DIET@NET: Best Practice Guidelines for dietary assessment in health research

BACKGROUND

Dietary assessment is complex, and strategies to select the most appropriate dietary assessment tool (DAT) in epidemiological research are needed. The DIETary Assessment Tool NETwork (DIET@NET) aimed to establish expert consensus on Best Practice Guidelines (BPGs) for dietary assessment using self-report.

METHODS

The BPGs were developed using the Delphi technique. Two Delphi rounds were conducted. A total of 131 experts were invited, and of these 65 accepted, with 48 completing Delphi round I and 51 completing Delphi round II. In all, a total of 57 experts from North America, Europe, Asia and Australia commented on the 47 suggested guidelines.

RESULTS

Forty-three guidelines were generated, grouped into the following four stages: Stage I. Define what is to be measured in terms of dietary intake (what? who? and when?); Stage II. Investigate different types of DATs; Stage III. Evaluate existing tools to select the most appropriate DAT by evaluating published validation studies; Stage IV. Think through the implementation of the chosen DAT and consider sources of potential biases.

CONCLUSIONS

The Delphi technique consolidated expert views on best practice in assessing dietary intake. The BPGs provide a valuable guide for health researchers to choose the most appropriate dietary assessment method for their studies. These guidelines will be accessible through the Nutritools website, www.nutritools.org .

Underweight among rural Indian adults: burden, and predictors of incidence and recovery

OBJECTIVE

To study the magnitude and predictors of underweight, incident underweight and recovery from underweight among rural Indian adults.

DESIGN

Prospective cohort study. Each participant's BMI was measured in 2008 and 2012 and categorized as underweight (BMI<18·5 kg/m2), normal (BMI=18·5-22·9 kg/m2) or overweight/obese (BMI ≥23·0 kg/m2). Incident underweight was defined as a transition from normal weight or overweight/obese in 2008 to underweight in 2012, and recovery from underweight as a transition from underweight in 2008 to normal weight in 2012. Bivariate and multivariable logistic regression analyses were employed.

SETTING

The Birbhum Health and Demographic Surveillance System, West Bengal, India.

SUBJECTS

Predominantly rural individuals (n 6732) aged ≥18 years enrolled in 2008 were followed up in 2012.

RESULTS

In 2008, the prevalence of underweight was 46·5 %. From 2008 to 2012, 25·8 % of underweight persons transitioned to normal BMI, 12·9 % of normal-weight persons became underweight and 0·1 % of overweight/obese persons became underweight. Multivariable models reveal that people aged 25-49 years, educated and wealthier people, and non-smokers had lower odds of underweight in 2008 and lower odds of incident underweight. Odds of recovery from underweight were lower among people aged ≥36 years and higher among educated (Grade 6 or higher) individuals.

CONCLUSIONS

The current study highlights a high incidence of underweight and important risk factors and modifiable predictors of underweight in rural India, which may inform the design of local nutrition interventions.

Interventions to reduce consumption of sugar-sweetened beverages or increase water intake: evidence from a systematic review and meta-analysis

A systematic review and meta-analyses were conducted to evaluate the effects of interventions to reduce sugar-sweetened beverages (SSB) or increase water intakes and to examine the impact of behaviour change techniques (BCTs) in consumption patterns. Randomized and nonrandomized controlled trials published after January 1990 and until December 2016 reporting daily changes in intakes of SSB or water in volumetric measurements (mL d^-1 ) were included. References were retrieved through searches of electronic databases and quality appraisal followed Cochrane principles. We calculated mean differences (MD) and synthesized data with random-effects models. Forty studies with 16 505 participants were meta-analysed. Interventions significantly decreased consumption of SSB in children by 76 mL d^-1 (95% confidence interval [CI] -105 to -46; 23 studies, P < 0.01), and in adolescents (-66 mL d^-1 , 95% CI -130 to -2; 5 studies, P = 0.04) but not in adults (-13 mL d^-1 , 95% CI -44 to 18; 12 studies, P = 0.16). Pooled estimates of water intakes were only possible for interventions in children, and results were indicative of increases in water intake (MD +67 mL d^-1 , 95% CI 6 to 128; 7 studies, P = 0.04). For children, there was evidence to suggest that modelling/demonstrating the behaviour helped to reduce SSB intake and that interventions within the home environment had greater effects than school-based interventions. In conclusion, public health interventions - mainly via nutritional education/counselling - are moderately successful at reducing intakes of SSB and increasing water intakes in children. However, on average, only small reductions in SSBs have been achieved by interventions targeting adolescents and adults. Complementary measures may be needed to achieve greater improvements in both dietary behaviours across all age groups.

