Statistical Aspects of the Use of Biomarkers in Nutritional Epidemiology Research

Enhancing Capacity for Food and Nutrient Intake Assessment in Population Sciences Research

Nutrition influences health throughout the life course. Good nutrition increases the probability of good pregnancy outcomes, proper childhood development, and healthy aging, and it lowers the probability of developing common diet-related chronic diseases, including obesity, cardiovascular disease, cancer, and type 2 diabetes. Despite the importance of diet and health, studying these exposures is among the most challenging in population sciences research. US and global food supplies are complex; eating patterns have shifted such that half of meals are eaten away from home, and there are thousands of food ingredients with myriad combinations. These complexities make dietary assessment and links to health challenging both for population sciences research and for public health policy and practice. Furthermore, most studies evaluating nutrition and health usually rely on self-report instruments prone to random and systematic measurement error. Scientific advances involve developing nutritional biomarkers and then applying these biomarkers as stand-alone nutritional exposures or for calibrating self-reports using specialized statistics.

Demographic, Health and Lifestyle Factors Associated with the Metabolome in Older Women

Demographic and clinical factors influence the metabolome. The discovery and validation of disease biomarkers are often challenged by potential confounding effects from such factors. To address this challenge, we investigated the magnitude of the correlation between serum and urine metabolites and demographic and clinical parameters in a well-characterized observational cohort of 444 post-menopausal women participating in the Women’s Health Initiative (WHI). Using LC-MS and lipidomics, we measured 157 aqueous metabolites and 756 lipid species across 13 lipid classes in serum, along with 195 metabolites detected by GC-MS and NMR in urine and evaluated their correlations with 29 potential disease risk factors, including demographic, dietary and lifestyle factors, and medication use. After controlling for multiple testing (FDR < 0.01), we found that log-transformed metabolites were mainly associated with age, BMI, alcohol intake, race, sample storage time (urine only), and dietary supplement use. Statistically significant correlations were in the absolute range of 0.2–0.6, with the majority falling below 0.4. Incorporation of important potential confounding factors in metabolite and disease association analyses may lead to improved statistical power as well as reduced false discovery rates in a variety of data analysis settings.

Evaluation of nutritional status and adherence to dietary monitoring among patients with human T-cell leukemia virus type 1 infection

BACKGROUND & AIMS

The nutritional status of people infected with Human T-Cell Leukemia Virus type 1 (HTLV-1) is rarely studied. Malnutrition affects the immune system, but the consequences of this remain unknown in HTLV-1 infection. This study aimed to analyze the nutritional status, food intake, and ability to adhere to dietary counseling in people with symptomatic HTLV-1 infection.

METHODS

This was a longitudinal, prospective, observational study in which the nutritional status of people with symptomatic HTLV-1 was monitored between June 2016 and August 2019 at a referral center. At the baseline (day 0) and last stage (day 90) anthropometric, bioimpedance, and dietary assessments were performed.

RESULTS

In total, 71 participants with symptomatic HTLV-1 infection were enrolled, the majority (43, 60.6%) of whom were female. The mean age was 59.2 (±11.1) years. Participants were often overweight or obese (40, 56.3%) had a BMI above 25 kg/m². The average consumption of recommended fibers and liquids was positively associated with percent lean body mass (%LM). Most participants who were overweight (36.6%) or obese (19.70%) and their nutritional status and eating behavior did not change during the follow-up period. Dietary counseling was unable to promote a change in the monotony of the existing eating habits. Participants with good adherence to nutritional guidelines had higher dietary diversity scores.

CONCLUSION

A high proportion of participants were overweight or obese. Adherence to dietary counseling was low, and its success was limited in improving dietary diversity/nutritional status. Loss to follow-up was a limitation of this study.

Biomarker-Based Methods and Study Designs to Calibrate Dietary Intake for Assessing Diet-Disease Associations

BACKGROUND

Dietary biomarkers measured in biospecimens can play an important role in correcting for random and systematic measurement error in self-reported nutrient intake when assessing diet-disease associations. To date, high-quality biomarkers for calibrating self-reported dietary intake have only been developed for a few nutrients.

