Table 1.
The effect of replacing SFA with other dietary macronutrients on cardiovascular outcomes.
| Study | Design | n | Mean Follow-Up Time (Years) | Outcome | Substitution | Result | Effect Size (95% CI) | Covariates Included in Analyses |
|---|---|---|---|---|---|---|---|---|
| Substitution of Saturated Fat for MUFA | ||||||||
| Jakobsen 2009 [42] | Pooled analysis of prospective cohort studies | 11 studies (n = 344,696) | Range 4 to 10 | Coronary events | 5% of energy from SFA → MUFA | ↔ | HR 1.19 (1.00–1.42) | Age; BMI; year survey completed; percentage of energy from MUFA, PUFA, trans-fat, protein and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension |
| Coronary deaths | ↔ | HR 1.01 (0.73–1.41) | ||||||
| Guasch-Ferré 2015 [43] [PREDIMED] | Prospective cohort | 7038 | 6 | CVD | 5% of energy from SFA → MUFA | ↓ | HR 0.63 (0.43–0.94) | Age; sex; BMI; intake of subtypes of fat, protein, and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension; intervention group; diabetes; hyper-cholesterolemia; family history of CHD; antihypertensive medication; oral antidiabetic agents; lipid lowering drugs |
| All-cause death | ↔ | HR 0.91 (0.65–1.26) | ||||||
| Li 2015 [44] [NHS; HPFS] | Prospective cohort | 127,536 | Range 24–30 | CHD | 5% of energy from SFA → MUFA | ↓ | HR 0.85 (0.74–0.97) | BMI, percentage of energy from protein; energy intake; smoking; physical activity; alcohol intake; cholesterol intake; hypertension at baseline; hypercholesterolemia at baseline; family history of myocardial infarction and diabetes; use of vitamins and aspirin |
| Praagman 2016 [31] [EPIC-Netherlands] | Prospective cohort | 35,597 | 12 | IHD | 5% of energy from SFA → cis-MUFA | ↑ | HR 1.30 (1.02–1.65) | Age, sex, BMI, waist circumference; intake of carbohydrate, cis-MUFA, PUFA, trans-fat, animal protein and vegetable (per 5% of energy); energy intake (excluding alcohol); smoking, physical activity; education; alcohol intake; fiber intake (energy adjusted); cholesterol intake (energy adjusted); vitamin c (energy adjusted) |
| Wang 2016 [45] [NHS; HPFS] | Prospective cohort | 126,233 | NHS ≤ 32; HPFS ≤ 26 | CVD mortality | 5% of energy from SFA → MUFA | ↔ | HR 0.96 (0.84–1.09) | Age; BMI, percentage of energy intake from protein, remaining fatty acids (saturated fat, PUFA, MUFA, trans-fat, ω-6 PUFAs, ω-3 PUFAs, linoleic acid, arachidonic acid, α-linolenic acid, and marine ω-3 fats); energy intake; smoking; physical activity; alcohol intake; cholesterol intake; family history of myocardial infarction, diabetes, cancer, hypertension, hyper-cholesterolemia; multivitamin use; vitamin E supplement; aspirin use; white race; marital status; menopausal status and hormone use in women |
| Total mortality | ↓ | HR 0.87 (0.82–0.93) | ||||||
| Zong 2016 [17] [NHS; HPFS] | Prospective cohort | 115,782 | NHS 25.8; HPFS 21.2 | CHD | 1% of energy from 12:0–18:0 SFA → MUFA | ↔ | HR 0.95 (0.90, 1.01) | Age; BMI; ethnicity; total energy; energy from trans-fat; energy from carbohydrates of non-whole grain sources; energy from non-plant sources; smoking status; physical activity; alcohol intake; family history of MI; menopausal status; postmenopausal hormone use; aspirin use; multivitamin use; baseline hypertension; baseline hypercholesterolemia; PUFA intake; whole grains intake; plant proteins intake; intake of other SFA |
| Hooper 2015 [46] Cochrane review | Meta-analysis of randomized controlled trials | 15 studies (n > 59,000) | >2 | CVD events | SFA → MUFA | ↔ | RR 1.00 (0.53–1.89) | Aggregate meta-analysis—no overall adjustment |
| Substitution of saturated fat for PUFA | ||||||||
| Mozaffarian 2010 [47] | Meta-analysis of randomized controlled trials | 8 studies (n = 13,614) | Median of all trials 4.25 | CHD | 5% of energy from SFA → total PUFA | ↓ | RR 0.90 (0.83–0.97) | Aggregate meta-analysis—no overall adjustment |
