Abstract
Background: Replacement of saturated fatty acids (SFAs) with unsaturated fatty acids (UFAs), especially polyunsaturated fatty acids (PUFAs), has been associated with a lower risk of ischemic heart disease (IHD). Whether this replacement is beneficial for drug-treated patients with cardiac disease is not yet clear.
Objective: In a prospective study of Dutch patients with cardiac disease (Alpha Omega Cohort), we examined the risk of cardiovascular disease (CVD) and IHD mortality when the sum of SFAs and trans fatty acids (TFAs) was theoretically replaced by total UFAs, PUFAs, or cis monounsaturated fatty acids (MUFAs).
Design: We included 4146 state-of-the-art drug-treated patients aged 60–80 y with a history of myocardial infarction (79% male patients) and reliable dietary data at baseline (2002–2006). Cause-specific mortality was monitored until 1 January 2013. HRs for CVD mortality and IHD mortality for theoretical, isocaloric replacement of dietary fatty acids (FAs) in quintiles (1–5) and continuously (per 5% of energy) were obtained from Cox regression models, adjusting for demographic factors, medication use, and lifestyle and dietary factors.
Results: Patients consumed, on average, 17.5% of energy of total UFAs, 13.0% of energy of SFAs, and <1% of energy of TFAs. During ∼7 y of follow-up, 372 CVD deaths and 249 IHD deaths occurred. Substitution modeling yielded significantly lower risks of CVD mortality when replacing SFAs plus TFAs with total UFAs [HR in quintile 5 compared with quintile 1: 0.45 (95% CI: 0.28, 0.72)] or PUFAs [HR: 0.66 (95% CI: 0.44, 0.98)], whereas HRs in cis MUFA quintiles were nonsignificant. HRs were similar for IHD mortality. In continuous analyses, replacement of SFAs plus TFAs with total UFAs, PUFAs, or cis MUFAs (per 5% of energy) was associated with significantly lower risks of CVD mortality (HRs between 0.68 and 0.75) and IHD mortality (HRs between 0.55 and 0.70).
Conclusion: Shifting the FA composition of the diet toward a higher proportion of UFAs may lower CVD mortality risk in drug-treated patients with cardiac disease. This study was registered at clinicaltrials.gov as NCT03192410.
Keywords: dietary fatty acids, cardiac patients, cardiovascular disease, coronary heart disease, substitution analysis, prospective cohort study
INTRODUCTION
Avoiding trans fatty acids (TFAs) and consuming fewer calories from SFAs by replacing them with PUFAs and cis MUFAs are recommended for the prevention of ischemic heart disease (IHD) (1–3). These recommendations are based on observational and experimental studies with clinical endpoints (4–6) and on supporting evidence from randomized controlled trials (RCTs) of fatty acid (FA) intake and blood lipids (6–8). Specifying the replacement nutrient is essential because substituting SFAs with carbohydrates, typically the most abundant nutrient in the diet, is not associated with lower IHD risk (4, 7, 9).
Despite advances in medical treatment (10), life expectancy is substantially reduced in patients who are at risk for cardiovascular disease (CVD), especially in those with diabetes (11), chronic kidney disease (12), or an unhealthy lifestyle (13). An analysis in the Prevención con Dieta Mediterránea cohort of 7038 participants at high risk for CVD showed that PUFAs and MUFAs were inversely associated with CVD risk, whereas SFAs and TFAs were positively associated with CVD risk (14). Whether replacing the sum of SFAs and TFAs with unsaturated FAs (UFAs) affects long-term mortality after an IHD event on top of cardiovascular medication is largely unknown. The present study examined the theoretical replacement of energy from SFAs plus TFAs with UFAs in relation to CVD mortality and IHD mortality in patients with cardiac disease from the Alpha Omega Cohort (clinicaltrials.gov; NCT03192410).
METHODS
Patients and study design
The Alpha Omega Cohort is a prospective study of 4837 state-of-the-art drug-treated Dutch patients aged 60–80 y with a clinically diagnosed myocardial infarction (MI) <10 y before enrollment. During the first 3 y of follow-up, patients took part in an experimental study of low doses of n–3 FAs (Alpha Omega Trial), as described elsewhere (15, 16). At baseline (2002–2006), data on demographics, medical history, medication use, diet, and lifestyle were collected through the use of questionnaires. Patients were physically examined, including blood sampling, by trained research nurses. The medical ethics committee at the Haga Hospital (The Hague, Netherlands) approved the study, and all of the patients provided written informed consent. The cohort was followed for cause-specific mortality until 1 January 2013. For the present analysis (Supplemental Figure 1) we excluded 453 patients with missing dietary data and 19 patients with implausible energy intakes (<800 or >8000 kcal/d for men, <600 or >6000 kcal/d for women). We additionally excluded 219 patients with extreme UFA intakes (<2.5th or >97.5th percentile) to obtain reliable risk estimates when analyzing UFA intake on a continuous scale. A total of 4146 patients remained for analysis.
Dietary data
Dietary data were collected by a 203-item food-frequency questionnaire (FFQ), which was an extended version of a reproducible, biomarker-validated questionnaire designed to estimate FAs and cholesterol intake (17, 18). Patients were asked to report their usual food intake during the past month, including type of food, frequency, amount, and preparation methods. For example, when patients indicated that they consumed baked potatoes, they were also asked which type of fat, from a list of 16 options, was used for baking. To ensure that brand names were reported correctly for the quantification of specific types of fat, a list of branded products commonly used in the Netherlands was included. Trained dietitians checked the returned questionnaires and obtained additional information on unclear or missing items by telephone. Double data entry of FFQ data was performed and inconsistencies were solved. Food consumption data were converted into energy and nutrient intakes by use of the 2006 Dutch food composition database (NEVO) (19). Total energy intake (in kilocalories per day) was calculated without energy from alcohol. Criteria based on Dutch dietary guidelines were used to cluster specific foods consumed in the Alpha Omega Cohort into 24 mutually exclusive food groups (20).
