Abstract
Objectives:
This study explores relationships betweencardiometabolic measures of antioxidant capacity or inflammationand diet quality assessed by the Healthy Eating Index (HEI)-2010 which measures conformity to Dietary Guidelines for Americans. This cross-sectional study was an ancillary analysis of baseline data for a randomized controlled trial with older adults at risk for cardiometabolic disease (ClinicalTrials.gov #NCT00955903).
Methods:
Community-dwelling older adults(n=133, 49% male, 70.4 + 4.8 years) with a body mass index of 30–40 kg/m2 provided a fasted blood sample for measurement of antioxidant capacity, high-sensitivity C-reactive protein, tumor necrosis factor-alpha, and interleukin-6. Dietary data were generated from the mean of three 24-hour recalls.
Results:
After adjustment for potential confounders,HEI-2010 composite scores were not significantly associated with decreased inflammation or greater antioxidant capacity. In analysis of the twelve components comprising the HEI-2010, significant positive association was observed between total dairy and total serum antioxidant capacity (0.039, 95% CI= 0.008, 0.069). Significant associations observed in inflammatory markers were between total vegetable and tumor necrosis factor-alpha (−0.078, 95% CI= −0.151, −0.005), sodium and interleukin-6 (0.091, 95% CI= 0.023, 0.158), andscores for combined calories from solid fats, alcoholic beverages, and added sugars and interleukin-6 (0.139, 95% CI= 0.027, 0.252). In models adjusting for HEI-2010 composite score when significant associations were observed betweencomponent scores and biomarkers, two of six associations were strengthened by adding the composite score as a potential confounder.
Conclusions:
Largely null findings along with those inconsistent with scientific expectations suggest caution in extrapolating adherence to the HEI-2010 with an individual’s inflammatory or antioxidant status. Results merit additional investigation with other biomarkers of chronic disease and emphasis on dietary patterns given potential synergy within food combinations.
Keywords: Healthy Eating Index, Dietary Guidelines for Americans, antioxidant capacity, inflammation, cardiometabolic, oxidative stress
INTRODUCTION
Adherence to healthy dietary and physical activity patterns may assist in reducing the prevalence of chronic disease throughout the lifespan. Among the more commonly investigated dietary patterns, the Healthy Eating Index measures conformity to evidence-based recommendations set forth by the Dietary Guidelines for Americans (DGA)[1]. HEI-2010 estimates the elements of an optimal diet as a density-based measure (amounts per 1,000 kcals) rather than an absolute measure such as diet quantity[2]. Twelve food group components comprise the HEI-2010 including nine adequacy components (total fruit, whole fruit, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, and fatty acids) and three moderation components (refined grains, sodium, and combined calories from solid fats, alcoholic beverages, and added sugars (SoFAAS)). Scoring of each component is truncated at point values of 5, 10, and 20 points, and these component values are summed to form a composite adherence score of up to 100 points.
Recently, a meta-analysis of fifteen studies (n=1,020,642) reported on the diet-disease relationship and adherence to established patterns of high dietary quality [3].Greater adherence to the HEI was significantly associated with reduction in all-cause mortality, cardiovascular disease, cancer, and type 2 diabetes (p<0.00001). Despite these positive findings, few studies have investigated the relationship between adherence to the DGAand mechanistic biomarkers underpinning cardiometabolic disease, including inflammation and oxidative stress [4]. Among the few studies investigating HEI and inflammation, results were conflicting. For example, data from the National Health and Nutrition Examination Survey (1988–2004) suggest an inverse relationship between HEI and serum C-reactive protein (CRP) [5]. In contrast, data from the Nurses’ Health Study suggest that HEI was not significantly associated with markers of inflammation including CRP and interleukin-6 (IL-6) [6].
To date, relationships between HEI adherence and oxidative stress or antioxidant capacity measures have not been investigated. However, randomized controlled trials implementing healthy dietary patterns such as the DASH and Mediterranean diets have demonstrated significant positive increases in antioxidant capacity biomarkers post-intervention [7–10], [11, 12]. As such, it is plausible that greater adherence to the HEI may result in increased circulating antioxidants with the potential to promote redox and inflammatory balance. Despite the potential for diet to mediate mechanisms underpinning cardiometabolic disease, diet quality, whether due to physiological, psychological, or socioeconomic reasons, often declines with age [13]. For example, a recent study examining dietary patterns among a diverse group of 416 men and women ages 65 years and older reported that fewer than 50% of participants were meeting nutrient recommendations [14]. Given the concomitant reduction in diet quality and accumulation of oxidative stress with aging[15], the purpose of this study was to investigate the relationships between diet quality assessed by the HEI-2010 and serum antioxidant capacity or biomarkers of inflammation in adults ages 65 years and older who were at risk for cardiometabolic disease. It was hypothesized that higher HEI-2010 composite and component scores would be significantly associated with greater antioxidant capacity and lower circulating levels of inflammatory cytokines.
MATERIALS AND METHODS
Participants
This cross-sectional study was an ancillary analysis of baseline data collected as part of a randomized controlled trial investigating the effects of a 12-month diet and exercise intervention among older adults at risk for cardiometabolic disease (ClinicalTrials.gov #NCT00955903) [16]. The parent study enrolled community-dwelling men and women ages 65 years and older from the Birmingham, Alabama area. By study design, all participants were obese (body mass index 30–40 kg/m2), without cognitive impairment, and taking at least one medication to control lipids, blood pressure, or blood glucose. Exclusion criteria included weight change greater than +/− 4.5 kg in the previous year as well as any psychiatric or physical limitations that would preclude participation. All participants provided written informed consent for the parent study and a subset provided additional written informed consent for this ancillary study. Dietary intake data and serum collected from participants (n=133) at baseline of the parent study were used for this ancillary investigation which was approved by the Institutional Review Boards at the University of Alabama at Birmingham and the University of Alabama.