Vitamin D insufficiency/deficiency is associated with insulin resistance in Brazilian children, regardless of body fat distribution

OBJECTIVE

To assess the prevalence of vitamin D insufficiency and deficiency and its association with cardiometabolic risk factors, controlled by adiposity, in a representative sample of prepubescent children.

DESIGN

Cross-sectional population-based study. Body composition was evaluated by dual-energy X-ray absorptiometry. Anthropometric measures and blood pressure were performed. Laboratory analyses were performed to determine the levels of vitamin D (25-hydroxyitamin D; 25(OH)D), glucose, insulin, serum lipids and intact parathyroid hormone. Dietary intake was assessed by three 24 h recalls.

SETTING

Viçosa, Minas Gerais, Brazil, 2015.

SUBJECTS

Representative sample of 378 children aged 8 and 9 years from urban schools.

RESULTS

Inadequate serum concentrations of 25(OH)D were diagnosed in more than half of the children and none of them met the recommended vitamin D intake. After adjusting for confounding factors in the multiple regression analysis, lower prevalence of insulin resistance and hypertriacylglycerolaemia was found in children with serum 25(OH)D levels ≥75 nmol/l (prevalence ratio=0·25; 95 % CI 0·08, 0·85) and ≥50 nmol/l (prevalence ratio=0·61; 95 % CI 0·37, 0·99), respectively. However, after adjusting for different indicators of adiposity, insulin resistance remained independently associated and the association with hypertriacylglycerolaemia was lost after adjusting for central adiposity. The prevalence of vitamin D insufficiency/deficiency was associated with the number of cardiometabolic alterations in children.

CONCLUSIONS

The study results showed that prevalence of vitamin D insufficiency/deficiency was high among the children and insulin resistance was the main cardiometabolic alteration associated with this condition, even in a tropical climate country such as Brazil.

The Influence of Pre-natal Supplement Initiation on Preterm Birth Among Majority Hispanic Women in Los Angeles County: The Role of Nativity

Objectives Despite being encouraged to take pre-natal supplements, suboptimal micronutrient intake is common in pregnancy, especially among Hispanic women. In this study, we assessed whether initiation and timing of pre-natal vitamin use influences the risk of preterm birth. Methods Women who gave birth to singletons in 2003 in Los Angeles County, California, were enrolled into a population-based case-control study. Focusing on non-Hispanic white and Hispanic women, associations between timing of pre-natal supplement use and preterm birth were assessed using logistic regression. Results Among Hispanic mothers, the odds of preterm birth increased the later a woman initiated pre-natal supplement use in pregnancy. The magnitude of this association was larger in US-born compared to foreign-born Hispanic women. Conclusions These findings suggest that nativity may modify the relationship between pre-natal supplement use and preterm birth possibly due to underlying differences in diet composition of Hispanic women by place of birth.