The dietary intake of FAs was expressed as a percentage of total energy intake. Total UFA intake comprised MUFAs and total PUFAs. MUFA intake was confined to cis MUFAs. PUFA intake included n–6 FAs (18:2n–6) and n–3 FAs [18:2n–3; α-linolenic acid (ALA), EPA, and DHA]. The number of patients who used supplements with n–3 FAs at baseline was negligible because it was an exclusion criterion for the Alpha Omega Trial (16), on which the present cohort study was based. Supplemental n–3 FA intake during the trial phase (ALA, EPA-DHA, both, or placebo) was unrelated to baseline FA intakes because of randomization.
Other measurements
Data collection has been described in detail elsewhere (15, 16). BMI was calculated as weight divided by squared height (kg/m2). Medication was coded according to the Anatomical Therapeutic Chemical classification system. The codes were C02, C03, C07, C08, and C09 for antihypertensive medication; C10AA and C10B for statins; and B01 for antithrombotic medication. Blood pressure was measured twice and the mean was taken. We analyzed nonfasting blood for lipids and glucose using standard laboratory methods. The presence of diabetes mellitus was based on self-report of a physician’s diagnosis, use of antidiabetic medication, or elevated glucose levels (≥7.0 mmol/L if fasting glucose, or ≥11.1 mmol/L if nonfasting glucose). Smoking status was assessed in 3 categories (never, former, or current). Alcohol intake, estimated from the FFQ, was categorized as 0, >0–10, >10–20, or >20 g/d. Physical activity, assessed by a validated questionnaire (21), was categorized as low (no activity or only light activity, defined as ≤3 metabolic equivalents), intermediate [>0 to <5 d/wk of moderate or vigorous activity (>3 metabolic equivalents)], or high (≥5 d/wk of moderate or vigorous activity). Educational level was defined as low (primary or lower secondary education), intermediate (higher secondary or lower tertiary education), or high (higher tertiary education).
Ascertainment and classification of mortality
Vital status and causes of death were monitored through a computerized link with municipal registries and through updates of the Dutch National Mortality Registry [(Statistics Netherlands (CBS)]. One patient was lost to follow-up and censored after 2.9 y. Causes of death were coded according to the International Classification of Diseases, 10th Revision by the WHO. CVD mortality included IHD (I20–I25), cardiac arrest (I46), sudden death undefined (R96), heart failure (I50), and stroke (I60–I69) as primary or secondary causes of death. IHD mortality included I20–I25, I46, and R96.
Statistical analysis
Baseline characteristics in energy-adjusted quintiles of total UFA intake (i.e., PUFAs plus cis MUFAs) are presented as unadjusted means with SDs for normally distributed variables, medians with IQRs for skewed variables, or percentages for categorical variables. Age- and sex-adjusted Pearson correlations between intakes of FAs were calculated. Survival time (in years) was calculated from entry into the cohort until the date of death, loss to follow-up, or 1 January 2013, whichever came first. Cox regression models were used to obtain HRs with 95% CIs for CVD mortality and IHD mortality for theoretical, isocaloric replacement of SFAs plus TFAs with total UFAs, PUFAs, or cis MUFAs. Proportional hazards assumptions were examined by log−log plots, and assumptions were met. Analyses were undertaken with FAs in quintiles, with the lowest quintile as the reference. The linear trend across HRs was tested by entering the median intakes in quintiles as a continuous variable in the Cox regression model. All of the analyses were repeated with FA intakes on a continuous scale (per 5% of energy). Missing data for BMI (n = 5), educational level (n = 21), smoking (n = 1), and physical activity (n = 22) were imputed by sex-specific medians to retain these covariables in the multivariable Cox models. Potential thresholds or nonlinear relations were identified with restricted cubic spline (RCS) analyses, including 3 knots placed at the 5th, 50th, and 95th percentiles.
The basic Cox model (model 1) included total energy intake (continuous; kilocalories per day) and intake of protein, carbohydrates, and the replacement FA (percentage of energy; quintiles). PUFA analyses also were adjusted for MUFAs (percentage of energy; quintiles), and MUFA analyses were adjusted for PUFAs (percentage of energy; quintiles). Model 2 additionally included age (in years), sex, Alpha Omega Trial treatment code (4 categories), BMI (in kg/m2), prevalent diabetes (yes or no), antithrombotic drugs (yes or no), antihypertensive drugs (yes or no), lipid-lowering drugs (yes or no), educational level (3 categories), smoking status (3 categories), alcohol intake (4 categories), and physical activity (3 categories). In model 3 further adjustments were made for dietary cholesterol (milligrams per day; quintiles) and dietary fiber (grams per day; quintiles). The analyses were repeated with TFAs added to the multivariable model, yielding HRs for SFA replacement only.
HRs for CVD mortality and IHD mortality were computed in predefined subgroups by age (<65 or ≥65 y), sex (men or women), BMI (<30 or ≥30), prevalent diabetes (yes or no), and smoking status (never, former, or current), by use of model 3. The interaction terms for subgroups were tested for statistical significance. In sensitivity analyses we excluded 1) nonstatin users, 2) patients who had an MI <2 y before entry into the study, and 3) patients who died during the first 2 y of follow-up, also extracting 2 y of follow-up time for the remaining cohort. Data analysis was performed with SAS version 9.3 (SAS Institute Inc.). The SAS Macro RCS based on Cox regression models (22) was used to conduct the RCS analyses. Two-sided P values <0.05 were considered statistically significant.
