Diet Quality and Mortality Risk in Metabolically Obese Normal-Weight Adults

ORIGINAL ARTICLE

Diet Quality and Mortality Risk in Metabolically Obese Normal-Weight Adults Yong-Moon Mark Park, MD, PhD; Teresa T. Fung, ScD; Susan E. Steck, PhD; Jiajia Zhang, PhD; Linda J. Hazlett, PhD; Kyungdo Han, PhD; Seung-Hwan Lee, MD, PhD; and Anwar T. Merchant, ScD, DMD Abstract Objective: To examine the associations among the Dietary Approaches to Stop Hypertension (DASH)e style diet, the Healthy Eating Index (HEI), and mortality risk in metabolically obese normal-weight (MONW) adults. Patients and Methods: Data were from normal-weight (body mass index of 18.5 to <25) adults aged 30 to 90 years at baseline in the Third National Health and Nutrition Examination Survey, October 18, 1988, through October 15, 1994, followed up for deaths (all-cause, cardiovascular, and cancer related) until December 31, 2011. A total of 2103 participants without known cardiovascular disease and cancer at baseline were included in this prospective cohort study. Metabolic obesity was defined as having 2 or more of the following: high glucose, blood pressure, triglyceride, C-reactive protein, and insulin resistance values and low high-density lipoprotein cholesterol levels; metabolic healthy status was defined as having 0 or 1 of these metabolic derangements. Results: During median follow-up of 18.6 years, there were 344 and 296 deaths in the MONW and metabolically healthy normal-weight (MHNW) phenotypes, respectively. In MONW individuals, a 1-SD increment in adherence to a DASH diet (2 points) or HEI (14 points) was significantly associated with reductions (17% [hazard ratio (HR), 0.83; 95% CI, 0.72-0.97] and 22% [HR, 0.78; 95% CI, 0.68-0.90], respectively) in the risk of all-cause mortality, after adjustment for potential confounders. The corresponding HRs for cardiovascular disease mortality were 0.72 (95% CI, 0.55-0.94) and 0.79 (95% CI, 0.650.97), respectively. In addition, reduction of cancer mortality was observed with 1-SD increment of HEI (HR, 0.63; 95% CI, 0.46-0.88). However, no association was observed in the MHNW phenotype. Sensitivity analyses suggested relationships robust to different definitions of MONW and also dose responses with the number of metabolic derangements. Conclusion: Higher diet quality scores were associated with lower risk of mortality in normal-weight individuals with metabolic abnormalities. ª 2016 Mayo Foundation for Medical Education and Research

From the Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, Columbia (Y.M.M.P., S.E.S., J.Z., L.J.H., A.T.M.); Epidemiology Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC (Y.-M.M.P.); Department of Nutrition, Simmons College, Boston, MA (T.T.F.); DepartAffiliations continued at the end of this article.

1372

D

espite being apparently nonobese, a subset of normal-weight individuals seems to be more susceptible to insulin resistance, type 2 diabetes mellitus, and cardiovascular disease (CVD), which are all metabolic conditions associated with obesity.1-3 These individuals display a metabolically obese normal weight (MONW) phenotype characterized by higher visceral adiposity, impaired insulin sensitivity, and a more atherogenic lipid profile compared with their metabolically healthy normal-weight (MHNW) counterparts.3 It has been reported that the prevalence of the MONW phenotype ranges from 7.1% to 30.1% in the US

n

Mayo Clin Proc. 2016;91(10):1372-1383

population depending on the criteria used.4,5 Identification of modifiable risk factors in MONW individuals who are apparently healthy but at high risk for cardiometabolic disease could be beneficial to preventing cardiometabolic morbidity and mortality.6 In addition, the MONW phenotype may relate to cancer morbidity and mortality, although there has been less research in this area.7 Healthy dietary patterns are associated with a reduced risk of CVD and cancer.8 In the United States, the Dietary Approaches to Stop Hypertension (DASH) score and the Healthy Eating Index (HEI) are used to assess diet quality.9 The DASH diet was developed to

Mayo Clin Proc. n October 2016;91(10):1372-1383 n http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org n ª 2016 Mayo Foundation for Medical Education and Research

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

prevent hypertension,10 and it has been reported that adherence to the DASH diet is related to a reduced risk of CVD11 and type 2 diabetes mellitus.12 The HEI was developed by US Department of Agriculture researchers to measure adherence to the Dietary Guidelines for Americans and the Food Guide Pyramid13 and has been associated with lower inflammation and risk of chronic disease.14,15 To date, intervention studies assessing the effect of healthy dietary patterns on reducing the risk of chronic disease have been largely focused on an obese population.16,17 Because the MONW phenotype presents risks for the development of diseases leading to avoidable morbidity and premature mortality, it is important to understand the role that modifiable factors, such as dietary pattern, may play in modifying that risk.9,11 Furthermore, the role of a dietary pattern on mortality in MONW individuals is unknown. Therefore, we addressed whether a highquality diet, measured by DASH or HEI criteria, relates to cardiovascular, cancer, and all-cause mortality risk in MONW individuals in a nationally representative normal-weight US population.

sensitivity C-reactive protein (hs-CRP). We excluded individuals who reported a history of myocardial infarction (n¼112), stroke (n¼94), congestive heart failure (n¼83), or cancer (other than skin cancer) (n¼136). To reduce the possibility of reverse causation with diets being modified due to diseases elevating mortality risk, we excluded individuals who died during the first year of follow-up (n¼33). We also excluded those who reported implausible extreme energy intakes (n¼48), pregnant (n¼14) or lactating (n¼14) women, and those with an hs-CRP level greater than 10 mg/L (to convert to nmol/L, multiply by 9.542) (n¼2). Finally, a total of 2103 individuals were included in these analyses because some of the participants had multiple exclusion criteria (for instance some people might have both myocardial infarction and cancer), the sum of those with each exclusion criterion should be greater than those excluded) (Figure). We did not include individuals younger than 30 years at baseline because the prevalence of the MONW

2509 Participants aged 30-90 y

PATIENTS AND METHODS Study Population We used data from the Third National Health and Nutrition Examination Survey (NHANES III), October 18, 1988, through October 15, 1994, followed up for deaths until December 31, 2011, in this prospective cohort analysis. NHANES III was conducted using a complex multistage stratified clustered probability sample design to achieve a nationally representative sample of the civilian, noninstitutionalized US population. The survey included personal interviews, physical examinations, and laboratory measurements. The primary analysis included 2,509 normal-weight (body mass index [BMI] of 18.5 to <25 [calculated as the weight in kilograms divided by the height in meters squared]) adults aged 30 to 90 years at baseline who were eligible for mortality follow-up and had complete dietary data and cardiometabolic parameters, including fasting glucose, insulin, triglycerides, high-density lipoprotein cholesterol (HDL-C), blood pressure (BP), and highMayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