Dietary Assessment and Calculation of HEI-2010
Dietary intake data were generated from the mean of three 24-hour recalls using the Nutrition Data System for Research (NDSR) software program (Nutrition Coordinating Center, Minneapolis, MN, 2013). This software system standardizes the collection of dietary intake data using a multi-pass interview approach which has been validated for use in older adults [17]. The five distinct passes of the interview include a first pass of all foods and beverages consumed within the previous 24-hour period, a second pass to review the foods listed, third pass to gather more details about each food, a fourth pass to query for forgotten foods, and a final fifth pass to review the record for correctness. All 24-hour diet recalls were conducted by phone by interviewers trained in the use of NDSR at the University of Alabama at Birmingham Survey Research Unit within the School of Public Health. HEI-2010 composite and food group component scores were calculated using publicly available SAS 9.4® macros[18].
Antioxidant Capacity Testing
A fasting blood sample was obtained for determination of serum biomarkers through routine venipuncture. Samples were centrifuged at 3,000 rpm and 4°C for 10 minutes. Serum was aliquoted into cryovials and stored at −80°C until time of analysis. Serum was deproteinated according to a validated method in order to allow for detection of small molecular weight antioxidants (<6kDa in size) with primary antioxidant functionality[19]. Hydrophilic and lipophilic antioxidant capacity (H-AOX and L-AOX, respectively) of deproteinated serum were measured using the oxygen radical absorbance assay on a FLUOstar Optima plate reader (BMG Labtech, Cary, NC) in accordance with the method by Prior et. al[20]. The compound 2,2-azobis(2-amidino-propane) dihydrochloride served as the peroxyl radical generator and Trolox, a water-soluble analogue of vitamin E, was the reference antioxidant standard. Total antioxidant capacity (T-AOX) was calculated as the sum of H-AOX and L-AOX.Intra- and inter-assay CV% for H-AOX was 3.61 and 4.97, respectively, and 4.15 and 5.3, respectively, for L-AOX.
Inflammatory Cytokines
Markers of inflammation included high-sensitivity C-reactive protein (hs-CRP), tumor necrosis factor-alpha (TNF-α), and IL-6. Circulating hs-CRP was quantified by a turbidometric assay using StanbioSirrus Clinical Chemistry Analyzer (Pointe Scientific, Canton, MI). Minimum sensitivity was 0.05 mg/l, inter-assay CV was 2.13%, and intra-assay CV was 7.49%. TNF-α and IL-6 concentrations were assessed by chemiluminescence using a MSD imager (MSD, Gaithersburg, MD). Minimum sensitivity for TNF-α was 0.507 pg/ml, inter-assay CV was 5.47%, and intra-assay CV was 7.61%. Minimum sensitivity for IL-6 was 0.27 pg/ml, inter-assay CV% was 11.26, and intra-assay CV% was 3.73.
Additional Covariates
To control for physical activity as a possible confounding variable, Actiwatch accelerometers (Philips Company, Amsterdam, Netherlands) were used as an objective measure of physical activity. Prior to assessment, each participant wore a monitor on the wrist of the dominant hand for seven consecutive days. The monitors quantified the minutes spent by each individual in light, moderate, or vigorous physical activity. The combined intensity and duration of physical activity allowed for a variable that could be incorporated into statistical models.
Statistical Analyses
Preliminary analyses included bivariate correlations and linear regression models. To account for violations in the assumptions of linear regression models, generalized linear regression models with log link function and Gaussian error distribution were applied. The application of the log link function available within the generalized linear regression model framework allows for addressing skewness in the distributions of dietary antioxidants and inflammatory markers as well as the non-normality of residuals from standard linear regression models. Parameter estimates and 95% confidence intervals for the association between HEI-2010 scores, serum antioxidant capacity, and inflammatory markers were estimated. The multivariable regression models controlled for age, gender, ethnicity, physical activity, and energy intake. Additionally, the interaction between HEI-2010 scores and gender was also analyzed, and gender-stratified estimates are presented when the interaction term was significant. When associations between individual component scores and antioxidant or inflammatory markers were statistically significant (p<0.05), additional analyses were conducted to further adjust the models for the HEI-2010 composite score as a confounder. These analyses assessed whether the model fit improved based on Akaike Information Criterion (AIC) and assessed the robustness of these findings where the AIC for the model with the composite score was lower than the model without the composite score (i.e. when adding the HEI-2010 composite score improved the model quality). Statistical analysis was conducted using SAS software, Version 9.4. Copyright © 2002–2012 SAS Institute Inc. (Cary, NC, USA). Given the exploratory nature of these analyses, exact p-values are presented along with highlighting of associations/estimates that have p-values less than 0.05.
The HEI-2010 score is used to provide assessment of adherence to dietary guidelines with individual component scores truncated at a determined threshold value which varies across the food group components. Thus, participants’ scoring above a certain threshold score are assigned the same score as those participants meeting the threshold score. Hence, the robustness of the findings were assessed by conducting sensitivity analyses in which the truncation of scores was eliminated, thereby allowing more spread of food group component scores.
RESULTS
Demographic data as well as distribution of variables used in these cross-sectional analyses are presented in Table 1. Of the 133 participants in this cross-sectional study, the average age was 70.4 + 4.8 years with 49% males and 94% of white ethnicity. Among the dietary components assessed for adequacy, over 80% of participants were adherent to the DGA for intake of five of the nine components including total vegetables, greens and beans, whole fruit, total protein foods, and seafood and plant proteins. Intake results of the three moderation components (refined grains, sodium, and calories from solid fats, alcoholic beverages, and added sugars) suggest much lower rates of adherence.