RESULTS
Patient characteristics
Table 1 shows the baseline characteristics of the cohort overall and in the extreme and middle quintiles of total UFA intake. Patients were on average 69.0 ± 5.6 y old and 79% were male. Sixteen percent currently smoked and 20% had diabetes. Most of the patients used antithrombotic drugs (98%), antihypertensive drugs (90%), or lipid-lowering drugs (86%). Mean FA intakes were 13.0 ± 3.1% of energy for SFAs, 7.6 ± 2.3% of energy for total PUFAs, and 9.9 ± 2.3% of energy for cis MUFAs. The intake of TFAs was low (0.75 ± 0.19% of energy). SFA intake (percentage of energy) was strongly correlated with TFA intake (r = 0.75, P < 0.001) and modestly correlated with cis MUFA intake (r = 0.42, P < 0.001). Other correlations between FAs were <0.40 (data not shown). The main dietary sources (Supplemental Figure 2) of SFAs were dairy (27%), edible fats and oils (24%), snacks (19%), and meats (13%). For total UFAs (i.e., PUFA plus cis MUFA), the main dietary sources were edible fats and oils (39%), snacks (14%), meats (12%), and grains (9%). Median UFA intake ranged from 12.9% of energy in the lowest quintile to 23.4% of energy in the highest quintile of UFAs (Table 1). Those with a higher UFA intake were slightly younger and more often male patients and current smokers. They consumed more alcohol, alcohol, total energy, cholesterol, and fat (all types); and less protein, carbohydrates, and fiber.
TABLE 1.
Baseline characteristics of 4146 patients with cardiac disease from the Alpha Omega Cohort, overall and by quintiles of energy-adjusted total UFA intake1
| Quintiles of total UFA intake |
|||||
| Total population (n = 4146) | 1 (n = 829) | 3 (n = 830) | 5 (n = 829) | P-trend2 | |
| Median intake (range), en% | 12.9 (10.2–14.3) | 17.2 (16.2–18.2) | 23.4 (20.7–27.9) | ||
| Age, y | 69.0 ± 5.63 | 69.7 ± 5.6 | 68.9 ± 5.4 | 68.8 ± 5.6 | <0.001 |
| Male sex, n (%) | 3274 (79) | 569 (69) | 674 (81) | 708 (85) | <0.001 |
| BMI, kg/m2 | 27.8 ± 3.8 | 27.8 ± 3.9 | 27.8 ± 3.9 | 27.7 ± 3.8 | 0.16 |
| Time since MI, y | 3.7 (1.7–6.3)4 | 3.8 (1.7–6.3) | 3.6 (1.7–6.4) | 3.3 (1.5–6.3) | 0.77 |
| Prevalent diabetes, n (%) | 841 (20) | 164 (20) | 174 (21) | 172 (21) | 0.94 |
| Use of antithrombotic drugs,5 n (%) | 4054 (98) | 807 (97) | 810 (98) | 814 (98) | 0.76 |
| Use of antihypertensive drugs,6 n (%) | 3726 (90) | 744 (90) | 749 (90) | 755 (91) | 0.64 |
| Use of statins,7 n (%) | 3566 (86) | 723 (87) | 721 (87) | 708 (85) | 0.25 |
| Blood pressure, mm Hg | |||||
| Systolic | 141.6 ± 21.7 | 142.0 ± 21.8 | 142.0 ± 21.5 | 141.5 ± 21.3 | 0.38 |
| Diastolic | 80.3 ± 11.1 | 79.8 ± 11.5 | 80.3 ± 11.1 | 80.9 ± 10.9 | 0.14 |
| Serum lipids,8 mmol/L | |||||
| Total cholesterol | 4.74 ± 0.96 | 4.77 ± 0.92 | 4.66 ± 1.00 | 4.71 ± 0.95 | 0.70 |
| LDL cholesterol | 2.61 ± 0.83 | 2.58 ± 0.80 | 2.55 ± 0.83 | 2.56 ± 0.83 | 0.86 |
| HDL cholesterol | 1.28 ± 0.34 | 1.32 ± 0.37 | 1.25 ± 0.31 | 1.30 ± 0.36 | 0.29 |
| Triglycerides | 1.65 (1.21–2.31) | 1.67 (1.21–2.37) | 1.63 (1.18–2.31) | 1.70 (1.20–2.31) | 0.57 |
| Education,9 n (%) | 0.06 | ||||
| Low | 2306 (56) | 471 (57) | 434 (53) | 481 (58) | |
| Intermediate | 1303 (32) | 263 (32) | 266 (32) | 256 (31) | |
| High | 516 (13) | 87 (11) | 125 (15) | 87 (11) | |
| Smoking, n (%) | <0.001 | ||||
| Never | 676 (16) | 191 (23) | 124 (15) | 98 (12) | |
| Former | 2786 (67) | 535 (65) | 581 (70) | 556 (67) | |
| Current | 683 (16) | 102 (12) | 125 (15) | 175 (21) | |
| Alcohol use, g/d, n (%) | <0.001 | ||||
| 0 | 823 (20) | 196 (24) | 152 (18) | 161 (19) | |