406 Participants excluded

2103 Participants included in the final analytical data set

Inclusion criteria • 18.5 ≤ body mass index < 25a • Eligible for mortality follow-up • Complete information on 24-h dietary recall, fasting glucose, insulin, blood pressure, triglycerides, high-density lipoprotein cholesterol, and hs-CRP

Exclusion criteria Number • A history of myocardial infarction 112 • A history of stroke 94 • A history of congestive heart failure 83 • A history of cancer (other than skin cancer) 136 • <1st and >99th percentiles of energy intake 48 • hs-CRP >10 mg/L (to convert to nmol/L, multiply by 9.542) 2 • Pregnant women 14 • Lactating women 14 • Death in the first year of follow-up 33

FIGURE. Participant flow diagram. Because some of the participants had multiple exclusion criteria (for instance some people might have both myocardial infarction and cancer), the number of sum of those with each exclusion criterion should be greater than the number of those excluded. a For sensitivity analyses, alternative definitions of normal weight were applied using (1) waist circumference less than 102 cm in men and less than 88 cm in women or (2) body mass index of 18.5 to less than 25 and waist circumference less than 102 cm in men and less than 88 cm in women. hs-CRP ¼ high-sensitivity C-reactive protein.

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1373

MAYO CLINIC PROCEEDINGS

phenotype and the proportion of deaths are quite low in those younger than 30 years_ENREF_17. However, these individuals take substantial sampling weights in the NHANES III, and, therefore, any conclusions would be largely influenced by relatively small numbers of the MONW phenotype and deaths in this age group. Assessment of Dietary Intake A single 24-hour dietary recall was used to calculate DASH and HEI scores in the present study because the NHANES III food frequency questionnaire was not designed to produce population nutrient intake estimates. Information on dietary intake was collected using an interactive interview with automated data entry, validated by the Nutrition Methodology Working Group, and the US Department of Agriculture Survey Nutrient Database was used to assign nutrient values.18 The DASH diet is characterized by high intake of fruits, vegetables, nuts and legumes, low-fat dairy products, and whole grains and low intake of fat (total/saturated), sodium, sweets, and red meats.19 There have been different approaches to evaluate adherence to the DASH diet based on the composition of food and nutrients.20 We calculated the Mellen index, which is one of the established DASH scores from the 24-hour dietary recall data.20 The Mellen index, a nutrient-based index with 9 components, including target nutrient values used in clinical trials,21 applied absolute targets on the basis of a 2100-kcal diet for both men and women. Individuals satisfying the goal for each component received 1 point; those who met an intermediate goal, defined as the midpoint between the DASH diet goal and the nutrient content of the DASH control diet, received 0.5 points; and those who met neither goal received 0 points.10 For instance, if someone consumed 100 mg/ 1000 kcal of total fat, a value that is between the cutoff DASH score target of 71.4 mg/ 1000 kcal and the intermediate target of 107.1 mg/1000 kcal, she or he would receive 0.5 points in the category of total fat. The possible DASH score ranged from 0 to 9, with higher values indicating greater adherence to the DASH diet. Ten main factors are used to calculate the HEI: grains, fruits, vegetables, dairy, meats, 1374

Mayo Clin Proc.

n

total fat, saturated fat, cholesterol, sodium, and dietary variety.13 Scoring methods for the HEI were based on the 1989-1990 HEI.22 The HEI scores range from 0 to 100, with a score on each component ranging from 0 to 10. Among the HEI components, grains, fruits, vegetables, meats, and dairy products were based on the 1992 Food Guide Pyramid’s recommended number of servings in which the maximum score was given to the recommended servings.23 The remaining components complied with the 1995 Dietary Guidelines for Americans24 considering the recommended intakes of total fat, saturated fat, cholesterol, and sodium. Assessment of Metabolic Health We assessed metabolic health using the metabolic parameters that were measured under the quality control standards of the Centers for Disease Control and Prevention. Waist circumference was measured at the level of the right iliac crest. The BP was averaged over 5 separate measurements. Glucose level was measured in serum using a modified hexokinase enzymatic method. Serum insulin level was measured using a radioimmnuoassay (Pharmacia Diagnostics). The HDL-C and triglyceride levels were measured using a Hitachi 704 analyzer (Boehringer-Mannheim Diagnostics). Serum hs-CRP concentrations were measured by latex-enhanced nephlometry (Department of Laboratory Medicine, Immunology Division, University of Washington, Seattle). We defined a participant as being metabolically obese if she or he had 2 or more cardiometabolic abnormalities (systolic/ diastolic BP 130/85 mm Hg or antihypertensive medication use, fasting glucose level 100 mg/dL [to convert to mmol/L, multiply by 0.0555] or antidiabetes medication use, homoeostasis model assessment of insulin resistance [HOMA-IR ¼ fasting glucose (mg/ dL)  fasting insulin (mlU/mL [to convert to pmol/L, multiply by 6.945]) / 405] level greater than the 90th percentile, hs-CRP level greater than the 90th percentile, triglyceride levels 150 mg/dL [to convert to mmol/L, multiply by 0.0113] or taking cholesterol-lowering medication, and HDL-C level <40 mg/dL [to convert to mmol/L, multiply by 0.0259] in men or <50 mg/dL in women or taking cholesterol-lowering

October 2016;91(10):1372-1383

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

medication).4 Individuals who were taking cholesterol-lowering medications were considered to have high triglyceride and low HDL-C levels to capture all the possible derangements of dyslipidemia in the metabolic unhealthy status.25 In addition, metabolic healthy status was defined as having 0 or 1 of the metabolic derangements. In this way, insulin resistance and low-grade chronic inflammation, as well as other metabolic parameters, such as abnormal glucose, BP, triglyceride, and HDLC levels, could be considered simultaneously for defining metabolic healthy or unhealthy status compared with alternative definitions such as metabolic syndrome26 or highest quartile of HOMA-IR.27 Ascertainment of Mortality To determine the vital status and cause of death, the National Center for Health Statistics linked all participants to the National Death Index through December 31, 2011. The underlying cause recorded on the death certificate was applied to identify cause of death based on the underlying Cause of Death-113 groups (International Classification of Diseases, Tenth Revision). All-cause mortality was defined as deaths due to any underlying cause of death; CVD and cancer mortality were defined as deaths due to underlying cause of death codes I00-I69 and C00-C97, respectively.28 Statistical Analyses We used the appropriate survey procedures to account for the complex sampling design and weights. Descriptive results were presented as unweighted counts (n) and weighted percentages. For the subgroup analysis, domain analysis was applied to preserve the complex sampling design in which the entire samples were used for estimating the variance of subpopulations. Continuous variables are expressed as mean  SE and were compared using linear regression analyses. Categorical variables are presented as percentages with SE and were compared by Rao-Scott c2 tests. The Cox proportional hazards model was used to estimate the hazard ratios (HRs) and 95% CIs for all-cause, CVD, and cancer mortality. The proportional hazards assumptions were evaluated by the logarithm of cumulative hazards function based on Kaplan-Meier estimates for DASH or HEI tertile group as well as age, sex, Mayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

