ABSTRACT
Background
Metabolic syndrome (MetS) is associated with increased mortality independent of BMI, resulting in discordant metabolic phenotypes, such as metabolically healthy obese and metabolically unhealthy normal-weight individuals. Studies investigating dietary intake in MetS have reported mixed results, due in part to the limitations of self-reported measures.
Objectives
To investigate the role of biomarker-calibrated estimates of energy and protein in MetS and metabolic phenotypes.
Methods
Postmenopausal participants from the Women's Health Initiative (WHI) study who were free of MetS at baseline, had available data from FFQs at baseline, and had components of MetS at Year 3 (n = 3963) were included. Dietary energy and protein intakes were estimated using biomarker calibration methods. MetS was defined as 3 or more of the following: elevated serum triglycerides (≥150 mg/dL), low HDL cholesterol (<50 mg/dL), hypertension [systolic blood pressure (BP) ≥130 or diastolic BP ≥85 mmHg], elevated serum glucose (>100 mg/dL), and abdominal adiposity (waist circumference > 89 cm). Models were adjusted for age, WHI study component, race/ethnicity, education, income, smoking, recreational physical activity, disease history, and parity.
Results
For every 10% increment in total calibrated energy intake, women were at a 1.37-fold elevated risk of MetS (95% CI, 1.15–1.63); a 10% increment in calibrated total protein intake was associated with a 1.21-fold elevated risk of MetS (95% CI, 1.00–1.47). Specifically, animal protein intake was associated with MetS (OR, 1.08; 95% CI, 1.02–1.14), whereas vegetable protein intake was not (OR, 0.99; 95% CI, 0.95–1.03). No differences were seen when examining metabolic phenotypes.
Conclusions
We found that higher calibrated total energy, total protein, and total animal protein intakes were strongly associated with MetS. If replicated in clinical trials, these results will have implications for the promotion of energy and animal protein restrictions for the reduction of MetS risks.
Keywords: metabolic syndrome, energy intake, diet, biomarker, body composition, protein intake
Introduction
Metabolic syndrome (MetS), a cluster of biometric indicators, including obesity, hypertension, hyperglycemia, and hyperlipidemia, is associated with an increased risk of cardiovascular disease, Type II diabetes, stroke, and all-cause mortality (1, 2). MetS is prevalent, occurring in approximately one-third of the adult US population (3–5). While obesity is thought to be a major contributor to MetS, it is not always present. In the United States, 31.7% of obese adults were identified as metabolically healthy obese (MH-O) and 25.5% of normal-weight individuals were identified as metabolically unhealthy normal weight (MUH-NW) (6). Due to this discordance, metabolic health is increasingly being investigated through separate metabolic phenotypes, which include MH-O, MUH-NW, metabolically healthy normal weight (MH-NW), and metabolically unhealthy obese (MUH-O).
Dietary intake, and its role in MetS, have received much attention. Multiple studies have shown that higher total energy intake is associated with an elevated risk of MetS. In contrast, reports on the impact of protein intake on MetS are less consistent (7). Multiple epidemiologic studies found that self-reported diets high in total protein and protein density were associated with an elevated risk of MetS (8–12); nonetheless, some studies report diets high in protein may be a protective factor against MetS (13, 14).
Potential discrepancies in these studies could be the result of self-reported diet measures, which have been shown to consistently, and with systematic bias, underreport protein intake (15–17). In an attempt to reduce the measurement error associated with self-reported dietary measures, methods have been developed to combine self-report data with recovery biomarkers for dietary intakes using statistical calibration equations (18). Thus, biomarker calibrations could provide more accurate and objective estimations of dietary intakes. In addition, these discrepancies may be attributed to whether protein comes from animal or vegetable sources.
Additionally, few studies have examined the associations of energy and protein intakes with metabolic phenotypes. Of the few that have, findings have shown that lower protein intake, in relation to total caloric intake, is associated with an elevated risk of MH-O in females (19, 20). However, research examining the association between energy and protein intakes and MetS has been sparse in normal-weight individuals. While poor dietary quality has been implicated with MetS in normal-weight women, research is limited on the role of the protein source on MetS (21). Given that a substantial proportion of MetS cases are MUH-NW individuals, it is important to identify the dietary risk factors associated with MUH-NW and examine whether the association with energy intake is restricted to MUH-O individuals. Lastly, race/ethnicity may play a role in the development of MetS and metabolic phenotypes, as there are multiple paradoxes between MetS and race/ethnicity (22, 23).
This study's purpose was to use biomarker calibrations to investigate the roles of energy intake, total protein intake, and protein density on MetS and metabolic phenotypes. We hypothesized that higher biomarker-calibrated estimates of total energy intake and lower estimates of total protein intake would be associated with MetS, regardless of BMI. Additionally, we explored whether these results differed by race/ethnicity.
Methods
Study population
The Women's Health Initiative (WHI) is a longitudinal, nationwide study of 161 808 post-menopausal women. As has been described in detail elsewhere, women between the ages of 50 and 79 years were enrolled in 1 or more randomized clinical trials (CTs) or an observational study (OS) at 40 clinical centers across the United States between 1993 and 1998 (24). There were 3 CTs in the WHI: the Hormone Therapy Trial (HT), the Dietary Modification Trial (DM), and the Calcium and Vitamin D Trial. For this analysis, women were included if they had completed a FFQ 2–3 years prior to the ascertainment of MetS criteria. FFQs were measured at follow-up Year 1 for all women in the DM. For women not in the DM, but in the OS or a different CT, FFQs were measured at baseline. In addition, women must have had data available for an assessment of MetS at WHI follow-up Year 3, which included fasting serum analyzed for triglycerides, HDL cholesterol, and/or fasting glucose obtained from 3 different substudies meeting our inclusion criteria. Women were excluded from this analysis if 1) they met the criteria for MetS at baseline (n = 1412); 2) they had no valid FFQ (defined as <600 or >5000 kcal/d) for analysis (n = 468); or 3) there was an inability to code MetS due to missing data (n = 32), resulting in a total sample size of 3963 women (Supplementary Figure 1).
Ethics
All participants provided informed consent prior to enrollment in the WHI. This study received approval through the Institutional Review Boards at all 40 clinical centers and through the Fred Hutchinson Cancer Research Center, the Clinical Coordinating Center of the WHI.
Blood collection and analysis
Fasting (≥8 h) blood specimens were collected on all WHI participants at baseline (with a subset analyzed) and on repeating subsets of WHI participants at various time points, including Year 3. This subset of participants with blood specimens was oversampled for minorities, with 47.4% of this subset being Black, Hispanic, American Indian, or Asian American/Pacific Islander. All blood was processed within 2 h of collection, then shipped and stored at −70°C at a central biorepository (Fisher BioServices) until analysis. Fasting serum glucose, triglycerides, and HDL cholesterol levels were measured as previously described (25). Each batch included 5% blinded duplicate quality assurance samples. The laboratory coefficients of variation were <2.5% for each of the assays.
Data collection
At enrollment, participants were asked to self-report their birth date; level of education; income; race/ethnicity; smoking status; recreational physical activity; history of cardiovascular disease, including diagnosis of hypertension; and parity. Annual questionnaires asked participants to report medications for diabetes and hypertension that were prescribed since the previous questionnaire.
All WHI participants completed in-person clinic visits at baseline and at 3 y, which ascertained the participant height and weight, waist circumference, and blood pressure (BP), and all participants also completed an FFQ at baseline. The methods describing ascertainment procedures have been previously published (25). Participants in the DM completed another clinic visit as described above and another FFQ at follow-up Year 1. Given the baseline FFQs were used to determine eligibility into the DM based on dietary fat intake, the clinic visit and FFQ data collected at follow-up Year 1, rather than baseline, were used to avoid the truncated distribution of the DM FFQ at baseline for the DM participants. For the purposes of this analysis, we grouped together participants from the DM and the non-DM OS and CTs; for ease, we henceforth refer to this as the baseline data collection for all participants.