| >0–10 | 1629 (39) | 351 (42) | 334 (40) | 289 (35) | |
| >10–20 | 732 (18) | 134 (16) | 148 (18) | 145 (17) | |
| >20 | 962 (23) | 148 (18) | 195 (23) | 234 (28) | |
| Physical activity,9 n (%) | 0.07 | ||||
| Low | 1684 (41) | 342 (42) | 346 (42) | 340 (41) | |
| Intermediate | 1558 (38) | 278 (34) | 306 (37) | 330 (40) | |
| High | 882 (21) | 201 (24) | 175 (21) | 156 (19) | |
| Dietary intake | |||||
| Energy, kcal/d | 1815 ± 494 | 1691 ± 477 | 1832 ± 503 | 1883 ± 492 | <0.001 |
| Protein, en% | 15.8 ± 2.9 | 16.2 ± 3.2 | 15.8 ± 2.7 | 15.2 ± 2.7 | <0.001 |
| Total fat, en% | 33.2 ± 6.0 | 26.1 ± 3.4 | 33.5 ± 3.3 | 40.1 ± 4.2 | <0.001 |
| SFAs, en% | 13.0 ± 3.1 | 11.2 ± 2.7 | 13.5 ± 2.9 | 14.1 ± 3.3 | <0.001 |
| PUFAs, en% | 7.6 ± 2.3 | 5.3 ± 1.1 | 7.3 ± 1.3 | 10.5 ± 2.1 | <0.001 |
| cis MUFAs, en% | 9.9 ± 2.3 | 7.5 ± 1.1 | 9.9 ± 1.3 | 12.6 ± 2.1 | <0.001 |
| TFAs, en% | 0.75 ± 0.19 | 0.67 ± 0.18 | 0.78 ± 0.20 | 0.79 ± 0.18 | <0.001 |
| TFAs, g/d | 1.54 ± 0.63 | 1.27 ± 0.55 | 1.61 ± 0.64 | 1.68 ± 0.64 | <0.001 |
| Cholesterol, mg/d | 182 ± 68 | 156 ± 56 | 192 ± 71 | 192 ± 73 | <0.001 |
| Carbohydrates, en% | 48.9 ± 6.4 | 55.9 ± 4.8 | 48.6 ± 4.0 | 42.3 ± 4.8 | <0.001 |
| Mono- and disaccharides, en% | 24.9 ± 6.9 | 30.9 ± 7.0 | 24.4 ± 5.6 | 20.0 ± 5.4 | <0.001 |
| Polysaccharides, en% | 23.9 ± 4.7 | 24.9 ± 5.5 | 24.1 ± 4.4 | 22.3 ± 3.8 | <0.001 |
| Fiber, g/d | 21.4 ± 6.7 | 22.3 ± 7.2 | 21.4 ± 7.1 | 20.2 ± 6.1 | <0.001 |
ATC, Anatomical Therapeutical Chemical; en%, percentage of total energy intake; MI, myocardial infarction; TFA, trans fatty acid; UFA, unsaturated fatty acid.
P values for differences between quintiles were obtained from ANOVA (normally distributed continuous variables), Kruskal-Wallis test (skewed variables), or chi-square test (categorical variables).
Mean ± SD (all such values).
Median; IQR in parentheses (all such values).
ATC classification system code: B01.
ATC classification system codes: C02, C03, C07, C08, and C09.
ATC classification system codes: C10AA and C10B.
To convert to milligrams per deciliter, divide by 0.02586 for cholesterol fractions and by 0.01129 for triglycerides.
Classification described in text (Methods).
Dietary FAs in relation to CVD mortality and IHD mortality
During a median follow-up time of 7.3 y (29,608 person-years), there were 888 deaths, including 372 from CVD and 249 from IHD. Table 2 shows HRs for CVD mortality and IHD mortality when isocalorically replacing SFAs plus TFAs with total UFAs, PUFAs, or cis MUFAs. The risk of CVD and IHD mortality decreased significantly across quintiles of total UFA intake. The HR in the highest quintile compared with the lowest quintile was 0.45 (95% CI: 0.28, 0.72; P-trend = 0.003) for CVD and 0.37 (95% CI: 0.21, 0.67; P-trend = 0.002) for IHD mortality when adjusting for demographic factors, medication use, lifestyle factors, and dietary factors (model 3). A significant inverse association was observed for PUFAs, with an HR of 0.66 (95% CI: 0.44, 0.98; P-trend = 0.011) for CVD and 0.53 (95% CI: 0.33, 0.88; P-trend = 0.002) for IHD mortality in the highest quintiles (model 3). HRs in quintiles of MUFA intake were not statistically significant (Table 2). When analyzing FAs on a continuous scale (per 5% of energy), replacement of SFAs plus TFAs with UFAs, PUFAs, or MUFAs yielded a significantly lower risk of CVD mortality (HR: 0.68–0.75) and IHD mortality (HR: 0.55–0.70; Table 2). When analyzing replacement of SFAs only, adjusting for TFA, the results for both outcomes remained essentially similar (Supplemental Table 1).
TABLE 2.