and race/ethnicity. The multivariable-adjusted proportional hazards model included the following confounders identified using the previous literature and contributing to at least a 5% change in the exposure effect estimate: age, sex, race/ethnicity (non-Hispanic white vs others), educational attainment (>12 years of education vs others), income (middle or high [poverty income ratio >1.3] vs others), smoking status (current vs others), alcohol consumption (moderate vs others),29 physical activity (inactive [no reported leisure-time physical activity] vs others), and total energy intake. Missing values for income were treated as a separate category in the multivariable models. The potential effect modification of the DASH or HEI scores by sociodemographic and lifestyle characteristics was evaluated through the stratified analysis and interaction testing. The interactions of the DASH index and the HEI with the variables mentioned previously herein were tested using the Satterthwaiteadjusted F test based on the Cox proportional hazards models for all-cause mortality with interaction terms such as age group  continuous DASH score. Furthermore, we conducted several sensitivity analyses: (1) alternative definitions for normal weight using waist circumference less than 102 cm in men and less than 88 cm in women or BMI of 18.5 to less than 25 and waist circumference less than 102 cm in men and less than 88 cm in women; (2) alternative definitions for metabolic obesity using metabolic syndrome defined when the individual had fewer than 3 cardiometabolic abnormalities (systolic/diastolic BP 130/85 mm Hg or antihypertensive medication use, triglyceride level 150 mg/L or taking cholesterol-lowering medication, fasting plasma glucose level 100 mg/dL or antidiabetes medication use, HDL-C level <40 mg/dL in men or <50 mg/dL in women or taking cholesterol-lowering medication, and waist circumference 102 cm in men or 88 cm in women)25,26 or highest quartile of HOMA-IR in nondiabetic adults in a whole population27; and (3) after exclusion of participants who died during the first 5-year follow-up. In addition, we explored whether there would be a dose-response relationship between increasing number of metabolic derangements in normal weight (defined by BMI, waist circumference, or both) and increased risk of mortality and whether there would be an interaction of

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1375

MAYO CLINIC PROCEEDINGS

DASH and HEI scores in this association after adjusting for age, sex, and race/ethnicity. All the statistical analyses were performed using SAS 9.4 software (SAS Institute Inc), and P<.05 was considered statistically significant. RESULTS The MONW phenotype (n¼663) was observed in 24.3% (SE, 1.6%) of the normal-weight sample from the present study by primary criteria using only BMI after considering the sampling weight. In MONW individuals, those with the higher tertile in both DASH and HEI scores were more likely to be older and female; were less likely to be current smokers; and had a higher HDL-C level. In MHNW individuals, those with the higher tertile in DASH and HEI scores were more likely to be older; had higher educational attainment and more income; were less likely to be current smokers; and had a lower hs-CRP level (Table 1 and Supplemental Table 1, available online at http://www. mayoclinicproceedings.org). The distribution of components of DASH diet and HEI scores according to each tertile in the MONW and MHNW phenotypes is shown in Table 2 and Supplemental Table 2 (available online at http://www.mayoclinicproceedings. org). The distribution of each component of the DASH diet and the HEI is also shown in Supplemental Table 3 (available online at http://www.mayoclinicproceedings.org). Total DASH diet and HEI scores were higher in MONW individuals (P<.001 and P¼.05, respectively). For the DASH diet, there were significant increasing trends of all nutrients in the DASH diet with tertile of DASH scores except sodium in MHNW individuals. DASH scores from percentage of energy in saturated fat and total fat and potassium levels were higher in MONW individuals. For HEI scores, there were significant increasing trends of all components in the HEI with tertile of HEI scores except meats in both MONW and MHNW individuals. In MONW individuals, HEI scores for fats, saturated fat, and cholesterol were higher, and for dairy and dietary variety were lower, compared with MHNW individuals. During median follow-up of 18.6 years (range, 1-21.6 years), there were 344 and 296 deaths in the MONW and MHNW phenotypes, respectively. Overall, higher mortality risk was observed in the MONW phenotype 1376

Mayo Clin Proc.

n

compared with the MHNW phenotype for all-cause mortality (HR, 1.31; 95% CI, 1.081.59) and CVD mortality (HR, 2.18; 95% CI, 1.53-3.10) after adjusting for potential confounders, including age, sex, race/ethnicity, educational attainment, income, smoking status, alcohol consumption, physical activity, and total energy intake. Table 3 shows multivariable-adjusted HRs and 95% CIs for all-cause, cardiovascular, and cancer mortality according to the tertile categories and a 1-SD increment in DASH and HEI scores in MONW and MHNW individuals. Significant inverse associations of increasing DASH and HEI scores with all-cause and CVD mortality were observed in MONW individuals after adjustment for potential confounders. For all-cause mortality, the HRs between the highest and lowest tertile were 0.83 (95% CI, 0.720.97) (P for trend¼.04) for the DASH score and 0.54 (95% CI, 0.39-0.75) (P for trend<.001) for the HEI score in the highest tertile. For CVD mortality, HRs were 0.41 (95% CI, 0.26-0.66) (P for trend<.001) for the DASH score and 0.51 (95% CI, 0.300.85) (P for trend¼.01) for the HEI score in the highest tertile compared with the lowest tertile. A 1-SD increment (2 points for the DASH score and 14 points for the HEI score) in adherence to the DASH diet and the HEI was inversely associated with a reduction in the risk of all-cause mortality (HR, 0.83 [95% CI, 0.72-0.97]; HR, 0.78 [95% CI, 0.68-0.90], respectively), CVD mortality (HR, 0.72 [95% CI, 0.55-0.94]; HR, 0.79 [95% CI, 0.65-0.97], respectively), and cancer mortality (HR, 0.69 [95% CI, 0.46-1.04]; HR, 0.63 [95% CI, 0.46-0.88], respectively). However, no association was observed in the MHNW phenotype with increasing DASH and HEI scores. In subgroup analyses for all-cause mortality, the inverse associations for each 1-SD increment of HEI score were stronger for MONW individuals 65 years and younger at baseline assessment and for current smokers (P interaction<.001 and P interaction¼.01, respectively). Although interaction terms did not reach statistical significance, the inverse associations for both indexes were stronger among those who were not non-Hispanic white and physically active individuals (Table 4). In MHNW individuals, the inverse association was stronger in those who were

October 2016;91(10):1372-1383

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

TABLE 1. Comparison of General Characteristics With Increasing Tertiles of DASH and HEI Scores in the MONW and MHNW Phenotypes at Baselinea,b DASH Characteristic