OBJECTIVES

To investigate new study designs and regression calibration approaches for calibrating self-reported nutrient intake for use in disease association analyses.

METHODS

We studied 3 regression calibration approaches: 1) an existing approach built on a calibration cohort assuming the existence of an objective biomarker (i.e., biomarker with random independent measurement error), 2) a proposed approach using a biomarker development cohort, and 3) a proposed 2-stage approach using both cohorts. We conducted simulation studies to compare the performance of different study designs/methods for estimating diet-disease associations and applied suitable methods to examine the association of sodium and potassium intake with cardiovascular disease (CVD) risk in Women's Health Initiative cohorts.

RESULTS

Simulation studies showed that the first approach can lead to biased association estimation when the objective biomarker assumption is violated; the second and third proposed approaches obviate the need for such an objective biomarker. Precision for estimating the association depends critically on sample size of the biomarker development cohort and the strength of the self-reported nutrient intake. Analyses based on the second and third approaches support previously reported significant findings using the first approach about associations of the ratio of sodium to potassium intake with CVD risk while providing efficiency gain for some outcomes.

CONCLUSIONS

Self-reported dietary intake needs to be calibrated for measurement error correction in diet-disease association analyses. When there are no existing objective biomarkers that can be used for calibration purpose, controlled feeding studies can be used to develop new biomarkers for use in calibration or can be used to calibrate self-reported dietary intake directly.

The Associations of Plasma Carotenoids and Vitamins With Risk of Age-Related Macular Degeneration: Results From a Matched Case-Control Study in China and Meta-Analysis

Background and Purpose

Data from studies support a beneficial effect of carotenoids and vitamins on an age-related macular degeneration (AMD) risk. However, studies on the relations between blood levels of these nutrients and AMD are limited and provided conflicting results. The objective of this case-control study and meta-analysis was to examine whether the blood concentrations of carotenoids and vitamins were associated with the risk of AMD.

Methods

A total of 164 cases of AMD and an equal number of controls are individually matched according to age and gender among the participants, who provided blood samples in the Xi'an Eye Study. Plasma carotenoids and vitamins were measured using reversed-phase high-performance liquid chromatography. Bonferroni-corrected covariate-adjusted conditional logistic regression were used to estimate AMD risk by category of these nutrients in the multivariable-adjusted model. Nine studies were identified for the meta-analysis and calculated pooled risk estimates by means of a random-effects model.

Results

Plasma concentrations of examined carotenoids and vitamins were significantly lower in patients with AMD than those in controls. Plasma concentrations of examined carotenoids and vitamins were significantly lower in patients with AMD than those in controls. After a multivariate adjustment for body mass index, blood cholesterol, and other lifestyle risk factors, higher lutein/zeaxanthin content in plasma was significantly associated with a decreased risk of AMD, and the odds ratio (OR) comparing the top and bottom tertiles was 0.21 (95% CI: 0.05, 0.84; P_trend = 0.024). Associations for β-carotenes (OR: 0.11; 95% CI: 0.02, 0.50; P_trend < 0.001), and β-cryptoxanthin (OR: 0.08, 95% CI: 0.02, 0.39; P_trend < 0.001) were similar to that for lutein/zeaxanthin. Inverse associations were also observed for a higher level of retinol (OR: 0.14, 95% CI: 0.03, 0.61; P_trend = 0.006) and α-tocopherol (OR: 0.25, 95% CI: 0.06, 0.98; P_trend = 0.006). In the meta-analysis, a protective effect was detected for AMD among the participants with high blood lutein/zeaxanthin level (OR: 0.53, 95% CI: 0.40, 0.72, P < 0.001), compared to those with low level. Similar inverse associations were seen for β-carotene (OR: 0.48, 95% CI: 0.28, 0.84, P = 0.010), β-cryptoxanthin (OR: 0.48, 95% CI: 0.23, 1, P = 0.049), lycopene (OR: 0.70, 95% CI: 0.54, 0.90, P = 0.006) and α-tocopherol (OR: 0.50, 95% CI: 0.31, 0.81, P = 0.005).