Biomarker-calibrated estimates of energy intake and protein
The exposure of interest was baseline biomarker-calibrated estimates of dietary intakes. Biomarker calibration equations for total energy, total protein, and protein density have been developed through the WHI Nutrition and Physical Activity Assessment Study (NPAAS; n = 450), which aimed to understand the measurement errors in self-reported diets using recovery biomarkers for energy and protein intakes. The NPAAS oversampled for Black and Hispanic women, as well as women in the extremes of the BMI range (17). We identified participant characteristics that were associated with underreporting of nutrient consumption, including BMI, age, and race/ethnicity. Biomarker calibration equations were developed using linear regression, where the log of the nutrient recovery biomarkers was regressed on the log of the respective self-reported dietary intake (FFQ), along with variables identified to be associated with underreporting of nutrient consumption. (17, 18, 26). In this analysis, “uncalibrated” intake refers to the estimated self-reported protein and energy intake based on the FFQ data. Animal and vegetable protein consumptions were estimated by multiplying the biomarker-based calibrated total protein intake by the self-reported protein subtype proportion.
Definition of MetS and overweight and obese status
The primary outcome was MetS at follow-up Year 3. According to the Third Report of the National Cholesterol Education Program's Adult Treatment Panel, MetS was defined as 3 or more of the following: elevated serum triglycerides (≥150 mg/dL), low HDL cholesterol (<50 mg/dL), hypertension (systolic BP ≥130 or diastolic BP ≥85 mmHg), elevated serum glucose (>100 mg/dL), and abdominal adiposity (waist circumference > 89 cm). BMI was calculated as weight/height2 using the height (m) and weight (kg) from clinic visits. BMI was categorized as underweight (<18.5), normal weight (18.5–24.9), overweight (25–29.9), or obese (≥30). For analyses stratified by metabolic phenotype, MetS and weight status were determined at follow-up Year 3. MH-NW was defined as having a normal BMI (18.5–24.9) without MetS, whereas MUH-NW was defined as having a normal BMI with MetS as defined above. For the purposes of this analysis, overweight and obese participants were combined. As such, MH-O was defined as having a BMI in the overweight or obese categories without MetS and MUH-O was defined as having an overweight or obese BMI with MetS.
Statistics
Baseline characteristics of the overall study sample and stratified by total energy and protein quartiles are reported using means and standard deviations for continuous variables or counts and frequencies for categorical variables. Descriptive statistics of both uncalibrated and calibrated nutrient intake are reported using geometric means and 95% CIs.
Logistic regression models were used to examine the association of dietary intake at baseline and MetS at follow-up Year 3 as a dichotomous outcome. All ORs and 95% CIs are reported for 10% increments in consumption of energy (kcal/d), protein (g/d), and protein density (% energy from protein) as part of the calibration equation approach. In addition, we ran exploratory logistic regression models for total animal and vegetable protein intakes (g/d). All models were adjusted for baseline covariates, including age (years), study component (CT or OS), HT and DM arms of the CT, race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic/Latina, Asian/Pacific Islander, American Indian/Alaskan Native, Unknown), education [≤ high school (HS), some post-HS, college degree, or higher], yearly income (<$20 000, $20 000–$49 999, $50 000–$74 999, or ≥$75 000), smoking (never, past, current), recreational physical activity (no activity, some activity, 2 to <4 episodes/week, ≥4 episodes/week), history of treated diabetes (yes/no), history of hypertension (yes/no), and parity (never pregnant/no term pregnancies, 1, 2, 3, 4, ≥5). In evaluating protein density, the models additionally adjusted for total energy. To account for the sampling characteristics of the study sample (as not all WHI participants had blood analyzed for components of MetS), inverse probability weighting was applied. Women with missing data on MetS criteria at Year 3 were excluded; however, this was only a small proportion of women (n = 32).
Logistic regression models were also used to examine the association of baseline dietary intake and MetS, stratified by weight status, at follow-up Year 3, to assess possible discordance between metabolic phenotypes. Lastly, ORs were estimated separately by race/ethnicity group and P values for interactions were calculated using a 5-degree-of-freedom partial F test to compare ORs across racial/ethnic categories.
As Prentice and Huang (26) point out, BMI is an important component of the biomarker calibration equation; though exclusion of BMI from the models may underestimate the contribution of BMI, inclusion of BMI within the regression models using biomarker-calibrated energy intake may overadjust for the impact of BMI. Thus, we did not include BMI in our regression models.
Sensitivity analyses were conducted that excluded those who reported taking medication for hypertension, diabetes, and/or high cholesterol at baseline. In addition, we examined the impact that weight gain (between baseline and Year 3) may have had on the associations by including a variable measuring weight gain between the baseline and Year 3 clinic visits. Lastly, we examined the associations of baseline dietary factors and the individual components of MetS at follow-up Year 3.
To account for random variation in estimating calibration equation coefficients, the standard errors for ORs from the models with calibrated estimates were estimated by a 2-step procedure where, in each of 1000 iterations, the nutrient biomarker sample is resampled with replacement and the calibration equation coefficients are then used to compute the calibrated intake in the resampled (with replacement) study sample. The P values for statistical tests were 2 tailed and considered statistically significant at an alpha level of 0.05. All analyses were conducted using SAS version 9.4 (SAS Institute) and R version 3.4 (R Foundation for Statistical Computing).
Results
A total of 3963 postmenopausal women were included in this analysis, with a majority older than 60 years of age (71.4%). The majority of women were White (48.9%); however, 22.6% and 12.7% of women identified as non-Hispanic Black or Hispanic/Latina, respectively. Uncalibrated intakes were approximately 31% and 19% lower for energy and total protein intakes, respectively, compared to calibrated intakes (Table 1). Missingness on covariate data is indicated in the tables.
TABLE 1.
Baseline characteristics of the eligible study sample
| Total sample, n = 3963 | |
|---|---|
| n (%) | |
| Age, years | |
| 50–59 | 972 (28.6) |
| 60–69 | 1538 (45.2) |
| 70–79 | 892 (26.2) |
| Race/ethnicity | |
| Non-Hispanic White | 2011 (50.7) |
| Non-Hispanic Black | 864 (21.8) |
| Hispanic/Latina | 489 (12.3) |
| Asian/Pacific Islander | 404 (10.2) |
| American Indian/Alaskan Native | 148 (3.7) |
| Unknown | 47 (1.1) |
| BMI, kg/m2 | |
| Underweight, <18.5 | 55 (1.4) |
| Normal weight, 18.5–24.9 | 1304 (32.9) |
| Overweight, 25–29.9 | 1395 (35.2) |
| Obese, ≥30 | 1158 (29.2) |
| Missing | 51 (1.3) |
| Smoking | |
| Current | 285 (7.3) |
| Past | 1457 (37.2) |
| Never | 2172 (55.5) |
| Missing | 49 (1.2) |
| Education | |
| High school/GED or less | 991 (25.0) |
| School after high school | 1472 (37.2) |
| College degree or higher | 1484 (37.4) |
| Missing | 16 (0.4) |
| Parity | |
| Never pregnant/no term pregnancy | 445 (11.2) |
| 1 | 385 (9.7) |
| 2 | 898 (22.7) |
| 3 | 927 (23.4) |
| 4 | 613 (15.5) |
| ≥5 | 668 (16.9) |
| Missing | 27 (0.6) |
| Recreational physical activity, moderate or strenuous, times/week | |
| No activity | 640 (16.1) |
| Some activity | 1586 (40.0) |
| 2 to <4 episodes | 629 (15.9) |
| ≥4 episodes | 879 (22.2) |
| Missing | 229 (5.8) |
| Study component | |
| Observational study | 1425 (36.0) |
| Clinical trial | 2538 (64.0) |
| DM | 1579 (39.8) |
| HT | 1445 (36.5) |
| History of cardiovascular disease | 75 (2.2) |
| Yes | 88 (2.2) |
| No | 3875 (97.8) |
| History of diabetes | |
| Yes | 148 (3.7) |
| No | 3811 (96.2) |
| Missing | 4 (0.1) |
| History of hypertension1 | |
| Yes | 1557 (39.3) |
| No | 2406 (60.7) |
| History of dyslipidemia2 | |
| Yes | 1538 (39.0) |
| No | 2403 (60.4) |
| Missing | 22 (0.6) |
| Biometric, mean (SD) | |
| Triglycerides, mg/dL | 127.5 (60.5) |
| HDL, mg/dL | 60.1 (14.1) |
| Fasting blood glucose, mg/dL | 96.4 (23.2) |
| Waist: hip ratio | 0.81 (0.08) |
| Waist circumference, cm | 85.5 (13.1) |
| Energy, kcal/d, geometric mean (95% CI) | |
| Uncalibrated | 1444.3 (1427.5–1461.3) |
| Calibrated | 2083.8 (2075.4–2092.3) |
| Total protein, g/d, geometric mean (95% CI) | |
| Uncalibrated | 59.4 (58.6–60.2) |
| Calibrated | 73.5 (73.1–73.9) |
| Animal protein, g/d, geometric mean (95% CI) | |
| Uncalibrated | 39.3 (38.7–40.0) |
| Estimated3 | 48.5 (48.1–49.0) |
| Vegetable protein, g/d, geometric mean (95% CI) | |
| Uncalibrated | 18.3 (18.1–18.5) |
| Estimated3 | 22.8 (22.5–23.0) |
For categorical variables, percentages are based on known values. Variables with no missing values have no “missing” row. Abbreviation: DM, Dietary Modification Trial; HT, Hormone Therapy Trial.