Overall

MONW phenotypes Total population (No.) 663 Male sex (%) 51.0 Non-Hispanic white race (%) 75.4 Educational attainment (>12 y) (%) 69.0 Income (PIR>1.3) (%) 81.6 Current smoker (%) 31.6 Moderate drinker (%) 46.0 Physical activity (inactive) (%) 17.6 Diabetes mellitus (%) 87.4 Hypertension (%) 51.7 Age (y), mean 56.2 BMI, mean 22.9 Waist circumference (cm), mean 86.8 Fasting glucose (mg/dL), mean 102 HOMA-IR score 2.53 Systolic BP, mm Hg, mean 133 Diastolic BP, mm Hg, mean 77 hs-CRP (mg/dL), mean 0.45 Triglycerides (mg/dL), mean 171 HDL-C (mg/dL), mean 47 MHNW phenotypes Total population (No.) 1440 Male sex (%) 38.6 Non-Hispanic white race (%) 81.3 Educational attainment (>12 y) (%) 85.1 Income (PIR>1.3) (%) 87.8 Current smoker (%) 27.5 Moderate drinker (%) 60.0 Physical activity (inactive) (%) 9.8 Diabetes mellitus (%) 97.4 Hypertension (%) 88.9 Age (y), mean 44.3 BMI, mean 22.3 Waist circumference (cm), mean 80.7 Fasting glucose (mg/dL), mean 91 HOMA-IR score, mean 1.36 Systolic BP (mm Hg), mean 116 Diastolic BP (mm Hg), mean 72 hs-CRP (mg/dL), mean 0.26 Triglycerides (mg/dL), mean 85 HDL-C (mg/dL), mean 58

HEI

c

Tertile 1 Tertile 2 Tertile 3 P value

P (MONW Tertile 1 Tertile 2 Tertile 3 P valuec vs. MHNW)c

222 58.0 79.2 70.8 81.5 40.5 53.7 19.9 8.0 47.8 52.9 23.0 88.5 98 2.16 130 78 0.44 159 46

200 61.4 70.2 64.5 81.5 36.4 53.0 14.6 11.8 34.6 53.7 23.1 87.1 101 2.53 131 77 0.51 176 45

241 35.4 75.9 70.8 81.8 18.8 32.5 17.9 18.0 59.9 61.6 22.7 85.0 107 2.90 138 77 0.41 177 50

.002 .46 .53 .99 .002 .01 .62 .02 .004 <.001 .25 .001 .01 .22 <.001 .40 .38 .20 .03

213 59.6 69.6 60.2 75.5 45.8 54.8 16.2 11.5 44.0 49.5 22.9 87.1 100 2.33 130 79 0.47 179 45

210 53.2 76.7 71.6 81.3 38.4 43.7 23.6 11.0 48.7 57.3 23.0 87.7 105 2.40 135 78 0.47 163 48

240 41.8 79.4 74.4 87.5 13.9 40.1 14.2 14.9 51.6 61.1 22.9 86.0 102 2.81 135 76 0.41 170 48

.02 .18 .08 .06 <.001 .14 .15 .48 .53 <.001 .82 .18 .29 .74 .10 .01 .005 .94 .03

564 40.8 80.2 81.4 82.8 35.3 61.8 12.0 1.8 9.1 42.7 22.3 80.9 91 1.38 115 72 0.28 80 57

468 41.7 80.0 87.7 91.5 27.6 64.8 8.6 3.4 11.8 43.3 22.4 81.3 90 1.34 117 73 0.25 85 60

408 31.9 84.2 86.9 90.1 16.7 51.6 8.2 2.8 12.8 47.8 22.2 79.8 91 1.34 116 71 0.25 90 58

.12 .33 .02 <.001 <.001 .02 .21 .44 .24 <.001 .15 .06 .49 .02 .33 .21 .01 <.001 .34

487 45.3 79.1 79.4 81.3 44.7 62.9 14.0 1.4 9.2 41.5 22.4 81.4 91 1.35 115 72 0.28 84 57

497 34.5 81.4 84.5 89.6 23.5 59.4 10.3 2.6 11.1 44.1 22.1 80.0 90 1.32 115 71 0.26 82 61

456 35.7 83.3 91.0 92.6 14.2 57.6 5.3 3.8 12.8 47.4 22.3 80.7 91 1.39 117 72 0.24 87 58

.02 .27 <.001 <.001 <.001 .53 .002 .16 .32 <.001 .50 .13 .59 .47 .19 .79 <.001 .21 .37

<.001 .02 <.001 .003 .11 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001

BMI ¼ body mass index; BP ¼ blood pressure; DASH ¼ Dietary Approaches to Stop Hypertension; HDL-C, high-density lipoprotein cholesterol; HEI ¼ Healthy Eating Index; HOMA-IR ¼ homoeostasis model assessment of insulin resistance; hs-CRP ¼ high-sensitivity C-reactive protein; MHNW ¼ metabolically healthy normal weight; MONW ¼ metabolically obese normal weight; PIR ¼ poverty income ratio. b SI conversion factors: To convert fasting glucose values to mmol/L, multiply by 0.0555; to convert hs-CRP values to nmol/L, multiply by 9.542; to convert triglyceride values to mmol/L, multiply by 0.0113; to convert HDL-C values to mmol/L, multiply by 0.0259. c P values represent P trend for continuous variables and the difference between tertiles in each index. P (MONW vs MHNW) represents the difference in overall value for each variable between MONW and MHNW individuals. Variables with a skewed distribution, such as HOMA-IR, hs-CRP, and triglycerides, were compared after log transformation. a

Mayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1377

MAYO CLINIC PROCEEDINGS

TABLE 2. Comparison of DASH Diet Scores and HEI Scores in the MONW and MHNW Phenotypes at Baselinea,b MONW phenotype Score

Overall

Tertile 1

Tertile 2

MHNW phenotype

Tertile 3 P trend

Overall

Tertile 1

Tertile 2

P (MONW Tertile 3 P trend vs MHNW)