HRs for CVD mortality and IHD mortality in quintiles of total UFA, PUFA, and cis MUFA intake for theoretical, isocaloric replacement of SFAs plus TFAs in 4146 patients from the Alpha Omega Cohort1
| Quintiles of dietary FA intake2 |
||||||||
| 1 (n = 829) | 2 (n = 829) | 3 (n = 830) | 4 (n = 829) | 5 (n = 829) | P-trend3 | Per 5% of energy intake | P | |
| CVD mortality | ||||||||
| Total UFAs | ||||||||
| Median, en% | 12.9 | 15.3 | 17.2 | 19.3 | 22.7 | |||
| Cases, n | 96 | 69 | 66 | 65 | 76 | |||
| Model 14 | 1.00 | 0.59 (0.42, 0.83) | 0.47 (0.32, 0.69) | 0.38 (0.25, 0.59) | 0.37 (0.23, 0.58) | <0.001 | 0.65 (0.53, 0.82) | <0.001 |
| Model 25 | 1.00 | 0.63 (0.45, 0.89) | 0.50 (0.34, 0.75) | 0.43 (0.28, 0.66) | 0.42 (0.26, 0.67) | 0.001 | 0.69 (0.55, 0.86) | 0.001 |
| Model 36 | 1.00 | 0.64 (0.46, 0.91) | 0.51 (0.35, 0.76) | 0.45 (0.29, 0.69) | 0.45 (0.28, 0.72) | 0.003 | 0.71 (0.57, 0.89) | 0.003 |
| PUFAs | ||||||||
| Median, en% | 4.9 | 6.2 | 7.3 | 8.5 | 10.8 | |||
| Cases, n | 89 | 87 | 64 | 62 | 70 | |||
| Model 14 | 1.00 | 0.93 (0.68, 1.25) | 0.63 (0.45, 0.89) | 0.54 (0.38, 0.77) | 0.52 (0.36, 0.76) | <0.001 | 0.62 (0.47, 0.82) | 0.001 |
| Model 25 | 1.00 | 1.00 (0.74, 1.35) | 0.67 (0.48, 0.93) | 0.61 (0.43, 0.88) | 0.62 (0.42, 0.90) | 0.003 | 0.70 (0.53, 0.93) | 0.014 |
| Model 36 | 1.00 | 1.01 (0.75, 1.38) | 0.68 (0.48, 0.96) | 0.64 (0.44, 0.92) | 0.66 (0.44, 0.98) | 0.011 | 0.68 (0.51, 0.90) | 0.007 |
| cis MUFAs | ||||||||
| Median, en% | 7.2 | 8.6 | 9.8 | 11.0 | 12.8 | |||
| Cases, n | 80 | 75 | 61 | 81 | 75 | |||
| Model 14 | 1.00 | 0.93 (0.66, 1.30) | 0.70 (0.48, 1.03) | 0.86 (0.58, 1.28) | 0.69 (0.45, 1.07) | 0.11 | 0.71 (0.52, 0.96) | 0.027 |
| Model 25 | 1.00 | 1.10 (0.78, 1.55) | 0.85 (0.58, 1.25) | 1.00 (0.67, 1.50) | 0.75 (0.48, 1.16) | 0.15 | 0.68 (0.50, 0.93) | 0.015 |
| Model 36 | 1.00 | 1.10 (0.78, 1.55) | 0.84 (0.57, 1.24) | 0.98 (0.65, 1.47) | 0.74 (0.48, 1.15) | 0.13 | 0.75 (0.57, 1.00) | 0.048 |
| IHD mortality | ||||||||
| Total UFAs | ||||||||
| Median, en% | 12.9 | 15.3 | 17.2 | 19.3 | 22.7 | |||
| Cases, n | 61 | 44 | 50 | 47 | 47 | |||
| Model 14 | 1.00 | 0.55 (0.36, 0.84) | 0.50 (0.32, 0.80) | 0.38 (0.23, 0.64) | 0.30 (0.17, 0.54) | <0.001 | 0.57 (0.44, 0.75) | <0.001 |
| Model 25 | 1.00 | 0.59 (0.38, 0.90) | 0.54 (0.34, 0.87) | 0.44 (0.26, 0.73) | 0.35 (0.20, 0.63) | 0.001 | 0.61 (0.46, 0.80) | <0.001 |
| Model 36 | 1.00 | 0.60 (0.39, 0.92) | 0.56 (0.35, 0.89) | 0.45 (0.27, 0.76) | 0.37 (0.21, 0.67) | 0.002 | 0.62 (0.47, 0.82) | 0.001 |
| PUFAs | ||||||||
| Median, en% | 4.9 | 6.2 | 7.3 | 8.5 | 10.8 | |||
| Cases, n | 58 | 60 | 49 | 38 | 44 | |||
| Model 14 | 1.00 | 0.94 (0.65, 1.35) | 0.69 (0.46, 1.03) | 0.46 (0.29, 0.72) | 0.44 (0.28, 0.71) | <0.001 | 0.53 (0.38, 0.74) | <0.001 |
| Model 25 | 1.00 | 1.00 (0.69, 1.44) | 0.72 (0.48, 1.08) | 0.51 (0.33, 0.81) | 0.52 (0.33, 0.84) | 0.001 | 0.60 (0.42, 0.85) | 0.004 |
| Model 36 | 1.00 | 0.99 (0.68, 1.44) | 0.72 (0.48, 1.08) | 0.52 (0.33, 0.83) | 0.53 (0.33, 0.88) | 0.002 | 0.55 (0.39, 0.78) | 0.001 |
| cis MUFAs | ||||||||
| Median, en% | 7.2 | 8.6 | 9.8 | 11.0 | 12.8 | |||
| Cases, n | 50 | 48 | 44 | 57 | 50 | |||
| Model 14 | 1.00 | 0.89 (0.58, 1.36) | 0.73 (0.46, 1.17) | 0.86 (0.53, 1.41) | 0.65 (0.38, 1.11) | 0.14 | 0.65 (0.44, 0.94) | 0.024 |
| Model 25 | 1.00 | 1.04 (0.68, 1.60) | 0.89 (0.55, 1.43) | 1.02 (0.62, 1.68) | 0.73 (0.42, 1.25) | 0.21 | 0.64 (0.44, 0.93) | 0.020 |
| Model 36 | 1.00 | 1.06 (0.69, 1.64) | 0.89 (0.56, 1.44) | 1.03 (0.62, 1.69) | 0.74 (0.43, 1.27) | 0.23 | 0.70 (0.49, 0.99) | 0.045 |
CVD, cardiovascular disease; en%, percentage of total energy intake; FA, fatty acid; IHD, ischemic heart disease; TFA, trans fatty acid; UFA, unsaturated fatty acid.