Hypertension, or high blood pressure, was defined as systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg.
Dyslipidemia was defined as triglycerides ≥150 mg/dL or HDL <50 mg/dL.
Estimated by multiplying the biomarker-based calibrated total protein intake by the self-reported protein subtype proportion.
Comparing baseline characteristics between quartiles of total energy (kcal/d) and total protein (g/d) intakes, we observed those in the highest energy quartile tended to be ages 50–59, Black, and obese (BMI ≥ 30), and reported no physical activity (Table 2). Those in the highest protein quartile tended to be ages 50–59, White, and obese (BMI ≥ 30), and had higher parity (Table 2).
TABLE 2.
Baseline characteristics of sample stratified by baseline calibrated energy (kcal/d) and total protein (g/d) quartiles
| Calibrated energy <1917.3 kcal/d, n = 841 | Calibrated energy 1917.3 to <2071.2 kcal/d, n = 842 | Calibrated energy 2071.2 to <2245.3 kcal/d, n = 842 | Calibrated energy ≥2245.3 kcal/d, n = 842 | Calibrated total protein <66.3 g/d, n = 841 | Calibrated total protein 66.3 to <73.4 g/d, n = 842 | Calibrated total protein 73.4 to <81.7 g/d, n = 843 | Calibrated total protein ≥81.7 g/d, n = 841 | |
|---|---|---|---|---|---|---|---|---|
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| Age, years | ||||||||
| 50–59 | 23 (2.7) | 127 (15.1) | 321 (38.1) | 494 (58.7) | 70 (8.3) | 154 (18.3) | 307 (36.4) | 434 (51.6) |
| 60–69 | 326 (38.8) | 457 (54.3) | 427 (50.7) | 309 (36.7) | 346 (41.1) | 429 (51.0) | 395 (46.9) | 349 (41.5) |
| 70–79 | 492 (58.5) | 258 (30.6) | 94 (11.2) | 39 (4.6) | 425 (50.5) | 259 (30.8) | 141 (16.7) | 58 (6.9) |
| Race/ethnicity | ||||||||
| Non-Hispanic White | 327 (38.9) | 447 (53.1) | 453 (53.8) | 421 (50.0) | 259 (30.8) | 399 (47.4) | 460 (54.6) | 530 (63.0) |
| Non-Hispanic Black/African American | 91 (10.8) | 160 (19.0) | 187 (22.2) | 321 (38.1) | 284 (33.8) | 178 (21.1) | 167 (19.8) | 130 (15.5) |
| Hispanic/Latina | 171 (20.3) | 115 (13.7) | 95 (11.3) | 43 (5.1) | 118 (14.0) | 116 (13.8) | 102 (12.1) | 88 (10.5) |
| Asian/Pacific Islander | 217 (25.8) | 89 (10.6) | 53 (6.3) | 7 (0.8) | 155 (18.4) | 113 (13.4) | 68 (8.1) | 30 (3.6) |
| American Indian/Alaskan Native | 27 (3.2) | 26 (3.1) | 37 (4.4) | 40 (4.8) | 20 (2.4) | 30 (3.6) | 31 (3.7) | 49 (5.8) |
| Unknown | 8 (1.0) | 5 (0.6) | 17 (2.0) | 10 (1.2) | 5 (0.6) | 6 (0.7) | 15 (1.8) | 14 (1.7) |
| BMI, kg/m2 | ||||||||
| Underweight, <18.5 | 39 (4.6) | 8 (1.0) | 3 (0.4) | 0 (0.0) | 28 (3.3) | 13 (1.5) | 6 (0.7) | 3 (0.4) |
| Normal weight, 18.5–24.9 | 550 (65.4) | 345 (41.0) | 202 (24.0) | 36 (4.3) | 441 (52.4) | 335 (39.8) | 235 (27.9) | 122 (14.5) |
| Overweight, 25–29.9 | 235 (17.9) | 390 (46.3) | 394 (46.8) | 194 (23.0) | 280 (33.3) | 333 (39.5) | 344 (40.8) | 256 (30.4) |
| Obese, ≥30 | 17 (2.0) | 99 (11.8) | 243 (28.9) | 612 (72.7) | 92 (10.9) | 161 (19.1) | 258 (30.6) | 460 (54.7) |
| Smoking | ||||||||
| Current | 77 (9.2) | 71 (8.4) | 58 (6.9) | 39 (4.6) | 71 (8.4) | 66 (7.8) | 63 (7.5) | 45 (5.4) |
| Past | 243 (28.9) | 300 (35.6) | 344 (40.9) | 348 (41.3) | 278 (33.1) | 308 (36.6) | 313 (37.1) | 336 (40.0) |
| Never | 521 (62.0) | 471 (55.9) | 440 (52.3) | 455 (54.0) | 492 (58.5) | 468 (55.6) | 467 (55.4) | 460 (54.7) |
| Education | ||||||||
| High school/GED or less | 220 (26.2) | 213 (25.3) | 173 (20.5) | 212 (25.2) | 270 (32.1) | 208 (24.7) | 183 (21.7) | 157 (18.7) |
| School after high school | 306 (36.4) | 297 (35.3) | 342 (40.6) | 323 (38.4) | 248 (29.5) | 280 (33.3) | 332 (39.4) | 408 (48.5) |
| College degree or higher | 315 (37.5) | 332 (39.4) | 327 (38.8) | 307 (36.5) | 323 (38.4) | 354 (42.0) | 328 (38.9) | 276 (32.8) |
| Parity | ||||||||
| Never pregnant/no term pregnancy | 131 (15.6) | 82 (9.7) | 84 (10.0) | 83 (9.9) | 152 (18.1) | 87 (10.3) | 88 (10.4) | 53 (6.3) |
| 1 | 55 (6.5) | 72 (8.6) | 80 (9.5) | 114 (13.5) | 133 (15.8) | 91 (10.8) | 66 (7.8) | 31 (3.7) |
| 2 | 227 (27.0) | 200 (23.8) | 183 (21.7) | 165 (19.6) | 261 (31.0) | 221 (26.2) | 176 (20.9) | 117 (13.9_ |
| 3 | 251 (29.8) | 207 (24.6) | 188 (22.3) | 151 (17.9) | 181 (21.5) | 211 (25.1) | 209 (24.8) | 196 (23.3) |
| 4 | 74 (8.8) | 118 (14.0) | 164 (19.5) | 169 (20.1) | 78 (9.3) | 137 (16.3) | 132 (15.7) | 178 (21.2) |
| ≥5 | 103 (12.2) | 163 (19.4) | 143 (17.0) | 160 (19.0) | 36 (4.3) | 95 (11.3) | 172 (20.4) | 266 (31.6) |
| Recreational physical activity, moderate or strenuous, times/week | ||||||||
| No activity | 122 (14.5) | 144 (17.1) | 128 (15.2) | 172 (20.4) | 155 (18.4) | 127 (15.1) | 142 (16.8) | 142 (16.9) |