Measuring diet in the 21st century: use of new technologies

The advent of the internet and smartphone technology has allowed dietary assessment to reach the 21st century! The variety of foods available on the supermarket shelf is now greater than ever before. New approaches to measuring diet may help to reduce measurement error and advance our understanding of nutritional determinants of disease. This advance provides the potential to capture detailed dietary data on large numbers of individuals without the need for costly and time-consuming manual nutrition coding. This aim of the present paper is to review the need for new technologies to measure diet with an overview of tools available. The three main areas will be addressed: (1) development of web-based tools to measure diet; (2) use of smartphone apps to self-monitor diet; (3) improving the quality of dietary assessment through development of an online library of tools. A practical example of the development of a web-based tool to assess diet myfood24 (www.myfood24.org) will be given exploring its potential, limitations and challenges. The development of a new food composition database using back-of-pack information will be described. Smartphone apps used to measure diet with a focus on obesity will be reviewed. Many apps are unreliable in terms of tracking, and most are not evaluated. Accurate and consistent measurement of diet is needed for public health and epidemiology. The choice of the most appropriate dietary assessment method tends to rely on experience. The DIET@NET partnership has developed best practice guidelines for selection of dietary assessment tools, which aim to improve the quality, consistency and comparability of dietary data. These developments provide us with a step-change in our ability to reliably characterise food and nutrient intake in population studies. The need for high-quality, validated systems will be important to fully realise the benefits of new technologies.

Dietary polychlorinated biphenyls, long-chain n-3 polyunsaturated fatty acids and incidence of malignant melanoma

BACKGROUND

For malignant melanoma, other risk factors aside from sun exposure have been hardly explored. Polychlorinated biphenyls (PCBs)-mainly from fatty fish- may affect melanogenesis and promote melanoma progression, while long-chain n-3 polyunsaturated fatty acids seem to exert antineoplastic actions in melanoma cells.

OBJECTIVES

We aimed to assess the association of validated estimates of dietary PCB exposure as well as the intake of eicosapentaenoic acid and docosahexaenoic acid (EPA-DHA), accounting for sun habits and skin type, with the risk of malignant melanoma in middle-aged and elderly women.

METHODS

We included 20,785 women at baseline in 2009 from the prospective population-based Swedish Mammography Cohort. Validated estimates of dietary PCB exposure and EPA-DHA intake were obtained via a food frequency questionnaire. Incident melanoma cases were ascertained through register-linkage.

RESULTS

During 4.5 years of follow-up, we ascertained 67 incident cases of melanoma. After multivariable adjustments, exposure to dietary PCBs was associated with four-fold increased risk of malignant melanoma (hazard ratio [HR], 4.0 [95% confidence interval {CI}, 1.2-13; P for trend = 0.02]), while EPA-DHA intake was associated with 80% lower risk (HR, 0.2 [95% CI, 0.1-0.8; P for trend = 0.03]), comparing the highest exposure tertiles with the lowest.

CONCLUSION

While we found a direct association between dietary PCB exposure and risk of melanoma, EPA-DHA intake showed to have a substantial protective association. Question of benefits and risk from fish consumption is very relevant and further prospective studies in the general population verifying these findings are warranted.

Sugars and health: a review of current evidence and future policy

The automation of the process of extracting sugars in the 1900s reduced cost and increased availability of sugars leading to a dramatic rise in consumption, which reached a peak in the 1970s. There are different definitions for sugars not naturally available in foods, and free sugars is the term used by WHO. The epidemiological evidence of the associations between sugars and obesity and type 2 diabetes mellitus is fairly strong and consistent, particularly for sugar sweetened drinks in adults. The Department of Health in the UK and many other countries have recently updated their recommendations for free sugars as a result of this scientific evidence. In the UK the recommended amount of free sugars is currently 5 % of energy (reduced from 10 %), which is difficult to meet and very different from current British dietary patterns. Reducing intakes of free sugars is a challenge and will necessitate a range of different actions and policies. Public Health England has put forward eight suggestions but the four most likely to improve dietary behaviour based on available evidence are social marketing, reduction of marketing of high sugar foods and drinks to children, reformulation and reductions in portion size and a sugar excise tax. Any action taken needs to be evaluated to check inequalities are not widened. The new childhood obesity strategy has incorporated some but not all of these strategies and may not go far enough. It is likely that government policies alone will not be sufficient and a change in the food culture is necessary to see real progress.