DASH Total population (No.) 663 222 200 241 1440 564 468 408 Score range 0-8.5 0-1.5 2-3 3.5-8.5 0-8.5 0-1.5 2-3 3.5-8.5 Total score 2.74 1.03 2.44 4.67 <.001 2.45 0.94 2.45 4.56 Saturated fat 0.38 0.14 0.35 0.63 <.001 0.31 0.12 0.32 0.56 Total fat 0.44 0.07 0.45 0.80 <.001 0.33 0.06 0.32 0.71 Protein 0.30 0.16 0.31 0.42 <.001 0.28 0.18 0.25 0.45 Cholesterol 0.40 0.25 0.34 0.59 <.001 0.40 0.18 0.50 0.58 Fiber 0.22 0.03 0.13 0.47 <.001 0.20 0.03 0.17 0.46 Magnesium 0.23 0.03 0.15 0.49 <.001 0.21 0.03 0.16 0.53 Calcium 0.29 0.19 0.23 0.44 <.001 0.27 0.16 0.26 0.42 Potassium 0.24 0.06 0.16 0.48 <.001 0.21 0.04 0.17 0.50 Sodium 0.25 0.10 0.32 0.35 <.001 0.25 0.13 0.30 0.35 HEI Total population (No.) 663 213 210 240 1440 487 497 456 Score range 21.3-97.7 21.3-58.5 58.6-72.8 71.9-97.7 19.1-96.7 19.1-58.5 58.6-71.8 71.9-96.7 Total score 66.1 50.2 65.4 80.5 <.001 65.0 50.1 65.1 79.5 Grains 6.5 5.8 6.7 7.0 <.001 6.6 5.8 6.5 7.3 Fruits 4.4 1.5 3.2 7.8 <.001 4.2 1.5 3.8 7.1 Vegetables 6.2 4.6 6.0 7.6 <.001 6.2 4.8 6.4 7.4 Dairy 6.0 5.2 6.1 6.6 .007 6.5 6.0 6.0 7.5 Meats 6.8 6.7 6.8 6.8 .79 7.0 7.1 7.0 6.9 Fats 7.3 5.1 7.2 9.3 <.001 6.4 3.9 6.7 8.6 Saturated fat 7.1 4.5 7.3 9.2 <.001 6.5 3.8 6.9 8.9 Cholesterol 8.1 6.1 8.4 9.5 <.001 7.6 5.6 7.9 9.3 Sodium 6.4 5.6 6.1 7.2 <.001 6.0 5.4 5.8 6.8 Dietary variety 7.5 5.1 7.6 9.5 <.001 8.0 6.1 8.1 9.7

<.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001 <.001

<.001 <.001 <.001 .39 .92 .23 .45 .25 .03 .80

<.001 <.001 <.001 <.001 <.001 0.48 <.001 <.001 <.001 <.001 <.001

.05 .80 .34 .88 .01 .10 <.001 .005 .01 .09 .01

DASH ¼ Dietary Approaches to Stop Hypertension; HEI ¼ Healthy Eating Index; MHNW ¼ metabolically healthy normal weight; MONW ¼ Metabolically Obese Normal Weight. b Data are presented as means. a

physically inactive (P interaction¼.005 for DASH score and P interaction¼.06 for HEI score). There was a significant interaction by race/ethnicity, but the direction of the association was not consistent in DASH and HEI scores (Supplemental Table 4, available online at http://www.mayoclinicproceedings.org). When alternative definitions of normal weight were applied, significant associations of DASH and HEI scores with risk of all-cause mortality and CVD mortality in MONW individuals persisted, and the size of the estimates tended to increase (Supplemental Tables 5 and 6, available online at http://www.mayoclinicproceedings. org). Using different criteria for defining metabolic obesity, the prevalence of the MONW phenotype was 10.0% (SE, 0.9%) by metabolic syndrome criteria and 7.1% (SE, 0.8%) for the highest quartile of HOMA-IR as alternative definitions, both of which were 1378

Mayo Clin Proc.

n

much lower than the prevalence by BMI. There was a reduced risk of all-cause mortality with increasing HEI scores in both MONW and MHNW individuals when metabolic syndrome was defined as metabolic obesity and in MHNW individuals when the highest quartile of HOMA-IR was defined as metabolic obesity (Supplemental Tables 7 and 8, available online at http://www.mayoclinicproceedings.org). When we analyzed the data after excluding participants who died at follow-up during the first 5 years, the overall results were not materially different from the main analysis (Supplemental Table 9, available online at http://www.mayoclinicproceedings.org). With increasing numbers of metabolic derangements, there was an increased risk of all-cause and CVD mortality regardless of the definition of normal weight (all P trend<.001), although a linear trend was more prominent

October 2016;91(10):1372-1383

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

TABLE 3. HRs and 95% CIs of All-cause, Cardiovascular, and Cancer Mortality According to the Tertile Categories and a Linear Increment in the DASH Diet and HEI Scores in the MONW and MHNW Phenotypes During Median Follow-up of 18.6 Yearsa All-cause mortality Parameter MONW phenotype DASH diet score Tertile 1 Tertile 2 Tertile 3 P trend HR for a 1-SD increase in DASH diet score HEI score Tertile 1 Tertile 2 Tertile 3 P trend HR for a 1-SD increase in HEI score MHNW phenotype DASH diet score Tertile 1 Tertile 2 Tertile 3 P trend HR for a 1-SD increase in DASH diet score HEI score Tertile 1 Tertile 2 Tertile 3 P trend HR for a 1-SD increase in HEI score

CVD mortality

Cancer mortalityc

Person-years

No. of deaths

Multivariable-adjusted HR (95% CI)b

No. of deaths

Multivariable-adjusted HR (95% CI)b

No. of deaths

Multivariable-adjusted HR (95% CI)b

3198 2828 3539

109 103 132

1 (Reference) 0.87 (0.69-1.11) 0.72 (0.53-0.98) .04 0.83 (0.72-0.97)

35 38 41

1 (Reference) 0.82 (0.57-1.17) 0.41 (0.26-0.66) <.001 0.72 (0.55-0.94)

26 17 19

1 (Reference) 0.67 (0.36-1.27) 0.70 (0.33-1.49) .35 0.69 (0.46-1.04)

3055 3009 3501

105 106 133

1 (Reference) 0.59 (0.44-0.79) 0.54 (0.39-0.75) <.001 0.78 (0.68-0.90)

31 35 48

1 (Reference) 0.74 (0.47-1.17) 0.51 (0.30-0.84) .01 0.79 (0.65-0.97)

22 19 21

1 (Reference) 0.24 (0.11-0.54) 0.52 (0.24-1.12) .14 0.63 (0.46-0.88)

9969 8277 6986

110 86 100

1 (Reference) 0.86 (0.59-1.25) 1.07 (0.79-1.46) .67 1.08 (0.92-1.25)

22 18 27

1 (Reference) 0.65 (0.32-1.33) 1.17 (0.51-2.70) .67 1.07 (0.77-1.48)

34 21 24

1 (Reference) 0.58 (0.27-1.21) 0.93 (0.57-1.53) .68 1.07 (0.74-1.54)

8582 8736 7913

100 96 100

1 (Reference) 0.64 (0.39-1.05) 0.68 (0.44-1.05) .09 0.83 (0.70-1.00)

19 21 27

1 (Reference) 0.66 (0.19-2.36) 0.91 (0.22-3.73) .44 0.77 (0.48-1.23)

31 25 23

1 (Reference) 0.73 (0.30-1.79) 0.54 (0.20-1.44) .21 0.93 (0.64-1.36)

