Abstract
Background
The relation between dietary fat intake and body weight remains controversial. Few studies have examined long-term changes in types of dietary fat and weight change in longitudinal studies.
Objective
The objective of this study was to examine associations between intake of different types of fat and long-term weight change in US women and men.
Methods
The association between changes in consumption of varying types of fat and weight change was examined every 4 y through the use of multivariate models adjusted for age, baseline body mass index, and change in percentage energy from protein, intake of cereal fiber, fruits, and vegetables, alcohol use, and other lifestyle covariates in 3 prospective US cohorts, including 121,335 men and women free of diabetes, cardiovascular disease, cancer, or obesity over a 20- to 24-y follow-up. Dietary intakes and body weight were assessed via validated questionnaires. Cohort-specific results were pooled with the use of a random-effect meta-analysis.
Results
Compared with equivalent changes in carbohydrate intake, a 5% increase in energy from saturated fatty acid (SFA) and a 1% increase in energy from trans-fat were associated with 0.61 kg (95% CI: 0.54, 0.68 kg) and 0.69 kg (95% CI: 0.56, 0.84 kg) greater weight gain per 4-y period, respectively. A 5% increase in energy from polyunsaturated fatty acid (PUFA) was associated with less weight gain (−0.55 kg; 95% CI: −0.81, −0.29 kg). Increased intake of monounsaturated fatty acid (MUFA) from animal sources by 1% was associated with weight gain of 0.29 kg (95% CI: 0.25, 0.33 kg), whereas MUFA from plant sources was not associated with weight gain.
Conclusions
Different dietary fats have divergent associations with long-term weight change in US men and women. Replacing saturated and trans-fats with unsaturated fats, especially PUFAs, contributes to the prevention of age-related weight gain.
These trials were registered at clinicaltrials.gov as NCT00005152 and NCT00005182.
Keywords: dietary fat, weight change, weight gain, obesity, obesity prevention, long-term, SFAs, trans-fat, MUFAs, PUFAs
Introduction
More than one-third (36.5%) of US adults have obesity (1). Obesity is a major risk factor for the development of cardiovascular disease (CVD), type 2 diabetes, and several cancers (2). On average, adults gain <0.5 kg of body weight per year which makes it challenging for people to perceive the cumulative effect of weight gain over a long period of time (3). Prevention of gradual weight gain is therefore an effective strategy to prevent obesity.
To date, low-fat diets are still commonly recommended as a popular weight-loss approach (4). A reduction in fat intake may lead to a compensatory increased intake of refined carbohydrates and added sugars (4). Low-fat diets (20–30% of calories from fat) did not provide long-term beneficial effects in cardiovascular health or clinically meaningful weight loss when compared with usual diets in postmenopausal women (5) and patients with type 2 diabetes (6). A previous study demonstrated that the amount of calories from fat has only a weak positive relation with weight gain, and increases in monounsaturated and polyunsaturated fat were not associated with weight gain in 41,518 women from the Nurses’ Health Study (NHS) during an 8-y follow-up (7). A recent meta-analysis reported that weight-loss diets without fat restriction led to an average 1.15-kg greater weight loss compared with low-fat diets (10–30% of calories from fat) (8).
An important limitation of any recommended diet focusing on a single macronutrient (e.g., low-fat or low-carbohydrate diets) is that it ignores the fact that the types or quality of macronutrients may have different effects on body weight and other health outcomes. Previous weight-loss trials have examined the effectiveness of varying amounts of macronutrients on weight loss and maintenance over short periods of time. However, these trials cannot be generalized to long-term weight gain prevention in the general population. Therefore, the present study aimed to investigate the associations between changes in intakes of varying types of dietary fat and long-term weight gain in women and men from 3 independent prospective cohort studies.
Methods
Population characteristics
We analyzed 3 prospective cohorts including the Health Professionals Follow-Up Study (HPFS), the NHS, and the NHS II. The HPFS includes 51,529 male health professionals between 40 and 75 y of age at enrollment in 1986 (9). The NHS includes 121,700 female nurses resident in 11 US states aged 30–55 y when first enrolled in 1976 (10). The NHS II enrolled 116,686 nurses from 14 states at younger ages (25–42 y) when first enrolled in 1989 (10). Approximately 97% of participants who have followed up were white and of European descent (11). The baseline used in the present study was the first year for which detailed information on diet, physical activity, and lifestyle was available. For this analysis, baseline was 1986 for HPFS and NHS and 1991 for NHS II. Participants were followed up through mailed biennial validated questionnaires that ascertained medical history, lifestyle factors, and other health-related behaviors as previously described (9). Follow-up questionnaires have been sent every 2 y to the cohort participants, updating information on a broad range of risk factors and identifying newly diagnosed cases of various diseases. Dietary intake was assessed every 4 y via validated FFQs. The study protocol was approved by the institutional review boards of Brigham and Women's Hospital and Harvard T.H. Chan School of Public Health.
Assessment of dietary fat intake
In each cohort, usual dietary intake was assessed every 4 y via a 130-item validated FFQ from which 4-y dietary changes were calculated (12, 13). Respondents were asked the frequency and quantity of consumption of each type of food or beverage within 9 possible frequency categories ranging from never to ≥6 times/d. The number of listed foods was expanded from 61 to 150 from 1980 to 1984; the additional frequently consumed foods were reported in an open-ended section. Detailed information on the types of fat, oils used in food preparation, and brand or type of margarines was collected on the FFQs. Nutrient intake was calculated by multiplying a weight proportional to the frequency of consumption of each food item by its nutrient composition and summing across all foods, taking into account the specific brand and type of margarines and the types of fat used in food preparation. Fatty acid and other nutrient contents of foods were obtained from the USDA food composition database. The assessment of total and specific types of fat has been validated by comparison with multiple 1-wk dietary records collected over 1 y, and with fatty acid measured in plasma and adipose tissue. Correlation coefficients between FFQs and dietary records were 0.57–0.62 for total fat, 0.68–0.75 for SFAs, 0.48–0.51 for PUFAs, and 0.58–0.60 for MUFAs in women and men (14–16). The correlation between dietary fatty acid intake assessed by FFQs and fatty acids composition in adipose tissue was 0.35–0.48 for linoleic acid in women and men (17).
Assessment of weight changes
Weight and height were assessed by questionnaire at enrollment, and weight was requested on every follow-up questionnaire. Weight change was calculated as the differences in weight over each 4-y interval. In a validation study, weight reported by a subsample of 184 women who were weighed 6–12 mo after completing the mailed questionnaire showed that self-reported body weight was highly correlated with measured body weight (r = 0.97) (18). We analyzed the concurrent changes in fat intake and changes in weight because this type of comparison produced more robust, consistent, and biologically plausible associations by mimicking the effects of a dietary intervention study on body weight (19).