Conclusions

Results of the case-control study were consistent with findings from the meta-analysis, indicating that higher concentrations of carotenoids and vitamins were inversely associated with the AMD risk. Our finding supports the current notion that these nutrients are likely to affect the development of AMD and may help to refine the strategies for the prevention of age-related eye diseases.

Perspective: Dietary Biomarkers of Intake and Exposure-Exploration with Omics Approaches

While conventional nutrition research has yielded biomarkers such as doubly labeled water for energy metabolism and 24-h urinary nitrogen for protein intake, a critical need exists for additional, equally robust biomarkers that allow for objective assessment of specific food intake and dietary exposure. Recent advances in high-throughput MS combined with improved metabolomics techniques and bioinformatic tools provide new opportunities for dietary biomarker development. In September 2018, the NIH organized a 2-d workshop to engage nutrition and omics researchers and explore the potential of multiomics approaches in nutritional biomarker research. The current Perspective summarizes key gaps and challenges identified, as well as the recommendations from the workshop that could serve as a guide for scientists interested in dietary biomarkers research. Topics addressed included study designs for biomarker development, analytical and bioinformatic considerations, and integration of dietary biomarkers with other omics techniques. Several clear needs were identified, including larger controlled feeding studies, testing a variety of foods and dietary patterns across diverse populations, improved reporting standards to support study replication, more chemical standards covering a broader range of food constituents and human metabolites, standardized approaches for biomarker validation, comprehensive and accessible food composition databases, a common ontology for dietary biomarker literature, and methodologic work on statistical procedures for intake biomarker discovery. Multidisciplinary research teams with appropriate expertise are critical to moving forward the field of dietary biomarkers and producing robust, reproducible biomarkers that can be used in public health and clinical research.

Biomarker of food intake for assessing the consumption of dairy and egg products

Dairy and egg products constitute an important part of Western diets as they represent an excellent source of high-quality proteins, vitamins, minerals and fats. Dairy and egg products are highly diverse and their associations with a range of nutritional and health outcomes are therefore heterogeneous. Such associations are also often weak or debated due to the difficulty in establishing correct assessments of dietary intake. Therefore, in order to better characterize associations between the consumption of these foods and health outcomes, it is important to identify reliable biomarkers of their intake. Biomarkers of food intake (BFIs) provide an accurate measure of intake, which is independent of the memory and sincerity of the subjects as well as of their knowledge about the consumed foods. We have, therefore, conducted a systematic search of the scientific literature to evaluate the current status of potential BFIs for dairy products and BFIs for egg products commonly consumed in Europe. Strikingly, only a limited number of compounds have been reported as markers for the intake of these products and none of them have been sufficiently validated. A series of challenges hinders the identification and validation of BFI for dairy and egg products, in particular, the heterogeneous composition of these foods and the lack of specificity of the markers identified so far. Further studies are, therefore, necessary to validate these compounds and to discover new candidate BFIs. Untargeted metabolomic strategies may allow the identification of novel biomarkers, which, when taken separately or in combination, could be used to assess the intake of dairy and egg products.

