Marcia Scazufca

Global prevalence of dementia: a Delphi consensus study.

BACKGROUND 100 years after the first description, Alzheimers disease is one of the most disabling and burdensome health conditions worldwide. We used the Delphi consensus method to determine dementia prevalence for each world region. METHODS 12 international experts were provided with a systematic review of published studies on dementia and were asked to provide prevalence estimates for every WHO world region, for men and women combined, in 5-year age bands from 60 to 84 years, and for those aged 85 years and older. UN population estimates and projections were used to estimate numbers of people with dementia in 2001, 2020, and 2040. We estimated incidence rates from prevalence, remission, and mortality. FINDINGS Evidence from well-planned, representative epidemiological surveys is scarce in many regions. We estimate that 24.3 million people have dementia today, with 4.6 million new cases of dementia every year (one new case every 7 seconds). The number of people affected will double every 20 years to 81.1 million by 2040. Most people with dementia live in developing countries (60% in 2001, rising to 71% by 2040). Rates of increase are not uniform; numbers in developed countries are forecast to increase by 100% between 2001 and 2040, but by more than 300% in India, China, and their south Asian and western Pacific neighbours. INTERPRETATION We believe that the detailed estimates in this paper constitute the best currently available basis for policymaking, planning, and allocation of health and welfare resources.

Dementia diagnosis in developing countries: a cross-cultural validation study

BACKGROUND Research into dementia is needed in developing countries. Assessment of variations in disease frequency between regions might enhance our understanding of the disease, but methodological difficulties need to be addressed. We aimed to develop and test a culturally and educationally unbiased diagnostic instrument for dementia. METHODS In a multicentre study, the 10/66 Dementia Research Group interviewed 2885 people aged 60 years and older in 25 centres, most in Universities, in India, China and southeast Asia, Latin America and the Caribbean, and Africa. 729 had dementia and three groups were free of dementia: 702 had depression, 694 had high education (as defined by each centre), and 760 had low education (as defined by each centre). Local clinicians diagnosed dementia and depression. An interviewer, masked to dementia diagnosis, administered the geriatric mental state, the community screening instrument for dementia, and the modified Consortium to Establish a Registry of Alzheimers Disease (CERAD) ten-word list-learning task. FINDINGS Each measure independently predicted a diagnosis of dementia. In an analysis of half the sample, an algorithm derived from all three measures gave better results than any individual measure. Applied to the other half of the sample, this algorithm identified 94% of dementia cases with false-positive rates of 15%, 3%, and 6% in the depression, high education, and low education groups, respectively. INTERPRETATION Our algorithm is a sound basis for culturally and educationally sensitive dementia diagnosis in clinical and population-based research, supported by translations of its constituent measures into most languages used in the developing world.

