Prepared by Michael B. First, M.D., DSM Consultant to the American Psychiatric Institute for Research and Education (APIRE), a subsidiary of the American Psychiatric Association
The fourth diagnosis-related research planning conference in the "Future of Psychiatric Diagnosis: Refining the Research Agenda" series, focusing on diagnostic issues in dementia, was held on September 16th and 17th at the Hôpital de l'Université de Genève - Belle Idée. Conference co-chairs were Trey Sunderland, MD, from the Geriatric Psychiatry Branch of the National Institute of Mental Health in Bethesda, MD, Dilip Jeste, MD, from the University of California at San Diego, San Diego CA, and Olusagun Baiyewu, MD, from the University of Ibadan, Ibadan, Nigeria, and invited participants included 14 scientists from around the world.
The conference began with a presentation by Robert Terry, MD (La Jolla, CA) on the topic of the neuropathology of dementia. Dr. Terry reported that 60% of dementia cases show evidence of Alzheimer's disease at autopsy, 20% show evidence of Lewy body disease in combination with Alzheimer's pathology, 5% have pure Lewy body disease, 5% show evidence of fronto-temporal disease including Pick's disease, 5% show pathology characteristic of progressive vascular dementia, and 5% miscellaneous. Although the "gold standard" for a diagnosis of Alzheimer's disease has been considered to be the presence of amyloid plaques and neurofibrillary tangles at autopsy, the gold has become tarnished in recent years--there is no strong correlation between the number of plaques and tangles and the severity of dementia. For example, after age 70, 20% of dementia cases do not have neocortical tangles. The best structural correlate of dementia is loss of synapses but there is currently no method of determining synapse count in vivo. Overall, the diagnosis of dementia remains a clinico-pathological diagnosis, which requires input from both clinicians and pathologists.
The next presentation was by Masatoshi Takeda, MD, PhD (Osaka, Japan) on the differential diagnosis of neurodegenerative dementia. He noted several problems with the current diagnostic criteria for Dementia of the Alzheimer's type: its lack of positive criteria for making a diagnosis, the inability to make a diagnosis of early Alzheimer's disease because the criteria require full-blown symptoms of impairment in social and occupational functioning, and the lack of guidelines for making a differential diagnosis with other types of dementia. Dr. Takeda then reviewed the biological markers that have been investigated to assist in the diagnosis of Alzheimer's disease (AD). Certain markers are helpful in screening (e.g., genetic markers, CSF total tau, serum homocysteine), in differential diagnosis (e.g., CSF phosphorylated tau, amyloid-42) and in evaluating severity (e.g., serum CoQ10 oxidation). He noted that while CSF tau may hold the most promise as a biological marker (e.g., it is elevated in many types of dementia and is typically elevated in early stages of Alzheimer's disease), it is not currently possible to integrate the neurobiology of tau, neuropathological classification of AD, and clinico-pathological classification because of a gap between tau deposition and neuronal loss.
John Saunders, MD (Brisbane, Australia), one of the co-chairs of the substance use disorders conference, provided a substance use perspective on dementia and cognitive impairment. He noted that, given the established evidence that alcohol is a major cause of brain damage and recent evidence linking inhalants, cannabis, sedative-hypnotics and psycho-stimulants with cognitive and other forms of brain impairment, substance use must be regarded as a key environmental variable for the development of dementia. The two most common alcohol-related cognitive impairments are alcohol-induced amnestic disorder, which is characterized primarily by impairment in short-term memory, and alcohol-induced frontal lobe syndrome, characterized primarily by impairments in executive functioning (e.g., abstract thinking) and personality change but no apraxia, agnosia, or aphasia, and sometimes no memory impairment, raising questions about whether the DSM-IV criteria for dementia are appropriate for alcohol-induced cognitive impairment.
Ronald C. Petersen, PhD, MD (Rochester, MN) then rresented on the current state of Mild Cognitive Impairment (MCI) as a diagnostic entity. By definition, Alzheimer's disease evolves over years and perhaps decade and, therefore, there must be an intermediate state during which time individuals are mildly symptomatic but do not manifest the full syndrome of dementia. This state has been designated as MCI and a great deal of research is being generated concerning its utility. The original criteria for MCI focused on memory impairment, but newer criteria have expanded the entity to involve memory impairment (amnestic MCI) as well as non-memory cognitive impairment (non-amnestic MCI). Most individuals with amnestic MCI progress to Alzheimer's disease at a rate of 10 to 15% per year with longitudinal data indicating that approximately 80% have progressed by approximately 6 years. The outcome of individuals with non-amnestic MCI remains a research issue. The diagnosis of MCI is being used clinically and the American Academy of Neurology has endorsed the construct in an evidence-based medicine review of the literature producing a practice parameter. From a research perspective, the topic of MCI is being investigated in virtually every aspect of the study of dementia. Data are coalescing on its sensitivity and specificity for predicting the future development of dementia. John O'Brien, MD (Newcastle, UK), the discussant, noted that the main advantage of the amnestic MCI label is that it is predictive of a high-risk group in clinic samples who subsequently develop AD. It has not yet been shown to perform well in population studies, with up to 40% reverting to normal in 3 years. Furthermore, non-amnestic subtypes have not been well-validated. While amnestic MCI will likely enable many cases of AD to be detected at an earlier stage, the diagnostic process will need to be broadened beyond cognitive testing to include behavioral, motor and other aspects (e.g., imaging, biomarkers) when considering non-AD dementias such as Vascular Dementia and dementia with Lewy bodies.