Table 1.
Characteristics of Study Participants
| Demographics | |||
| Age (years): Mean (SD) | 70.4 (4.8) | ||
| Male Gender: N (%) | 49 (40.2) | ||
| White Ethnicity: N (%) | 94 (77.0) | ||
| Energy Intake(kcal/day): Mean (SD) | 1,679.5 (482.2) | ||
| Antioxidant Capacity and Inflammatory Biomarkers | |||
| H-AOX (uM TE1): Mean (SD) | 605.2 (364.7) | ||
| Median (Range) | 559.9 (2035.4) | ||
| L-AOX (uM TE1): Mean (SD) | 695.7 (309.6) | ||
| Median (Range) | 685.2 (1801.1) | ||
| T-AOX(uM TE1): Mean (SD) | 1296.0 (600.3) | ||
| Median (Range) | 1167.9 (3338.3) | ||
| TNF-α (pg/mL): Mean (SD) | 4.9 (2.0) | ||
| Median (Range) | 4.0 (13.0) | ||
| hs-CRP (mg/L): Mean (SD) | 4.5 (4.5) | ||
| Median (Range) | 3.0 (30.0) | ||
| IL-6(pg/mL): Mean (SD) | 3.1 (3.4) | ||
| Median (Range) | 2.0 (23.0) | ||
| Dietary Components | |||
| HEI-2010 | Non-truncated Score3 | ||
|---|---|---|---|
| Adequacy | Score Mean (SD) | DGA Adherence2 N (%) | Score Mean (SD) |
| Total Vegetables | 4.7 (0.8) | 103 (83.1) | 10.3 (6.3) |
| Greens & Beans | 4.7 (1.2) | 111 (89.5) | 28.1 (25.2) |
| Total Fruit | 4.2 (1.5) | 86 (69.4) | 10.2 (7.8) |
| Whole Fruit | 4.5 (1.3) | 102 (82.3) | 18.0 (15.0) |
| Whole Grains | 4.1 (4.2) | 30 (24.2) | 6.1 (8.3) |
| Total Dairy | 0.4 (1.7) | 76 (61.3) | 15.1 (11.5) |
| Total Protein | 7.7 (3.4) | 122 (98.4) | 16.0 (8.9) |
| Sea & Plant Protein | 5.0 (0.1) | 103 (83.1) | 16.3 (13.6) |
| Fatty Acids | 4.3 (1.6) | 13 (10.5) | 1.9 (0.5) |
| Moderation | |||
| Sodium | 5.3 (2.9) | 2 (1.6) | 1.7 (0.4) |
| Refined Grains | 3.6 (3.0) | 1 (0.8) | 9.2 (3.3) |
| SoFAAS | 13.4 (5.0) | 13 (10.5) | 28.7 (8.6) |
| Composite Score | 62.0 (11.1) | 161.7 (43.7) | |
Trolox equivalents.
DGA - 2010 Dietary Guidelines for Americans (1).
Truncation of scores was eliminated, thereby allowing more spread of scores.
H-AOX – hydrophilic antioxidant capacity; L-AOX – lipophilic antioxidant capacity; T-AOX – total antioxidant capacity; TNF-α – tumor necrosis factor-alpha; hs-CRP – high sensitivity C-reactive protein; IL-6 – interleukin-6; SoFAAS - Calories from solid fats, alcoholic beverages, and added sugars
Table 2presents the point and interval estimates describing the relationship between each HEI-2010 composite and component scores and serum antioxidant capacity biomarkers. After adjustment for age, gender, ethnicity, physical activity, and energy intake, the models for greens and beans did not converge. Hence, adjusted estimates for greens and beans are not provided. HEI-2010 composite scores were not significantly associated with greater antioxidant capacity. However, H-AOX was found to be significantly associated with three HEI-2010 dietary component scores including total dairy, total fruit, and whole fruit (p= 0.039, 0.007, and 0.041, respectively). Among these three components, only total dairy exhibited positive correlation. For every unit increase in total dairy, the log-transformed H-AOX, L-AOX and T-AOX increased by 0.043 (p=0.039), 0.035 (p=0.015) and 0.039 (p=0.013), respectively. This implies that one unit increase in the component score for total dairy was estimated to be associated with a 4.4% increase in mean H-AOX, 3.6% increase in mean L-AOX, and 4.0% increase in mean T-AOX.In contrast to these significant positive associations between antioxidant capacity and scores for total dairy intake, negative associations were found between H-AOX and total fruit as well as whole fruit. It was estimated that for every unit increase in the component scores for total fruit or whole fruit, the log-transformed H-AOX decreased by 0.096 or 0.08, respectively. This implies that each one-unit increase in the component scores for total fruit or whole fruit was associated with a 9.2% or 7.7% decrease in mean H-AOX, respectively. In gender-stratified estimates, the interaction between HEI-2010 component score for fatty acids and gender was significant for L-AOX (p=0.03) and T-AOX (p=0.03). Additionally, the interaction between the component score for total dairy and gender was significant for L-AOX (p=0.02) as was the interaction between the component score for whole fruit and H-AOX (p<0.0001). Among females only, there was a statistically significant inverse relationship between fattyacids component score and L-AOX equating to a 4% decrease in L-AOX with each unit increase in the fatty-acids score. With regard to T-AOX, a one unit increase in the fatty acids component score was associated with 6.4% increase in T-AOX among males. Next, among females only, a unit increase in the component score for total dairy was associated with a 6% increase in L-AOX; furthermore, a one unit increase in the component score for whole fruit was associated with a 16.5% decrease in H-AOX among females.
TABLE 2.