Métodos para estimar razão de prevalência em estudos de corte transversal

OBJETIVO: Comparar empiricamente las regresiones de Cox, log-binomial, Poisson y logistica para estimar razones de prevalencia en estudios de cohorte transversal. METODOS: Fueron utilizados datos de un estudio epidemiologico transversal (n=2.072), de base poblacional, realizado con ancianos en la ciudad de Sao Paulo (Sureste de Brasil), entre mayo de 2003 y abril de 2005. Diagnosticos de demencia, posibles casos de trastorno mental comun y autopercepcion de salud pesima fueron escogidos como resultados con prevalencia baja, intermedia y alta, respectivamente. Fueron utilizadas variables de confusion con dos o mas categorias o valores continuos. Valores de referencia para estimaciones por punto y por intervalo para las razones de prevalencia (RP) fueron obtenidos por el metodo de estratificacion de Mantel-Haenszel. Estimaciones ajustadas fueron calculadas utilizando regresiones de Cox y Poisson con varianza robusta, y regresion log-binomial. Odds ratios (OR) brutos y ajustados fueron obtenidos por la regresion logistica. RESULTADOS: Las estimaciones por punto y por intervalo obtenidas por las regresiones de Cox y Poisson fueron semejantes a la obtenida por la estratificacion de Mantel-Haenszel, independientemente de la prevalencia del resultado y de las covariables del modelo. El modelo log-binomial presento dificultad de convergencia cuando el resultado tenia prevalencia alta y habia convariable continua en el modelo. La regresion logistica produjo estimaciones por punto y por intervalo mayores de las obtenidas por los otros metodos, principalmente para los resultados con mayores prevalencia iniciales. Si se interpretaban como estimaciones de RP, los OR superestimarian las asociaciones para los resultados con prevalencia inicial baja, intermedia y alta en 13%, casi 100% y cuatro veces mas, respectivamente. CONCLUSIONES: En analisis de datos de estudios de cohorte transversal, los modelos Cox y Poisson con varianza robusta son mejores alternativas que la regresion logistica. El modelo de regresion log-binomial produjo estimaciones sin sesgo de la RP, pero puede presentar dificultad de convergencia cuando el resultado es muy frecuente y la variable de confusion es continua.OBJECTIVE To empirically compare the Cox, log-binomial, Poisson and logistic regressions to obtain estimates of prevalence ratios (PR) in cross-sectional studies. METHODS Data from a population-based cross-sectional epidemiological study (n = 2072) on elderly people in Sao Paulo (Southeastern Brazil), conducted between May 2003 and April 2005, were used. Diagnoses of dementia, possible cases of common mental disorders and self-rated poor health were chosen as outcomes with low, intermediate and high prevalence, respectively. Confounding variables with two or more categories or continuous values were used. Reference values for point and interval estimates of prevalence ratio (PR) were obtained by means of the Mantel-Haenszel stratification method. Adjusted PR estimates were calculated using Cox and Poisson regressions with robust variance, and using log-binomial regression. Crude and adjusted odds ratios (ORs) were obtained using logistic regression. RESULTS The point and interval estimates obtained using Cox and Poisson regressions were very similar to those obtained using Mantel-Haenszel stratification, independent of the outcome prevalence and the covariates in the model. The log-binomial model presented convergence difficulties when the outcome had high prevalence and there was a continuous covariate in the model. Logistic regression produced point and interval estimates that were higher than those obtained using the other methods, particularly when for outcomes with high initial prevalence. If interpreted as PR estimates, the ORs would overestimate the associations for outcomes with low, intermediate and high prevalence by 13%, almost by 100% and fourfold, respectively. CONCLUSIONS In analyses of data from cross-sectional studies, the Cox and Poisson models with robust variance are better alternatives than logistic regression is. The log-binomial regression model produces unbiased PR estimates, but may present convergence difficulties when the outcome is very prevalent and the confounding variable is continuous.

Methods for estimating prevalence ratios in cross-sectional studies.