Statistical analysis
We assessed the association between changes in percentage energy of fat intake and weight changes in 4-y intervals over follow-up periods of 24 y in HPFS and NHS, and 20 y in NHS II. Participants who reached 65 y of age at baseline and during the follow-up periods were censored to minimize confounding from age-related weight loss. We excluded participants who reported diagnoses of cancer, cardiovascular disease, obesity (BMI ≥30), or diabetes, those with >9 blank responses on the FFQs, and those with missing energy intake at baseline. We excluded participants with implausible energy intake (<600 or >3500 kcal/d for women or <800 or >4200 kcal/d for men) or with missing data on body weight or dietary intake at baseline and during follow-up. Individuals who were or became pregnant during follow-up were also excluded. For the change analyses, individuals who did not consecutively complete FFQs were excluded from the intervals with the missing data but re-entered the analyses when those data were available. The final analyses include 25,052 men in the HPFS, 49,932 women in the NHS, and 46,371 women in the NHS II (Supplemental Figure 1).
Multivariate linear models with an unstructured correlation matrix to account for repeated measures were used to examine the associations between changes in dietary fat intakes and changes in body weight. Models were adjusted for age, baseline BMI in each 4-y period, sleeping duration and hours of sitting, changes in dietary factors including percentage of energy from protein, intake of cereal fiber, fruits, and vegetables, and alcohol consumption, and changes in lifestyle covariates including smoking status (never, former, current: 1–14, 15–24, or ≥25 cigarettes/d) and hours of physical activity (quintiles), which were coded as categoric variables. Changes in percentage energy of fat intake and changes in dietary factors were censored at the 0.5th and 99.5th percentiles to minimize the influence of outliers (3). Missing data from categoric variables were assigned a missing indicator.
Total energy intake was not included as a covariable in the primary model (Model 2) because it could mediate (in causal pathways) the effects of dietary fat intake on weight change. In additional analysis (Model 3), we included energy intake as a covariate to examine whether the associations were independent from energy intake.
We included energy from protein as a continuous variable in the multivariable-adjusted model; therefore, an increase in energy from fat will largely represent an expense of energy from carbohydrates. The β coefficient represents the absolute weight change during a 4-y period when participants increased a certain percentage of energy from different types of fat as substitution for the same percentage of energy from carbohydrates (14). Different types of fat were mutually adjusted to examine their independent contributions to weight changes. In the multivariate models, animal fat and vegetable fat were mutually adjusted; SFA, trans-fat, MUFA, and PUFA were mutually adjusted; plant- and animal-source MUFAs were adjusted for trans-fat and PUFA; PUFA, linolenic acid, and n–3 PUFA were adjusted for SFA, trans-fat, and MUFA, respectively; and marine-origin n–3 and plant n–3 PUFAs (α-linolenic acid—ALA) were adjusted for SFA, trans-fat, and MUFA.
To evaluate the effect of substituting between specific types of fat, we treated fat intake as a continuous variable and calculated the difference in coefficients. We performed separate analyses on the association between intake of MUFA from animal sources compared with plant sources and weight change because the food sources contributing to MUFA intake shifted from animal sources such as red meat and dairy before 1994 to plant sources such as olive oil and nuts after 1994 (20). For MUFA analyses, we divided the time period into 2 categories, with 1994 as the cutoff. Findings across cohorts were pooled with a random-effect model meta-analysis. The Q statistic was used to assess heterogeneity. Spearman correlation was used to examine the associations between changes in dietary and food intake. SAS version 9.3, SAS Institute was used to analyze the data. Statistical significance was set at a 2-tailed P < 0.05.
Results
Baseline characteristics and weight change during follow-up
Table 1 presents the anthropometric and lifestyle characteristics of participants at baseline and their changes at 4-y intervals during the 20- to 24-y follow-up. The mean baseline BMI (kg/m2) was 24.8 for HPFS (mean ± SD age: 50.3 ± 7.58 y), 23.7 for NHS (52.1 ± 6.22 y), and 23.0 for NHS II (37.8 ± 4.01 y). Total fat intake was 32.1% for HPFS, 32.6% for NHS, and 31.5% for NHS II. The mean ± SD weight gain per 4 y across all follow-up periods was 0.78 ± 4.18 kg for men, 1.22 ± 4.61 kg for women in the NHS, and 1.87 ± 5.71 kg for women in the NHS II. The corresponding estimate of weight gain averaged across the 3 cohorts was 1.27 kg over 4 y, equivalent to a weight gain of 0.32 kg/y. The correlations between changes in specific foods and changes in intake of types of fat are presented in Supplemental Table 1. The time trend of fat intake is presented in Supplemental Figure 2. The consumption of animal fat decreased from 18.2% to 12.3% (1986–2010) in HPFS, from 18.1% to 12.1% (1986–2010) in NHS, and from 17.0% to 11.9% (1991–2011) in NHS II. The consumption of fat from plant sources increased from 13.2% to 19.5% in HPFS, from 14.1% to 19.2% in NHS, and from 13.9% to 20.3% in NHS II. More than half of the study population had an increase in consumption of PUFA, particularly n–3 PUFA over the last decade (Supplemental Figure 3). The associations between average 4-y changes in total and varying types of fat intake and changes in food items are presented in Supplemental Table 2. Weight, total energy intake, and physical activity during follow-up (1986–2010) are presented in Supplemental Figure 4.