| Jakobsen 2009 [42] | Pooled analysis of prospective cohort studies | 11 studies (n = 344,696) | Range 4 to 10 | Coronary events | 5% of energy from SFA → total PUFA | ↓ | HR 0.87 (0.77–0.97) | Age; BMI; year survey completed; percentage of energy from MUFA, PUFA, trans-fat, protein and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension |
| Coronary deaths | ↓ | HR 0.74 (0.61–0.89) | ||||||
| Farvid 2014 [48] | Meta-analysis of prospective cohort studies | 13 studies (n = 310,602) | Range 5.3 to 30 | Coronary events | 5% of energy from SFA → linoleic acid | ↓ | RR 0.91 (0.87–0.96) | Aggregate meta-analysis—analyses in the individuals studies adjusted but no overall adjustment |
| Coronary deaths | ↓ | RR 0.87 (0.82–0.94) | ||||||
| Li 2015 [44] [NHS; HPFS] | Prospective cohort | 127,536 | Range 24–30 | CHD | 5% of energy from SFA → total PUFA | ↓ | HR 0.75 (0.67–0.84) | BMI, percentage of energy from protein; energy intake; smoking; physical activity; alcohol intake; cholesterol intake; hypertension at baseline; hypercholesterolemia at baseline; family history of myocardial infarction and diabetes; use of vitamins and aspirin |
| Guasch-Ferré 2015 [43] [PREDIMED] | Prospective cohort | 7038 | 6 | CVD | 5% of energy from SFA → PUFA | ↓ | HR 0.67 (0.45–0.98) | Age; sex; BMI; intake of subtypes of fat, protein, and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension; intervention group; diabetes; hyper-cholesterolemia; family history of CHD; antihypertensive medication; oral antidiabetic agents; lipid lowering drugs |
| All-cause mortality | ↓ | HR 0.61 (0.39–0.97) | ||||||
| Chen 2016 [41] [NHS; NHS II; HPFS] | Prospective cohort | 134,327 | NHS ≤ 32; NHS II ≤; HPFS ≤ 24 | CVD | 5% of energy from dairy fat → total PUFA | ↓ | HR 0.76 (0.71–0.81) | Age, BMI, intake of protein; energy intake; smoking; physical activity; intake of fruit, vegetables, coffee; alcohol intake; baseline hypertension; baseline hyper-cholesterolemia; race; menopausal status and menopausal hormone use (NHS and NHS II); oral contraceptive use (NHS II only) |
| CHD | ↓ | HR 0.74 (0.68–0.81) | ||||||
| Stroke | ↓ | HR 0.78 (0.70–0.88) | ||||||
| CVD | 5% of energy from dairy fat → n-6 PUFA | ↓ | HR 0.75 (0.70–0.81) | |||||
| CHD | ↓ | HR 0.75 (0.69–0.82) | ||||||
| Stroke | ↓ | HR 0.76 (0.68–0.86) | ||||||
| CVD | 0.3% of energy from dairy fat → α-linolenic acid | ↓ | HR 0.86 (0.82–0.90) | |||||
| CHD | ↓ | HR 0.83 (0.78–0.88) | ||||||
| Stroke | ↓ | HR 0.89 (0.83–0.96) | ||||||
| CVD | 0.3% of energy from dairy fat → marine n-3 | ↓ | HR 0.89 (0.84–0.94) | |||||
| CHD | ↓ | HR 0.87 (0.81–0.93) | ||||||
| Stroke | ↔ | HR 0.92 (0.84–1.01) | ||||||
| Praagman 2016 [31] [EPIC-Netherlands] | Prospective cohort | 35,597 | 12 | IHD | 5% of energy from SFA → PUFA | ↑ | HR 1.35 (1.14–1.61) | Age, sex, BMI, waist circumference; intake of carbohydrate, cis-MUFA, PUFA, trans-fat, animal protein and vegetable (per 5% of energy); energy intake (excluding alcohol); smoking, physical activity; education; alcohol intake; fiber intake (energy adjusted); cholesterol intake (energy adjusted); vitamin c (energy adjusted) |
| Wang 2016 [45] [NHS; HPFS] | Prospective cohort | 126,233 | NHS ≤ 32; HPFS ≤ 26 | CVD mortality | 5% of energy from SFA → total PUFA | ↓ | HR 0.72 (0.65–0.80) | Age; BMI, percentage of energy intake from protein, remaining fatty acids (saturated fat, PUFA, MUFA, trans-fat, ω-6 PUFAs, ω-3 PUFAs, linoleic acid, arachidonic acid, α-linolenic acid, and marine ω-3 fats); energy intake; smoking; physical activity; alcohol intake; cholesterol intake; family history of myocardial infarction, diabetes, cancer, hypertension, hyper-cholesterolemia; multivitamin use; vitamin E supplement; aspirin use; white race; marital status; menopausal status and hormone use in women |