Parameter estimates (95% Wald Confidence Interval) of antioxidant capacity measures associated with a unit change in HEI-2010 score.
| H-AOX(uMTrolox equivalents) |
L-AOX(uMTrolox equivalents) |
T-AOX(uMTrolox equivalents) |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Adjusted Estimatea (95% CI) | P value | Adjusted Estimatea (95% CI) | P value | Adjusted Estimatea (95% CI) | P value | ||||
| Total Vegetables | 0.032 (−0.127, 0.190) | 0.696 | 0.086 (−0.042, 0.215) | 0.189 | 0.059 (−0.069, 0.187) | 0.367 | |||
| Greens & Beans | |||||||||
| Total Fruit | −.096 (−0.167, −0.026) | 0.007 | b | 0.028 (−0.042, 0.098) | 0.428 | −.034 (−0.097, 0.029) | 0.29 | ||
| Whole Fruit | −0.08 (−0.156, −0.003) | 0.041 | b | 0.032 (−0.048, 0.112) | 0.434 | −0.025 (−0.095, 0.045) | 0.483 | ||
| Whole Grains | −0.018 (−0.046, 0.01) | 0.201 | −0.011 (−0.031, 0.009) | 0.291 | −0.014 (−0.036, 0.007) | 0.194 | |||
| Total Dairy | 0.043 (0.002, 0.085) | 0.039 | b | 0.035 (0.007, 0.063) | 0.015 | b | 0.039 (0.008, 0.069) | 0.013 | b |
| Total Protein | 1.852 (−3.143, 6.847) | 0.4674 | 0.151 (−0.854, 1.157) | 0.768 | 0.729 (−0.871, 2.33) | 0.372 | |||
| Seafood & Plant Protein | 0.031 (−0.052, 0.114) | 0.459 | −0.02 (−0.073, 0.033) | 0.464 | 0.003 (−0.056, 0.062 | 0.916 | |||
| Fatty Acids | 0.036 (−0.007, 0.078) | 0.0977 | −0.011 (−0.042, 0.019) | 0.47 | 0.01 (−0.022, 0.043) | 0.541 | |||
| Sodium | 0.033 (−0.006, 0.072) | 0.0963 | 0.022 (−0.006, 0.049) | 0.125 | 0.026 (−0.003, 0.056 | 0.079 | |||
| Refined Grains | −0.007 (−0.084, 0.07) | 0.8546 | −0.006 (−0.062, 0.05) | 0.831 | −0.007 (−0.066, 0.052) | 0.825 | |||
| SoFAAS | 0.013 (−0.014, 0.041) | 0.3527 | 0.012 (−0.007, 0.031) | 0.23 | 0.012 (−0.009, 0.032) | 0.258 | |||
| Composite Score | 0.006 (−0.005, 0.017) | 0.2885 | 0.005 (−0.002, 0.013) | 0.177 | 0.006 (−0.003, 0.014) | 0.193 | |||
Multivariable models were adjusted for age, gender, ethnicity, physical activity, and energy intake.
p < 0.05
H-AOX – hydrophilic antioxidant capacity; L-AOX – lipophilic antioxidant capacity; T-AOX – total antioxidant capacity; SoFAAS - combined calories from solid fats, alcoholic beverages, and added sugars.
Table 3presents the point and interval estimates of the parameters that describe the relationship between HEI-2010 composite and component scores and inflammatory biomarkers. HEI-2010 composite scores were not significantly associated with inflammatory status. Significant association was found between the component score for total vegetables and TNF-α(p=0.037) such that each unit increase in the total vegetables score was estimated to be associated with 7.5% unit decrease in mean TNF-α levels. Significant positive association was found between IL-6 and sodium (p=0.008) as well asSoFAAS (p=0.015). For every unit increase in the component scores for sodium or SoFAAS, the log-transformedIL-6 increased by 0.091 or 0.139, respectively. This implies that for every unit increase in sodium or SoFAAS scores was estimated to be associated with a 9.5% or 14.5% increase in mean IL-6, respectively. In gender-stratified estimates, significant interactions between HEI-2010 component scores and inflammatory biomarkers were only observed when modeled with IL-6. More specifically, significant interactions were observed between HEI-2010 component scores for fatty acids, sodium, orSoFAASand IL-6 (p=0.008, p=0.004, and p=0.02, respectively). Among males, all three component scores were significantly associated with increased levels of IL-6.
TABLE 3.
Parameter estimates (95% Wald Confidence Interval) of inflammatory biomarkers associated with a unit change in HEI-2010 score.
| TNF-α(pg/mL) |
hs-CRP(mg/L) |
IL-6(pg/mL) |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | ||||
| Total Vegetables | −0.078 (−0.151, −0.005) | 0.037 | b | −0.054 (−0.184, 0.075) | 0.413 | −0.035 (−0.358, 0.288) | 0.831 | ||
| Greens & Beans | |||||||||
| Total Fruit | −0.009 (−0.065, 0.047) | 0.756 | −0.08 (−0.188, 0.029) | 0.150 | −0.062 (−0.203, 0.08) | 0.393 | |||
| Whole Fruit | 0.009 (−0.056, 0.074) | 0.782 | 0.058 (−0.135, 0.25) | 0.557 | 0.006 (−0.175, 0.188) | 0.947 | |||
| Whole Grains | 0.007 (−0.011, 0.025) | 0.433 | 0.018 >(−0.02, 0.055) | 0.358 | −0.041 (−0.096, 0.014) | 0.147 | |||
| Total Dairy | −0.010 (−0.032, 0.013) | 0.393 | 0.039 >(−0.014, 0.092) | 0.147 | 0.046 (−0.031, 0.124) | 0.242 | |||
| Total Protein | 0.205 (−0.638, 1.048) | 0.634 | 0.070 (−1.624, 1.764) | 0.935 | −0.104 (−1.777, 1.569) | 0.903 | |||
| Seafood & Plant Protein |
−0.014 (−0.060, 0.032) | 0.555 | 0.051 (−0.08, 0.182) | 0.446 | −0.023 (−0.138, 0.092) | 0.697 | |||
| Fatty Acids | 0.002 (−0.025, 0.029) |
0.881 | 0.005 >(−0.058, 0.068) | 0.881 | 0.077 (−0.015, 0.169) | 0.102 | |||
| Sodium | 0.007 (−0.019, 0.033) | 0.621 | 0.004 (−0.052, 0.061) | 0.878 | 0.091 (0.023, 0.158) | 0.008 | b | ||
| Refined Grains | 0.009 (−0.032, 0.05) | 0.683 | −0.165 (−0.363, 0.032) | 0.101 | −0.115 (−0.397, 0.167) | 0.423 | |||
| SoFAAS | −0.006 (−0.023, 0.011) | 0.516 | −0.007 (−0.039, 0.024) | 0.646 | 0.139 (0.027, 0.252) | 0.015 | b | ||
| Composite Score | −0.001 (−0.009, 0.006) | 0.685 | 0.001 (−0.013, 0.015) | 0.924 | 0.016 (−0.005, 0.038) | 0.140 | |||
Multivariable models were adjusted for age, gender, ethnicity, physical activity, and energy intake.