| Some activity | 385 (45.8) | 332 (39.4) | 352 (41.8) | 371 (44.1) | 379 (45.1) | 342 (40.6) | 358 (42.5) | 361 (42.9) |
| 2 to <4 episodes | 138 (16.4) | 143 (17.0) | 153 (18.2) | 139 (16.5) | 134 (15.9) | 148 (17.6) | 140 (16.6) | 151 (18.0) |
| ≥4 episodes | 196 (23.3) | 223 (26.5) | 209 (24.8) | 160 (19.0) | 173 (20.6) | 225 (26.7) | 203 (24.1) | 187 (22.2) |
| Study component | ||||||||
| Observational study | 449 (53.4) | 342 (40.6) | 272 (32.3) | 206 (24.5) | 413 (49.1) | 350 (41.6) | 276 (32.7) | 230 (27.3) |
| Clinical trial | 392 (46.6) | 500 (59.4) | 570 (67.7) | 636 (75.5) | 428 (50.9) | 492 (58.4) | 567 (67.3) | 611 (72.7) |
| DM | 220 (26.2) | 293 (34.8) | 348 (41.3) | 397 (47.1) | 249 (29.6) | 314 (37.3) | 340 (40.3) | 355 (42.2) |
| HT | 233 (27.7) | 283 (33.6) | 334 (39.7) | 391 (46.4) | 255 (30.3) | 261 (31.0) | 332 (39.4) | 393 (46.7) |
| Cardiovascular disease | ||||||||
| Yes | 21 (2.5) | 16 (1.9) | 15 (1.8) | 22 (2.6) | 16 (1.9) | 21 (2.5) | 17 (2.0) | 20 (2.4) |
| No | 820 (97.5) | 826 (98.1) | 827 (98.2) | 820 (97.4) | 825 (98.1) | 821 (97.5) | 826 (98.0) | 821 (97.6) |
| Diabetes | ||||||||
| Yes | 21 (2.5) | 29 (3.4) | 31 (3.7) | 40 (4.8) | 7 (0.8) | 19 (2.3) | 28 (3.3) | 67 (8.0) |
| No | 820 (97.5) | 813 (96.6) | 811 (96.3) | 920 (95.2) | 834 (92.8) | 823 (97.7) | 815 (96.7) | 774 (92.0) |
| Hypertension1 | ||||||||
| Yes | 359 (42.7) | 326 (38.7) | 323 (38.4) | 337 (40.0) | 394 (46.8) | 330 (39.2) | 294 (34.9) | 327 (38.9) |
| No | 482 (57.3) | 516 (61.3) | 519 (61.6) | 505 (60.0) | 447 (53.2) | 512 (60.8) | 549 (65.1) | 514 (61.1) |
| Dyslipidemia2 | ||||||||
| Yes | 302 (36.2) | 306 (36.5) | 328 (39.2) | 369 (44.0) | 279 (33.3) | 310 (37.0) | 322 (38.4) | 394 (47.0) |
| No | 539 (63.8) | 536 (63.5) | 514 (60.8) | 473 (56.0) | 562 (66.7) | 532 (63.0) | 521 (61.6) | 447 (53.0) |
| Missing | 6 (0.7) | 3 (0.4) | 5 (0.6) | 3 (0.4) | 4 (0.5) | 5 (0.6) | 5 (0.6) | 3 (0.4) |
| Biometrics, mean ± SD | ||||||||
| Triglycerides, mg/dL | 127.7 ± 58.9 | 124.9 ± 61.5 | 126.2 ± 62.2 | 125.7 ± 57.9 | 118.6 ± 53.3 | 126.6 ± 62.7 | 124.8 ± 58.6 | 134.5 ± 64.3 |
| HDL, mg/dL | 63.0 ± 15.1 | 61.3 ± 14.9 | 59.6 ± 13.7 | 56.1 ± 12.0 | 62.9 ± 15.3 | 61.2 ± 14.6 | 59.3 ± 13.6 | 56.6 ± 12.3 |
| Fasting blood glucose, mg/dL | 95.4 ± 19.5 | 94.9 ± 19.3 | 95.9 ± 23.7 | 99.0 ± 27.7 | 93.8 ± 14.1 | 95.3 ± 17.3 | 95.9 ± 24.5 | 100.3 ± 31.1 |
| Waist: hip ratio | 0.80 ± 0.10 | 0.80 ± 0.07 | 0.81 ± 0.07 | 0.82 ± 0.08 | 0.80 ± 0.10 | 0.80 ± 0.07 | 0.80 ± 0.07 | 0.82 ± 0.08 |
| Waist circumference, cm | 77.1 ± 9.4 | 81.8 ± 9.4 | 85.7 ± 10.9 | 96.4 ± 13.8 | 79.6 ± 10.5 | 82.7 ± 11.3 | 85.6 ± 11.8 | 93.1 ± 14.6 |
| Energy, kcal/d, geometric mean (95% CI) | ||||||||
| Uncalibrated | 1276.1 (1246.1–1306.8) | 1403.8 (1369.9–1438.6) | 1494.1 (1457.8–1531.4) | 1618.6 (1576.7–1661.7) | 1148.9 (1122.6–1175.8) | 1380.4 (1349.5–1412.0) | 1520.6 (1486.1–1555.8) | 1796.7 (1755.2–1839.2) |
| Calibrated | 1802.5 (1796.7–1808.4) | 1993.2 (1990.2–1996.2) | 2151.9 (2148.5–2155.2) | 2438.6 (2426.1–2451.1) | 1871.9 (1861.1–1882.8) | 2012.6 (2001.9–2023.4) | 2140.3 (2128.1–2152.5) | 2338.5 (2322.3–2354.7) |
| Total protein, g/d, geometric mean (95% CI) | ||||||||
| Uncalibrated | 51.7 (50.3–53.1) | 57.7 (56.1–59.3) | 61.8 (60.2–63.5) | 66.5 (64.6–68.5) | 44.3 (43.1–45.5) | 56.1 (54.7–57.5) | 63.4 (61.9–65.0) | 77.9 (76.0–79.8) |
| Calibrated | 63.4 (63.0–63.9) | 70.7 (70.2–71.2) | 76.7 (76.1–77.2) | 84.9 (84.1–85.6) | 60.3 (59.9–60.6) | 69.8 (69.7–70.0) | 77.3 (77.1–77.4) | 89.8 (89.3–90.3) |
| Animal protein, g/d, geometric mean (95% CI) | ||||||||
| Uncalibrated | 32.1 (30.9–33.3) | 38.1 (36.8–39.5) | 41.5 (40.1–42.8) | 45.6 (44.4–47.5) | 27.1 (26.1–28.1) | 36.8 (35.8–37.9) | 42.7 (41.5–44.0) | 54.7 (53.1–56.2) |
| Estimated3 | 39.4 (38.6–40.2) | 46.7 (45.9–47.5) | 51.4 (50.6–52.2) | 58.6 (57.7–59.4) | 36.9 (36.1–37.7) | 45.9 (45.3–46.4) | 52.0 (51.4–52.7) | 63.0 (62.3–63.8) |
| Vegetable protein, g/d, geometric mean (95% CI) | ||||||||
| Uncalibrated | 17.6 (17.1–18.1) | 17.9 (17.4–18.4) | 18.8 (18.3–19.3) | 19.1 (18.6–19.7) | 15.3 (14.9–15.8) | 17.8 (17.3–18.3) | 19.1 (18.6–19.6) | 21.7 (21.1–22.3) |
| Estimated3 | 21.6 (21.1–22.1) | 21.9 (21.5–22.4) | 23.3 (22.8–23.8) | 24.4 (23.9–24.9) | 20.9 (20.4–21.4) | 22.2 (21.7–22.7) | 23.3 (22.8–23.8) | 25.0 (24.5–25.5) |
For categorical variables, percentages are based on known values. Variables with no missing values do not have a “missing” row. Abbreviations: DM, Dietary Modification Trial; HT, Hormone Therapy Trial.