John Breitner, MD (Seattle, WA) discussed epidemiologic considerations in dementia and cross-cultural differences. Dementia is defined in terms of being a syndrome (i.e., a constellation of signs and symptoms without specified etiology), that is global and represents a decline in cognition, but some diagnostic criteria have confused this syndrome with the variety of brain diseases that often provoke it. Dr. Breitner cautioned that such "classification by presumed etiology" was premature in most areas of psychiatry, and in dementing illness particularly, when little is known about the root causes of the disorder. He noted that weaknesses in the DSM-IV definition include the lack of a simple definition of dementia per se, and that "executive function" remains an intuitive construct that is still inadequately defined. The increase in the prevalence of dementia by age levels off after age 90, and the incidence of dementia increases exponentially through age 85, but the increase may slow thereafter. The prevalence of dementia is undoubtedly increasing with the increasing longevity of the population, with the greatest change in the developing world. There are differences in rates of AD among populations of similar ethnic status in different regions; for example, rates of AD are much lower in Ibadan, Nigeria, than in Indianapolis. Dr. Breitner cited such findings as evidence that, despite its very strong genetic determinants, AD has important environmental risk factors that may be modifiable. He concluded by noting that there is no known risk factor for Alzheimer's disease that operates by a mechanism other than by altering time of onset. Olusagun Baiyewu, MD (Ibadan, Nigeria), in his discussion, noted that some prevalence studies report lower rates in developing countries than in developed countries whereas others report similar rates. Incidence rates, which are more reliable, have been reported to be lower in two studies which compared rates in Ibadan, Nigeria, and Ballabgarh, India, to sites in the United States. He also pointed out that instruments that assess cognitive performance and functioning, which were developed for middle class populations in the US, may be applicable in developing countries only after appropriate modifications. Joao Carlos Barbosa Machodo (Minas Gerais, Brazil) noted that considering the definition of dementia is critical for understanding the results of cross-cultural comparisons.
Barry Reisberg, MD (New York, NY) gave a presentation the focused on the strengths and weaknesses of the current DSM-IV and ICD-10 criteria. He suggested adopting the parsimonious term "Cognitive Disorders" for the diagnostic grouping currently referred to as "Dementia, Delirium. Or Other Cognitive Disorders" or as "Organic Mental Disorders". Regarding the definition of dementia, he suggested that dementia is characterized by multiple cognitive deficits, functional deficits and, commonly, personality changes, and that memory deficit, per se, is not the only presentation of dementia. He noted that the current subtyping of dementia into early onset and late onset is anachronistic, given that AD is the same illness whether it occurs before or after age 65 and most patients with dementia under age 65 also have AD. Regarding vascular dementia, there is now agreement that cognitive impairments associated with cerebrovascular disease extend well beyond the traditional concept of multi-infarct dementia and that memory impairment is not necessarily a prime symptom in vascular dementia. The definition of Pick's disease should be broadened to include current concepts of fronto-temporal dementia (e.g., involving executive functions and language comprehension) and should not require memory impairment since many common presentations of FTD have relatively preserved memory function, and many individuals with FTD only develop memory impairment late in their course. Dr. Reisberg suggested that in order to improve the criteria for dementia there should be an exploration of community physicians' capacity to assess dementia on a simple severity dimension and an assessment of cognitive and functional relationships in different dementia subtypes.
Norman Sartorius, MD, PhD (Geneva, Switzerland), in his discussion, raised a number of questions about the classification of dementia, including a consideration of the problems associated with using social functioning as a criterion in the definition of dementia; the need to consider the complexity of the environment in assessing failure of cognitive functioning; and difficulties in defining the natural course of illness. Dr. Sartorius also questioned whether dementia is one disease or whether it should be conceptualized as several diseases given reports from several African countries about the rarity of dementia; differences between the East Asian populations and the populations in Europe and Australia in the relative frequency of different types of dementia; and differences in the progression of dementia among patients in different cultures.