Determinants of consumption-day amounts applicable for the estimation of usual dietary intake with a short 24-h food list

Next to the information on frequency of food consumption, information on consumption-day amounts is important to estimate usual dietary intake in epidemiological studies. Our objective was to identify determinants of consumption-day amounts to derive person-specific standard consumption-day amounts applicable for the estimation of usual dietary intake using separate sources to assesss information on consumption probability and amount consumed. 24-h Dietary recall data from the German National Nutrition Survey II (n = 8522; aged 20–80 years) conducted between 2005 and 2007 were analysed for determinants of consumption-day amounts of thirty-eight food and beverage groups using LASSO variable selection for linear mixed-effects models. Determinants included sex, age, BMI, smoking status, years of education, household net income, living status and employment status. Most often, sex, age and smoking status were selected as predictors for consumption-day amounts across thirty-eight food groups. In contrast, living with a partner, employment status and household net income were less frequently chosen. Overall, different determinants were of relevance for different food groups. The number of selected determinants ranged from eight for coffee and juice to zero for cabbage, tea, root vegetables, leafy vegetables, fruit vegetables, legumes, offal, vegetable oils, and other fats. For the estimation of usual dietary intake in a combined approach with a 24-h food list, person-specific standard consumption-day amounts could be used. Sex, age and smoking status were shown to be the most relevant predictors in our analysis. Their impact on the estimation of usual dietary intake needs to be evaluated in future studies.

Moment reconstruction and moment-adjusted imputation when exposure is generated by a complex, nonlinear random effects modeling process

For the classical, homoscedastic measurement error model, moment reconstruction (Freedman et al., 2004, 2008) and moment-adjusted imputation (Thomas et al., 2011) are appealing, computationally simple imputation-like methods for general model fitting. Like classical regression calibration, the idea is to replace the unobserved variable subject to measurement error with a proxy that can be used in a variety of analyses. Moment reconstruction and moment-adjusted imputation differ from regression calibration in that they attempt to match multiple features of the latent variable, and also to match some of the latent variable's relationships with the response and additional covariates. In this note, we consider a problem where true exposure is generated by a complex, nonlinear random effects modeling process, and develop analogues of moment reconstruction and moment-adjusted imputation for this case. This general model includes classical measurement errors, Berkson measurement errors, mixtures of Berkson and classical errors and problems that are not measurement error problems, but also cases where the data-generating process for true exposure is a complex, nonlinear random effects modeling process. The methods are illustrated using the National Institutes of Health-AARP Diet and Health Study where the latent variable is a dietary pattern score called the Healthy Eating Index-2005. We also show how our general model includes methods used in radiation epidemiology as a special case. Simulations are used to illustrate the methods.

Plasma fatty acid biomarkers are associated with gait speed in community-dwelling older adults: The Three-City-Bordeaux study

BACKGROUND & AIMS

Nutritional factors, such as fatty acids (FA), could modulate physical performance in the elderly. In particular, the opposite properties of long-chain n-3 and n-6 polyunsaturated FAs (LC PUFAs) could impact muscle function. We aimed to assess the cross-sectional association between plasma FAs and gait speed in French elderly community-dwellers.

METHODS

Elderly participants from the Bordeaux centre of the Three-City Study were included. The proportion of 12 FAs, and gait speed (m/s) were measured concomitantly at enrollment. Low gait speed (LGS) was defined as below the first quartile of gait speed. FA patterns were derived from the 12 individual FAs using principal component analysis. Multivariate logistic regression models were used and odds-ratios (OR) were expressed per one additional standard-deviation unit of each plasma FA or per one additional unit of pattern score.

RESULTS

Among 982 participants, 239 (24.3%) had a low gait speed (<0.63 m/s) at baseline. Regarding individually each FA, a higher proportion of eicosapentaenoic acid (EPA) and of docosahexaenoic acid (DHA) were associated with lower odds of LGS (OR = 0.76; 95% CI: 0.63-0.93 and OR = 0.79; 95% CI: 0.67-0.95 respectively). Conversely, a higher arachidonic acid (AA):(EPA + DHA) ratio was associated with higher odds of LGS. Three main FA patterns were identified. A higher score on the FA pattern characterized by higher proportions of LC n-3 PUFAs was associated with lower odds of LGS (OR = 0.78; 95% CI: 0.67-0.90).

