Alleviating spatial confounding in frailty models.

The confounding between fixed effects and (spatial) random effects in a regression setup is termed spatial confounding. This topic continues to gain attention and has been studied extensively in recent years, given that failure to account for this may lead to a suboptimal inference. To mitigate this, a variety of projection-based approaches under the class of restricted spatial models are available in the context of generalized linear mixed models. However, these projection approaches cannot be directly extended to the spatial survival context via frailty models due to dimension incompatibility between the fixed and spatial random effects. In this work, we introduce a two-step approach to handle this, which involves (i) projecting the design matrix to the dimension of the spatial effect (via dimension reduction) and (ii) assuring that the random effect is orthogonal to this new design matrix (confounding alleviation). Under a fully Bayesian paradigm, we conduct fast estimation and inference using integrated nested Laplace approximation. Both simulation studies and application to a motivating data evaluating respiratory cancer survival in the US state of California reveal the advantages of our proposal in terms of model performance and confounding alleviation, compared to alternatives.

Volume de Gordura Epicárdica está Associada com Disfunção Endotelial, mas Não com Calcificação Coronariana: Do ELSA-Brasil / Epicardial Fat Volume Is Associated with Endothelial Dysfunction, but not with Coronary Calcification: From the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil)

Resumo Fundamento O aumento no volume de gordura epicárdica (VGE) está relacionado com doença arterial coronariana (DAC), independentemente de gordura visceral ou subcutânea. O mecanismo dessa associação não é claro. O escore de cálcio coronariano (CC) e a disfunção endotelial estão relacionados com eventos coronarianos, mas não está bem esclarecido se o VGE está relacionado com esses marcadores. Objetivos Avaliar a associação entre VGE medido por método automatizado, fatores de risco cardiovasculares, escore de CC, e função endotelial. Métodos: Em 470 participantes do Estudo Longitudinal de Saúde do Adulto LSA-Brasil com medidas de VGE, escore de CC e função endotelial, realizamos modelos multivariados para avaliar a relação entre fatore de risco cardiovascular e VGE (variável resposta), e entre VGE (variável explicativa), e função endotelial ou escore de CC. Valor de p<0,05 bilateral foi considerado estatisticamente significativo. Resultados A idade média foi 55 ± 8 anos, e 52,3% dos pacientes eram homens. O VGE médio foi 111mL (86-144), e a prevalência de escore de CC igual a zero foi 55%. Nas análises multivariadas, um VGE mais alto relacionou-se com sexo feminino, idade mais avançada, circunferência da cintura, e triglicerídeos (p<0,001 para todos). Um VGE mais alto foi associado com pior função endotelial: em comparação ao primeiro quartil, os valores de odds ratio para a amplitude de pulso basal foram (q2=1,22; IC95% 1,07-1,40; q3=1,50, IC95% 1,30-1,74; q4=1,50, IC95% 1,28-1,79) e para a razão de tonometria arterial periférica foram (q2=0,87; IC95% 0,81-0,95; q3=0,86, IC95% 0,79-0,94; q4=0,80, IC95% 0,73-0,89), mas não com escore de CC maior que zero. Conclusão Um VGE mais alto associou-se com comprometimento da função endotelial, mas não com escore de CC. Os resultados sugerem que o VGE esteja relacionado ao desenvolvimento de DAC por uma via diferente da via do CC, possivelmente pela piora da disfunção endotelial e doença microvascular.

Abstract Background The increase in epicardial fat volume (EFV) is related to coronary artery disease (CAD), independent of visceral or subcutaneous fat. The mechanism underlying this association is unclear. Coronary artery calcium (CAC) score and endothelial dysfunction are related to coronary events, but whether EFV is related to these markers needs further clarification. Objectives To evaluate the association between automatically measured EFV, cardiovascular risk factors, CAC, and endothelial function. Methods In 470 participants from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) with measures of EFV, CAC score and endothelial function, we performed multivariable models to evaluate the relation between cardiovascular risk factors and EFV (response variable), and between EFV (explanatory variable) and endothelial function variables or CAC score. Two-sided p <0.05 was considered statistically significant. Results Mean age was 55 ± 8 years, 52.3% of patients were men. Mean EFV was 111mL (IQ 86-144), and the prevalence of CAC score=0 was 55%. In the multivariable analyses, increased EFV was related to female sex, older age, waist circumference, and triglycerides (p<0.001 for all). Higher EFV was associated with worse endothelial function: as compared with the first quartile, the odds ratio for basal pulse amplitude were (q2=1.22, 95%CI 1.07-1.40; q3=1.50, 95%CI 1.30-1.74; q4=1.50, 95%CI 1.28-1.79) and for peripheral arterial tonometry ratio were (q2=0.87, 95%CI 0.81-0.95; q3=0.86, 95%CI 0.79-0.94; q4=0.80, 95%CI 0.73-0.89), but not with CAC score>0. Conclusion Higher EFV was associated with impaired endothelial function, but not with CAC. The results suggest that EFV is related to the development of CAD through a pathway different from the CAC pathway, possibly through aggravation of endothelial dysfunction and microvascular disease.