Mary Sano, PhD (New York, NY) gave a presentation on neuropsychological testing in dementia. Rationales for employing neuropsychological testing include documenting cognitive deficits, identifying patterns of deficit, addressing factors that confound the diagnosis, evaluating functional consequences, predicting outcomes, and documenting disease progression. The main strength of the current diagnostic criteria is that the specification of memory as a key cognitive domain has yielded good normative data on memory and its prognostic value is well-documented. For example, age-based norms are available in many languages and cultures providing a standard against which to measure performance and identify deficit (e.g., memory decline with age is well characterized and distinct). Weaknesses of the current criteria include limited characterization of lesser states of cognitive loss which represent significant clinical complaint and limited specification of non-memory functions. In addition, research criteria for entities such as Age Associated Memory Impairment, (AAMI) Age Related Cognitive Deficit (ARCD) and Vascular Cognitive Impairment (VCI) need to be evaluated to determine if they have utility as clinical diagnoses. From a neuropsychological perspective, there is limited characterization of executive function which may have importance in diagnostic specificity, limited ability to measure social and occupational functioning and to correlate it with specific neuropsychological deficits. Future needs include adopting the technology and culture to make neuropsychological testing a more widely available tool and to develop better clinician-administered tests. The widely used Mini-Mental Status Exam has no measure of executive functioning and too limited memory assessment to capture early changes.
Dilip Jeste, MD (San Diego, CA), presented on behavioral syndromes of dementia. Behavioral disturbances in Alzheimer's disease have a high incidence and prevalence, are typically chronic or recurrent, and commonly produce disruption in functioning which is a primary reason for institutionalization. There are a number of problems with the current DSM-IV/ICD-10 nosology which is mostly based on symptoms, including the lack of empirical support for the distinctiveness, reliability, or validity of categories, lack of a neurobiological basis, and limited utility for therapeutic or prognostic purposes. Dr. Jeste suggested that specific behavioral syndromes of dementia be defined in order to allow better studies of prevalence and incidence, facilitate recognition and treatment, promote hypothesis-based research leading to a better understanding of the conditions, and enable treatment research in diagnostically homogeneous populations. Specific behavioral syndromes were suggested for psychosis of AD, depression of AD sleep-wake cycle disturbance of AD, anxiety disorders of AD, personality disorders of AD, symptoms (e.g., physical aggression) or symptom complexes (e.g., agitation) associated with AD, and conditions associated with non-AD dementias. Additional research is needed for syndrome refinement, disease-specific instrument development, accurate estimation of prevalence and incidence, determination of longitudinal course, treatment validation, and determination of cost and service needs.
Deborah Blacker MD, ScD (Charlestown, MA), presented on genetic data and dementia nosology. There is a 2-3 fold increased risk of AD in the first-degree relatives of patients with AD, with the age of onset correlated within families. In the vast majority of cases, there is complex inheritance pattern, with only rare families having an autosomal dominant genetic transmission. Although monozygotic twins are much more likely to be concordant for the illness than dizygotic twins, the concordance for monozygotic twins is less than 100% and the age of onset for concordant monozygotic twins may differ indicating a role for environmental factors in the development of AD. Genetic factors are associated with late onset as well as early onset AD. Early-onset genes include amyloid precursor protein (APP), presenilin 1, and presenilin 2; mutations in these genes have been identified in fewer than 500 families worldwide. In late-onset cases, APOE is a susceptibility gene that affects AD risk and age at onset, which falls according to the dose of the APOE-4 gene. There is wide variation in risk estimates associated with APOE-4 (2-8 fold) and the risk varies with age (peaks in 60s), gender (greater in women), and ethnicity (greater in whites). For fronto-temporal dementia, 25-50% have a first-degree relative with FTD, many with an autosomal dominant transmission. For Creuzfeldt-Jakob dementia, 15% of cases are familial. With respect to the relationship between nosology and genetics, from a research perspective, genetically defined subtypes (e.g., based on family history, age of onset, specific symptoms) may be useful, but there is insufficient data to make them nosologic entities. Similarly, there is no evidence supporting a change in the clinical criteria for AD based on genetics and no evidence that genetic testing should be required for (or substantially helps with) making a diagnosis. For fronto-temporal dementia, family data confirms distinctive symptoms, variable neuropathology, and earlier age at onset. There may be sufficient evidence to consider an autosomal dominant subtype of FTD, possibly pulling out FTDP-17. Simon Lovestone, PhD (London, UK), in his discussion, notes that for familial autosomal dominant AD, improved nosology may 'capture' the remaining genes and genetics may usefully inform the classification. For late-onset cases of AD, however, improved nosology is unlikely to be helpful for genetic research in quite the same way, as late-onset dementia does not have a causative relationship with genes, rather genes are susceptibility factors. In fact, it is possible that an increasingly restrictive and well-defined nosology may hinder genetic research as it is unlikely that genetic susceptibility factors will map exactly onto clinical diagnoses. From a clinical perspective, it is doubtful that genetics will ever provide more than probabilistic profiling in late onset dementia.