CVD ¼ cardiovascular disease; DASH ¼ Dietary Approaches to Stop Hypertension; HEI ¼ Healthy Eating Index; HR ¼ hazard ratio; MHNW ¼ Metabolically Healthy Normal Weight; MONW ¼ Metabolically Obese Normal Weight. b Adjusted for age at baseline, sex, race/ethnicity, educational attainment, income, smoking status, alcohol consumption, level of physical activity, and total calorie intake. c Those who had a history of skin cancer were additionally excluded. a

in the BMI-defined normal weight (Supplemental Table 10, available online at http://www.mayoclinicproceedings.org). In addition, there was significant interaction by tertiles of DASH or HEI scores in this association regardless of the definition of normal weight when those with highest tertiles of DASH or HEI scores and no metabolic derangement were treated as a reference (all P interaction<.001) (Supplemental Table 11, available online at http://www. Mayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

mayoclinicproceedings.org). When we did not consider those who were taking cholesterollowering medications but did not have abnormal blood triglyceride or reduced HDLC levels, the overall results did not materially change (Supplemental Table 12, available online at http://www.mayoclinicproceedings.org). DISCUSSION The present findings suggest that high diet quality, according to DASH and HEI criteria,

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1379

MAYO CLINIC PROCEEDINGS

TABLE 4. Subgroup Analyses of the Association of 1-SD Increment in DASH Diet and HEI Scores With the Risk of All-cause Mortality in the MONW Phenotypea No. of participants (deaths)

Factor Age (y) <65 65 Sex Men Women Race/ethnicity Non-Hispanic white Others Smoking status Current smoker Nonsmoker or former smoker Drinking alcohol Moderate drinker Nondrinker or heavy drinker Physical activityc Physically inactive Physically active

Multivariable-adjusted HR (95% CI)b for a 1-SD increase in DASH diet score P interaction

Multivariable-adjusted HR (95% CI)b for a 1-SD increase in HEI score

<.001

.20 354 (98) 309 (246)

0.86 (0.66-1.11) 0.84 (0.74-0.96)

343 (193) 320 (151)

0.85 (0.64-1.12) 0.84 (0.72-0.99)

358 (211) 305 (133)

0.86 (0.70-1.05) 0.77 (0.61-0.98)

199 (102) 464 (242)

0.92 (0.61-1.40) 0.83 (0.73-0.95)

270 (124) 393 (220)

0.87 (0.68-1.12) 0.87 (0.71-1.06)

154 (92) 509 (252)

1.08 (0.71-1.64) 0.79 (0.70-0.89)

P interaction

0.71 (0.56-0.90) 0.85 (0.76-0.95) .49

.11 0.78 (0.62-0.99) 0.83 (0.69-1.01)

.31

.10 0.84 (0.71-1.00) 0.65 (0.51-0.83)

.20

.01 0.74 (0.53-1.03) 0.85 (0.76-0.95)

.58

.78 0.87 (0.68-1.12) 0.73 (0.60-0.88)

.91

.96 0.91 (0.74-1.13) 0.74 (0.65-0.85)

DASH ¼ Dietary Approaches to Stop Hypertension; HEI ¼ Healthy Eating Index; HR ¼ hazard ratio; MONW ¼ metabolically obese normal weight. Adjusted for the same covariates used in Table 3, except for the stratifying factor. c Individuals with recommended physical activity were defined as those who had self-reported leisure time moderate activity (3  metabolic equivalents [METs] <6) of 5 or more times per week or leisure time vigorous activity (METs 6) 3 or more times per week. Physically inactive individuals were those with no reported leisure time physical activity. a

b

is associated with a lower risk of all-cause, CVD, and possibly cancer mortality in MONW individuals, independent of potential confounders, based on the nationally representative sample of normal-weight US adults. We observed 17% and 22% reductions in all-cause mortality with each 1-SD increment in DASH and HEI scores, respectively, in individuals with MONW defined by BMI criteria and even greater reductions when alternative criteria for defining MONW were used. We also observed risk reductions in CVD mortality for both DASH and HEI scores and in cancer mortality especially for HEI scores. However, these beneficial effects of highquality diet on the reduction of mortality were not observed in the MHNW phenotype. To our knowledge, this is the first study to report that high diet quality may reduce mortality risk specifically in MONW individuals. The inverse associations of DASH and HEI scores with all-cause and CVD mortality were consistent with previous studies showing that higher-quality diets are associated with reductions in the risks of all-cause 30-33 and 1380

Mayo Clin Proc.

n

CVD30,31,33,34 mortality for DASH scores and in all-cause33,35 and CVD33,35,36 mortality for HEI scores. Of these previous studies without regard for metabolic obesity or weight, only 1 study demonstrated that the inverse association between the DASH diet score and all-cause mortality was obvious in nonobese individuals (BMI <30) in their subgroup analysis.32 Underlying mechanisms for the beneficial effect of adherence to the DASH diet and HEI on reducing the risk of mortality in MONW individuals might be complex. Those with the MONW phenotype are at high risk for diseases marked by inflammation in which levels of CRP, interleukins, and tumor necrosis factor are elevated.37,38 It is known that the DASH diet is associated with a reduction in plasma CRP and fibrinogen levels39 and that the HEI score is inversely associated with serum CRP concentrations,15 suggesting that the DASH diet and the HEI could have antiinflammatory effects. Although the interaction terms were not significant, the observation that the physically active MONW individuals had a risk reduction in mortality on DASH

October 2016;91(10):1372-1383

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

and HEI scores reflect the possible synergistic effects of healthy lifestyles.40-42 Interestingly, high-quality dietary patterns were not associated with mortality reduction in MHNW individuals. This finding might suggest that high diet quality does not operate independently in reducing the risk of mortality for those at lower risk for dying. Instead, a high-quality diet may contribute to decreasing the mortality risk in MHNW individuals who had unfavorable lifestyle characteristics. In a stratified analysis, a large reduction in mortality risk was observed only in physically inactive individuals with the MHNW phenotype for both the DASH and HEI scores. Similar findings were found in current smokers. In the sensitivity analysis using different criteria for normal weight, the inverse association of DASH and HEI scores with risk of mortality tended to be more prominent when normal range of waist circumference was used alone or with normal BMI compared with normal BMI alone. This finding may support that abdominal obesity could better explain “obesity-related” health risk rather than body weight.43-45 Additional sensitivity analyses using different criteria for metabolic health suggested findings consistent with the main results, although findings were not always statistically significant. A possible explanation could include that the small sample size of the MONW phenotype defined by metabolic syndrome and highest quartile of HOMA-IR might contribute to lack of differentiation in metabolic healthy status. In the present study, the Mellen DASH index was focused on a nutritional target approach in which some of the complexity of diet that could have been captured through a food groupsebased approach might not be reflected.46 However, it could directly determine the adherence to the recommendation of each component in the DASH diet instead of assessing indirectly through foods rich in specific components. The HEI used in the present study was the earliest version primarily focused on adherence to the Dietary Guidelines for Americans and the Food Guide Pyramid. Thus, it may not be an adequate predictor of chronic disease risk compared with the currently used HEI-2010 or Alternate HEI-2010 in both of Mayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