Association of pesticide exposure with human congenital abnormalities

Human pesticide exposure can occur both occupationally and environmentally during manufacture and after the application of indoor and outdoor pesticides, as well as through consumption via residues in food and water. There is evidence from experimental studies that numerous pesticides, either in isolation or in combination, act as endocrine disruptors, neurodevelopmental toxicants, immunotoxicants, and carcinogens. We reviewed the international literature on this subject for the years between 1990 and 2017. The studies were considered in this review through MEDLINE and WHO resources. Out of the n = 1817 studies identified, n = 94 were reviewed because they fulfilled criteria of validity and addressed associations of interest. Epidemiological studies have provided limited evidence linking pre- and post-natal exposure to pesticides with cancers in childhood, neurological deficits, fetal death, intrauterine growth restriction, preterm birth, and congenital abnormalities (CAs). In this review, the potential association between pesticide exposure and the appearance of some human CAs (including among others musculoskeletal abnormalities; neural tube defects; urogenital and cardiovascular abnormalities) was investigated. A trend towards a positive association between environmental or occupational exposure to some pesticides and some CAs was detected, but this association remains to be substantiated. Main limitations of the review include inadequate exposure assessment and limited sample size. Adequately powered studies with precise exposure assessments such as biomonitoring, are warranted to clarify with certainty the potential association between pesticide exposure and human CAs.

Empirically Derived Dietary Patterns in UK Adults Are Associated with Sociodemographic Characteristics, Lifestyle, and Diet Quality

The aim of this study was to examine empirical dietary patterns in UK adults and their association with sociodemographic characteristics, lifestyle factors, self-reported nutrient intake, nutrient biomarkers, and the Nutrient-based Diet Quality Score (NDQS) using National Diet and Nutrition Survey data 2008–2012 (n = 2083; mean age 49 years; 43.3% male). Four patterns explained 13.6% of the total variance: ‘Snacks, fast food, fizzy drinks’ (SFFFD), ‘Fruit, vegetables, oily fish’ (FVOF), ‘Meat, potatoes, beer’ (MPB), and ‘Sugary foods, dairy’ (SFD). ‘SFFFD’ was associated positively with: being male; smoking; body mass index (BMI); urinary sodium; intake of non-milk extrinsic sugars (NMES), fat and starch; and negatively with: age; plasma carotenoids; and NDQS. ‘FVOF’ was associated positively with: being non-white; age; income; socioeconomic classification (National Statistics Socio-economic Classifications; NSSEC); plasma carotenoids; intake of non-starch polysaccharides and polyunsaturated fatty acids. It was negatively associated with: being male, smoking, BMI, urinary sodium, intake of saturated fat; and NMES and NDQS. Whilst the patterns explained only 13.6% of the total variance, they were associated with self-reported nutrient intake, biomarkers of nutrient intake, sociodemographic and lifestyle variables, and the NDQS. These findings provide support for dietary patterns analyses as a means of exploring dietary intake in the UK population to inform public health nutrition policy and guidance.

Invited Commentary: The Contribution to the Field of Nutritional Epidemiology of the Landmark 1985 Publication by Willett et al

The semiquantitative food frequency questionnaire (FFQ) has been the primary source of dietary exposure data in epidemiology for decades. Although frequency instruments had been evaluated before the 1985 publication "Reproducibility and Validity of a Semiquantitative Food Frequency Questionnaire" by Willett et al. (Am J Epidemiol. 1985;122(1):51-65), that paper was the prototype for the development and validation of what was then a highly innovative method for collecting dietary data. This approach was adopted in nearly all subsequent cohort studies of diet and disease. The paper also catalyzed an extended scientific discourse regarding methods for validation, energy adjustment, and measurement error. It is now well established that data from FFQs and other self-reported dietary assessment instruments have both value and error and that this error should be considered in the analysis and interpretation of findings, including sensitivity analyses in which adjustment for measurement error is explored. Advances in technology make it feasible to consider collecting multiple granular short-term instruments such as recalls or records over time in addition to FFQs among all participants in large cohort studies; both provide valuable information. Without a doubt, the 1985 publication by Willett et al. provided the foundation that propelled the field of nutritional epidemiology forward, and it continues to be relevant today.

Biomarker-predicted sugars intake compared with self-reported measures in US Hispanics/Latinos: results from the HCHS/SOL SOLNAS study

OBJECTIVE

Measurement error in self-reported total sugars intake may obscure associations between sugars consumption and health outcomes, and the sum of 24 h urinary sucrose and fructose may serve as a predictive biomarker of total sugars intake.