| Total mortality | ↓ | HR 0.73 (0.70–0.77) | ||||||
| CVD mortality | 2% of energy from SFA → n-6 PUFA | ↓ | HR 0.89 (0.85–0.94) | |||||
| Total mortality | ↓ | HR 0.93 (0.91–0.96) | ||||||
| CVD mortality | 0.3% of energy from SFA → n-3 PUFA | ↔ | HR 1.01 (0.97–1.05) | |||||
| Total mortality | ↓ | HR 0.95 (0.93-0.96) | ||||||
| Zong 2016 [17] [NHS; HPFS] | Prospective cohort | 115,782 | NHS 25.8; HPFS 21.2 | CHD | 1% of energy from 12:0–18:0 SFA → PUFA | ↓ | HR 0.92 (0.89, 0.96) | Age; BMI; ethnicity; total energy; energy from trans-fat; energy from carbohydrates of non-whole grain sources; energy from non-plant sources; smoking status; physical activity; alcohol intake; family history of MI; menopausal status; postmenopausal hormone use; aspirin use; multivitamin use; baseline hypertension; baseline hypercholesterolemia; MUFA intake; whole grain intake; plant protein intake; intake of other SFA |
| Hooper 2015 [46] Cochrane review | Meta-analysis of randomized controlled trials | 15 studies (n > 59,000) | >2 | CVD events | SFA → PUFA | ↓ | RR 0.73 (0.58–0.92) | Aggregate meta-analysis—no overall adjustment |
| Substitution of Saturated Fat for Carbohydrate | ||||||||
| Jakobsen 2009 [42] | Pooled analysis of prospective cohort studies | 11 studies (n = 344,696) | Range 4 to 10 | Coronary events | 5% of energy from SFA → total carbohydrate | ↑ | HR 1.07 (1.01–1.14) | Age; BMI; year survey completed; percentage of energy from MUFA, PUFA, trans-fat, protein and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension |
| Coronary deaths | 5% of energy from SFA → total carbohydrate | ↔ | HR 0.96 (0.82–1.13) | |||||
| Jakobsen 2010 [49] | Prospective cohort | 53,644 | Median 12 | MI | 5% of energy from SFA → total carbohydrates | ↔ | HR 1.04 (0.92–1.17) | Age, sex, BMI; percentage of energy from glycemic carbohydrates, proteins, MUFA, PUFA; energy intake; smoking; physical activity; education; alcohol consumer; intake of alcohol; hypertension |
| 5% of energy from SFA → carbohydrates with low-GI (median GI 82) | ↔ | HR 0.88 (0.72–1.07) | ||||||
| 5% of energy from SFA → carbohydrates with medium-GI (median GI 88) | ↔ | HR 0.98 (0.80–1.21) | ||||||
| 5% of energy from SFA → carbohydrates with high-GI (median GI 93) | ↑ | HR 1.33 (1.08–1.64) | ||||||
| Guasch-Ferré 2015 [43] [PREDIMED] | Prospective cohort | 7038 | 6 | CVD | 5% of energy from SFA→ total carbohydrate | ↔ | HR 0.83 (0.63–1.10) | Age; sex; BMI; intake of subtypes of fat, protein, and carbohydrates; energy intake; smoking; physical activity; education; alcohol intake; fiber intake; cholesterol intake; hypertension; intervention group; diabetes; hyper-cholesterolemia; family history of CHD; antihypertensive medication; oral antidiabetic agents; lipid lowering drugs |
| All-cause death | ↔ | HR 1.04 (0.81–1.33) | ||||||
| Li 2015 [44] [NHS; HPFS] | Prospective cohort | 127,536 | Range 24–30 | CHD | 5% of energy from SFA → whole grains | ↓ | HR 0.91 (0.85–0.98) | BMI, percentage of energy from protein; energy intake; smoking; physical activity; alcohol intake; cholesterol intake; hypertension at baseline; hypercholesterolemia at baseline; family history of myocardial infarction and diabetes; use of vitamins and aspirin |
| 5% of energy from SFA → refined starches/added sugar | ↔ | Not reported | ||||||
| Zong 2016 [17] [NHS; HPFS] | Prospective cohort | 115,782 | NHS 25.8; HPFS 21.2 | CHD | 1% of energy from 12:0–18:0 SFA → whole grains | ↓ | HR 0.94 (0.91, 0.97) | Age; BMI; ethnicity; total energy; energy from trans-fat; energy from carbohydrates of non-whole grain sources; energy from non-plant sources; smoking status; physical activity; alcohol intake; family history of MI; menopausal status; postmenopausal hormone use; aspirin use; multivitamin use; baseline hypertension; baseline hypercholesterolemia; MUFA intake; PUFA intake; plant protein intake; intake of other SFA |