which the role of whole grain, added sugar, alcohol, and polyunsaturated fat was further considered.47 This study has several strengths. First, a prospective study design with 18 years of follow-up for mortality based on the representative US population allowed us to evaluate a long-term effect of high diet quality on the reduction of mortality risk.48 Second, data were collected based on extensive laboratory and physical examinations using standardized protocols to minimize the influence of measurement errors. Finally, we were able to assess potential effect modification and to replicate the findings using a sensitivity analysis. There are also several limitations. First, the data are based on single 24-hour dietary recalls, which are subject to misclassification using a categorical approach with tertiles of DASH and HEI scores and might be discordant with the recommendations that a diet quality index be based on usual intake, although a 24-hour recall adequately groups a population into levels of intake.49 Second, although the risk of cancer mortality tended to decrease with each 1-SD increment in the DASH and HEI scores, we could not confirm the findings using tertiles of DASH or HEI scores and P for trend, which may be due to the limited number of cancer deaths. In addition, there may be overfit in the multivariable models investigating the association of DASH and HEI scores with cancer mortality in MONW and MHNW individuals as well as CVD mortality in MHNW individuals. This might result in increasing type 1 error, relative bias, and unreliable CI bounds.50 Third, because only a single measure of diet was collected at baseline, we could not account for changes in dietary intake over time. Finally, there may be residual confounding due to not measuring the covariates in an objective way, such as using self-reported physical activity and smoking status.

CONCLUSION This study found that adherence to a highquality diet such as defined by DASH and HEI criteria is associated with a lower risk of allcause and CVD mortality in the MONW phenotype, based on a nationally representative US adult population. High-quality diets may be

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1381

MAYO CLINIC PROCEEDINGS

particularly beneficial for normal-weight individuals with metabolic abnormalities.

SUPPLEMENTAL ONLINE MATERIAL Supplemental material can be found online at http://www.mayoclinicproceedings.org. Supplemental material attached to journal articles has not been edited, and the authors take responsibility for the accuracy of all data.

9.

10.

11.

12.

Abbreviations and Acronyms: BMI = body mass index; BP = blood pressure; CVD = cardiovascular disease; DASH = Dietary Approaches to Stop Hypertension; HDL-C = highdensity lipoprotein cholesterol; HEI = Healthy Eating Index; HOMA-IR = homoeostasis model assessment of insulin resistance; HR = hazard ratio; hs-CRP = high-sensitivity Creactive protein; MHNW = metabolically healthy normal weight; MONW = metabolically obese normal weight; NHANES III = Third National Health and Nutrition Examination Survey; PIR = poverty income ratio Affiliations (Continued from the first page of this article.): ment of Nutrition, Harvard TH Chan School of Public Health, Boston, MA (T.T.F.); and Department of Biostatistics (K.H.) and Division of Endocrinology and Metabolism, Department of Internal Medicine (S.-H.L.), College of Medicine, The Catholic University of Korea, Seoul. Correspondence: Address to Anwar T. Merchant, ScD, DMD, Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, 915 Greene St, Columbia, SC 29208 (merchant@ mailbox.sc.edu).

14.

15.

16.

17.

18.

19.

REFERENCES 1. Kim MK, Han K, Kwon HS, et al. Normal weight obesity in Korean adults. Clin Endocrinol (Oxf). 2014;80(2):214-220. 2. Conus F, Rabasa-Lhoret R, Peronnet F. Characteristics of metabolically obese normal-weight (MONW) subjects. Appl Physiol Nutr Metab. 2007;32(1):4-12. 3. Ruderman N, Chisholm D, Pi-Sunyer X, Schneider S. The metabolically obese, normal-weight individual revisited. Diabetes. 1998;47(5):699-713. 4. Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999-2004). Arch Intern Med. 2008;168(15):1617-1624. 5. Meigs JB, Wilson PW, Fox CS, et al. Body mass index, metabolic syndrome, and risk of type 2 diabetes or cardiovascular disease. J Clin Endocrinol Metab. 2006;91(8):2906-2912. 6. Oliveros E, Somers VK, Sochor O, Goel K, Lopez-Jimenez F. The concept of normal weight obesity. Prog Cardiovasc Dis. 2014;56(4):426-433. 7. Moore LL, Chadid S, Singer MR, Kreger BE, Denis GV. Metabolic health reduces risk of obesity-related cancer in framingham study adults. Cancer Epidemiol Biomarkers Prev. 2014; 23(10):2057-2065. 8. Heidemann C, Schulze MB, Franco OH, van Dam RM, Mantzoros CS, Hu FB. Dietary patterns and risk of mortality

1382

13.

Mayo Clin Proc.

n

20.

21.

22.

23.

24.

25.

from cardiovascular disease, cancer, and all causes in a prospective cohort of women. Circulation. 2008;118(3):230-237. 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. 2015;115(5):780-800e5. Appel LJ, Moore TJ, Obarzanek E, et al; DASH Collaborative Research Group. A clinical trial of the effects of dietary patterns on blood pressure. N Engl J Med. 1997;336(16):1117-1124. Siervo M, Lara J, Chowdhury S, Ashor A, Oggioni C, Mathers JC. Effects of the Dietary Approach to Stop Hypertension (DASH) diet on cardiovascular risk factors: a systematic review and meta-analysis. Br J Nutr. 2015;113(1):1-15. Shirani F, Salehi-Abargouei A, Azadbakht L. Effects of Dietary Approaches to Stop Hypertension (DASH) diet on some risk for developing type 2 diabetes: a systematic review and meta-analysis on controlled clinical trials. Nutrition. 2013;29(78):939-947. Kennedy ET, Ohls J, Carlson S, Fleming K. The Healthy Eating Index: design and applications. J Am Diet Assoc. 1995;95(10): 1103-1108. McCullough ML, Feskanich D, Rimm EB, et al. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in men. Am J Clin Nutr. 2000;72(5):1223-1231. 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. 2005;59(2):278-283. Dalzill C, Nigam A, Juneau M, et al. Intensive lifestyle intervention improves cardiometabolic and exercise parameters in metabolically healthy obese and metabolically unhealthy obese individuals. Can J Cardiol. 2014;30(4):434-440. Liu RH, Wharton S, Sharma AM, Ardern CI, Kuk JL. Influence of a clinical lifestyle-based weight loss program on the metabolic risk profile of metabolically normal and abnormal obese adults. Obesity (Silver Spring). 2013;21(8):1533-1539. US National Center for Health Statistics. Plan and Operation of the Third National Health and Nutrition Examination Survey, 1988e94. Washington, DC: US Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention; 1994. Sacks FM, Svetkey LP, Vollmer WM, et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. N Engl J Med. 2001;344(1):3-10. Miller PE, Cross AJ, Subar AF, et al. Comparison of 4 established DASH diet indexes: examining associations of index scores and colorectal cancer. Am J Clin Nutr. 2013;98(3):794803. Mellen PB, Gao SK, Vitolins MZ, Goff DC Jr. Deteriorating dietary habits among adults with hypertension: DASH dietary accordance, NHANES 1988-1994 and 1999-2004. Arch Intern Med. 2008;168(3):308-314. Third National Health and Nutrition Examination Survey 19881994, NHANES III Healthy Eating Index Data File. Series 11, No. 6A. Hyattsville, MD: National Center for Health Statistics, Centers for Disease Control and Prevention; 1999. US Department of Agriculture. Food Guide Pyramid: a guide to daily food choices. In: . Human Nutrition Information Service, Vol. 252. Washington, DC: US Dept of Agriculture; 1992:341-342. US Department of Agriculture. Nutrition and Your Health: Dietary Guidelines for Americans. Washington, DC: US Dept of Agriculture; 1995. Alberti KG, Eckel RH, Grundy SM, et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International