CONCLUSIONS

A FA pattern mainly driven by high plasma concentrations of LC n-3 PUFAs is cross-sectionally associated with higher gait speed in community-dwelling older adults, while a higher AA:(EPA + DHA) ratio is associated with lower gait speed. These findings suggest a potential protective effect of n-3 PUFA on physical performance decline.

[Factors associated with abdominal obesity in children]

Objective:

To identify the association of dietary, socioeconomic factors, sedentary behaviors and maternal nutritional status with abdominal obesity in children.

Methods:

A cross-sectional study with household-based survey, in 36 randomly selected census tracts in the city of Santos, SP. 357 families were interviewed and questionnaires and anthropometric measurements were applied in mothers and their 3-10 years-old children. Assessment of abdominal obesity was made by maternal and child's waist circumference measurement; for classification used cut-off points proposed by World Health Organization (1998) and Taylor et al. (2000) were applied. The association between variables was performed by multiple logistic regression analysis.

Results:

30.5% of children had abdominal obesity. Associations with children's and maternal nutritional status and high socioeconomic status were shown in the univariate analysis. In the regression model, children's body mass index for age (OR=93.7; 95%CI 39.3-223.3), female gender (OR=4.1; 95%CI 1.8-9.3) and maternal abdominal obesity (OR=2.7; 95%CI 1.2-6.0) were significantly associated with children's abdominal obesity, regardless of the socioeconomic status.

Conclusions:

Abdominal obesity in children seems to be associated with maternal nutritional status, other indicators of their own nutritional status and female gender. Intervention programs for control of childhood obesity and prevention of metabolic syndrome should consider the interaction of the nutritional status of mothers and their children.

Depression and serum 25-hydroxyvitamin D in older adults living at northern latitudes - AGES-Reykjavik Study

Low vitamin D status may be associated with depression. Few studies have examined vitamin D and depression in older adults living at northern latitudes. The present study cross-sectionally investigated serum 25-hydroxyvitamin D (25(OH)D) status and depression among 5006 community-dwelling older persons (66-96 years) living in Iceland (latitudes 64-66°N). Depressive symptoms were measured by the fifteen-item Geriatric Depression Scale (GDS-15). Current major depressive disorder was assessed according to Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criteria. Serum 25(OH)D was analysed using chemiluminescence immunoassay and categorised into three groups: deficient (<30 nmol/l); inadequate (30-49·9 nmol/l); and adequate (≥50 nmol/l). There were twenty-eight (2 %) men and fifty (1 %) women with current major depressive disorder. Mean GDS-15 scores for men and women with adequate vitamin D concentrations were 2·1 and 2·2, respectively. Men and women with deficient v. adequate vitamin D status had more depressive symptoms (higher GDS-15 scores) (difference 0·7 (95 % CI 0·4, 0·9) and 0·4 (95 % CI 0·1, 0·6), respectively). Furthermore, men with deficient vitamin D status were more likely to have current major depressive disorder (adjusted OR 2·51; 95 % CI 1·03, 6·13) compared with men with adequate vitamin D status. Associations among women were not significant. In this older population living at northern latitudes, deficient vitamin D status may be associated with depression. Further investigations are warranted to evaluate the pathways that may be associated with risk of depression among older adults.

Measurement error models with interactions

An important use of measurement error models is to correct regression models for bias due to covariate measurement error. Most measurement error models assume that the observed error-prone covariate (WW ) is a linear function of the unobserved true covariate (X) plus other covariates (Z) in the regression model. In this paper, we consider models for W that include interactions between X and Z. We derive the conditional distribution of X given W and Z and use it to extend the method of regression calibration to this class of measurement error models. We apply the model to dietary data and test whether self-reported dietary intake includes an interaction between true intake and body mass index. We also perform simulations to compare the model to simpler approximate calibration models.

Length of urban residence and obesity among within-country rural-to-urban Andean migrants

Objective

To evaluate the association between length of residence in an urban area and obesity among Peruvian rural-to-urban migrants.