TABLE 1.
Baseline characteristics and average 4-y lifestyle changes and changes in fat intake among 121,355 US women and men in 3 prospective cohorts1
| HPFS, n = 25,052 | NHS, n = 49,932 | NHS II, n = 46,371 | ||||
|---|---|---|---|---|---|---|
| Baseline | Changes within each 4-y period | Baseline | Changes within each 4-y period | Baseline | Changes within each 4-y period | |
| Age2, y | 50.3 ± 7.58 | 52.1 ± 6.22 | 37.8 ± 4.01 | |||
| BMI, kg/m2 | 24.8 ± 2.31 | 0.24 ± 1.29 | 23.7 ± 2.84 | 0.32 ± 1.26 | 23.0 ± 2.87 | 0.69 ± 2.09 |
| Weight, kg | 79.4 ± 9.30 | 0.79 ± 4.18 | 64.0 ± 8.85 | 1.22 ± 4.61 | 62.6 ± 9.12 | 1.87 ± 5.71 |
| Physical activity, MET-h/wk | 19.7 ± 26.0 | 4.25 ± 28.2 | 14.8 ± 20.6 | 1.62 ± 23.0 | 21.5 ± 27.6 | −0.06 ± 28.3 |
| Alcohol, g/d | 11.7 ± 15.2 | 0.44 ± 10.3 | 6.71 ± 10.9 | −0.11 ± 7.63 | 3.47 ± 6.61 | 0.71 ± 6.73 |
| Energy, kcal | 2016 ± 605 | −5.74 ± 506 | 1763 ± 509 | −0.49 ± 441 | 1740 ± 521 | −6.11 ± 483 |
| Carbohydrate, % energy | 47.0 ± 8.26 | 0.44 ± 7.59 | 48.3 ± 7.62 | 0.96 ± 7.82 | 49.8 ± 7.57 | −0.29 ± 8.62 |
| Protein, % energy | 18.3 ± 3.22 | −0.24 ± 3.13 | 18.5 ± 3.22 | −0.23 ± 3.28 | 19.3 ± 3.49 | −0.28 ± 3.53 |
| Total fat intake, % energy | 32.1 ± 6.13 | −0.14 ± 5.90 | 32.6 ± 5.59 | −0.62 ± 6.13 | 31.5 ± 5.64 | 0.26 ± 6.81 |
| Animal fat, % energy | 18.4 ± 5.45 | −1.20 ± 4.59 | 18.2 ± 4.97 | −1.22 ± 4.66 | 17.2 ± 4.60 | −0.94 ± 5.05 |
| Vegetable fat, % energy | 13.7 ± 4.46 | 1.06 ± 4.95 | 14.4 ± 4.45 | 0.60 ± 5.08 | 14.3 ± 4.07 | 1.20 ± 5.65 |
| SFA, % energy | 11.6 ± 2.64 | −0.29 ± 2.25 | 11.6 ± 2.55 | −0.55 ± 2.43 | 11.1 ± 2.42 | −0.20 ± 2.65 |
| trans-Fat, % energy | 1.28 ± 0.49 | −0.07 ± 0.53 | 1.38 ± 0.50 | −0.10 ± 0.49 | 1.63 ± 0.60 | −0.23 ± 0.43 |
| MUFA, % energy | 12.2 ± 2.66 | 0.01 ± 2.83 | 12.0 ± 2.39 | −0.05 ± 2.90 | 11.9 ± 2.45 | 0.13 ± 3.33 |
| Animal MUFA, % energy | 6.12 ± 2.00 | −1.20 ± 4.59 | 6.49 ± 1.89 | −1.22 ± 4.66 | 5.69 ± 1.71 | −0.94 ± 5.05 |
| Plant MUFA, % energy | 5.57 ± 2.09 | 1.06 ± 4.95 | 5.33 ± 1.90 | 0.60 ± 5.08 | 5.93 ± 1.89 | 1.20 ± 5.65 |
| PUFA, % energy | 5.94 ± 1.55 | 0.18 ± 1.70 | 6.15 ± 1.61 | 0.00 ± 1.82 | 5.70 ± 1.39 | 0.31 ± 1.94 |
| n–6 PUFA, % energy | 5.27 ± 1.58 | 0.11 ± 1.59 | 5.37 ± 1.56 | −0.01 ± 1.64 | 4.96 ± 1.36 | 0.24 ± 1.68 |
| n–3 PUFA, % energy | 0.63 ± 0.20 | 0.06 ± 0.28 | 0.67 ± 0.21 | 0.03 ± 0.28 | 0.61 ± 0.19 | 0.09 ± 0.33 |
| ALA, % energy | 0.50 ± 0.13 | 0.03 ± 0.20 | 0.57 ± 0.15 | 0.01 ± 0.22 | 0.51 ± 0.13 | 0.06 ± 0.28 |
| Marine n–3 PUFA, % energy | 0.13 ± 0.11 | 0.02 ± 0.15 | 0.12 ± 0.10 | 0.01 ± 0.13 | 0.10 ± 0.09 | 0.02 ± 0.14 |
1Baseline values are means ± SDs, and are standardized to the age distribution of the study population. Data are based on 24 y of follow-up (1986–2010) in the HPFS, 24 y of follow-up (1986–2010) in the NHS, and 16 y of follow-up (1991–2007) in the NHS II. ALA, α-linolenic acid; HPFS, Health Professionals Follow-up Study; MET-h, metabolic task equivalent-hours; NHS, Nurses’ Health Study.
2Values are not age-adjusted.
Associations between total fat, animal fat, vegetable fat, and weight changes
Table 2 presents the age- and multivariable-adjusted associations between changes in intake of dietary fat and corresponding weight changes from pooled analyses. An increase in energy from total fat by 5% at the expense of carbohydrate was associated with 0.19 kg (95% CI: 0.15, 0.22 kg) of weight gain during a 4-y period (Figure 1). Similarly, 5% greater energy from animal fat was associated with 0.57 kg (95% CI: 0.47, 0.67 kg) of weight gain per 4 y (Figure 1). In contrast, an increase in energy by 5% of fat from vegetable sources was not associated with weight gain (−0.04 kg; 95% CI: −0.18, 0.10 kg). Age- and multivariable-adjusted associations between changes in intake of total fat, animal fat, or vegetable fat and the corresponding weight changes from each cohort are presented in Supplemental Table 3.
TABLE 2.
Association between weight changes and changes in intake of total fat, animal fat, vegetable fat, SFAs, trans-fat, MUFAs, and PUFAs over 4-y periods among US women and men in 3 prospective cohorts1
| Changes in dietary fat intake (percentage of total energy) in quintiles,2 kg | ||||||
|---|---|---|---|---|---|---|
| Pooled | Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | P for trend |
| Total fat | ||||||
| Median, % E | −7.87 | −3.16 | −0.23 | 2.72 | 7.51 | |
| Model 13 | 0.85 (0.21, 1.49) | 1.21 (0.67, 1.76) | 1.36 (0.77, 1.96) | 1.42 (0.83, 2.01) | 1.44 (0.87, 2.01) | <0.0001 |
| Model 24 | 0.87 (0.23, 1.51) | 1.23 (0.67, 1.79) | 1.38 (0.77, 1.99) | 1.43 (0.84, 2.03) | 1.45 (0.87, 2.02) | <0.0001 |
| Model 35 | 0.96 (0.29, 1.62) | 1.27 (0.70, 1.84) | 1.38 (0.77, 1.99) | 1.40 (0.81, 1.99) | 1.36 (0.81, 1.91) | <0.0001 |
| Animal source fat6 | ||||||
| Median, % E | −6.99 | −3.29 | −1.04 | 1.15 | 4.68 | |
| Model 13 | 0.86 (0.29, 1.43) | 1.12 (0.58, 1.67) | 1.25 (0.70, 1.81) | 1.41 (0.80, 2.01) | 1.65 (0.96, 2.33) | <0.0001 |
| Model 24 | 0.61 (0.10, 1.13) | 1.02 (0.48, 1.55) | 1.27 (0.69, 1.86) | 1.53 (0.89, 2.17) | 1.93 (1.20, 2.66) | <0.0001 |
| Model 35 | 0.70 (0.16, 1.23) | 1.27 (0.68, 1.86) | 1.27 (0.68, 1.86) | 1.50 (0.87, 2.13) | 1.85 (1.15, 2.56) | <0.0001 |
| Vegetable source fat6 | ||||||
| Median, % E | −5.31 | −1.61 | 0.78 | 3.22 | 7.37 | |
| Model 13 | 1.16 (0.46, 1.87) | 1.29 (0.68, 1.90) | 1.32 (0.77, 1.87) | 1.31 (0.73, 1.90) | 1.20 (0.68, 1.73) | 0.79 |