Epicardial Fat Volume Is Associated with Endothelial Dysfunction, but not with Coronary Calcification: From the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). / Volume de Gordura Epicárdica está Associada com Disfunção Endotelial, mas Não com Calcificação Coronariana: Do ELSA-Brasil.

BACKGROUND: The increase in epicardial fat volume (EFV) is related to coronary artery disease (CAD), independent of visceral or subcutaneous fat. The mechanism underlying this association is unclear. Coronary artery calcium (CAC) score and endothelial dysfunction are related to coronary events, but whether EFV is related to these markers needs further clarification. OBJECTIVES: To evaluate the association between automatically measured EFV, cardiovascular risk factors, CAC, and endothelial function. METHODS: In 470 participants from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) with measures of EFV, CAC score and endothelial function, we performed multivariable models to evaluate the relation between cardiovascular risk factors and EFV (response variable), and between EFV (explanatory variable) and endothelial function variables or CAC score. Two-sided p <0.05 was considered statistically significant. RESULTS: Mean age was 55 ± 8 years, 52.3% of patients were men. Mean EFV was 111mL (IQ 86-144), and the prevalence of CAC score=0 was 55%. In the multivariable analyses, increased EFV was related to female sex, older age, waist circumference, and triglycerides (p<0.001 for all). Higher EFV was associated with worse endothelial function: as compared with the first quartile, the odds ratio for basal pulse amplitude were (q2=1.22, 95%CI 1.07-1.40; q3=1.50, 95%CI 1.30-1.74; q4=1.50, 95%CI 1.28-1.79) and for peripheral arterial tonometry ratio were (q2=0.87, 95%CI 0.81-0.95; q3=0.86, 95%CI 0.79-0.94; q4=0.80, 95%CI 0.73-0.89), but not with CAC score>0. CONCLUSION: Higher EFV was associated with impaired endothelial function, but not with CAC. The results suggest that EFV is related to the development of CAD through a pathway different from the CAC pathway, possibly through aggravation of endothelial dysfunction and microvascular disease.

FUNDAMENTO: O aumento no volume de gordura epicárdica (VGE) está relacionado com doença arterial coronariana (DAC), independentemente de gordura visceral ou subcutânea. O mecanismo dessa associação não é claro. O escore de cálcio coronariano (CC) e a disfunção endotelial estão relacionados com eventos coronarianos, mas não está bem esclarecido se o VGE está relacionado com esses marcadores. OBJETIVOS: Avaliar a associação entre VGE medido por método automatizado, fatores de risco cardiovasculares, escore de CC, e função endotelial. Métodos: Em 470 participantes do Estudo Longitudinal de Saúde do Adulto LSA-Brasil com medidas de VGE, escore de CC e função endotelial, realizamos modelos multivariados para avaliar a relação entre fatore de risco cardiovascular e VGE (variável resposta), e entre VGE (variável explicativa), e função endotelial ou escore de CC. Valor de p<0,05 bilateral foi considerado estatisticamente significativo. RESULTADOS: A idade média foi 55 ± 8 anos, e 52,3% dos pacientes eram homens. O VGE médio foi 111mL (86-144), e a prevalência de escore de CC igual a zero foi 55%. Nas análises multivariadas, um VGE mais alto relacionou-se com sexo feminino, idade mais avançada, circunferência da cintura, e triglicerídeos (p<0,001 para todos). Um VGE mais alto foi associado com pior função endotelial: em comparação ao primeiro quartil, os valores de odds ratio para a amplitude de pulso basal foram (q2=1,22; IC95% 1,07-1,40; q3=1,50, IC95% 1,30-1,74; q4=1,50, IC95% 1,28-1,79) e para a razão de tonometria arterial periférica foram (q2=0,87; IC95% 0,81-0,95; q3=0,86, IC95% 0,79-0,94; q4=0,80, IC95% 0,73-0,89), mas não com escore de CC maior que zero. CONCLUSÃO: Um VGE mais alto associou-se com comprometimento da função endotelial, mas não com escore de CC. Os resultados sugerem que o VGE esteja relacionado ao desenvolvimento de DAC por uma via diferente da via do CC, possivelmente pela piora da disfunção endotelial e doença microvascular.