Hypertension, or high blood pressure, was defined as systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg.
Dyslipidemia was defined as triglycerides ≥150 mg/dL or HDL <50 mg/dL.
Estimated by multiplying the biomarker-based calibrated total protein intake by the self-reported protein subtype proportion.
In this subsample, 429 women (10.8%) met the criteria for MetS at the 3-year WHI clinic visit. The remaining 3534 women did not meet the definition of MetS at Year 3, and are henceforth referred to as metabolically healthy. Elevated total energy intake was associated with MetS, but only when using calibrated estimates. For every 10% increment in calibrated energy intake, women were at a 1.37-fold elevated risk of MetS (95% CI, 1.15–1.63; Table 3). Higher calibrated total protein intake was also associated with an elevated risk of MetS, such that for a 10% increment in calibrated protein intake, women were at a 1.21-fold elevated risk of MetS (95% CI, 1.00–1.47); however, this association was trending towards statistical significance (P = 0.053). No association was found between uncalibrated total protein intake and MetS. Additionally, no association was observed for a 10% increment in calibrated protein density and MetS (OR, 1.02; 95% CI, 0.77–1.35). When looking at protein type, higher total animal protein intake was associated with a 1.08-fold elevated risk of MetS (95% CI, 1.02–1.14), whereas no association was seen for total vegetable protein intake (OR, 0.99; 95% CI, 0.95–1.03). When looking at the association of baseline dietary intakes and the individual components of MetS, similar results were seen for elevated triglycerides, low HDL cholesterol, and elevated blood glucose (Supplementary Table 1). However, we did not observe any significant association between any baseline dietary factor and hypertension at follow-up Year 3.
TABLE 3.
Associations between 10% increments in consumption of baseline dietary intake measures and metabolic syndrome at Year 3
| Uncalibrated | Calibrated | |||||
|---|---|---|---|---|---|---|
| Dietary intake, modeled as 10% increments in: | Metabolic syndrome, n = 429 | |||||
| OR1 | 95% CI | P value | OR1 | 95% CI | P value | |
| Energy, kcal/d | 0.99 | 0.96–1.01 | 0.466 | 1.37 | 1.15–1.63 | <0.001 |
| Total protein, g/d | 0.99 | 0.96–1.01 | 0.659 | 1.21 | 1.00–1.47 | 0.053 |
| Protein density, % energy from protein2 | 1.02 | 0.94–1.10 | 0.698 | 1.02 | 0.77–1.35 | 0.880 |
| Animal protein, g/d3 | 1.00 | 0.97–1.03 | 0.962 | 1.08 | 1.02–1.14 | 0.006 |
| Vegetable protein, g/d3 | 0.98 | 0.94–1.01 | 0.201 | 0.99 | 0.95–1.03 | 0.531 |
Logistic regression was used to calculate ORs and 95% CIs. Abbreviations: CT, clinical trial; DM, Dietary Modification Trial; HS, high school; HT, Hormone Therapy Trial; OS, observational study.
Adjusted for age (y), OS/CT, HT/DM, race/ethnicity (non-Hispanic White, Non-Hispanic Black/African American, Hispanic/Latina, Asian/Pacific Islander, American Indian/Alaskan Native, Unknown), education (≤HS, some post-HS, college degree, or higher), income (<$20K, $20 to <$35K, $35 to <$50K, $50 to <$75K, ≥$75K), smoking (never, past, current), total recreational physical activity, history of treated diabetes (no/yes), history of hypertension (no/yes), and parity (never pregnant/no term pregnancies, 1, 2, 3, 4, ≥5).
Additional adjustment for total energy.
Estimated by multiplying the biomarker-based calibrated total protein intake by the self-reported protein subtype proportion.
We also investigated the roles of calibrated energy, total protein, and protein density intakes by metabolic phenotype. We did not observe a significant association between calibrated total energy intake, calibrated total protein intake, or calibrated protein density and the MetS risk in normal-weight or overweight/obese women (Table 4). Higher animal protein intake was associated with a modestly elevated risk of MetS in normal-weight women; however, this association was not significant (OR, 1.11; 95% CI, 1.00–1.32; P = 0.134).
TABLE 4.
Associations between baseline biomarker calibrated dietary intakes and metabolic syndrome, stratified by weight status at Year 3
| Dietary intake, modeled as 10% increments in: | OR1, 2 | 95% CI | P value |
|---|---|---|---|
| MUH-NW2 (n = 62) versus MH-NW2 (n = 1147) | |||
| Energy, kcal/d | 1.07 | 0.62–1.84 | 0.813 |
| Total protein, g/d | 0.94 | 0.62–1.42 | 0.762 |
| Protein density, % energy from protein3 | 0.87 | 0.56–1.34 | 0.516 |
| Animal protein, g/d4 | 1.11 | 1.00–1.32 | 0.134 |
| Vegetable protein, g/d4 | 0.95 | 0.84–1.04 | 0.292 |
| MUH-O2 (n = 360) versus MH-O2 (n = 2290) | |||
| Energy, kcal/d | 1.09 | 0.89–1.35 | 0.404 |
| Protein, g/d | 1.01 | 0.81–1.25 | 0.960 |
| Protein density, % energy from protein3 | 0.99 | 0.73–1.35 | 0.973 |
| Animal protein, g/d4 | 1.01 | 0.93–1.09 | 0.871 |
| Vegetable protein, g/d4 | 0.99 | 0.94–1.05 | 0.744 |
Logistic regression was used to calculate ORs and 95% CIs. Abbreviations: CT, clinical trial; DM, Dietary Modification Trial; HS, high school; HT, Hormone Therapy Trial; MetS, metabolic syndrome; MH-NW, metabolically healthy normal weight; MH-O: metabolically healthy overweight/obese; MUH-NW, metabolically unhealthy normal weight; MUH-O, metabolically unhealthy obese/overweight; OS, observational study.
Adjusted for age (y), OS/CT, HT/DM, race/ethnicity (Non-Hispanic White, Non-Hispanic Black/African American, Hispanic/Latina, Asian/Pacific Islander, American Indian/Alaskan Native, Unknown), education (≤HS, some post-HS, college degree, or higher), income (<$20K, $20 to <$35K, $35 to <$50K, $50 to <$75K, ≥$75K), smoking (never, past, current), total recreational physical activity, history of treated diabetes (no/yes), history of hypertension (no/yes), and parity (never pregnant/no term pregnancies, 1, 2, 3, 4, ≥5).
2MUH-NW was defined as MetS and a BMI 18.5 to <25; MH-NW was defined as no MetS and a BMI 18.5 to <25; MH-O was defined as no MetS and a BMI ≥25; and MUH-O was defined as MetS and a BMI ≥25.
Additional adjustment for total energy.
Estimated by multiplying the biomarker-based calibrated total protein intake by the self-reported protein subtype proportion.
In an exploratory analysis, we did not detect evidence of a difference by race/ethnicity in the risk of MetS associated with calibrated total energy, total protein intake, or protein density (Supplementary Table 2).