Gary Small, MD (Los Angeles, CA) presented on neuroimaging as a surrogate marker of disease. Currently, structural imaging is recommended in all cases of dementia to identify space-occupying lesions but is not recommended to rule-in a diagnosis AD or other degenerative dementias. Functional imaging with FDG-PET may assist with the differential diagnosis of degenerative dementia for those individuals who meet criteria for both AD and FTD, with parietal and temporal deficits suggesting AD and frontal and temporal deficits suggesting FTD. Additional studies are underway to determine the usefulness of FDG-PET in patients with MCI. Promising new technologies include cognitive stress testing during fMRI scans to elicit abnormalities not observed during mental rest and PET scanning using small molecule probes (e.g., 11C-BTA or 18F-DDNP) to measure plaques and tangles which are indicative of AD. Future diagnostic approaches might combine several imaging approaches with APOE-4 genetic determination and cognitive stress tests, in order to accurately predict conversion of cases normal aging to MCI or of MCI to AD. FDG-PET imaging, which is becoming more widely available, has been used as a surrogate marker for clinical trials, and amyloid-PET imaging will likely accelerate anti-amyloid drug discovery and prevention strategies.
In the final presentation, Trey Sunderland, MD (Bethesda, MD) described current and future biomarker strategies in the diagnosis of AD. The ideal AD biomarker would detect a fundamental feature of AD, be validated in autopsy-proven cases, be specific for AD compared to other dementias, be reliable across different laboratories, be non-invasive and simple to perform, and be inexpensive. Currently the best candidate markers include CSF beta-amyloid 1-42, CSF total tau, and CSF phosphorylated-Tau231. Other biomarkers, like peripheral blood amyloid and serum homocysteine have little or no accuracy data or evidence of replication. A meta-analysis of the effect size for CSF beta-amyloid studies shows that the effect size is relatively robust but a severe lack of reliability in this assay worldwide is problematic. CSF total tau is elevated in AD patients as a group but overlap with older controls makes it difficult to use CSF total tau as a diagnostic tool. Using CSF total tau and CSF beta-amyloid together gives 92% specificity and 89% sensitivity but these accuracy rates involve comparing AD to controls rather than other forms of degenerative dementia and is thus not useful in differentiating AD from other types of dementia. Phosphorylated tau, however, shows good differentiation between AD and other types of dementia. Biomarkers may be helpful prognostically because they become elevated prior to the development of full-blown dementia. In particular, phosphorylated tau may be a way to differentiate MCI from normal controls. Harald Hampel, MD, PhD (Munich, Germany), in his discussion, reported that independent clinical research has shown that phosphorylated tau protein is a core, feasible biomarker candidate for early diagnosis and differential diagnosis, prediction, and potentially for progression of Alzheimer's disease (AD). He proposed multimodal development of biomarkers in AD to address relevant clinical questions (e.g. mapping of specific biologic activity): stage one would be methodological study and establishing the technical characteristics of the biomarkers; stage two would determine sensitivity and specificity and normative values in patient populations; and stage three would determine prevalence and positive/negative predictive values, and validate normative values. It is important to differentiate biomarkers by clinical utility, i.e., whether they are diagnostic, classificatory, or prognostic. Biomarker development must parallel the evaluation of novel secondary prevention drugs in order to select patient populations, define meaningful endpoints and support to create new clinical study designs.
Two breakout groups convened after the presentations to discuss them in further detail and to make recommendations for future research and for possible changes to be considered for DSM-V/ICD-11. These recommendations included: 1) the fact that instrumental activities of daily living (ADL) are the most important arena of impairment in dementia and metrics need to be developed that are culturally non-specific; 2) the conclusion that genetic markers do not add much to the diagnosis of AD but may be useful research tools and possibly of prognostic or predictive value; 3) the recognition that both imaging and biomarkers have tremendous research promise but are not yet ready for clinical diagnostic use, with the possible exception of 14-13 protein for Creutzfeldt-Jakob dementia; 4) the proposal that the definition of dementia should be reconfigured to not require memory impairment and adopt a definition based on a decline in two or more cognitive abilities lasting at least one year; 5) the recommendation that criteria for AD should be changed to make it a positive diagnosis, rather than a diagnosis by exclusion; 6) a call for continuing research to identify an at-risk group for developing dementia using multiple technologies, and 7) a call for autopsy verification of the clinical diagnosis as the only way to improve clinical accuracy for purposes of therapy.
The full presentations were published in a monograph by American Psychiatric Publishing, Inc., which can be purchased here.