HRs are presented with 95% CIs in quintiles of the replacing fatty acid intake, with the use of the lowest quintile as the reference.
The linear trend across HRs was tested by entering the median intakes in quintiles as a continuous variable in the Cox regression model.
Model includes total energy intake (kilocalories per day) and intake (en%; quintiles) of protein, carbohydrates, and MUFAs (in PUFA analysis) or PUFAs (in MUFA analysis).
Includes in model 1, with the addition of age, sex, Alpha Omega Trial treatment code (4 categories), BMI, prevalent diabetes (yes or no), antithrombotic drugs (yes or no), antihypertensive drugs (yes or no), lipid-lowering drugs (yes or no), level of education (3 categories), smoking status (3 categories), alcohol intake (4 categories), and physical activity (3 categories).
Includes in model 2, with the addition of dietary cholesterol (milligrams per day; quintiles) and dietary fiber (grams per day; quintiles).
Figure 1 shows changes in HRs for CVD and IHD mortality, obtained by RCS analysis, for the theoretical replacement of SFAs plus TFAs with UFAs (model 3). The overall associations were inverse and statistically significant (P < 0.001). For CVD mortality the relation with UFA intake was nonlinear (P-nonlinearity = 0.014), with HRs flattening above the 50th percentile of intake. For IHD mortality HRs decreased across the entire range of UFA intake, with no evidence of a nonlinear relation (P-nonlinearity = 0.28; Figure 1). RCS analyses for PUFAs indicated a possible nonlinear inverse relation with CVD and IHD mortality (Supplemental Figure 3), whereas for MUFAs the HRs tended to decrease linearly for both outcomes (Supplemental Figure 4).
FIGURE 1.
Associations with CVD mortality (A) and IHD mortality (B) for the theoretical, isocaloric replacement of energy from SFAs plus TFAs with total UFAs in 4146 patients from the Alpha Omega Cohort. Lines are restricted cubic splines, showing the shape of associations on a continuous scale, with 3 knots located at the 5th, 50th, and 95th percentiles (11.9%, 17.2%, and 24.5% of energy intake, respectively). The y-axis shows the predicted HRs for CVD mortality for any value of intake, compared with the reference value set at the 5th percentile. Gray areas indicate 95% CIs. Results are presented for model 3 (details in text). CVD, cardiovascular disease; en%, percentage of total energy intake; IHD, ischemic heart disease; TFA, trans fatty acid; UFA, unsaturated fatty acid.
Subgroup and sensitivity analyses
Although interaction terms for stratification variables were not statistically significant (P > 0.10), the inverse associations of UFA intake (per 5% of energy) with CVD mortality and IHD mortality tended to be more pronounced in the subgroup of patients with diabetes (Supplemental Figures 5 and 6). When confined to statin users (86% of cohort), HRs per 5% of energy of UFA intake were 0.73 (95% CI: 0.57, 0.94) for CVD mortality and 0.67 (95% CI: 0.49, 0.91) for IHD mortality. Excluding 1324 patients with an MI <2 y before entry into the study yielded HRs (model 3) of 0.75 (95% CI: 0.58, 0.97) for CVD mortality and 0.65 (95% CI: 0.47, 0.89) for IHD mortality, and excluding the first 2 y of follow-up (149 cases of death) yielded HRs of 0.70 (95% CI: 0.55, 0.90) and 0.64 (95% CI: 0.47, 0.87), respectively.
DISCUSSION
In this theoretical substitution analysis in the Alpha Omega Cohort of Dutch patients with cardiac disease, we observed significant inverse relations with CVD mortality and IHD mortality when isocalorically replacing SFAs plus TFAs with UFAs. Because TFA intake was low, these associations resulted mainly from replacement of SFAs.
A growing body of evidence suggests that replacing SFAs with PUFAs can decrease IHD risk. Since 2009, meta-analyses of RCTs (6, 7, 23–27) have consistently shown a 5–10% lower IHD risk for each 5% of energy of SFAs replaced with PUFAs. Whether results were statistically significant depended on the number and quality of the studies. A pooled analysis of individual data from 11 cohort studies (4) also showed a lower IHD mortality risk when energy from SFAs was replaced with PUFAs by use of a substitution model. A meta-analysis of 13 cohort studies showed an inverse, dose-response relation with IHD mortality for the main PUFA in the diet (LA), irrespective of the nutrient that it was replacing (5). The meta-analysis of 16 cohort studies by Siri-Tarino et al. (28) showed no benefit in lowering SFA intake, but their study did not consider the replacement nutrient. Specifying the replacement nutrient is essential because no associations are found with IHD risk when substituting SFAs with total carbohydrates, typically the most abundant nutrient in the diet (4, 7, 9).
Only limited data exist on dietary FAs in the secondary prevention of IHD. A meta-analysis of RCTs of modified-fat diets showed no significant associations with CVD events when SFAs were replaced by PUFAs in patients with established IHD (29). Heterogeneity was substantial, however, and the overall quality of secondary prevention trials (mainly older studies) was only moderate (29). A prospective analysis in the high-risk Prevención con Dieta Mediterránea cohort showed a reduction in CVD events of at least one-third when replacing 5% of energy from SFAs with PUFAs or MUFAs (mainly from olive oil and nuts) in a substitution model (14). Although different food sources contributed to the intake of FAs in our cohort of Dutch patients with cardiac disease, we obtained risk estimates for CVD mortality of the same order of magnitude when replacing 5% of energy of SFA with UFAs on top of state-of-the-art drug treatment.