| Model 24 | 1.26 (0.51, 2.00) | 1.32 (0.69, 1.96) | 1.34 (0.78, 1.90) | 1.29 (0.72, 1.87) | 1.15 (0.68, 1.63) | 0.57 |
| Model 35 | 1.29 (0.53, 2.04) | 1.33 (0.69, 1.97) | 1.34 (0.78, 1.91) | 1.29 (0.71, 1.86) | 1.11 (0.64, 1.57) | 0.29 |
| SFAs7 | ||||||
| Median, % E | −3.40 | −1.50 | −0.33 | 0.81 | 2.63 | |
| Model 13 | 0.73 (0.12, 1.35) | 1.18 (0.66, 1.70) | 1.31 (0.70, 1.91) | 1.44 (0.86, 2.01) | 1.63 (0.99, 2.27) | <0.0001 |
| Model 24 | 0.84 (0.22, 1.47) | 1.21 (0.65, 1.77) | 1.31 (0.70, 1.93) | 1.43 (0.86, 2.00) | 1.59 (0.99, 2.18) | <0.0001 |
| Model 35 | 0.90 (0.26, 1.55) | 1.23 (0.67, 1.80) | 1.31 (0.69, 1.93) | 1.41 (0.84, 1.97) | 1.53 (0.96, 2.10) | <0.0001 |
| trans-Fat7 | ||||||
| Median, % E | −1.15 | −0.42 | −0.18 | 0.05 | 0.41 | |
| Model 13 | 0.75 (0.18, 1.33) | 0.98 (0.42, 1.54) | 1.21 (0.69, 1.72) | 1.48 (0.97, 1.99) | 1.91 (1.17, 2.66) | <0.0001 |
| Model 24 | 0.95 (0.38, 1.53) | 1.08 (0.49, 1.67) | 1.21 (0.67, 1.76) | 1.40 (0.88, 1.92) | 1.73 (1.01, 2.45) | <0.0001 |
| Model 35 | 0.95 (0.37, 1.53) | 1.08 (0.48, 1.68) | 1.21 (0.66, 1.76) | 1.40 (0.89, 1.92) | 1.73 (1.01, 2.45) | <0.0001 |
| MUFAs7 | ||||||
| Median, % E | −3.53 | −1.40 | −0.06 | 1.31 | 3.67 | |
| Model 13 | 0.85 (0.20, 1.50) | 1.23 (0.66, 1.79) | 1.34 (0.74, 1.94) | 1.46 (0.86, 2.06) | 1.42 (0.89, 1.95) | <0.0001 |
| Model 24 | 1.15 (0.55, 1.74) | 1.35 (0.78, 1.92) | 1.32 (0.70, 1.95) | 1.33 (0.71, 1.95) | 1.23 (0.67, 1.79) | 0.20 |
| Model 35 | 1.20 (0.60, 1.79) | 1.37 (0.80, 1.94) | 1.32 (0.70, 1.95) | 1.31 (0.69, 1.93) | 1.18 (0.62, 1.74) | 0.48 |
| PUFAs7 | ||||||
| Median, % E | −1.99 | −0.68 | 0.13 | 0.93 | 2.32 | |
| Model 13 | 1.28 (0.60, 1.96) | 1.30 (0.64, 1.95) | 1.30 (0.73, 1.88) | 1.32 (0.79, 1.84) | 1.10 (0.59, 1.61) | <0.0001 |
| Model 24 | 1.48 (0.77, 2.18) | 1.37 (0.70, 2.03) | 1.29 (0.70, 1.88) | 1.25 (0.72, 1.78) | 0.99 (0.51, 1.47) | <0.0001 |
| Model 35 | 1.46 (0.76, 2.16) | 1.36 (0.70, 2.02) | 1.30 (0.71, 1.88) | 1.26 (0.73, 1.79) | 1.00 (0.52, 1.49) | <0.0001 |
1Values are mean changes (95% CIs) unless otherwise indicated. Data are based on 24 y of follow-up (1986–2010) in the Health Professionals Follow-up Study, 24 y of follow-up in the NHS (1986–2010), and 20 y of follow-up in the NHS II (1991–2011). E, energy; NHS, Nurses’ Health Study.
2For total fat animal fat, vegetable fat, SFAs, trans-fat, MUFAs, and PUFAs, Quintiles 1–5 are nutrient-specific quintiles with an n of 79,761–79,763 for each quintile.
3Model 1: Age-adjusted.
4Model 2: Multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, and alcohol use.
5Model 3: Multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, alcohol use, and energy intake.
6Animal fat and vegetable fat were mutually adjusted in Models 2 and 3.
7SFA, trans-fat, MUFA, and PUFA were mutually adjusted in Models 2 and 3.
FIGURE 1.
Weight change (kilograms) per 4-y period associated with the increase in percentage of energy from total fat, animal-source fat, vegetable-source fat, SFAs, trans-fat, MUFAs, and PUFAs from a meta-analysis of 3 prospective cohorts of US women and men. HPFS, n = 25,052; NHS, n = 49,932; NHS II, n = 46,371; pooled, n = 121,355. Values are mean changes (95% CIs). Number of observations = 398,809. Multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, and alcohol use (Model 2). Animal fat and vegetable fat were mutually adjusted in Model 2. SFA, trans-fat, MUFA, and PUFA were mutually adjusted in Model 2. E, energy; HPFS, Health Professionals Follow-up Study; NHS, Nurses’ Health Study.
Associations between SFA, trans-fat, MUFA, PUFA, and weight changes
Increased intakes of SFA and trans-fat were associated with greater weight gain in men and women (Table 2). The pooled analysis illustrated that a 5% increase in energy from SFA was associated with 0.61 kg (95% CI: 0.54, 0.68 kg) of weight gain per 4-y interval (Figure 1) and the association did not attenuate after the model was adjusted for animal protein (0.64 kg; 95% CI: 0.58, 0.72 kg). The intake of trans-fat decreased over the years (Table 1). Thus a decrease in energy from trans-fat by 1% was associated with less weight gain (0.69 kg; 95% CI: 0.56, 0.84 kg) per 4-y interval (Figure 1). In contrast, greater energy intake from PUFA was inversely associated with weight gain (−0.55 kg, 95% CI: −0.81, −0.29 kg) within each 4-y period. Increased percentage of energy from MUFAs was associated with a small weight gain of 0.05 kg (95% CI: −0.02, 0.11 kg). The changes in fat intake reflect the changes in food consumption over time. Consumption of red and processed meat across the 3 cohorts was correlated with higher intake of SFA and trans-fat. Consumption of nuts and olive oil was associated with higher intake of plant source MUFA. Consumption of nuts was also positively correlated with intake of PUFA (Supplemental Table 1). Age- and multivariable-adjusted associations between changes in intake of SFA, trans-fat, MUFA, and PUFA, and the corresponding weight changes from each cohort are presented in Supplemental Table 3.
Before 1994, an increase of energy from MUFA of 5% was associated with greater weight gain by 0.33 kg (95% CI: 0.00, 0.67 kg), whereas after 1994, an increase by the same proportion of energy from MUFA had an inverse association with weight gain (−0.15 kg; 95% CI: −0.31, 0.00 kg, P < 0.0001 for interaction, Supplemental Figure 5).
Furthermore, we examined associations between MUFA from plant or animal sources and concurrent weight changes. Increased intake of MUFA from animal sources by 1% was associated with greater weight gain of 0.29 kg (95% CI: 0.25, 0.33 kg) per 4-y interval. There was no association between intake of MUFA from plant sources and weight gain (Figure 2).
FIGURE 2.
