion and lactate accumulation are associated with hypermetabolic neuronal necrosis. Excitotoxic mechanisms mediated by both N-methyl-D-aspartate (NMDA) and non-NMDA glutamate receptors open ionic channels permeable to calcium and play a major role in neuronal injury from SE. Hypoxia, systemic lactic acidosis, CO2 narcosis, hyperkalemia, hypoglycemia, shock, cardiac arrhythmias, pulmonary edema, acute renal tubular necrosis, high output failure, aspiration pneumonia, hyperpyrexia, blood leukocytosis and CSF pleocytosis are common and potentially serious complications of SE. Our improved understanding of the pathophysiology of brain damage in SE should lead to further improvement in treatment and outcome.

Neuropathological changes during generalized seizures in newborn monkeys.

The brains of four 2-week-old marmoset monkeys were perfusion-fixed immediately after bicuculline-induced seizures lasting 1.5-4.3 h and were later examined by light and electron microscopy. Mean arterial blood pressure and rectal temperature measurements during seizures did not differ significantly from baseline. Plasma glucose concentrations decreased to the 1.5 mM range at the end of seizures, and arterial pH and bicarbonate were lower than in control animals, although arterial pO2 and pCO2 were maintained. Neuropathological changes were minimal. Swollen astrocytic processes surrounded some capillaries and some neurons in cerebral cortex, hippocampus, putamen and thalamus. Almost all the neurons examined looked normal, but mitochondrial swelling was present in a few. All but the most severe mitochondrial swelling, which occurred very rarely in one of four animals, is potentially reversible. The virtual absence of neuronal necrosis in these neonatal monkeys is consistent with the resistance to seizure-induced brain damage found in immature rats, and stands in sharp contrast to the damage seen in older animals. Lack of neuronal damage, however, does not rule out potential adverse effects of prolonged seizure activity on subsequent brain growth and development.

Local cerebral glucose utilization during status epilepticus in newborn primates.

The effect of bicuculline-induced status epilepticus (SE) on local cerebral metabolic rates for glucose (LCMRglc) was studied in 2-wk-old ketamine-anesthetized marmoset monkeys, using the 2-[14C]-deoxy-D-glucose autoradiographical technique. To estimate LCMRglc in cerebral cortex and thalamus during SE, the lumped constant (LC) for 2-deoxy-D-glucose (2-DG) and the rate constants for 2-DG and glucose were calculated for these regions. The control LC was 0.43 in frontoparietal cortex, 0.51 in temporal cortex, and 0.50 in thalamus; it increased to 1.07 in frontoparietal cortex, 1.13 in temporal cortex, and 1.25 in thalamus after 30 min of seizures. With control LC values, LCMRglc in frontoparietal cortex, temporal cortex, and dorsomedial thalamus appeared to increase four to sixfold. With seizure LC values, LCMRglc increased 1.5- to 2-fold and only in cortex. During 45-min seizures, LCMRglc in cortex and thalamus probably increases 4- to 6-fold initially and later falls to the 1.5- to 2-fold level as tissue glucose concentrations decrease. Together with our previous results demonstrating depletion of high-energy phosphates and glucose in these regions, the data suggest that energy demands exceed glucose supply. The long-term effects of these metabolic changes on the developing brain remain to be determined.

Cerebral glucose metabolic patterns in Alzheimer's disease. Effect of gender and age at dementia onset.

No previous study of Alzheimer's disease has, to our knowledge, assessed the effect of both age at dementia onset and gender on cerebral glucose metabolic patterns. To this end, we used positron emission tomography (fludeoxyglucose F 18 method) to study 24 patients with clinical diagnoses of probable Alzheimer's disease. Comparisons of the 13 patients with early-onset dementia (less than 65 years of age) with the 11 patients with late-onset dementia (greater than 65 years of age) revealed significantly lower left parietal metabolic ratios (left posterior parietal region divided by the hemispheric average) in the early-onset group. The metabolic ratio of posterior parietal cortex divided by the relatively disease-stable average of caudate and thalamus also separated patients with early-onset dementia from those with late-onset dementia, but not men from women. Further comparisons between sexes showed that, in all brain regions studied, the 9 postmenopausal women had higher nonweighted mean metabolic rates than the 15 men from the same age group, with hemispheric sex differences of 9% on the right and 7% on the left. These results demonstrate decreased parietal ratios in early-onset dementia of Alzheimer's disease, independent of a gender effect.

Posthypoxic treatment with MK-801 reduces hypoxic-ischemic damage in the neonatal rat.

We evaluated the neuroprotective effect of MK-801, a noncompetitive, selective N-methyl-D-aspartate receptor antagonist, in a neonatal hypoxic-ischemic animal model. Seven-day-old rats underwent bilateral ligation of the carotid arteries followed by exposure to an 8% oxygen atmosphere for 1 hr. We sacrificed the animals 72 hrs later and assessed the hypoxic-ischemic brain damage histologically. MK-801 (10 mg/kg), administered IP 0.5 hr before the hypoxia, completely prevented hypoxic-ischemic infarction in cerebral cortex, while treatment immediately and 1 hr after the end of the hypoxia resulted in 76% and 52% reduction in the infarcted area, respectively. MK-801, given 0.5 hr before and immediately after the insult, reduced striatal damage and, given 0.5 hr before, attenuated neuronal necrosis in hippocampal regions. These results show that in neonates MK-801 is neuroprotective even when administered up to 1 hr after the end of a hypoxic-ischemic insult.

Clinical characterization of Alzheimer's disease: reliability of 'age at onset' and a new descriptor, 'age at shift'.

To determine the interrater reliability of clinical descriptors for Alzheimer's disease (AD), we assessed the degree of agreement among four clinicians who rated 21 patients during a longitudinal study. Despite variability in response patterns, degree of agreement for determining age at onset of dementia was statistically significant (P less than 0.005). We also found significant agreement (P less than 0.0001) among three clinicians for the clinical descriptor, "age at shift" from questionable to probable AD, according to the National Institutes of Health Consensus Criteria. These data demonstrate that both retrospective and prospective descriptors can be reliably determined in the clinical assessment of AD.