| Chen 2016 [41] [NHS; NHS II; HPFS] | Prospective cohort | 134,327 | NHS ≤ 32; NHS II ≤ 20; HPFS ≤ 24 | CVD | 5% of energy from dairy fat → carbohydrate from whole grains | ↓ | HR 0.72 (0.69–0.75) | Age, BMI, intake of protein; energy intake; smoking; physical activity; intake of fruit, vegetables, coffee; alcohol intake; baseline hypertension; baseline hyper-cholesterolemia; race; menopausal status and menopausal hormone use (NHS and NHS II); oral contraceptive use (NHS II only) |
| CHD | ↓ | HR 0.66 (0.62–0.70) | ||||||
| Stroke | ↓ | HR 0.84 (0.78–0.91) | ||||||
| CVD | 5% of energy from dairy fat → carbohydrate from refined starch and added sugar | ↔ | HR 0.97 (0.94–1.00) | |||||
| CHD | ↔ | HR 0.96 (0.93–1.00) | ||||||
| Stroke | ↔ | HR 0.98 (0.94–1.03) | ||||||
| Praagman 2016 [31] [EPIC-NL] | Prospective cohort | 35,597 | 12 | IHD | 5% of energy from SFA → total carbohydrates | ↑ | HR (1.23 (1.09–1.40) | Age, sex, BMI, waist circumference; intake of carbohydrate, cis-MUFA, PUFA, trans-fat, animal protein and vegetable (per 5% of energy); energy intake (excluding alcohol); smoking, physical activity; education; alcohol intake; fiber intake (energy adjusted); cholesterol intake (energy adjusted); vitamin c (energy adjusted) |
| 5% of energy from SFA → carbohydrates with low GI (GI < 53) | ↔ | HR 1.14 (0.91–1.43) | ||||||
| 5% of energy from SFA → carbohydrates with medium GI | ↑ | HR 1.35 (1.05–1.73) | ||||||
| 5% of energy from SFA → carbohydrates with high GI (GI > 56) | ↑ | HR 1.27 (1.03–1.56) | ||||||
| Hooper 2015 [46] Cochrane review | Meta-analysis of randomized controlled trials | 15 studies (n > 59,000) | >2 | CVD events | SFA → carbohydrate | ↔ | RR 0.93 (0.79–1.08) | Aggregate meta-analysis—no overall adjustment |
| Substitution of Saturated Fat for Protein | ||||||||
| Larsson 2012 [50] | Prospective cohort | 34,670 | Median 10.4 | Stroke | 5% of energy from SFA → protein | ↓ | 13% lower risk (0–26%) | Age, BMI; intake of fat; energy intake; smoking status and smoking pack years; physical activity; education; alcohol intake; intake of cholesterol, calcium, fruits and vegetables; hypertension; diabetes; aspirin use; family history of myocardial infarction |
| Praagman 2016 [31] [EPIC-NL] | Prospective cohort | 35,597 | 12 | IHD | 5% of energy from SFA → total protein | ↑ | HR 1.29 (1.08–1.54) | Age, sex, BMI, waist circumference; intake of carbohydrate, cis-MUFA, PUFA, trans-fat, animal protein and vegetable (per 5% of energy); energy intake (excluding alcohol); smoking, physical activity; education; alcohol intake; fiber intake (energy adjusted); cholesterol intake (energy adjusted); vitamin c (energy adjusted) |
| 5% of energy from SFA → animal protein | ↑ | HR 1.37 (1.14–1.65) | ||||||
| 5% of energy from SFA → vegetable protein | ↔ | HR 0.81 (0.57–1.17) | ||||||
| Zong 2016 [17] [NHS; HPFS] | Prospective cohort | 115,782 | NHS 25.8; HPFS 21.2 | CHD | 1% of energy from 12:0–18:0 SFA → plant protein | ↓ | HR 0.93 (0.89, 0.97) | Age; BMI; ethnicity; total energy; energy from trans-fat; energy from carbohydrates of non-whole grain sources; energy from non-plant sources; smoking status; physical activity; alcohol intake; family history of MI; menopausal status; postmenopausal hormone use; aspirin use; multivitamin use; baseline hypertension; baseline hypercholesterolemia; MUFA intake; whole grain intake; intake of other SFA |
| Hooper 2015 [46] Cochrane review | Meta-analysis of randomized controlled trials | 15 studies (n > 59,000) | >2 | CVD events | SFA → protein | ↔ | RR 0.98 (0.90–1.06) | Aggregate meta-analysis—no overall adjustment |