Weight change (kilograms) per 4-y period associated with the increase in percentage of energy from animal-source MUFA, plant-source MUFA, n–3 PUFAs, and n–6 PUFAs from a meta-analysis of 3 prospective cohorts of US women and men. HPFS, n = 25,052; NHS, n = 49,932; NHS II, n = 46,371; pooled, n = 121,355. Values are mean changes (95% CIs). Number of observations = 398,809. Multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, and alcohol use (Model 2). Plant-source MUFA and animal-source MUFA were mutually adjusted for trans-fat and PUFA in Model 2. n–6 PUFA was adjusted for SFA, trans-fat, and MUFA in Model 2. Linoleic acid was adjusted for SFA, trans-fat, and MUFA in Model 2. n–3 PUFA was adjusted for SFA, trans-fat, and MUFA in Model 2. Marine origin n–3 and plant n–3 PUFA (ALA) were mutually adjusted for SFA, trans-fat, and MUFA in Model 2. ALA, α-linolenic acid; HPFS, Health Professionals Follow-up Study; NHS, Nurses’ Health Study.
Associations between n–3 and n–6 PUFAs and weight changes
The age- and multivariable-adjusted associations between weight changes and changes in intake of n–6 or n–3 PUFAs are presented in Table 3. An increase in total energy from n–6 PUFA by 5% at the expense of energy from carbohydrate was inversely associated with weight gain (−0.27 kg, 95% CI: −0.40, −0.14 kg) within each 4-y period (Figure 2). The inverse association between n–6 PUFA intake and weight gain was mainly driven by linoleic acid. A 5% higher energy intake from linoleic acid was associated with weight loss over 4 y (−0.54 kg; 95% CI: −0.80, −0.28 kg). For marine origin n–3 PUFA, a 0.3% increment in energy was inversely associated with weight gain (−0.60 kg; 95% CI: −0.70, −0.50 kg). Each 1% increase in ALA was associated with less weight gain (−0.93 kg; 95% CI: −1.04, −0.82 kg) at a 4-y interval. Age- and multivariable-adjusted associations between changes in intake of n–3 and n–6 PUFAs and the corresponding weight changes from each cohort are presented in Supplemental Table 4.
TABLE 3.
Association between weight changes and changes in intake of plant source MUFA, animal source MUFA, n–6 PUFAs, and n–3 PUFAs over 4-y periods among US women and men in 3 prospective cohorts1
| Changes in dietary fat intake (percentage of total energy) in quintiles,2 kg | ||||||
|---|---|---|---|---|---|---|
| Pooled | Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | P for trend |
| Animal source MUFA6 | ||||||
| Median, % E | −2.55 | −1.17 | −0.35 | 0.45 | 1.79 | |
| Model 13 | 0.84 (0.29, 1.39) | 1.08 (0.54, 1.63) | 1.22 (0.65, 1.79) | 1.42 (0.80, 2.03) | 1.74 (1.07, 2.40) | <0.0001 |
| Model 24 | 0.68 (0.19, 1.17) | 1.02 (0.48, 1.57) | 1.24 (0.64, 1.84) | 1.51 (0.88, 2.14) | 1.74 (1.07, 2.40) | <0.0001 |
| Model 35 | 0.75 (0.24, 1.25) | 1.05 (0.50, 1.59) | 1.24 (0.64, 1.85) | 1.49 (0.86, 2.12) | 1.86 (1.19, 2.52) | <0.0001 |
| Plant source MUFA6 | ||||||
| Median, % E | −2.68 | −0.83 | 0.32 | 1.52 | 3.74 | |
| Model 13 | 1.15 (0.45, 1.85) | 1.30 (0.66, 1.95) | 1.29 (0.68, 1.89) | 1.36 (0.81, 1.91) | 1.20 (0.73, 1.67) | 0.82 |
| Model 24 | 1.32 (0.61, 2.03) | 1.36 (0.72, 2.01) | 1.28 (0.69, 1.88) | 1.29 (0.73, 1.84) | 1.13 (0.67, 1.60) | 0.23 |
| Model 35 | 1.35 (0.64, 2.06) | 1.38 (0.73, 2.03) | 1.29 (0.69, 1.89) | 1.28 (0.72, 1.83) | 1.09 (0.63, 1.55) | 0.10 |
| Total n–6 PUFAs7 | ||||||
| Median, % E | −1.81 | −0.62 | 0.09 | 0.82 | 2.05 | |
| Model 13 | 1.15 (0.55, 1.76) | 1.24 (0.61, 1.87) | 1.35 (0.77, 1.93) | 1.34 (0.75, 1.93) | 1.20 (0.66, 1.74) | 0.02 |
| Model 24 | 1.33 (0.71, 1.95) | 1.31 (0.66, 1.95) | 1.34 (0.76, 1.93) | 1.29 (0.70, 1.88) | 1.11 (0.59, 1.63) | <0.0001 |
| Model 35 | 1.32 (0.70, 1.93) | 1.30 (0.66, 1.95) | 1.35 (0.76, 1.93) | 1.30 (0.71, 1.90) | 1.11 (0.59, 1.63) | <0.0001 |
| Linoleic acid8 | ||||||
| Median, % E | −1.83 | −0.64 | 0.08 | 0.80 | 2.02 | |
| Model 13 | 1.22 (0.57, 1.88) | 1.28 (0.64, 1.91) | 1.32 (0.74, 1.90) | 1.32 (0.77, 1.87) | 1.15 (0.63, 1.68) | 0.0005 |
| Model 24 | 1.22 (0.57, 1.88) | 1.28 (0.64, 1.91) | 1.32 (0.74, 1.90) | 1.32 (0.77, 1.87) | 1.15 (0.63, 1.68) | <0.0001 |
| Model 35 | 1.43 (0.74, 2.11) | 1.36 (0.71, 2.00) | 1.31 (0.72, 1.90) | 1.26 (0.71, 1.81) | 1.03 (0.53, 1.52) | <0.0001 |
| Total n–3 PUFAs9 | ||||||
| Median, % E | −0.24 | −0.07 | 0.03 | 0.14 | 0.39 | |
| Model 13 | 1.46 (0.79, 2.13) | 1.39 (0.78, 2.00) | 1.31 (0.69, 1.92) | 1.25 (0.68, 1.81) | 0.90 (0.42, 1.37) | <0.0001 |
| Model 24 | 1.45 (0.78, 2.12) | 1.38 (0.76, 1.99) | 1.32 (0.70, 1.94) | 1.27 (0.71, 1.82) | 0.97 (0.48, 1.46) | <0.0001 |
| Model 35 | 1.41 (0.75, 2.08) | 1.36 (0.75, 1.97) | 1.32 (0.70, 1.94) | 1.29 (0.72, 1.85) | 1.00 (0.50, 1.50) | <0.0001 |
| Marine n–3 PUFA10 | ||||||
| Median, % E | −0.10 | −0.03 | 0.00 | 0.03 | 0.15 | |
| Model 13 | 1.66 (1.04, 2.28) | 1.44 (0.81, 2.06) | 1.26 (0.65, 1.87) | 1.12 (0.52, 1.72) | 0.83 (0.33, 1.33) | <0.0001 |
| Model 24 | 1.51 (0.90, 2.11) | 1.37 (0.77, 1.97) | 1.27 (0.67, 1.87) | 1.22 (0.58, 1.86) | 1.02 (0.49, 1.54) | <0.0001 |
| Model 35 | 1.48 (0.88, 2.09) | 1.35 (0.76, 1.94) | 1.27 (0.67, 1.88) | 1.24 (0.60, 1.88) | 1.04 (0.51, 1.57) | <0.0001 |
| ALA10 | ||||||
| Median, % E | −0.19 | −0.06 | 0.02 | 0.10 | 0.27 | |
| Model 13 | 1.28 (0.61, 1.96) | 1.33 (0.68, 1.97) | 1.34 (0.74, 1.94) | 1.29 (0.75, 1.84) | 1.06 (0.58, 1.53) | <0.0001 |
| Model 24 | 1.40 (0.72, 2.08) | 1.38 (0.74, 2.02) | 1.32 (0.71, 1.92) | 1.25 (0.72, 1.79) | 1.04 (0.54, 1.53) | <0.0001 |
| Model 35 | 1.37 (0.69, 2.04) | 1.36 (0.72, 2.00) | 1.32 (0.72, 1.93) | 1.27 (0.73, 1.82) | 1.06 (0.56, 1.56) | <0.0001 |
1Values are mean changes (95% CIs) unless otherwise indicated. Data are based on 24 y of follow-up (1986–2010) in the Health Professionals Follow-up Study, 24 y of follow-up in the Nurses’ Health Study (NHS) (1986–2010), and 20 y of follow-up in the NHS II (1991–2011). ALA, α-linolenic acid; E, energy.