OBJETIVO: Comparar empiricamente las regresiones de Cox, log-binomial, Poisson y logistica para estimar razones de prevalencia en estudios de cohorte transversal. METODOS: Fueron utilizados datos de un estudio epidemiologico transversal (n=2.072), de base poblacional, realizado con ancianos en la ciudad de Sao Paulo (Sureste de Brasil), entre mayo de 2003 y abril de 2005. Diagnosticos de demencia, posibles casos de trastorno mental comun y autopercepcion de salud pesima fueron escogidos como resultados con prevalencia baja, intermedia y alta, respectivamente. Fueron utilizadas variables de confusion con dos o mas categorias o valores continuos. Valores de referencia para estimaciones por punto y por intervalo para las razones de prevalencia (RP) fueron obtenidos por el metodo de estratificacion de Mantel-Haenszel. Estimaciones ajustadas fueron calculadas utilizando regresiones de Cox y Poisson con varianza robusta, y regresion log-binomial. Odds ratios (OR) brutos y ajustados fueron obtenidos por la regresion logistica. RESULTADOS: Las estimaciones por punto y por intervalo obtenidas por las regresiones de Cox y Poisson fueron semejantes a la obtenida por la estratificacion de Mantel-Haenszel, independientemente de la prevalencia del resultado y de las covariables del modelo. El modelo log-binomial presento dificultad de convergencia cuando el resultado tenia prevalencia alta y habia convariable continua en el modelo. La regresion logistica produjo estimaciones por punto y por intervalo mayores de las obtenidas por los otros metodos, principalmente para los resultados con mayores prevalencia iniciales. Si se interpretaban como estimaciones de RP, los OR superestimarian las asociaciones para los resultados con prevalencia inicial baja, intermedia y alta en 13%, casi 100% y cuatro veces mas, respectivamente. CONCLUSIONES: En analisis de datos de estudios de cohorte transversal, los modelos Cox y Poisson con varianza robusta son mejores alternativas que la regresion logistica. El modelo de regresion log-binomial produjo estimaciones sin sesgo de la RP, pero puede presentar dificultad de convergencia cuando el resultado es muy frecuente y la variable de confusion es continua.OBJECTIVE To empirically compare the Cox, log-binomial, Poisson and logistic regressions to obtain estimates of prevalence ratios (PR) in cross-sectional studies. METHODS Data from a population-based cross-sectional epidemiological study (n = 2072) on elderly people in Sao Paulo (Southeastern Brazil), conducted between May 2003 and April 2005, were used. Diagnoses of dementia, possible cases of common mental disorders and self-rated poor health were chosen as outcomes with low, intermediate and high prevalence, respectively. Confounding variables with two or more categories or continuous values were used. Reference values for point and interval estimates of prevalence ratio (PR) were obtained by means of the Mantel-Haenszel stratification method. Adjusted PR estimates were calculated using Cox and Poisson regressions with robust variance, and using log-binomial regression. Crude and adjusted odds ratios (ORs) were obtained using logistic regression. RESULTS The point and interval estimates obtained using Cox and Poisson regressions were very similar to those obtained using Mantel-Haenszel stratification, independent of the outcome prevalence and the covariates in the model. The log-binomial model presented convergence difficulties when the outcome had high prevalence and there was a continuous covariate in the model. Logistic regression produced point and interval estimates that were higher than those obtained using the other methods, particularly when for outcomes with high initial prevalence. If interpreted as PR estimates, the ORs would overestimate the associations for outcomes with low, intermediate and high prevalence by 13%, almost by 100% and fourfold, respectively. CONCLUSIONS In analyses of data from cross-sectional studies, the Cox and Poisson models with robust variance are better alternatives than logistic regression is. The log-binomial regression model produces unbiased PR estimates, but may present convergence difficulties when the outcome is very prevalent and the confounding variable is continuous.

High prevalence of dementia among older adults from poor socioeconomic backgrounds in Sao Paulo, Brazil

BACKGROUND The aim of this study was to determine the prevalence of dementia in a socioeconomically disadvantaged population of older adults living in the city of São Paulo, Brazil. METHODS A cross-sectional one-phase population-based study was carried out among all residents aged > or = 65 in defined census sectors of an economically disadvantaged area of São Paulo. Identification of cases of dementia followed the protocol developed by the 10/66 Dementia Research Group. RESULTS Of 2072 individuals in the study, 105 met the criteria for a diagnosis of dementia, yielding a prevalence of 5.1%. Prevalence increased with age for both men and women after age 75 years, but was stable from 65 to 74 years. Low education and income were associated with increased risk of dementia. CONCLUSIONS The prevalence of dementia among older adults from low socioeconomic backgrounds is high. This may be partly due to adverse socioeconomic conditions and consequent failure to compress morbidity into the latter stages of life. The increasing survival of poorer older adults with dementia living in developing countries may lead to a rapid increase in the prevalence of dementia worldwide.

Age-related gray matter volume changes in the brain during non-elderly adulthood

Previous magnetic resonance imaging (MRI) studies described consistent age-related gray matter (GM) reductions in the fronto-parietal neocortex, insula and cerebellum in elderly subjects, but not as frequently in limbic/paralimbic structures. However, it is unclear whether such features are already present during earlier stages of adulthood, and if age-related GM changes may follow non-linear patterns at such age range. This voxel-based morphometry study investigated the relationship between GM volumes and age specifically during non-elderly life (18–50 years) in 89 healthy individuals (48 males and 41 females). Voxelwise analyses showed significant (p < 0.05, corrected) negative correlations in the right prefrontal cortex and left cerebellum, and positive correlations (indicating lack of GM loss) in the medial temporal region, cingulate gyrus, insula and temporal neocortex. Analyses using ROI masks showed that age-related dorsolateral prefrontal volume decrements followed non-linear patterns, and were less prominent in females compared to males at this age range. These findings further support for the notion of a heterogeneous and asynchronous pattern of age-related brain morphometric changes, with region-specific non-linear features.