p < 0.05
TNF-α – tumor necrosis factor-alpha; hs-CRP – high sensitivity C-reactive protein; IL-6 – interleukin-6; SoFAAS - Calories from solid fats, alcoholic beverages, and added sugars.
In order to assess the robustness of these findings, sensitivity analyses were conducted in which the truncation of scores was eliminated, thereby allowing more spread of food group component scores. According to point and interval estimates describing the relationship between non-truncated HEI-2010 composite and component scores and serum antioxidant capacity (Table 4), the only significant associationssensitive to truncation were between total dairy and L-AOX (p=0.039) along with fatty acids and H-AOX (p=0.025). In analysis of inflammatory biomarkers, associations between HEI-2010 scores and hs-CRP as well as IL-6 were sensitive to truncation(Table 5); however, no associations between scores and TNF-α exhibited sensitivity to truncation. The non-truncated HEI-2010 composite score as well as the component scores for greens and beans, whole grains, and total dairy were positively and significantly associated with hs-CRP.
Table 4.
Parameter Estimates (95% Wald Confidence Interval) of antioxidant capacity measures associated with a unit change in non-truncated HEI-2010 scores.
| H-AOX(uMTrolox equivalents) |
L-AOX(uMTrolox equivalents) |
T-AOX(uMTrolox equivalents) |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | ||||
| Total Vegetables | −0.003 (−0.023, 0.018) | 0.809 | −0.004 (−0.019, 0.011) | 0.560 | −0.004 (−0.019, 0.012) | 0.646 | |||
| Greens & Beans | −0.003 (−0.009, 0.004) | 0.390 | −0.001 (−0.005, 0.003) | 0.551 | −0.002 (−0.006, 0.003) | 0.407 | |||
| Total Fruit | −0.014 (−0.034, 0.006) | 0.163 | 0.005 (−0.008, 0.017) | 0.467 | −0.003 (−0.017, 0.011) | 0.636 | |||
| Whole Fruit | −0.005 (−0.016, 0.005) | 0.299 | 0.001 (−0.005, 0.007) | 0.742 | −0.002 (−0.009, 0.006) | 0.652 | |||
| Whole Grains | −0.012 (−0.027, 0.004) | 0.149 | −0.004 (−0.015, 0.007) | 0.479 | −0.008 (−0.019, 0.004) | 0.209 | |||
| Total Dairy | 0.006 (−0.004, 0.015) | 0.255 | 0.008 (0.000, 0.015) | 0.039 | b | 0.007 (−0.001, 0.014) | 0.084 | ||
| Total Protein | 0.005 (−0.009, 0.019) | 0.499 | −0.001 (−0.012, 0.011) | 0.901 | 0.002 (−0.009, 0.013) | 0.716 | |||
| Seafood & Plant Protein | 0.005 (−0.003, 0.012) | 0.248 | 0.000 (−0.006, 0.006) | 0.970 | 0.002 (−0.004, 0.008) | 0.493 | |||
| Fatty Acids | 0.29 (0.036, 0.545) | 0.025 | b | −0.057 (−0.261, 0.146) | 0.581 | 0.105 (−0.103, 0.313) | 0.322 | ||
| Sodium | −0.288 (−0.625, 0.048) | 0.093 | −0.173 (−0.414, 0.068) | 0.160 | −0.222 (−0.477, 0.034) | 0.089 | |||
| Refined Grains | −0.011 (−0.049, 0.028) | 0.582 | −0.005 (−0.035, 0.025) | 0.736 | −0.008 (−0.038, 0.023) | 0.622 | |||
| SoFAAS | −0.006 (−0.022, 0.01) | 0.462 | −0.007 (−0.018, 0.004) | 0.217 | −0.006 (−0.018, 0.006) | 0.304 | |||
| Composite Score | −0.001 (−0.005, 0.002) | 0.425 | 0.000 (−0.002, 0.002) | 0.919 | −0.001 (−0.003, 0.002) | 0.590 | |||
Multivariable models were adjusted for age, gender, ethnicity, physical activity, and energy intake.
p < 0.05
H-AOX – hydrophilic antioxidant capacity; L-AOX – lipophilic antioxidant capacity; T-AOX – total antioxidant capacity; SoFAAS - Calories from solid fats, alcoholic beverages, and added sugars
Table 5.