2For plant-source MUFA, animal-source MUFA, total n–6 PUFAs, total n–3 PUFAs, marine n–3 PUFA, and ALA, Quintiles 1–5 are nutrient-specific quintiles with an n of 79,761–79,763 for each quintile.
3Model 1: age-adjusted.
4Model 2: multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, and alcohol use.
5Model 3: multivariable model adjusted for age, baseline BMI at the beginning of each 4-y period, baseline hours of sitting, sleep duration, changes in intake of protein (percentage of total energy), cereal fiber, fruit, and vegetables, smoking status (no, current, or past), physical activity, alcohol use, and energy intake.
6Plant source MUFA and animal source MUFA were mutually adjusted for trans-fat and PUFA in Models 2 and 3.
7n–6 PUFA was adjusted for SFA, trans-fat, and MUFA in Models 2 and 3.
8Linoleic acid was adjusted for SFA, trans-fat, and MUFA in Models 2 and 3.
9n–3 PUFA was adjusted for SFA, trans-fat, and MUFA in Models 2 and 3.
10Marine origin n–3 and plant n–3 PUFA (ALA) were mutually adjusted for SFA, trans-fat, and MUFA in Models 2 and 3.
Discussion
In 3 large prospective cohorts of US men and women, we found that types of dietary fats have diverging effects on long-term weight change. In particular, increasing consumption of n–6 and n–3 PUFAs as well as plant-based MUFA at the expense of carbohydrate was associated with less weight gain, whereas increasing intake of SFA and trans-fat was associated with greater weight gain.
Although the overall diet quality of the US population has improved in recent decades, it remains far from optimal (21). In the present study, more than half of the study population increased their consumption of unsaturated fat, especially from plant sources, over the last decade (Supplemental Figure 3). Yet, we also observed that 47% of study participants increased their consumption of fat from animal sources. The average consumption of SFA and fat from animal sources at last analyzed follow-up (2010 for HPFS and NHS; 2011 for NHS II) was 10% and 12%, respectively (Supplemental Figure 2), for all 3 cohorts, indicating substantial room for improvement in the quality of dietary fat. A realistic modification in diet by replacing 5% of SFA with 2.5% PUFA and 2.5% of MUFA can mitigate weight gain from 0.61 kg to 0.27 kg (equivalent to a 56% reduction in weight gain) over 4 y (data not shown). Our findings highlight the important roles of different types and food sources of dietary fats in the prevention of weight gain in US adults. These changes would also improve blood lipids and reduce the risk of CVD (22) and total mortality independent of the weight change (20).
The average weight gain in these 3 cohorts was 0.32 kg/y, which was consistent with previous findings (3, 23). Weight gain in nonobese populations during adulthood is gradual. Even a modest weight gain of 0.4 kg/y is significantly associated with increased risk of CVD, cancer, diabetes, and metabolic syndrome (2, 24). Based on our findings, changes in the quality or types of fat can attenuate age-related weight gain. At the population level, replacing SFA by PUFAs or MUFA from olive oil or other plant sources could provide a long-term beneficial effect in weight management. Results from the present study suggest that modifying the types of fat but not necessarily the total amount of fat may mitigate weight gain during adulthood, and thus reduce obesity rates in the population.
Different types of fat possess different physiologic and mechanistic effects. Diets rich in SFA decrease total lipid oxidation and energy expenditure (25, 26), and decrease diet-induced thermogenesis, resulting in body fat accumulation when compared with a diet high in MUFAs (27). Kien et al. (25) reported that in healthy nonobese participants, consumption of diets high in SFAs (16.8% palmitic acid, 16.4% oleic acid) for 28 d reduced postprandial lipid oxidation and decreased daily energy expenditure when compared with diets high in MUFAs (31.4% oleic acid, 1.7% palmitic acid). In a 4-wk intervention study, although total energy from fat intake remained the same (40% of total fat), participants experienced a significant weight loss (−1.6 kg) when using high-MUFA diets (22% MUFA, 11% SFA) to replace high-SFA diets (24% SFA, 13% MUFA) (26). Different types of fat also have divergent effects on ectopic fat storage. Excessive intake of SFAs and PUFAs resulted in similar weight gain; however, participants who received SFAs had an increase in hepatic and visceral fat storage, whereas participants receiving PUFAs had greater lean tissue (28).
The association we observed between MUFAs and weight changes partially reflects the shift in participants’ food choices over time. After 1994, the major food sources of MUFA shifted from animal foods to plant foods. That greater energy intake from MUFA was associated with less weight gain may partially reflect this change because increased intakes of MUFA from plant sources are not associated with weight gain as reported in the Prevención con Dieta Mediterránea (PREDIMED) study, which found that long-term consumption of food items with high MUFA content (extra-virgin olive oil or mixed nuts) reduced body weight and waist circumference and provided cardiovascular benefits (29, 30). The underlying mechanism includes diets high in MUFAs increasing fat oxidation and en-ergy expenditure, thus leading to potential weight loss (31, 32).
A meta-analysis including 934 participants reported that consumption of EPA and DHA were associated with greater weight loss. Participants in the intervention groups with daily supplementation of EPA and DHA ranging from 157 to 3360 mg showed more weight loss (−0.59 kg, 95% CI: −0.96, −0.21 kg) compared with the control groups with no n–3 PUFA supplementation (33). In our analysis, intake of EPA and DHA was inversely associated with weight gain. The 0.3% energy from EPA and DHA is equivalent to 0.67 g of n–3 PUFA intake based on a 2000-kcal diet. This level of intake corresponds with the recommendation made by the American Heart Association to consume fish with ≥500 mg of EPA + DHA per 85 g for the general adult population (34). Several mechanisms have been proposed to explain the association between intakes of long chain n–3 PUFAs and weight loss, including that long chain n–3 PUFAs increase lipolysis and reduce lipogenesis (35), increase β-oxidation (36), and reduce plasma leptin concentrations in nonobese individuals (37). There is growing interest in the metabolism and health benefits of ALA, a plant-based n–3 PUFA. ALA has the highest rate of β-oxidation among unsaturated fatty acids, which may partially explain the potential effect in weight loss (38).