In addition to the main analyses, we conducted sensitivity analyses. First, we excluded participants taking antihypertensive, lipid-lowering, or diabetes medication at baseline from the models; results were unchanged (data not shown). Additionally, adjusting for the impact of weight gain, as defined as a ≥2.23 kg increase between baseline and Year 3, did not affect the results (data not shown). Lastly, we explored modeling calibrated energy and protein intakes as categorical variables based on quartiles, which showed significant increases in ORs for both energy and total protein intakes with increasing quartiles compared to the lowest quartile, suggesting a linear trend.
Discussion
The present study examined the associations between biomarker-calibrated daily energy and protein intakes and the risk of MetS. Greater calibrated energy and total protein intakes were associated with elevated risks of MetS; however, no association was found for protein density and the risk of MetS. When looking at protein type, total animal protein intake was associated with an elevated risk of MetS, whereas total vegetable protein intake was not. Additionally, when stratified by metabolic phenotypes, no associations were found for calibrated total energy intake, total protein intake, or protein density. However, we observed that higher consumption of animal protein was associated with MetS in normal-weight women as compared to overweight/obese women, although the association was not statistically significant.
Findings from our study demonstrate that higher biomarker-calibrated energy intake is associated with an elevated risk of MetS, which is consistent with the hypothesized mechanism of MetS (4, 27). Higher caloric intake leads to overnutrition, which results when more energy is consumed than required for metabolic function, leading to increased visceral adiposity (27, 28). While increased visceral adiposity is linked to MetS, not all obese individuals develop MetS, suggesting that higher caloric intake may be independent of body size. When we examined this association by metabolic phenotype, no differences were found when stratified by weight status. Thus, energy restriction may improve metabolic health regardless of body size.
Greater calibrated total protein intake was associated with an elevated risk of MetS in the current study; however, when we examined protein density there were no associations with the risk of MetS. The findings from this study are consistent with some previous reports (9–12, 14, 29–32) showing that higher protein intake was associated with the incidence of MetS, increased central obesity, greater weight gain, and incident Type II diabetes. However, it is important to consider the source of protein. In an exploratory analysis, we found a higher consumption of total animal protein to be associated with an elevated risk of MetS, whereas no association was found for total vegetable protein consumption and MetS. Given there are no recovery biomarkers for animal or vegetable protein sources, the values in this study are not biomarker-calibrated, but are estimates based on the biomarker-calibrated total protein and the self-reported proportion of protein type from the FFQ. This is congruent with other studies that found a detrimental influence of protein intake from red meat and other types of animal protein on metabolic health (10–12, 14, 29–31). In contrast, studies that examined the effects of protein intake from plant sources found a more favorable impact on metabolic health (14, 33). A potential mechanism proposed for these associations is that diets high in animal protein contain more saturated fats and can promote oxidative stress and inflammation, leading to an increased risk of components of MetS, whereas plant-based proteins have been associated with reduced cholesterol and lower blood pressure (20, 34). However, studies examining total protein intake regardless of source have found higher protein intake to be associated with greater weight gain (9, 10), although this has not been supported in all studies (7, 33, 35, 36). Given that animal protein intake is associated with a higher risk of MetS and vegetable protein intake is typically protective, these associations could be missed when total protein intake is investigated. While exploratory, this study provides support for the theory that protein source plays a role in MetS risk, with a higher consumption of animal protein conferring a higher risk of MetS, indicating that greater detail on protein type should be included in future studies. In addition, our findings also suggest that higher animal protein intake is associated with an elevated risk of MetS only among normal-weight women, although this association was not significant. Future studies should investigate the differences between protein types and MetS risks among metabolic phenotypes.
In general, high-protein diets are considered to consist of 25–35% energy from protein, whereas normal-protein diets are often considered to consist of 15% energy from protein (37). We did not observe any association between the MetS risk and protein density; however, this could be the result of relatively low protein intake in this sample. In both metabolically healthy and unhealthy women, the average protein density in our data set was approximately 14% of energy intake. Therefore, the average protein density in this sample may not be sufficiently high to detect an association with MetS. There is a further need to examine the role protein plays in the risk of MetS, particularly looking at the source of protein and the differences between total protein intake and protein density.
Associations between dietary intake and MetS were only found when examining biomarker-calibrated estimates of energy and protein intakes. Associations were not observed with uncalibrated estimates. The use of self-reported dietary measures in the literature could help explain the inconsistencies reported on associations of protein intake with MetS. A large meta-analysis found that less than 50% of the studies used a biomarker to validate self-reported protein intake (7). However, the results from this study are similar to previous studies using biomarker-calibrated estimates of dietary intake. Two prior studies found biomarker-calibrated protein to be significantly associated with weight gain, a component of MetS (12, 32). Self-reported dietary measures are often criticized for consistently underreporting energy and protein intakes, especially for individuals with higher BMIs, which is important when examining MetS (38, 39). A previous study in the WHI reported significant underreporting of energy and protein by up to 27–33% and 10–15%, respectively (16). This was also seen in our study, as energy and protein intakes were much lower when looking at uncalibrated estimates compared to calibrated estimates. This consistent underreporting of dietary intake can lead to systematic measurement errors, resulting in bias and erroneous results. Calibration methods have been developed and used successfully to correct for this type of measurement error (18). As observed in this study, by correcting for dietary intake measurement errors, the results provide strong support for the need to use objective measures of dietary intake in order to accurately assess associations between diet and MetS.
The strengths of this study include the prospective cohort design, a large sample size of metabolically healthy and unhealthy women, the use of biomarker-calibrated estimates of energy and protein intakes, and the wide representation of racial/ethnic minorities. The biomarker subsample, NPAAS, included diverse samplings of race/ethnicity and BMI status that were used in developing the biomarker calibrations, with potential variances in self-reporting bias when assessing calibrated intakes by subtypes. The use of biomarker-calibrated estimates reduces the risks of bias and misclassification that are often associated with self-reported dietary intake (18).
Despite these strengths, there are limitations of this study. First, while biomarker calibrations provide more objective assessments of dietary intake than uncalibrated, self-reported assessments, the calibrations are not a direct measure of recovery biomarkers. However, biomarker calibrations have been used successfully to account for dietary reporting error and provide a cost efficient and feasible way to measure dietary intake. Relatedly, we were only able to examine the associations between energy and protein intakes as components of dietary intake, since there are no currently established recovery biomarkers for fat or carbohydrate intakes. There is evidence that MetS or its components may be associated with varying ratios of carbohydrate and fat intake (40). Further research is needed to examine the associations between MetS and other macronutrients, as well as the distributions of macronutrients in relation to total intake. Lastly, as caloric needs are in part driven by the extent of physical activity and body size, our ascertainment of leisure-time physical activity only and the inability to include BMI in our models are limitations. However, as stated previously, BMI was an important factor in the biomarker calibration equations, and further inclusion in our models would overadjust for BMI. Thus, we stratified our analysis by body size to better overcome this limitation. To the extent that other forms of physical activity (e.g., vocational physical activity) are present or that physical activity is misclassified due to errors in self-reports, there will be residual confounding due to physical activity in our analysis.
In conclusion, this study highlights the importance of caloric intake, total protein intake, and protein type in the risk of MetS and in metabolic phenotypes, as well as the importance of using methods to minimize bias while estimating dietary intake. Exploratory data also suggest that animal protein consumption may be driving the results seen for total protein intake. Given the high prevalence of MetS, dietary interventions focused on energy and animal protein restrictions may be effective ways to reduce MetS.