October 2016;91(10):1372-1383

n

http://dx.doi.org/10.1016/j.mayocp.2016.06.022 www.mayoclinicproceedings.org

DIET QUALITY AND METABOLICALLY OBESE NORMAL WEIGHT

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009;120(16):1640-1645. Lee K. Metabolically obese but normal weight (MONW) and metabolically healthy but obese (MHO) phenotypes in Koreans: characteristics and health behaviors. Asia Pac J Clin Nutr. 2009;18(2):280-284. Lee SH, Ha HS, Park YJ, et al. Identifying metabolically obese but normal-weight (MONW) individuals in a nondiabetic Korean population: the Chungju Metabolic disease Cohort (CMC) study. Clin Endocrinol (Oxf). 2011;75(4):475-481. National Center for Health Statistics website. NCHS Surveys 2011 Linked Mortality Files Public-Use Data Dictionary. http:// www.cdc.gov/nchs/data_access/data_linkage/mortality/nhanes3_ linkage.htm. Accessed February 9, 2016. Liangpunsakul S, Crabb DW, Qi R. Relationship among alcohol intake, body fat, and physical activity: a population-based study. Ann Epidemiol. 2010;20(9):670-675. Liese AD, Krebs-Smith SM, Subar AF, et al. The Dietary Patterns Methods Project: synthesis of findings across cohorts and relevance to dietary guidance. J Nutr. 2015;145(3):393-402. Harmon BE, Boushey CJ, Shvetsov YB, et al. Associations of key diet-quality indexes with mortality in the Multiethnic Cohort: the Dietary Patterns Methods Project. Am J Clin Nutr. 2015; 101(3):587-597. Boggs DA, Ban Y, Palmer JR, Rosenberg L. Higher diet quality is inversely associated with mortality in African-American women. J Nutr. 2015;145(3):547-554. Reedy J, Krebs-Smith SM, Miller PE, et al. Higher diet quality is associated with decreased risk of all-cause, cardiovascular disease, and cancer mortality among older adults. J Nutr. 2014; 144(6):881-889. Fung TT, Chiuve SE, McCullough ML, Rexrode KM, Logroscino G, Hu FB. Adherence to a DASH-style diet and risk of coronary heart disease and stroke in women. Arch Intern Med. 2008;168(7):713-720. Kappeler R, Eichholzer M, Rohrmann S. Meat consumption and diet quality and mortality in NHANES III. Eur J Clin Nutr. 2013; 67(6):598-606. McCullough ML, Feskanich D, Stampfer MJ, et al. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in women. Am J Clin Nutr. 2000;72(5):1214-1222. Di Renzo L, Galvano F, Orlandi C, et al. Oxidative stress in normal-weight obese syndrome. Obesity (Silver Spring). 2010; 18(11):2125-2130.

Mayo Clin Proc. n October 2016;91(10):1372-1383 www.mayoclinicproceedings.org

n

38. De Lorenzo A, Del Gobbo V, Premrov MG, Bigioni M, Galvano F, Di Renzo L. Normal-weight obese syndrome: early inflammation? Am J Clin Nutr. 2007;85(1):40-45. 39. Azadbakht L, Surkan PJ, Esmaillzadeh A, Willett WC. The Dietary Approaches to Stop Hypertension eating plan affects C-reactive protein, coagulation abnormalities, and hepatic function tests among type 2 diabetic patients. J Nutr. 2011;141(6):1083-1088. 40. Park YM, Sui X, Liu J, et al. The effect of cardiorespiratory fitness on age-related lipids and lipoproteins. J Am Coll Cardiol. 2015;65(19):2091-2100. 41. Lee DH, Kim YM, Jekal Y, et al. Low levels of physical activity are associated with increased metabolic syndrome risk factors in Korean adults. Diabetes Metab J. 2013;37(2):132-139. 42. Lee S, Kim Y. Effects of exercise alone on insulin sensitivity and glucose tolerance in obese youth. Diabetes Metab J. 2013;37(4): 225-232. 43. Park YM, Kwon HS, Lim SY, et al. Optimal waist circumference cutoff value reflecting insulin resistance as a diagnostic criterion of metabolic syndrome in a nondiabetic Korean population aged 40 years and over: the Chungju Metabolic Disease Cohort (CMC) study. Yonsei Med J. 2010;51(4):511-518. 44. Janssen I, Katzmarzyk PT, Ross R. Waist circumference and not body mass index explains obesity-related health risk. Am J Clin Nutr. 2004;79(3):379-384. 45. Oh HG, Nallamshetty S, Rhee EJ. Increased risk of progression of coronary artery calcification in male subjects with high baseline waist-to-height ratio: the Kangbuk Samsung Health Study. Diabetes Metab J. 2016;40(1):54-61. 46. Jacques PF, Tucker KL. Are dietary patterns useful for understanding the role of diet in chronic disease? Am J Clin Nutr. 2001;73(1):1-2. 47. Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012; 142(6):1009-1018. 48. Davy BM, Estabrooks PA. The validity of self-reported dietary intake data: focus on the “what we eat in America” component of the National Health and Nutrition Examination Survey Research Initiative. Mayo Clin Proc. 2015;90(7):845-847. 49. Biro G, Hulshof KF, Ovesen L, Amorim Cruz JA. Selection of methodology to assess food intake. Eur J Clin Nutr. 2002; 56(suppl 2):S25-S32. 50. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol. 2007; 165(6):710-718.

http://dx.doi.org/10.1016/j.mayocp.2016.06.022

1383