DESIGN

The Study of Latinos: Nutrition & Physical Activity Assessment Study (SOLNAS) was an ancillary study to the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) cohort. Doubly labelled water and 24 h urinary sucrose and fructose were used as biomarkers of energy and sugars intake, respectively. Participants' diets were assessed by up to three 24 h recalls (88 % had two or more recalls). Procedures were repeated approximately 6 months after the initial visit among a subset of ninety-six participants.

SETTING

Four centres (Bronx, NY; Chicago, IL; Miami, FL; San Diego, CA) across the USA.

SUBJECTS

Men and women (n 477) aged 18-74 years.

RESULTS

The geometric mean of total sugars was 167·5 (95 % CI 154·4, 181·7) g/d for the biomarker-predicted and 90·6 (95 % CI 87·6, 93·6) g/d for the self-reported total sugars intake. Self-reported total sugars intake was not correlated with biomarker-predicted sugars intake (r=-0·06, P=0·20, n 450). Among the reliability sample (n 90), the reproducibility coefficient was 0·59 for biomarker-predicted and 0·20 for self-reported total sugars intake.

CONCLUSIONS

Possible explanations for the lack of association between biomarker-predicted and self-reported sugars intake include measurement error in self-reported diet, high intra-individual variability in sugars intake, and/or urinary sucrose and fructose may not be a suitable proxy for total sugars intake in this study population.

Diet, nutrition, and cancer: past, present and future

Despite the potentially important roles of diet and nutrition in cancer prevention, the evidence to support these roles is widely perceived by the public and health professionals as being inconsistent. In this Review, we present the issues and challenges in conducting and interpreting diet-cancer research, including those relating to the design of epidemiological studies, dietary data collection methods, and factors that affect the outcome of intervention trials. Approaches to improve effect estimates, such as the use of biomarkers to improve the accuracy of characterizing dietary exposures, are also discussed. Nutritional and dietary patterns are complex; therefore, the use of a reductionist approach to investigations, by focusing on specific nutrients, can produce misleading information. The effects of tumour heterogeneity and the failure to appreciate the nonlinear, U-shaped relationship between micronutrients and cancer in both observational studies and clinical trials are discussed. New technologies and investigational approaches are enabling the exploration of complex interactions between genetic, epigenetic, metabolic, and gut-microbial processes that will inform our knowledge of the diet-cancer relationship. Communicating the status of the evolving science in the context of the overall scientific evidence base, and evidence-based dietary recommendations for cancer prevention, should be emphasized in guidance for the public and for individual patients.

Addressing Current Criticism Regarding the Value of Self-Report Dietary Data

Recent reports have asserted that, because of energy underreporting, dietary self-report data suffer from measurement error so great that findings that rely on them are of no value. This commentary considers the amassed evidence that shows that self-report dietary intake data can successfully be used to inform dietary guidance and public health policy. Topics discussed include what is known and what can be done about the measurement error inherent in data collected by using self-report dietary assessment instruments and the extent and magnitude of underreporting energy compared with other nutrients and food groups. Also discussed is the overall impact of energy underreporting on dietary surveillance and nutritional epidemiology. In conclusion, 7 specific recommendations for collecting, analyzing, and interpreting self-report dietary data are provided: (1) continue to collect self-report dietary intake data because they contain valuable, rich, and critical information about foods and beverages consumed by populations that can be used to inform nutrition policy and assess diet-disease associations; (2) do not use self-reported energy intake as a measure of true energy intake; (3) do use self-reported energy intake for energy adjustment of other self-reported dietary constituents to improve risk estimation in studies of diet-health associations; (4) acknowledge the limitations of self-report dietary data and analyze and interpret them appropriately; (5) design studies and conduct analyses that allow adjustment for measurement error; (6) design new epidemiologic studies to collect dietary data from both short-term (recalls or food records) and long-term (food-frequency questionnaires) instruments on the entire study population to allow for maximizing the strengths of each instrument; and (7) continue to develop, evaluate, and further expand methods of dietary assessment, including dietary biomarkers and methods using new technologies.