Our study has an observational design, which has limitations. FAs are not consumed in isolation but in the context of a total diet and lifestyle pattern. This cannot be captured fully in theoretical substitution models, which are focused on the exchange of single macronutrients. Despite extensive data collection on lifestyle and diet, we cannot exclude the possibility of residual confounding (e.g., from salt intake that could not be assessed accurately by FFQs). Patients may have changed their diets after MI (e.g., through dietary counseling to lower intakes of SFAs and TFAs); however, habitual diets in our cohort of stable, noninstitutionalized patients with cardiac disease did not differ much from those of the general older Dutch population for which comparable intakes of SFAs (13% of energy) and TFAs (1% of energy) were reported in a 2010–2012 national survey (30). In the sensitivity analyses we excluded patients with an MI occurring <2 y ago and patients who died in the first 2 y of follow-up because their baseline diets may have changed as a result of clinical symptoms. This exclusion yielded results similar to those of the total cohort, and we consider reverse causality bias unlikely.
The strengths of the present study include the large, well-documented cohort of patients after MI, a growing segment of Western populations for which data on diet and clinical endpoints are limited. Detailed dietary information was collected by an extensive, validated FFQ that was designed specifically for the assessment of FA intake. Only 1 patient was lost to long-term follow-up for cause-specific mortality. The Alpha Omega Cohort, however, is rather homogenous (e.g., for age, race, dietary habits) and thus our findings may not be generalizable to other populations. Women were underrepresented and stratified analyses (e.g., by prevalent diabetes) lacked statistical power, which limits the interpretation of risk estimates in patient subgroups.
In conclusion, our data showed that the theoretical replacement of SFAs (and TFAs) with UFAs was associated with a lower CVD mortality and IHD mortality risk in patients with cardiac disease who received state-of-the-art drug treatment, including statins. More research on FA replacements after MI is needed in women, patients with diabetes, and nonwhites. Confirmation of our observational findings in secondary prevention trials with clinical endpoints is warranted.
Acknowledgments
We thank Eveline Waterham for logistic support and data management.
The authors’ responsibilities were as follows—JdG and JMG: designed the research; FJMM, JdG, DK, and JMG: conducted the research; FJMM and JdG: analyzed the data; FJMM and JMG: wrote the manuscript; JMG: had primary responsibility for the final content and handled funding and supervision; and all authors: interpreted the results, critically revised the manuscript for intellectual content, and read and approved the final manuscript. AJW and PLZ are employed by Unilever R&D. JMG received funding from Unilever R&D for epidemiological studies of dietary and circulating fatty acids and cardiovascular disease. None of the other authors reported a conflict of interest related to the study.
Footnotes
Abbreviations used: ALA, α-linolenic acid; CVD, cardiovascular disease; FA, fatty acid; FFQ, food-frequency questionnaire; IHD, ischemic heart disease; MI, myocardial infarction; RCS, restricted cubic spline; RCT, randomized controlled trial; TFA, trans fatty acid; UFA, unsaturated fatty acid.
REFERENCES
- 1.WHO. Diet, nutrition, and the prevention of chronic diseases: report of a Joint WHO/FAO Expert Consultation. Geneva (Switzerland): WHO; 2003. [Google Scholar]
- 2.Dietary Guidelines for Americans 2015–2020. 8th edition [Internet]. Washington (DC): US Department of Health and Human Services/USDA; 2015 [cited 2017 Mar 13]. Available from: https://health.gov/dietaryguidelines/2015/guidelines/.
- 3.Kromhout D, Spaaij CJ, de Goede J, Weggemans RM. The 2015 Dutch food-based dietary guidelines. Eur J Clin Nutr 2016;70:869–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jakobsen MU, O’Reilly EJ, Heitmann BL, Pereira MA, Balter K, Fraser GE, Goldbourt U, Hallmans G, Knekt P, Liu S, et al. Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies. Am J Clin Nutr 2009;89:1425–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Farvid MS, Ding M, Pan A, Sun Q, Chiuve SE, Steffen LM, Willet WC, Hu FB. Dietary linoleic acid and risk of coronary heart disease: a systematic review and meta-analysis of prospective cohort studies. Circulation 2014;130:1568–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mozaffarian D, Micha R, Wallace S. Effects on coronary heart disease of increasing polyunsaturated fat in place of saturated fat: a systematic review and meta-analysis of randomized controlled trials. PLoS Med 2010;7:e1000252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hooper L, Martin N, Abdelhamid A, Davey Smith G. Reduction in saturated fat intake for cardiovascular disease. Cochrane Database Syst Rev 2015;CD011737. [DOI] [PubMed] [Google Scholar]
- 8.Mensink RP. Effects of saturated fatty acids on serum lipids and lipoproteins: a systematic review and regression analysis. Geneva (Switzerland): WHO; 2016. [Google Scholar]