Design

Cross-sectional database analysis of the migrant group from the PERU MIGRANT Study (2007). Exposure was length of urban residence, analysed as both a continuous (10-year units) and a categorical variable. Four skinfold site measurements (biceps, triceps, subscapular and suprailiac) were used to calculate body fat percentage and obesity (body fat percentage >25% males, >33% females). We used Poisson generalized linear models to estimate adjusted prevalence ratios and 95 % confidence intervals. Multicollinearity between age and length of urban residence was assessed using conditional numbers and correlation tests.

Setting

A peri-urban shantytown in the south of Lima, Peru.

Subjects

Rural-to-urban migrants (n 526) living in Lima.

Results

Multivariable analyses showed that for each 10-year unit increase in residence in an urban area, rural-to-urban migrants had, on average, a 12 % (95 % CI 6, 18 %) higher prevalence of obesity. This association was also present when length of urban residence was analysed in categories. Sensitivity analyses, conducted with non-migrant groups, showed no evidence of an association between 10-year age units and obesity in rural (P=0·159) or urban populations (P=0·078). High correlation and a large conditional number between age and length of urban residence were found, suggesting a strong collinearity between both variables.

Conclusions

Longer lengths of urban residence are related to increased obesity in rural-to-urban migrant populations; therefore, interventions to prevent obesity in urban areas may benefit from targeting migrant groups.

A bivariate measurement error model for semicontinuous and continuous variables: Application to nutritional epidemiology

Semicontinuous data in the form of a mixture of a large portion of zero values and continuously distributed positive values frequently arise in many areas of biostatistics. This article is motivated by the analysis of relationships between disease outcomes and intakes of episodically consumed dietary components. An important aspect of studies in nutritional epidemiology is that true diet is unobservable and commonly evaluated by food frequency questionnaires with substantial measurement error. Following the regression calibration approach for measurement error correction, unknown individual intakes in the risk model are replaced by their conditional expectations given mismeasured intakes and other model covariates. Those regression calibration predictors are estimated using short-term unbiased reference measurements in a calibration substudy. Since dietary intakes are often "energy-adjusted," e.g., by using ratios of the intake of interest to total energy intake, the correct estimation of the regression calibration predictor for each energy-adjusted episodically consumed dietary component requires modeling short-term reference measurements of the component (a semicontinuous variable), and energy (a continuous variable) simultaneously in a bivariate model. In this article, we develop such a bivariate model, together with its application to regression calibration. We illustrate the new methodology using data from the NIH-AARP Diet and Health Study (Schatzkin et al., 2001, American Journal of Epidemiology 154, 1119-1125), and also evaluate its performance in a simulation study.

Associations between flavan-3-ol intake and CVD risk in the Norfolk cohort of the European Prospective Investigation into Cancer (EPIC-Norfolk)

Dietary intervention studies suggest that flavan-3-ol intake can improve vascular function and reduce the risk of cardiovascular diseases (CVD). However, results from prospective studies failed to show a consistent beneficial effect. Associations between flavan-3-ol intake and CVD risk in the Norfolk arm of the European Prospective Investigation into Cancer and Nutrition (EPIC-Norfolk) were investigated. Data were available from 24,885 (11,252 men; 13,633 women) participants, recruited between 1993 and 1997 into the EPIC-Norfolk study. Flavan-3-ol intake was assessed using 7-day food diaries and the FLAVIOLA Flavanol Food Composition database. Missing data for plasma cholesterol and vitamin C were imputed using multiple imputation. Associations between flavan-3-ol intake and blood pressure at baseline were determined using linear regression models. Associations with CVD risk were estimated using Cox regression analyses. Median intake of total flavan-3-ols was 1034mg/d (range: 0-8531mg/d) for men and 970mg/d (0-6695mg/d) for women, median intake of flavan-3-ol monomers was 233mg/d (0-3248mg/d) for men and 217 (0-2712mg/d) for women. There were no consistent associations between flavan-3-ol monomer intake and baseline systolic and diastolic blood pressure (BP). After 286,147 person-years of follow-up, there were 8463 cardiovascular events and 1987 CVD related deaths; no consistent association between flavan-3-ol intake and CVD risk (HR 0.93, 95% CI: 0.87; 1.00; Q1 vs Q5) or mortality was observed (HR 0.93, 95% CI: 0.84; 1.04). Flavan-3-ol intake in EPIC-Norfolk is not sufficient to achieve a statistically significant reduction in CVD risk.