Estimating hidden populations by transferring knowledge from geographically misaligned levels.

The estimation of hidden sub-populations is a hard task that appears in many fields. For example, public health planning in Brazil depends crucially of the number of people who holds a private health insurance plan and hence rarely uses the public services. Different sources of information about these sub-populations may be available at different geographical levels. The available information can be transferred between these different geographic levels to improve the estimation of the hidden population size. In this study, we propose a model that use individual level information to learn about the dependence between the response variable and explanatory variables by proposing a family of link functions with asymptotes that are flexible enough to represent the real aspects of the data and robust to departures from the model. We use the fitted model to estimate the size of the sub-population at any desired level. We illustrate our methodology estimating the sub-population that uses the public health system in each neighborhood of large cities in Brazil.

Assessing spatial confounding in cancer disease mapping using R.

BACKGROUND: Exploring spatial patterns in the context of cancer disease mapping (DM) is a decisive approach to bring evidence of geographical tendencies in assessing disease status and progression. However, this framework is not insulated from spatial confounding, a topic of significant interest in cancer epidemiology, where the latent correlation between the spatial random effects and fixed effects (such as covariates), often lead to misleading interpretation. AIMS: To introduce three popular approaches (RHZ, HH and SPOCK; details in paper) often employed to tackle spatial confounding, and illustrate their implementation in cancer research via the popular statistical software R. METHODS: As a solution to alleviate spatial confounding, restricted spatial regressions are constructed by either projecting the latent effect onto the orthogonal space of covariates, or by displacing the spatial locations. Popular parametric count data models, such as the Poisson, generalized Poisson and negative binomial, were considered for the areal count responses, while the spatial association is quantified via the conditional autoregressive (CAR) model. Our method of inference in Bayesian, sometimes aided by the integrated nested Laplace approximation (INLA) to accelerate computing. The methods are implemented in the R package RASCO available from the first author's GitHub page. RESULTS: The results reveal that all three methods perform well in alleviating the bias and variance inflation present in the spatial models. The effects of spatial confounding were also explored, which, if ignored in practice, may lead to wrong conclusions. CONCLUSION: Spatial confounding continues to remain a critical bottleneck in deriving precise inference from spatial DM models. Hence, its effects must be investigated, and mitigated. Several approaches are available in the literature, and they produce trustworthy results. The central contribution of this paper is providing the practitioners the R package RASCO, capable of fitting a large number of spatial models, as well as their restricted versions.

Association between C reactive protein and all-cause mortality in the ELSA-Brasil cohort.

BACKGROUND: High-sensitivity C reactive protein (hsCRP) has been proposed as a marker of incident cardiovascular disease and vascular mortality, and may also be a marker of non-vascular mortality. However, most evidence comes from either North American or European cohorts. The present proposal aims to investigate the association of hsCRP with the risk of all-cause mortality in a multiethnic Brazilian population. METHODS: Baseline data (2008-2010) of a cohort of 14 238 subjects participating in the Brazilian Longitudinal Study of Adult Health were used. hsCRP was assayed with immunochemistry. The association of baseline covariates with all-cause mortality was calculated by Cox regression for univariate model and adjusted for different confounders after a mean follow-up of 8.0±1.1 years. The final model was adjusted for age, sex, self-rated race/ethnicity, schooling, health behaviours and prevalent chronic disease. RESULTS: The risk of death increased steadily by quartiles of hsCRP, from 1.45 (95% CI 1.05 to 2.01) in quartile 2 to 1.95 (95% CI 1.42 to 2.69) in quartile 4, compared with quartile 1. Furthermore, the persistence of a significant graded association after the exclusion of deaths in the first year of follow-up suggests that these results are unlikely to be due to reverse causality. Finally, the HR was unaffected by the exclusion of participants who had self-reported medical history of diabetes, cancer and chronic obstructive pulmonary disease. CONCLUSIONS: Our study shows that hsCRP level is associated with mortality in a highly admixed population, independent of a large set of lifestyle and clinical variables.