Risk factors across the life course and dementia in a Brazilian population: results from the Sao Paulo Ageing & Health Study (SPAH)

BACKGROUND Several mechanisms have been suggested to explain the association between adversities across life and dementia. This study aimed to investigate the association between indicators of socioeconomic disadvantages throughout the life-course and dementia among older adults in Sao Paulo, Brazil and to explore possible causal pathways. METHODS We used baseline data from the SPAH study which involved participants aged 65 years and older (n = 2005). The outcome of interest was prevalent dementia. Exposures included in the analyses were socioeconomic position (SEP) indicators in childhood (place of birth and literacy) and adulthood (occupation and income), anthropometric measurements as markers of intrauterine and childhood environment (head circumference and leg length), smoking, diabetes and hypertension. Logistic regression models were used to test the hypothesized pathways and to assess whether there was an association between cumulative adversities across the life course and prevalent dementia. RESULTS Indicators of socioeconomic disadvantage in early life were associated with increased prevalence of dementia. This association was partially mediated through adulthood SEP. Head circumference and leg length were also clearly associated with dementia but there was no evidence that this association was mediated by early life socioeconomic disadvantage. There was an association between cumulative unfavourable conditions across the life course and dementia. CONCLUSIONS Early life disadvantages seem to operate through biological mechanisms associated with passive brain reserve and opportunities in life representing active cognitive reserve. Prevention of dementia should start early in life and continue through life span as seen with many other chronic diseases.

Investigações epidemiológicas sobre demência nos países em desenvolvimento

As the world population is ageing, dementia becomes an important public health problem, particularly in developing countries. Epidemiological research in these settings is scarce and present additional methodological difficulties, mainly regarding the socio-cultural adequacy of instruments used to identify cases of dementia. As a result of these concerns the 10/66 Dementia Research Group was founded to fill this gap. This is an international network of investigators, mostly from developing countries, and the groups name was based on the paradox that less than 10% of the population-based studies on dementia are directed to 2/3 or more cases of people with dementia living in developing countries. The aim of the paper is to update data in the literature regarding the differences in dementia prevalence and incidence seen in developed and developing countries.

Incidence of first-contact psychosis in São Paulo, Brazil

BACKGROUND Little is known about the incidence of first-episode psychosis in urban centres of low- or middle-income countries. AIMS To estimate the incidence of psychosis in São Paulo, a large metropolis of Brazil. METHOD Prospective survey of first-episode psychosis among residents aged 18-64 years resident in a defined area of São Paulo, over a 30-month period (July 20 2002-December 2004). Assessments were carried out with the SCID-I, and diagnoses given according to DSM-IV criteria. Population at risk was drawn from the 2000 Census data. RESULTS There were 367 first-episode cases identified (51% women), and almost 40% fulfilled criteria for schizophrenia or schizophreniform disorder. The incidence rate for any psychosis was 15.8/100 000 person-years at risk (95% CI 14.3-17.6). Incidence of non-affective psychoses was higher among younger males. CONCLUSIONS Incidence of psychosis in São Paulo was lower than expected for a large metropolis.

Grey matter abnormalities in Brazilians with first-episode psychosis

BACKGROUND In low- and middle-income countries people with schizophrenia are reported to experience better outcomes than those in high-income countries. AIMS To examine structural brain differences in people with first-episode psychosis and controls in Brazil. METHOD Magnetic resonance imaging using voxel-based morphometry was performed on 122 people with first-episode psychosis and 94 controls. RESULTS There were significant decreases in grey matter in the left superior temporal and inferior prefrontal cortices, insula bilaterally and the right hippocampal region in first-episode psychosis (P<0.05, corrected for multiple comparisons). The subgroup of people with schizophrenia (n=62) exhibited a similar pattern of decrease in grey matter relative to controls. CONCLUSIONS Structural abnormalities reported in psychosis in high-income countries are also present in first-episode psychosis in Brazil.