Secular trends and factors associated with overweight among Brazilian preschool children: PNSN-1989, PNDS-1996, and 2006/07

OBJECTIVE

to describe the secular trends in overweight among preschool children in the years 1989, 1996, and 2006, and to identify risk factors associated with this condition in 2006.

METHODS

anthropometric data from three surveys (1989, 1996, and 2006) with a representative sample of the population were analyzed. Overweight was defined as the weight-for-height Z-score. The multivariable models of overweight association with risk factors were generated by Poisson regression, and the estimates were shown as prevalence ratios with their respective 95% confidence intervals (PR [95% CI]).

RESULTS

throughout the 17-year period studied, the relative prevalence of overweight in preschoolers increased by 160% in Brazil, representing an increase of 9.4% per year. Based on data from the National Survey on Demography and Health of Women and Children - 2006/07, four multivariable models were created (macro-environmental, maternal, individual, and final model) assuming hierarchy among the risk factors. In the final model, only the following remained associated with overweight: regions South/Southeast (1.55 [1.17 to 2.06]), middle-class (1.35 [1.02 to 1.77]), maternal obesity (1.66 [1.22 to 2.27]), birth weight ≥ 3.9kg (1.87 [1.31 to 2.67]), and being an only child or having only one sibling (1.81 [1.31 to 2.49]).

CONCLUSION

the prevalence of overweight among preschool children in Brazil has increased dramatically over the past 17 years, and it was higher in the 1996-2006 period. Future strategies for prevention and control of overweight in public health should focus or intensify actions in communities that are characterized by the presence of the risks identified in the present study.

Estimating the Distribution of Dietary Consumption Patterns

In the United States the preferred method of obtaining dietary intake data is the 24-hour dietary recall, yet the measure of most interest is usual or long-term average daily intake, which is impossible to measure. Thus, usual dietary intake is assessed with considerable measurement error. We were interested in estimating the population distribution of the Healthy Eating Index-2005 (HEI-2005), a multi-component dietary quality index involving ratios of interrelated dietary components to energy, among children aged 2-8 in the United States, using a national survey and incorporating survey weights. We developed a highly nonlinear, multivariate zero-inflated data model with measurement error to address this question. Standard nonlinear mixed model software such as SAS NLMIXED cannot handle this problem. We found that taking a Bayesian approach, and using MCMC, resolved the computational issues and doing so enabled us to provide a realistic distribution estimate for the HEI-2005 total score. While our computation and thinking in solving this problem was Bayesian, we relied on the well-known close relationship between Bayesian posterior means and maximum likelihood, the latter not computationally feasible, and thus were able to develop standard errors using balanced repeated replication, a survey-sampling approach.

Use of genetic markers and gene-diet interactions for interrogating population-level causal influences of diet on health

Differences in diet appear to contribute substantially to the burden of disease in populations, and therefore changes in diet could lead to major improvements in public health. This is predicated on the reliable identification of causal effects of nutrition on health, and unfortunately nutritional epidemiology has deficiencies in terms of identifying these. This is reflected in the many cases where observational studies have suggested that a nutritional factor is protective against disease, and randomized controlled trials have failed to verify this. The use of genetic variants as proxy measures of nutritional exposure-an application of the Mendelian randomization principle-can contribute to strengthening causal inference in this field. Genetic variants are not subject to bias due to reverse causation (disease processes influencing exposure, rather than vice versa) or recall bias, and if obvious precautions are applied are not influenced by confounding or attenuation by errors. This is illustrated in the case of epidemiological studies of alcohol intake and various health outcomes, through the use of genetic variants related to alcohol metabolism (in ALDH2 and ADH1B). Examples from other areas of nutritional epidemiology and of the informative nature of gene-environment interactions interpreted within the Mendelian randomization framework are presented, and the potential limitations of the approach addressed.