Parameter estimates (95% Wald Confidence Intervals) of inflammatory biomarkers associated with a unit change in non-truncated HEI-2010 scores.
| TNF-α(pg/mL) |
hs-CRP(mg/L) |
IL-6m (pg/mL) |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | Adjusted Estimate a (95% CI) | P value | ||||||
| Total Vegetables | −0.007 (−0.021, 0.007) | 0.316 | −0.004 (−0.043, 0.035) | 0.826 | 0.036 (0.016, 0.056) | <.0001 | b | ||||
| Greens & Beans | −0.001 (−0.005, 0.002) | 0.430 | 0.006 (0.002, 0.011) | 0.009 | b | 0.009 (0.003, 0.014) | 0.001 | b | |||
| Total Fruit | −0.004 (−0.015, 0.006) | 0.426 | −0.035 (−0.062, −0.008) | 0.011 | b | 0.001 (−0.027, 0.03) | 0.920 | ||||
| Whole Fruit | −0.001 (−0.007, 0.004) | 0.631 | −0.012 (−0.028, 0.003) | 0.111 | 0.001 (−0.013, 0.015) | 0.926 | |||||
| Whole Grains | 0.003 (−0.006, 0.011) | 0.536 | 0.028 (0.017, 0.039) | <.0001 | b | −0.028 (−0.064, 0.007) | 0.120 | ||||
| Total Dairy | −0.001 (−0.008, 0.006) | 0.699 | 0.029 (0.015, 0.043) | <.0001 | b | 0.006 (−0.011, 0.022) | 0.505 | ||||
| Total Protein | 0.002 (−0.007, 0.011) | 0.636 | −0.018 (−0.041, 0.004) | 0.103 | −0.005 (−0.032, 0.023) | 0.746 | |||||
| Seafood & Plant Protein | −0.001 (−0.007, 0.004) | 0.654 | 0.003 (−0.005, 0.012) | 0.414 | −0.001 (−0.018, 0.016) | 0.900 | |||||
| Fatty Acids | 0.017 (−0.156, 0.19) | 0.849 | 0.004 (−0.404, 0.413) | 0.983 | 1.000 (0.532, 1.468) | <.0001 | b | ||||
| Sodium | 0.007 (−0.207, 0.22) | 0.951 | −0.011 (−0.451, 0.429) | 0.961 | −0.809 (−1.481, −0.138) | 0.018 | b | ||||
| Refined Grains | 0.013 (−0.013, 0.039) | 0.337 | 0.025 (−0.025, 0.076) | 0.324 | 0.056 (−0.018, 0.129) | 0.137 | |||||
| SoFAAS | 0.003 (−0.008, 0.013) | 0.635 | 0.006 (−0.012, 0.024) | 0.520 | −0.064 (−0.096, −0.033) | <.0001 | b | ||||
| Composite Score | −0.001 (−0.003, 0.001) | 0.383 | 0.005 (0.001, 0.009) | 0.014 | b | 0.003 (−0.004, 0.01) | 0.379 | ||||
Multivariable models were adjusted for age, gender, ethnicity, physical activity, and energy intake.
p < 0.05
TNF-α – tumor necrosis factor-alpha; hs-CRP – high sensitivity C-reactive protein; IL-6 – interleukin-6; SoFAAS - Calories from solid fats, alcoholic beverages, and added sugars.
Additionally, IL-6 and the component scores for total vegetables, greens and beans, fatty acids, sodium, and calories from solid fats, alcoholic beverages, and added sugars(SoFAAS) were all sensitive to truncation. Among these five significant associations, only the component scores for sodium andSoFAASexhibited an inverse relationship. It was estimated that with every unit increase in sodium or SoFAAS score, IL-6 decreased by55.5% and 6.2%, respectively.
In other sensitivity analysis, additional statistical models wererun toadjust for the HEI-2010 composite scorewhen HEI-2010 component scores were significantly associated (p<0.05) with antioxidant or inflammatory markers. In these analyses, improvement in the model fit based on Akaike Information Criterion (AIC) was observed in only two of the significant associations detected in the primary analysis. Results suggest that theassociationbetweenthe component score for whole fruit and H-AOX and the component score for total fruit and H-AOX had lower AIC when the HEI-2010 composite score was included as compared to models that did not adjust for the composite score. In all other models, the AIC increased when the HEI-2010 composite score was added as a confounder. In these associations, a one unit increase in the whole fruit score was associated with a 13% decrease in H-AOX (p=0.006) and the total fruit score was associated with a 11% decrease in H-AOX (p=0.049).
DISCUSSION
This exploratory study investigated relationships between adherence to the DGA as assessed by HEI-2010 scores and serum biomarkers of antioxidant capacity and inflammation among a cohort of adults ages 65 years and older who were at risk for cardiometabolic disease. Updated every five years, the DGA provides evidence-based recommendations to help the general public adopt healthy eating patterns [21]. As such, we hypothesized that higher HEI-2010 composite and component scores would be significantly associated with greater antioxidant capacity and lower circulating levels of inflammatory cytokines. However, in analysis of truncated HEI-2010 composite and adequacy component scores, the only findings supporting our hypothesis were the significant positive associations between total dairy and antioxidant capacity and the significant inverse relationship between total vegetable intake and TNF-α. Among the moderation components, the only findings supporting our hypothesis were the significant positive associations between IL-6 and sodium or SoFAAS. Contrary to our hypothesis, several relationships between truncated and non-truncated component scores and antioxidant and inflammatory markers exhibited significant associations that are not supported by existing literature. For example, among truncated scores, a significant inverse association was observed between the component scores for total fruit or whole fruit and H-AOX. Nevertheless, the non-intuitive inverse relationships between total fruit and whole fruit with H-AOX were slightly stronger when the models were adjusted for the HEI-2010 composite score.