- 9.Li Y, Hruby A, Bernstein AM, Ley SH, Wang DD, Chiuve SE, Sampson L, Rexrode KM, Rimm EB, Willet WC, et al. Saturated fats compared with unsaturated fats and sources of carbohydrates in relation to risk of coronary heart disease: a prospective cohort study. J Am Coll Cardiol 2015;66:1538–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Soedamah-Muthu SS, Geleijnse JM, Giltay EJ, Kromhout D; Alpha Omega Trial Group. Cardiovascular risk factor management of myocardial infarction patients with and without diabetes in the Netherlands between 2002 and 2006: a cross-sectional analysis of baseline data. BMJ Open 2012;2:e001360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Seshasai SR, Kaptoge S, Thomson A, Di Angelantonio E, Gao P, Sarwar N, Whincup PH, Mukamal KJ, Gillum RF, Holme I, et al. ; The Emerging Risk Factors Collaboration. Diabetes mellitus, fasting glucose, and risk of cause-specific death. N Engl J Med 2011;364:829–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hallan SI, Matsushita K, Sang Y, Mahmoodi BK, Black C, Ishani A, Kleefstra N, Naimark D, Roderick P, Tonelli M, et al. Age and association of kidney measures with mortality and end-stage renal disease. JAMA 2012;308:2349–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Smith SC Jr, Benjamin EJ, Bonow RO, Braun LT, Creager MA, Franklin BA, Gibbons RJ, Grundy SM, Hiratzka LF, Jones DW, et al. AHA/ACCF secondary prevention and risk reduction therapy for patients with coronary and other atherosclerotic vascular disease: 2011 update: a guideline from the American Heart Association and American College of Cardiology Foundation. Circulation 2011;124:2458–73. [DOI] [PubMed] [Google Scholar]
- 14.Guasch-Ferré M, Babio N, Martínez-González MA, Corella D, Ros E, Martín-Peláez S, Estruch R, Arós F, Gómez-Garcia E, Fiol M, et al. Dietary fat intake and risk of cardiovascular disease and all-cause mortality in a population at high risk of cardiovascular disease. Am J Clin Nutr 2015;102:1563–73. [DOI] [PubMed] [Google Scholar]
- 15.Kromhout D, Giltay EJ, Geleijnse JM, Alpha Omega Trial Group. N-3 fatty acids and cardiovascular events after myocardial infarction. N Engl J Med 2010;363:2015–26. [DOI] [PubMed] [Google Scholar]
- 16.Geleijnse JM, Giltay EJ, Schouten EG, de Goede J, Oude Griep LM, Teitsma-Jansen AM, Katan MB, Kromthout D, Alpha Omega Trial Group. Effect of low doses of n-3 fatty acids on cardiovascular diseases in 4837 post-myocardial infarction patients: design and baseline characteristics of the Alpha Omega Trial. Am Heart J 2010;159:539–46 e2. [DOI] [PubMed] [Google Scholar]
- 17.Feunekes IJ, Van Staveren WA, Graveland F, De Vos J, Burema J. Reproducibility of a semiquantitative food frequency questionnaire to assess the intake of fats and cholesterol in The Netherlands. Int J Food Sci Nutr 1995;46:117–23. [DOI] [PubMed] [Google Scholar]
- 18.Feunekes GI, Van Staveren WA, De Vries JH, Burema J, Hautvast JG. Relative and biomarker-based validity of a food-frequency questionnaire estimating intake of fats and cholesterol. Am J Clin Nutr 1993;58:489–96. [DOI] [PubMed] [Google Scholar]
- 19.Stichting NEVO. Dutch food consumption table 2006 (in Dutch: NEVO-tabel 2006). The Hague (Netherlands): Dutch Nutrition Centre; 2006. [Google Scholar]
- 20.Sijtsma FP, Soedamah-Muthu SS, de Goede J, Oude Griep LM, Geleijnse JM, Giltay EJ, de Boer MJ, Jacobs DR, Kromhout D. Healthy eating and lower mortality risk in a large cohort of cardiac patients who received state-of-the-art drug treatment. Am J Clin Nutr 2015;102:1527–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schuit AJ, Schouten EG, Westerterp KR, Saris WH. Validity of the Physical Activity Scale for the Elderly (PASE): according to energy expenditure assessed by the doubly labeled water method. J Clin Epidemiol 1997;50:541–6. [DOI] [PubMed] [Google Scholar]
- 22.Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med 2010;29:1037–57. [DOI] [PubMed] [Google Scholar]
- 23.Skeaff CM, Miller J. Dietary fat and coronary heart disease: summary of evidence from prospective cohort and randomised controlled trials. Ann Nutr Metab 2009;55:173–201. [DOI] [PubMed] [Google Scholar]
- 24.Ramsden CE, Hibbeln JR, Majchrzak SF, Davis JM. n-6 fatty acid-specific and mixed polyunsaturate dietary interventions have different effects on CHD risk: a meta-analysis of randomised controlled trials. Br J Nutr 2010;104:1586–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hooper L, Summerbell CD, Thompson R, Sills D, Roberts FG, Moore HJ, Davey Smith G. Reduced or modified dietary fat for preventing cardiovascular disease. Cochrane Database Syst Rev 2012;CD002137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ramsden CE, Zamora D, Leelarthaepin B, Majchrzak-Hong SF, Faurot KR, Suchindran CM, Ringel A, Davis JM, Hibbeln JR. Use of dietary linoleic acid for secondary prevention of coronary heart disease and death: evaluation of recovered data from the Sydney Diet Heart Study and updated meta-analysis. BMJ 2013;346:e8707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chowdhury R, Warnakula S, Kunutsor S, Crowe F, Ward HA, Johnson L, Franco OH, Butterworth AS, Forouhi NG, Thompson SG, et al. Association of dietary, circulating, and supplement fatty acids with coronary risk: a systematic review and meta-analysis. Ann Intern Med 2014;160:398–406. [DOI] [PubMed] [Google Scholar]
- 28.Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease. Am J Clin Nutr 2010;91:535–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Schwingshackl L, Hoffmann G. Dietary fatty acids in the secondary prevention of coronary heart disease: a systematic review, meta-analysis and meta-regression. BMJ Open 2014;4:e004487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ocké MC, Buurma-Rethans EJM, de Boer EJ, Wilson-van den Hooven C, Etemad-Ghameshlou Z, Drijvers JJMM, van Rossum CTM. Diet of community-dwelling older adults. Dutch National Food Consumption Survey Older Adults 2010-2012. Bilthoven (Netherlands): National Institute for Public Health and the Environment (RIVM); 2013. (RIVM report 050413001/2013.). [Google Scholar]