Several potential limitations should be considered. First, although we adjusted for many dietary and lifestyle factors, residual confounding cannot be completely ruled out. It is possible that people who consumed diets high in SFA and trans-fat were more likely to follow a “Western diet,” and consumed more sugar-sweetened beverages and fast foods. However, adjustment for correlated dietary factors including sweet sugary beverages and consumption of pizza did not appreciably alter the results. Second, measurement error in self-report of diet is inevitable. However, our previous validation studies in a subsample of participants showed a reasonable degree of correlation between our FFQ and multiple dietary records for fat intake. In addition, several dietary fatty acids, including PUFA and trans-fat assessed by the FFQs, were also correlated with biomarkers of these fatty acids. Third, lifestyle changes were self-selected, and thus a person's weight change may correlate with other changes in lifestyle, including changes in physical activity or total calorie intake. The bias of reverse causation cannot be completely eliminated from analysis even in the present change compared with change analysis. Finally, our cohorts largely consisted of white health professionals with relatively high socioeconomic status, and thus the results may not be generalizable to other groups.
There are several strengths to our study. We comprehensively addressed the relation between changes in different types of dietary fat and weight change in 3 large cohorts with extended follow-up and repeated measurements of diet via a validated questionnaire, including 130 items over ≥20 y in >100,000 adults. The large sample size including both sexes and high follow-up rates provided us ample power to detect relatively small changes in weight within each 4-y period in relation to changes in fat intake. Weight changes occur gradually at the population level. Therefore, our 4-y assessment period is consistent with the time course of weight change in response to a change in diet (39). The repeated measures of diet and weight reduced potential biases due to reverse causation (19), and simulated a dietary intervention trial on weight change.
In conclusion, our results suggest that different types of fat have divergent associations with long-term weight changes. Not all fats are associated with weight gain. In particular, increasing consumption of n–6 and n–3 PUFAs as well as plant-based MUFA at the expense of carbohydrates was associated with less weight gain. Modifying types of dietary fat in the diet without lowering the total fat may help mitigate the gradual weight gain that is common during adulthood. Our findings provide further evidence to support food-approach recommendations, such as increasing consumption of unsaturated fats from plant oils, nuts, seeds, and seafood, at the expense of saturated fat from meats and dairy sourcess, and trans-fat from partially hydrogenated oils, to contribute to the prevention of obesity and reduction of risk of chronic diseases.
Supplementary Material
Acknowledgments
The authors’ responsibilities were as follows—XL and FBH: contributed to the design of the study and drafted the manuscript; XL, YL, and FBH: contributed to the analysis and interpretation of the data; XL and YL: developed the statistical analysis plan; XL: prepared the figures and tables; XL, YL, DKT, DDW, JEM, WCW, and FBH: revised the manuscript for critically important intellectual content; and all authors: read and approved the final manuscript. XL had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. FBH takes full responsibility for the work as a whole, the study design, access to data, and the decision to submit and publish the manuscript.
Notes
Supported by NIH grants UM1 CA186107, UM1 CA176726, P01 CA87969, UM1 CA167552, T32DK007703, HL34594, HL60712, and DK46200.
Author disclosures: FBH reports grants from California Walnut Commission and personal fees from Metagenics, outside the submitted work. XL, YL, DKT, DDW, JEM, and WCW, no conflicts of interest.
Supplemental Tables 1–4 and Supplemental Figures 1–5 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/jn/.
The funders of this study had no role in its design or conduct; in the collection, management, analysis, or interpretation of data; or in the preparation, review, or approval of the manuscript.
Abbreviations used: ALA, α-linolenic acid; CVD, cardiovascular disease; HPFS, Health Professionals Follow-Up Study; NHS, Nurse's Health Study.
References
- 1. WHO. Global health risks: mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization; 2009. [Google Scholar]
- 2. Zheng Y, Manson JE, Yuan C, Liang MH, Grodstein F, Stampfer MJ, Willett WC, Hu FB. Associations of weight gain from early to middle adulthood with major health outcomes later in life. JAMA 2017;318(3):255–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med 2011;364(25):2392–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ludwig DS. Lowering the bar on the low-fat diet. JAMA 2016;316(20):2087–8. [DOI] [PubMed] [Google Scholar]
- 5. Howard BV, Van Horn L, Hsia J, Manson JE, Stefanick ML, Wassertheil-Smoller S, Kuller LH, LaCroix AZ, Langer RD, Lasser NL et al. Low-fat dietary pattern and risk of cardiovascular disease: the Women's Health Initiative Randomized Controlled Dietary Modification Trial. JAMA 2006;295(6):655–66. [DOI] [PubMed] [Google Scholar]
- 6. The Look AHEAD Research Group. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N Engl J Med 2013;369(2):145–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Field AE, Willett WC, Lissner L, Colditz GA. Dietary fat and weight gain among women in the Nurses’ Health Study. Obesity (Silver Spring) 2007;15(4):967–76. [DOI] [PubMed] [Google Scholar]
- 8. Tobias DK, Chen M, Manson JE, Ludwig DS, Willett W, Hu FB. Effect of low-fat diet interventions versus other diet interventions on long-term weight change in adults: a systematic review and meta-analysis. Lancet Diabetes Endocrinol 2015;3(12):968–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rimm EB, Stampfer MJ, Colditz GA, Giovannucci E, Willett WC. Effectiveness of various mailing strategies among nonrespondents in a prospective cohort study. Am J Epidemiol 1990;131(6):1068–71. [DOI] [PubMed] [Google Scholar]
- 10. Bao Y, Bertoia ML, Lenart EB, Stampfer MJ, Willett WC, Speizer FE, Chavarro JE. Origin, methods, and evolution of the three Nurses’ Health Studies. Am J Public Health 2016;106(9):1573–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Rimm EB, Giovannucci EL, Willett WC, Colditz GA, Ascherio A, Rosner B, Stampfer MJ. Prospective study of alcohol consumption and risk of coronary disease in men. Lancet 1991;338(8765):464–8. [DOI] [PubMed] [Google Scholar]