Taken collectively, the HEI-2010 remains a well-validated metric of adherence to diet quality outlined by the DGA[22]. Furthermore, greater adherence to high-quality diets including the HEI have demonstrated significant inverse association with various chronic diseases[3]. Yet, in light of the largely null findings of this exploratory analysis along with those inconsistent with scientific explanation, results indicate that caution is warranted in extrapolating diet quality measured by the HEI-2010 to an individual’s inflammatory or antioxidant status.These results are supported by previous investigation reporting a lack of significant association between HEI and markers of inflammation including CRP and IL-6 [6]. It is possible that adherence to the HEI-2010 is reflected in other biomarkers not included in this study; however, in a recent publication investigating HEI-2010 diet quality measures and chronic disease risk among low-income youth, HEI-2010 scores were not related to other markers of disease risk including body mass index, percent body or abdominal fat, serum cholesterol, serum lipids, or impaired glucose tolerance [23].Nevertheless, the selection of appropriate biomarkers is fraught with challenges as evidenced by the recent report from the United States/Canadian-sponsored working group on Dietary Reference Intakes and chronic disease endpoints [24]. Results of this study underscore the complex nature of the diet-disease relationship. Thus, the relationships and trends reported herein warrant further investigation. Additionally, it may be prudent to investigate the relationship between chronic disease biomarkers and dietary patterns rather than metrics based on individual food groups given the potential synergy within specific food combinations[25].
This study provides insightful results for future research, yet it is not exempt from some limitations, namely the cross-sectional design that precludes ascribing causality. Also, because participants were mostly Caucasians recruited from the same geographical region in the southeastern United States, this sample may not be representative of the general older adult population. Despite the highly-standardized methods used to collect dietary data, inherent limitations of self-reported intake are also important to note as these recalls are reliant on memory and may not depict long-term dietary habits. However, use of 24-hour recalls has been validated in and frequently used with this population[26, 27]. It is also important to note that individuals with cognitive impairment were excluded from participating in the parent study, thus strengthening the validity of this dietary instrument in the study population. Additionally, an inherent limitation of the study is the lack of assessment of adherence to medication, thus adjustment for adherence is not included in the statistical analysis. Strengths of the study include the sample size for exploring these associations, and the statistical analyses were carefully selected to address the research question. Likewise, the study was strengthened by robust serum measures of antioxidant capacity and inflammation.
CONCLUSIONS
Diet quality undoubtedly plays an important role in lowering chronic disease risk through mitigating oxidative stress and inflammation. Despite the utility of HEI-2010 as a measure of adherence to a high-quality diet as outlined by the DGA, this study illustrates that caution is warranted in equating self-reported adherence to this diet pattern and reductions in biomarkers associated with cardiometabolic disease. Future controlled trials can shed light on whether recent revisions to the DGA may better equate to reductions in oxidative stress and systemic inflammation as well as other biomarkers associated with chronic disease outcomes.
ACKNOWLEDGEMENTS
The Authors would like to acknowledge Hunter Allman for assisting with statistical analysis as well as NavneetBaidwan and Fuchenchu Wang for independently replicating these analyses.
Financial: This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Footnotes
Conflict of Interest Disclosure:
The Authors have no conflicts of interest to report.
References
- 1.United States Department of Agriculture. 2015 – 2020 Dietary Guidelines for Americans. 8th Edition. Available from: http://health.gov/dietaryguidelines/2015/guidelines/. Updated December 2015. Accessed June 25, 2018.
- 2.Guenther PM, Casavale KO, Kirkpatrick SI, Reedy J, Hiza HAB, Kuczynski KJ, Kahle LL, Krebs-Smith SM. Update of the Healthy Eating Index: Hei-2010. J Acad Nutr Diet 113(4), 2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Schwingshackl L, Hoffmann G. Diet quality as assessed by the Healthy Eating Index, the Alternate Healthy Eating Index, the Dietary Approaches to Stop Hypertension score, and health outcomes: a systematic review and meta-analysis of cohort studies. J Acad Nutr Diet 115(5):780–800.e785, 2015. [DOI] [PubMed] [Google Scholar]
- 4.Schap T, Kuczynski K, Hiza H. Healthy Eating Index - beyond the score. J Acad Nutr Diet 117(4):519–521, 2017. [DOI] [PubMed] [Google Scholar]
- 5.Ford ES, Mokdad AH, Liu S. Healthy Eating Index and c-reactive protein concentration: findings from the National Health and Nutrition Examination Survey III, 1988–1994. Eur J Clin Nutr 59(2):278–283, 2004. [DOI] [PubMed] [Google Scholar]
- 6.Fung TT, McCullough ML, Newby P, Manson JE, Meigs JB, Rifai N, Willett WC, Hu FB. Diet-quality scores and plasma concentrations of markers of inflammation and endothelial dysfunction. Am J ClinNutr 82(1):163–173, 2005. [DOI] [PubMed] [Google Scholar]