Considerations in selection of diet assessment methods for examining the effect of nutrition on cognition

Older adults are the most rapidly growing age group in the United States, and it is estimated that 22.2 % of U.S. adults over 71 years of age have cognitive impairments without dementia and 13.9% have dementia. Cognitive impairment is associated with reduced quality of life, increased risk of hospitalization, inability to live independently, and increased health care costs; therefore, identification of modifiable risk factors for prevention and delay of cognitive decline is of increasing importance. There is a growing body of research and interest in the relationship between diet and cognitive function. Epidemiologic studies suggest that cognitive function may be improved and cognitive decline prevented as a function of a particular nutrient, food group or dietary pattern; however, results from these trials have failed to be replicated in randomized controlled trials. One possible reason for the equivocality of findings in the diet and cognitive function literature may be the methodological issues and limitations in the assessment of dietary patterns and nutritional intake. Self-reported dietary data can be biased by many factors such as age, gender, socioeconomic status, and education; yet, there is limited research on the impact of cognitive function on the integrity of self-reported dietary data. Cognitive function itself may bias diet assessment methods, subsequently obscuring the evaluation of the nutrition-cognition relationship. The present review summarizes methodological validation studies that provide insight into potential errors of diet assessment methods due to cognitive function, identifies research gaps and provides recommendations for improving diet assessment accuracy in studies of individuals with cognitive impairments.

Biomarker-calibrated dietary energy and protein intake associations with diabetes risk among postmenopausal women from the Women's Health Initiative

BACKGROUND

Self-report of dietary energy and protein intakes has been shown to be systematically and differentially underreported.

OBJECTIVE

We assessed and compared the association of diabetes among postmenopausal women with biomarker-calibrated and uncalibrated dietary energy and protein intakes from food-frequency questionnaires (FFQs).

DESIGN

The analyses were performed for 74,155 participants of various race-ethnicities from the Women's Health Initiative. Uncalibrated and calibrated energy and protein intakes from FFQs were assessed for associations with incident diabetes by using HR estimates based on Cox regression.

RESULTS

A 20% increment in uncalibrated energy consumption was associated with increased diabetes risk (HR) of 1.03 (95% CI: 1.01, 1.05), 2.41 (95% CI: 2.06, 2.82) with biomarker calibration, and 1.30 (95% CI: 0.96, 1.76) after adjustment for BMI. A 20% increment in uncalibrated protein (g/d) resulted in an HR of 1.05 (95% CI: 1.03, 1.07), 1.82 (95% CI: 1.56, 2.12) with calibration, and 1.16 (95% CI: 1.05, 1.28) with adjustment for BMI. A 20% increment in uncalibrated protein density (% of energy from protein) resulted in an HR of 1.13 (95% CI: 1.09, 1.17), 1.01 (95% CI: 0.75, 1.37) with calibration, and 1.19 (95% CI: 1.07, 1.32) with adjustment for BMI.

CONCLUSIONS

Higher protein and total energy intakes (calibrated) appear to be associated with a substantially increased diabetes risk that may be mediated by an increase in body mass over time. Diet-disease associations without correction of self-reported measurement error should be viewed with caution. This trial is registered at clinicaltrials.gov as NCT00000611.