Supplementary Material
ACKNOWLEDGEMENTS
The authors thank the Women's Health Initiative Investigators from the Program Office (National Heart, Lung, and Blood Institute, Bethesda, Maryland), Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, and Nancy Geller; and the Clinical Coordinating Center (Fred Hutchinson Cancer Research Center, Seattle, WA), Garnet Anderson, Ross Prentice, Andrea LaCroix, and Charles Kooperberg. They thank the following investigators and Academic Centers: (Brigham and Women's Hospital, Harvard Medical School, Boston, MA) JoAnn E Manson; (MedStar Health Research Institute/Howard University, Washington, DC) Barbara V Howard; (Stanford Prevention Research Center, Stanford, CA) Marcia L Stefanick; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Iowa, Iowa City/Davenport, IA) Jennifer Robinson; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker; and (University of Nevada, Reno, NV) Robert Brunner. A full list of WHI investigators can be found at www.whi.org.
The authors’ responsibilities were as follows – AV and KWR: designed the research and took primary responsibility for the final content; MP: conducted the data analysis; AV, LFT, MLN, LH, XL, LLH, and KWR: wrote the manuscript; and all authors: played a role in the interpretation of data and read and approved the final manuscript.
Author disclosures: LFT, MLN, MP, CD, OZ, LLH, XL, CE, DW, AS, MLS, WEB, and KWR, no conflicts of interest. AV is supported by the American Achievement for College Scientists Seattle Chapter, the National Institute of Nursing Research Omics and Symptom Science Training Program at the University of Washington (T32016913-01), and a Ruth L Kirschstein National Research Service Award Research Training Grant (F31NR018588-01) by the NIH/National Institute of Nursing Research. LH receives funding from the NIH for separate research from this project, and she also receives an honorarium from the National Sleep Foundation for her role as Editor-in-Chief of the journal Sleep Health.
Notes
The Women's Health Initiative is funded by the National Heart, Lung, and Blood Institute, NIH, US Department of Health and Human Services, through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C.
Supplemental Table 1 and Supplemental Figure 1 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.
Abbreviations used: BP, blood pressure; CT, clinical trial; DM, Dietary Modification Trial; HS, high school; HT, Hormone Therapy Trial; MetS, metabolic syndrome; MH-NW, metabolically healthy normal weight; MH-O, metabolically healthy overweight/obese; MUH-NW, metabolically unhealthy normal weight; MUH-O, metabolically unhealthy overweight/obese; NPAAS, Nutrition and Physical Activity Assessment Study; OS, observational study; WHI, Women's Health Initiative.
Contributor Information
Alexi Vasbinder, Biobehavioral Nursing and Health Informatics, University of Washington School of Nursing, Seattle, WA, USA.
Lesley F Tinker, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Marian L Neuhouser, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Mary Pettinger, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Lauren Hale, Program in Public Health, Department of Family, Population, and Preventive Medicine, Renaissance School of Medicine, Stony Brook University, Stony Brook, NY, USA.
Chongzhi Di, Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Oleg Zaslavsky, Biobehavioral Nursing and Health Informatics, University of Washington School of Nursing, Seattle, WA, USA.
Laura L Hayman, Department of Nursing, University of Massachusetts Boston, Boston, MA, USA; Department of Population and Quantitative Health Sciences, University of Massachusetts Medical School, Boston, MA, USA.
Xioachen Lin, Department of Epidemiology, Center for Global Cardiometabolic Health, Brown University, Providence, RI, USA.
Charles Eaton, Department of Family Medicine and Epidemiology, Alpert Medical School, Brown University, Providence, RI, USA.
Di Wang, Biobehavioral Nursing and Health Informatics, University of Washington School of Nursing, Seattle, WA, USA.
Ashley Scherman, Biobehavioral Nursing and Health Informatics, University of Washington School of Nursing, Seattle, WA, USA.
Marcia L Stefanick, Department of Medicine, Stanford Prevention Research Center, Stanford, CA, USA.
Wendy E Barrington, Child, Family, Population Health Nursing, University of Washington School of Nursing, Seattle, WA, USA.
Kerryn W Reding, Biobehavioral Nursing and Health Informatics, University of Washington School of Nursing, Seattle, WA, USA; Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Data Availability
Data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval.
Data are available in accordance with policies developed by the National Heart, Lung, and Blood Institute and Women's Health Initiative (WHI) in order to protect sensitive participant information and approved by the Fred Hutchinson Cancer Research Center, which currently serves as the Institutional Review Board of record for the WHI. Data requests may be made by emailing helpdesk@WHI.org.
References
- 1. Mottillo S, Filion KB, Genest J, Joseph L, Pilote L, Poirier P, Rinfret S, Schiffrin EL, Eisenberg MJ. The metabolic syndrome and cardiovascular risk: A systematic review and meta-analysis. J Am Coll Cardiol. 2010;56(14):1113–32.. doi: 10.1016/j.jacc.2010.05.034. [DOI] [PubMed] [Google Scholar]
- 2. Ford ES. Risks for all-cause mortality, cardiovascular disease, and diabetes associated with the metabolic syndrome. Diabetes Care. 2005;28(7):1769–78. [DOI] [PubMed] [Google Scholar]
- 3. Aguilar M, Bhuket T, Torres S, Liu B, Wong RJ. Prevalence of the metabolic syndrome in the United States, 2003–2012. JAMA. 2015;313(19):1973–4. [DOI] [PubMed] [Google Scholar]
- 4. Grundy SM. Metabolic syndrome update. Trends Cardiovasc Med. 2016;26(4):364–73.. doi: 10.1016/j.tcm.2015.10.004. [DOI] [PubMed] [Google Scholar]
- 5. Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens Rep. 2018;20(2):12–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wildman RP, Muntner P, Reynolds K, McGinn AP, Rajpathak S, Wylie-Rosett J, Sowers MR. 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):1617–24. [DOI] [PubMed] [Google Scholar]
- 7. Wycherley TP, Moran LJ, Clifton PM, Noakes M, Brinkworth GD. Effects of energy-restricted high-protein, low-fat compared with standard-protein, low-fat diets: A meta-analysis of randomized controlled trials. Am J Clin Nutr. 2012;96(6):1281–98. [DOI] [PubMed] [Google Scholar]
- 8. Baxter AJ, Coyne T, McClintock C. Dietary patterns and metabolic syndrome–A review of epidemiologic evidence. Asia Pac J Clin Nutr. 2006;15(2):134–42. [PubMed] [Google Scholar]
- 9. Cheng M, Wang H, Wang Z, Du W, Ouyang Y, Zhang B. Relationship between dietary factors and the number of altered metabolic syndrome components in Chinese adults: A cross-sectional study using data from the China Health and Nutrition Survey. BMJ Open. 2017;7(5):e014911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hernández-Alonso P, Salas-Salvadó J, Ruiz-Canela M, Corella D, Estruch R, Fitó M, Arós F, Gómez-Gracia E, Fiol M, Lapetra Jet al. High dietary protein intake is associated with an increased body weight and total death risk. Clin Nutr. 2016;35(2):496–506.. doi: 10.1016/j.clnu.2015.03.016. [DOI] [PubMed] [Google Scholar]
- 11. van Nielen M, Feskens EJM, Mensink M, Sluijs I, Molina E, Amiano P, Ardanaz E, Balkau B, Beulens JWJ, Boeing Het al. Dietary protein intake and incidence of type 2 diabetes in Europe: The EPIC-InterAct case-cohort study. Dia Care. 2014;37(7):1854–62. [DOI] [PubMed] [Google Scholar]
- 12. Vergnaud A-C, Norat T, Romaguera D, Mouw T, May AM, Travier N, Luan J, Wareham N, Slimani N, Rinaldi Set al. Meat consumption and prospective weight change in participants of the EPIC-PANACEA study. Am J Clin Nutr. 2010;92(2):398–407. [DOI] [PubMed] [Google Scholar]