- 12. Feskanich D, Marshall J, Rimm EB, Litin LB, Willett WC. Simulated validation of a brief food frequency questionnaire. Ann Epidemiol 1994;4(3):181–7. [DOI] [PubMed] [Google Scholar]
- 13. Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, Hennekens CH, Speizer FE. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 1985;122(1):51–65. [DOI] [PubMed] [Google Scholar]
- 14. Willett WC. Nutritional epidemiology. 3rd ed New York, NY: Oxford University Press; 2012. [Google Scholar]
- 15. Sun Q, Ma J, Campos H, Hankinson SE, Hu FB. Comparison between plasma and erythrocyte fatty acid content as biomarkers of fatty acid intake in US women. Am J Clin Nutr 2007;86(1):74–81. [DOI] [PubMed] [Google Scholar]
- 16. Yuan C, Spiegelman D, Rimm EB, Rosner BA, Stampfer MJ, Barnett JB, Chavarro JE, Subar AF, Sampson LK, Willett WC. Validity of a dietary questionnaire assessed by comparison with multiple weighed dietary records or 24-hour recalls. Am J Epidemiol 2017;185(7):1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Hunter DJ, Rimm EB, Sacks FM, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Comparison of measures of fatty acid intake by subcutaneous fat aspirate, food frequency questionnaire, and diet records in a free-living population of US men. Am J Epidemiol 1992;135(4):418–27. [DOI] [PubMed] [Google Scholar]
- 18. Willett W, Stampfer MJ, Bain C, Lipnick R, Speizer FE, Rosner B, Cramer D, Hennekens CH. Cigarette smoking, relative weight, and menopause. Am J Epidemiol 1983;117(6):651–8. [DOI] [PubMed] [Google Scholar]
- 19. Smith JD, Hou T, Hu FB, Rimm EB, Spiegelman D, Willett WC, Mozaffarian D. A comparison of different methods for evaluating diet, physical activity, and long-term weight gain in 3 prospective cohort studies. J Nutr 2015;145(11):2527–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Wang DD, Li Y, Chiuve SE, Stampfer MJ, Manson JE, Rimm EB, Willett WC, Hu FB. Association of specific dietary fats with total and cause-specific mortality. JAMA Intern Med 2016;176(8):1134–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wang DD, Leung CW, Li Y, Ding EL, Chiuve SE, Hu FB, Willett WC. Trends in dietary quality among adults in the United States, 1999 through 2010. JAMA Intern Med 2014;174(10):1587–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sacks FM, Lichtenstein AH, Wu JHY, Appel LJ, Creager MA, Kris-Etherton PM, Miller M, Rimm EB, Rudel LL, Robinson JG et al. Dietary fats and cardiovascular disease: a presidential advisory from the American Heart Association. Circulation 2017;136(3):e1–e23. [DOI] [PubMed] [Google Scholar]
- 23. Martinez-Gonzalez MA, Bes-Rastrollo M. Nut consumption, weight gain and obesity: epidemiological evidence. Nutr Metab Cardiovasc Dis 2011;21(Suppl 1):S40–5. [DOI] [PubMed] [Google Scholar]
- 24. Suzuki A, Akamatsu R. Long-term weight gain is related to risk of metabolic syndrome even in the non-obese. Diabetes Metab Syndr 2014;8(3):177–83. [DOI] [PubMed] [Google Scholar]
- 25. Kien CL, Bunn JY, Ugrasbul F. Increasing dietary palmitic acid decreases fat oxidation and daily energy expenditure. Am J Clin Nutr 2005;82(2):320–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Piers LS, Walker KZ, Stoney RM, Soares MJ, O'Dea K. Substitution of saturated with monounsaturated fat in a 4-week diet affects body weight and composition of overweight and obese men. Br J Nutr 2003;90(3):717–27. [DOI] [PubMed] [Google Scholar]
- 27. Takeuchi H, Matsuo T, Tokuyama K, Shimomura Y, Suzuki M. Diet-induced thermogenesis is lower in rats fed a lard diet than in those fed a high oleic acid safflower oil diet, a safflower oil diet or a linseed oil diet. J Nutr 1995;125(4):920–5. [DOI] [PubMed] [Google Scholar]
- 28. Rosqvist F, Iggman D, Kullberg J, Cedernaes J, Johansson HE, Larsson A, Johansson L, Ahlstrom H, Arner P, Dahlman I et al. Overfeeding polyunsaturated and saturated fat causes distinct effects on liver and visceral fat accumulation in humans. Diabetes 2014;63(7):2356–68. [DOI] [PubMed] [Google Scholar]
- 29. Damasceno NR, Sala-Vila A, Cofan M, Perez-Heras AM, Fito M, Ruiz-Gutierrez V, Martinez-Gonzalez MA, Corella D, Aros F, Estruch R et al. Mediterranean diet supplemented with nuts reduces waist circumference and shifts lipoprotein subfractions to a less atherogenic pattern in subjects at high cardiovascular risk. Atherosclerosis 2013;230(2):347–53. [DOI] [PubMed] [Google Scholar]
- 30. Estruch R, Martinez-Gonzalez MA, Corella D, Salas-Salvado J, Fito M, Chiva-Blanch G, Fiol M, Gomez-Gracia E, Aros F, Lapetra J et al. Effect of a high-fat Mediterranean diet on bodyweight and waist circumference: a prespecified secondary outcomes analysis of the PREDIMED randomised controlled trial. Lancet Diabetes Endocrinol 2016;4(8):666–76. [DOI] [PubMed] [Google Scholar]
- 31. Kien CL, Bunn JY. Gender alters the effects of palmitate and oleate on fat oxidation and energy expenditure. Obesity (Silver Spring) 2008;16(1):29–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Jones PJ, Jew S, AbuMweis S. The effect of dietary oleic, linoleic, and linolenic acids on fat oxidation and energy expenditure in healthy men. Metabolism 2008;57(9):1198–203. [DOI] [PubMed] [Google Scholar]
- 33. Bender N, Portmann M, Heg Z, Hofmann K, Zwahlen M, Egger M. Fish or n3-PUFA intake and body composition: a systematic review and meta-analysis. Obes Rev 2014;15(8):657–65. [DOI] [PubMed] [Google Scholar]
- 34. Mosca L, Benjamin EJ, Berra K, Bezanson JL, Dolor RJ, Lloyd-Jones DM, Newby LK, Pina IL, Roger VL, Shaw LJ et al. Effectiveness-based guidelines for the prevention of cardiovascular disease in women—2011 update: a guideline from the American Heart Association. Circulation 2011;123(11):1243–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Martinez-Fernandez L, Laiglesia LM, Huerta AE, Martinez JA, Moreno-Aliaga MJ. Omega-3 fatty acids and adipose tissue function in obesity and metabolic syndrome. Prostaglandins Other Lipid Mediat 2015;121(Pt A):24–41. [DOI] [PubMed] [Google Scholar]
- 36. Couet C, Delarue J, Ritz P, Antoine JM, Lamisse F. Effect of dietary fish oil on body fat mass and basal fat oxidation in healthy adults. Int J Obes Relat Metab Disord 1997;21(8):637–43. [DOI] [PubMed] [Google Scholar]
- 37. Hariri M, Ghiasvand R, Shiranian A, Askari G, Iraj B, Salehi-Abargouei A. Does omega-3 fatty acids supplementation affect circulating leptin levels? A systematic review and meta-analysis on randomized controlled clinical trials. Clin Endocrinol (Oxf) 2015;82(2):221–8. [DOI] [PubMed] [Google Scholar]
- 38. Nettleton JA. Omega-3 fatty acids: comparison of plant and seafood sources in human nutrition. J Am Diet Assoc 1991;91(3):331–7. [PubMed] [Google Scholar]
- 39. Hall KD, Sacks G, Chandramohan D, Chow CC, Wang YC, Gortmaker SL, Swinburn BA. Quantification of the effect of energy imbalance on bodyweight. Lancet 2011;378(9793):826–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.