Validity of a multipass, web-based, 24-hour self-administered recall for assessment of total energy intake in blacks and whites

To date, Web-based 24-hour recalls have not been validated using objective biomarkers. From 2006 to 2009, the validity of 6 Web-based DietDay 24-hour recalls was tested among 115 black and 118 white healthy adults from Los Angeles, California, by using the doubly labeled water method, and the results were compared with the results of the Diet History Questionnaire, a food frequency questionnaire developed by the National Cancer Institute. The authors performed repeated measurements in a subset of 53 subjects approximately 6 months later to estimate the stability of the doubly labeled water measurement. The attenuation factors for the DietDay recall were 0.30 for blacks and 0.26 for whites. For the Diet History Questionnaire, the attenuation factors were 0.15 and 0.17 for blacks and whites, respectively. Adjusted correlations between true energy intake and the recalls were 0.50 and 0.47 for blacks and whites, respectively, for the DietDay recall. For the Diet History Questionnaire, they were 0.34 and 0.36 for blacks and whites, respectively. The rate of underreporting of more than 30% of calories was lower with the recalls than with the questionnaire (25% and 41% vs. 34% and 52% for blacks and whites, respectively). These findings suggest that Web-based DietDay dietary recalls offer an inexpensive and widely accessible dietary assessment alternative, the validity of which is equally strong among black and white adults. The validity of the Web-administered recall was superior to that of the paper food frequency questionnaire.

Using regression calibration equations that combine self-reported intake and biomarker measures to obtain unbiased estimates and more powerful tests of dietary associations

The authors describe a statistical method of combining self-reports and biomarkers that, with adequate control for confounding, will provide nearly unbiased estimates of diet-disease associations and a valid test of the null hypothesis of no association. The method is based on regression calibration. In cases in which the diet-disease association is mediated by the biomarker, the association needs to be estimated as the total dietary effect in a mediation model. However, the hypothesis of no association is best tested through a marginal model that includes as the exposure the regression calibration-estimated intake but not the biomarker. The authors illustrate the method with data from the Carotenoids and Age-Related Eye Disease Study (2001--2004) and show that inclusion of the biomarker in the regression calibration-estimated intake increases the statistical power. This development sheds light on previous analyses of diet-disease associations reported in the literature.

Measurement error modeling and nutritional epidemiology association analyses

This paper summarizes the results of a Nutrient Biomarker Study in the Women's Health Initiative, and its application to studies of the association between energy and protein consumption and the risk of major cancers and cardiovascular diseases. The presentation emphasizes measurement error modeling and related data analysis methods, since addressing measurement issues appears to be central to these topics and to progress in nutritional epidemiology more generally. The manner in which body mass index is modeled in disease association analysis is particularly challenging, since it could serve as a mediator or as a confounder of the association, and at the same time contributes valuably to energy and protein consumption assessment. A hazard ratio parameter estimation procedure that acknowledges body mass index as a possible mediating variable is described and applied. Some aspects of the future nutritional epidemiology research agenda are briefly discussed, including an ongoing human feeding study to develop biomarkers for additional dietary components.

Dealing with dietary measurement error in nutritional cohort studies

Dietary measurement error creates serious challenges to reliably discovering new diet-disease associations in nutritional cohort studies. Such error causes substantial underestimation of relative risks and reduction of statistical power for detecting associations. On the basis of data from the Observing Protein and Energy Nutrition Study, we recommend the following approaches to deal with these problems. Regarding data analysis of cohort studies using food-frequency questionnaires, we recommend 1) using energy adjustment for relative risk estimation; 2) reporting estimates adjusted for measurement error along with the usual relative risk estimates, whenever possible (this requires data from a relevant, preferably internal, validation study in which participants report intakes using both the main instrument and a more detailed reference instrument such as a 24-hour recall or multiple-day food record); 3) performing statistical adjustment of relative risks, based on such validation data, if they exist, using univariate (only for energy-adjusted intakes such as densities or residuals) or multivariate regression calibration. We note that whereas unadjusted relative risk estimates are biased toward the null value, statistical significance tests of unadjusted relative risk estimates are approximately valid. Regarding study design, we recommend increasing the sample size to remedy loss of power; however, it is important to understand that this will often be an incomplete solution because the attenuated signal may be too small to distinguish from unmeasured confounding in the model relating disease to reported intake. Future work should be devoted to alleviating the problem of signal attenuation, possibly through the use of improved self-report instruments or by combining dietary biomarkers with self-report instruments.