- 13. Hankinson AL, Daviglus ML, Horn LV, Chan Q, Brown I, Holmes E, Elliott P, Stamler J. Diet composition and activity level of at risk and metabolically healthy obese American adults. Obesity. 2013;21(3):637–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Shang X, Scott D, Hodge A, English DR, Giles GG, Ebeling PR, Sanders KM. Dietary protein from different food sources, incident metabolic syndrome and changes in its components: An 11-year longitudinal study in healthy community-dwelling adults. Clin Nutr. 2017;36(6):1540–8. [DOI] [PubMed] [Google Scholar]
- 15. Subar AF, Kipnis V, Troiano RP, Midthune D, Schoeller DA, Bingham S, Sharbaugh CO, Trabulsi J, Runswick S, Ballard-Barbash Ret al. Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: The OPEN study. Am J Epidemiol. 2003;158(1):1–13. [DOI] [PubMed] [Google Scholar]
- 16. Neuhouser ML, Tinker L, Shaw PA, Schoeller D, Bingham SA, Horn LV, Beresford SAA, Caan B, Thomson C, Satterfield Set al. Use of recovery biomarkers to calibrate nutrient consumption self-reports in the Women's Health Initiative. Am J Epidemiol. 2008;167(10):1247–59. [DOI] [PubMed] [Google Scholar]
- 17. Prentice RL, Mossavar-Rahmani Y, Huang Y, Van Horn L, Beresford SAA, Caan B, Tinker L, Schoeller D, Bingham S, Eaton CBet al. Evaluation and comparison of food records, recalls, and frequencies for energy and protein assessment by using recovery biomarkers. Am J Epidemiol. 2011;174(5):591–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Prentice RL, Tinker LF, Huang Y, Neuhouser ML. Calibration of self-reported dietary measures using biomarkers: An approach to enhancing nutritional epidemiology reliability. Curr Atheroscler Rep. 2013;15(9):353–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Kim H-N, Song S-W.. Associations between macronutrient intakes and obesity/metabolic risk phenotypes: Findings of the Korean National Health and Nutrition Examination Survey. Nutrients. 2019;11(3):628–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Kim Y, Je Y.. Meat consumption and risk of metabolic syndrome: Results from the Korean population and a meta-analysis of observational studies. Nutrients. 2018;10(4):390–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Park Y-MM, Fung TT, Steck SE, Zhang J, Hazlett LJ, Han K, Lee S-H, Merchant AT. Diet quality and mortality risk in metabolically obese normal-weight adults. Mayo Clin Proc. 2016;91(10):1372–83. [DOI] [PubMed] [Google Scholar]
- 22. Gaillard TR. The metabolic syndrome and its components in African-American women: Emerging trends and implications. Front Endocrinol. 2018;8:383─88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Palaniappan LP, Wong EC, Shin JJ, Fortmann SP, Lauderdale DS. Asian Americans have greater prevalence of metabolic syndrome despite lower body mass index. Int J Obes. 2011;35(3):393–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. The Women's Health Initiative Study Group . Design of the Women's Health Initiative clinical trial and observational study. Control Clin Trials. 1998;19(1):61–109.. 10.1016/S0197-2456(97)00078-0. [DOI] [PubMed] [Google Scholar]
- 25. Neuhouser ML, Howard B, Lu J, Tinker LF, Van Horn L, Caan B, Rohan T, Stefanick ML, Thomson CA. A low-fat dietary pattern and risk of metabolic syndrome in postmenopausal women: The Women's Health Initiative. Metabolism. 2012;61(11):1572–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Prentice RL, Huang Y, Kuller LH, Tinker LF, Horn LV, Stefanick ML, Sarto G, Ockene J, Johnson KC. Biomarker-calibrated energy and protein consumption and cardiovascular disease risk among postmenopausal women. Epidemiology. 2011;22(2):170–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Rochlani Y, Pothineni NV, Kovelamudi S, Mehta JL. Metabolic syndrome: Pathophysiology, management, and modulation by natural compounds. Ther Adv Cardiovasc Dis. 2017;11(8):215–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Svacina S. Basic concepts in nutrition: Overnutrition—functional and clinical consequences. Eur J Clin Nutr Metab. 2008;3(4):e173–8. [Google Scholar]
- 29. Babio N, Sorlí M, Bulló M, Basora J, Ibarrola-Jurado N, Fernández-Ballart J, Martínez-González MA, Serra-Majem L, González-Pérez R, Salas-Salvadó J. Association between red meat consumption and metabolic syndrome in a Mediterranean population at high cardiovascular risk: Cross-sectional and 1-year follow-up assessment. Nutr Metab Cardiovasc Dis. 2012;22(3):200–7.. doi: 10.1016/j.numecd.2010.06.011. [DOI] [PubMed] [Google Scholar]
- 30. Sluijs I, Beulens JWJ, van der ADL, Spijkerman AMW, Grobbee DE, van der Schouw YT. Dietary intake of total, animal, and vegetable protein and risk of type 2 diabetes in the European Prospective Investigation Into Cancer and Nutrition (EPIC)-NL study. Diabetes Care. 2010;33(1):43–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. The InterAct Consortium, Bendinelli B, Palli D, Masala G, Sharp SJ, Schulze MB, Guevara M, van der AD, Sera F, Amiano Pet al. Association between dietary meat consumption and incident type 2 diabetes: The EPIC-InterAct study. Diabetologia. 2013;56(1):47–59. [DOI] [PubMed] [Google Scholar]
- 32. Vergnaud A-C, Norat T, Mouw T, Romaguera D, May AM, Bueno-de-Mesquita HB, van der AD, Agudo A, Wareham N, Khaw K-Tet al. Macronutrient composition of the diet and prospective weight change in participants of the EPIC-PANACEA study. PLOS One. 2013;8(3):e57300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hill AM, Harris Jackson KA, Roussell MA, West SG, Kris-Etherton PM. Type and amount of dietary protein in the treatment of metabolic syndrome: A randomized controlled trial. Am J Clin Nutr. 2015;102(4):757–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Chalvon-Demersay T, Azzout-Marniche D, Arfsten J, Egli L, Gaudichon C, Karagounis LG, Tomé D. A systematic review of the effects of plant compared with animal protein sources on features of metabolic syndrome. J Nutr. 2017;147(3):281–92. [DOI] [PubMed] [Google Scholar]
- 35. Santesso N, Akl EA, Bianchi M, Mente A, Mustafa R, Heels-Ansdell D, Schünemann HJ. Effects of higher- versus lower-protein diets on health outcomes: A systematic review and meta-analysis. Eur J Clin Nutr. 2012;66:780–8. [Internet]. Available from: https://www.nature.com/articles/ejcn201237#supplementary-information. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Te Morenga L, Mann J. The role of high-protein diets in body weight management and health. Br J Nutr. 2012;108(S2):S130–S8. [DOI] [PubMed] [Google Scholar]
- 37. Clifton PM, Keogh J.. Metabolic effects of high-protein diets. Curr Atheroscler Rep. 2007;9(6):472–8. [DOI] [PubMed] [Google Scholar]
- 38. Freedman LS, Commins JM, Moler JE, Arab L, Baer DJ, Kipnis V, Midthune D, Moshfegh AJ, Neuhouser ML, Prentice RLet al. Pooled results from 5 validation studies of dietary self-report instruments using recovery biomarkers for energy and protein intake. Am J Epidemiol. 2014;180(2):172–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Park Y, Dodd KW, Kipnis V, Thompson FE, Potischman N, Schoeller DA, Baer DJ, Midthune D, Troiano RP, Bowles Het al. Comparison of self-reported dietary intakes from the automated self-administered 24-h recall, 4-d food records, and food-frequency questionnaires against recovery biomarkers. Am J Clin Nutr. 2018;107(1):80–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Skilton MR, Laville M, Cust AE, Moulin P, Bonnet F. The association between dietary macronutrient intake and the prevalence of the metabolic syndrome. Br J Nutr. 2008;100(2):400–7. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval.
Data are available in accordance with policies developed by the National Heart, Lung, and Blood Institute and Women's Health Initiative (WHI) in order to protect sensitive participant information and approved by the Fred Hutchinson Cancer Research Center, which currently serves as the Institutional Review Board of record for the WHI. Data requests may be made by emailing helpdesk@WHI.org.