- 7.Asemi Z, Samimi M, Tabassi Z, Sabihi S-s, Esmaillzadeh A. A randomized controlled clinical trial investigating the effect of DASH diet on insulin resistance, inflammation, and oxidative stress in gestational diabetes. Nutr 29(4):619–624, 2013. [DOI] [PubMed] [Google Scholar]
- 8.Asemi Z, Samimi M, Tabassi Z, Shakeri H, Sabihi S-S, Esmaillzadeh A. Effects of DASH diet on lipid profiles and biomarkers of oxidative stress in overweight and obese women with polycystic ovary syndrome: a randomized clinical trial. Nutr 30(11):1287–1293, 2014. [DOI] [PubMed] [Google Scholar]
- 9.Miller ER III, Erlinger TP, Sacks FM, Svetkey LP, Charleston J, Lin P-H, Appel LJ. A dietary pattern that lowers oxidative stress increases antibodies to oxidized LDL: results from a randomized controlled feeding study. Atherosclerosis 183(1):175–182, 2005. [DOI] [PubMed] [Google Scholar]
- 10.Zamora-Ros R, Serafini M, Estruch R, Lamuela-Raventós RM, Martínez-González MA, Salas-Salvadó J, Fiol M, Lapetra J, Arós F, Covas MI, Andres-Lacueva C; PREDIMED Study Investigators.Mediterranean diet and non-enzymatic antioxidant capacity in the PREDIMEDStudy: evidence for a mechanism of antioxidant tuning. NutrMetab CardiovascDis 23(12):1167–1174, 2013. [DOI] [PubMed] [Google Scholar]
- 11.Ortner Hadžiabdić M, Vitali Čepo D, Rahelić D, Božikov V. The effect of the Mediterranean diet on serum total antioxidant capacity in obese patients: a randomized controlled trial. J AmColl Nutr 35(3):224–235, 2016. [DOI] [PubMed] [Google Scholar]
- 12.Kolomvotsou A, Rallidis L, Mountzouris K, Lekakis J, Koutelidakis A, Efstathiou S, Nana-Anastasiou M, Zampelas A.Adherence to Mediterranean diet and close dietetic supervision increase total dietary antioxidant intake and plasma antioxidant capacity in subjects with abdominal obesity. Eur J Nutr 52(1):37–48, 2013. [DOI] [PubMed] [Google Scholar]
- 13.Porter Starr KN, McDonald SR, Bales CW. Nutritional vulnerability in older adults: a continuum of concerns. CurrNutrReports 4(2):176–184, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hsiao PY, Mitchell DC, Coffman DL, Allman RM, Locher JL, Sawyer P, Jensen G, Hartman T. Dietary patterns and diet quality among diverse older adults: the University of Alabama at Birmingham Study of Aging. J Nutr Health Aging 17(1):19–25, 2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Terman Alexi BU. Oxidative stress, accumulation of biological “garbage”, and aging. Antiox Redox Signal 8:197–204, 2006. [DOI] [PubMed] [Google Scholar]
- 16.Haas MC, Bodner EV, Brown CJ, Bryan D, Buys DR, Keita AD, Flagg LA, Goss A, Gower B, Hovater M, Hunter G, Ritchie CS, Roth DL, Wingo BC, Ard J, Locher JL.Calorie restriction in overweight seniors: response of older adults to a dieting study: the CROSSROADS randomized controlled clinical trial. J Nutr Gerontol Geriatr 33(4):376–400, 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Harnack L, Stevens M, Van Heel N, Schakel S, Dwyer JT, Himes J. A computer-based approach for assessing dietary supplement use in conjunction with dietary recalls. J Food Compost Anal 21(Suppliment 1):S78–S82, 2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.National Cancer Institute. Diet history questionnaire II: calculating Healthy Eating Index (HEI) scores using diet*calc output. Available from: https://epi.grants.cancer.gov/dhq2/dietcalc/output.html. Updated April 17, 2018. Accessed June 11, 2018.
- 19.Crowe KM. Optimizing protein precipitation efficiency for assessing the contribution of low molecular weight compounds to serum antioxidant capacity. ClinBiochem 47(15):116–118, 2014. [DOI] [PubMed] [Google Scholar]
- 20.Prior RL, Hoang H, Gu L, Wu X, Bacchiocca M, Howard L, Hampsch-Woodill M, Huang D, Ou B, Jacob R.Assays for hydrophilic and lipophilic antioxidant capacity (oxygen radical absorbance capacity (ORACFL)) of plasma and other biological and food samples. J Ag Food Chem 51(11):3273–3279, 2003. [DOI] [PubMed] [Google Scholar]
- 21.Hauk L Dgac makes food-based recommendations in the 2015–2020 Dietary Guidelines for Americans. Am Fam Physician 93(6):525, 2016. [PubMed] [Google Scholar]
- 22.Guenther PM, Kirkpatrick SI, Reedy J, Krebs-Smith SM, Buckman DW, Dodd KW, Casavale KO, Carroll RJ. The Healthy Eating Index-2010 is a valid and reliable measure of diet quality according to the 2010 Dietary Guidelines for Americans. J Nutr 144(3):399–407, 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wrobleski MM, Parker EA, Hurley KM, Oberlander S, Merry BC, Black MM. Comparison of the HEI and HEI-2010 diet quality measures in association with chronic disease risk among low-income, African-American urban youth in Baltimore, Maryland. J Am Coll Nutr 37(3):201–208, 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yetley EA, MacFarlane AJ, Greene-Finestone LS, Garza C, Ard JD, Atkinson SA, Bier DM, Carriquiry AL, Harlan WR, Hattis D, King JC, Krewski D, O’Connor DL, Prentice RL, Rodricks JV, Wells GA. Options for basing Dietary Reference Intakes (DRIs) on chronic disease endpoints: report from a joint US/Canadian-sponsored working group. Am J Clin Nutr 105(1):249S–285S, 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Jacobs JDR, Steffen LM. Nutrients, foods, and dietary patterns as exposures in research: a framework for food synergy. Am J Clin Nutr 78(3):508S–513S, 2003. [DOI] [PubMed] [Google Scholar]
- 26.Sun Y, Roth DL, Ritchie CS, Burgio KL, Locher JL. Reliability and predictive validity of energy intake measures from the 24-hour dietary recalls of homebound older adults. J Am Diet Assoc 110(5):773–778, 2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Vaccaro JA, Huffman FG. Dietary risk factors by race/ethnicity, age-group, and gender in a representative sample of us older adults. J Nutr Health Aging 21(6):704–709, 2017. [DOI] [PubMed] [Google Scholar]