Table 2.
Prospective cohort studies published between 2008 and August 2018 that investigated the association between plasma and serum fatty acids and risk for diabetes
| Reference | Region | Study objective | Population | Adjustments | Time | Results | Risk |
|---|---|---|---|---|---|---|---|
| Patel et al. (2010) [22] | England | Investigate the association between FA composition and development of incident diabetes | 199 cases of incident diabetes and 184 non-cases among men and women, 40–79 years of age, who resided in and around Norwich, England, from the European Prospective Investigation into Cancer and Nutrition – Norfolk study | Fully adjusted model: age, sex, BMI, family history of diabetes, physical activity, smoking status, alcohol intake | Baseline 1993–1997 Measurements taken throughout, and up until 2005 |
There were stronger associations with diabetes risk when FA were measured in plasma vs. erythrocytes or by FFQ. Plasma FA only are reported here | |
| Third tertile vs. first tertile of total SFA (OR = 2.57; 95% CI 1.42, 4.66), palmitic acid (16:0) (OR = 2.47; 95% CI 1.37, 4.46), and Δ9 – SCD2 (18:1n-9/18:0) (OR = 2.01; 95% CI 1.12, 3.61) were associated with risk for diabetes | ↑ | ||||||
| Third tertile vs. first tertile of stearic acid (18:0) (OR = 0.43; 95% CI 0.24, 0.79), vaccenic acid (18:1n-7) (OR = 0.40; 95% CI 0.22, 0.72), eicosenoic acid (20:1n-9) (OR = 0.48; 95% CI 0.27, 0.87), linoleic acid (18:2n-6) (OR = 0.50; 95% CI 0.28, 0.91), dihomo-ϒ-linolenic acid (20:3n-6) (OR = 0.41; 95% CI 0.23, 0.74), and Δ5 – desaturase (D5D) (20:4n-6/20:3n-6) (OR = 0.47; 95% CI 0.26, 0.84) were inversely associated with risk for diabetes | ↓ | ||||||
| Mozaffarian et al. (2010) [23••] | USA | Investigate whether circulating trans-palmitoleic acid (trans-16:1n-7) was independently related to lower metabolic risk and incident T2D | 3736 men and women, 65 years of age and older, from the Cardiovascular Health Study | Age, sex, race, education, coronary heart disease, stroke, diabetes, smoking status, alcohol use, physical activity, BMI, dietary factors | 14 years | trans-palmitoleic acid (trans-16:1n-7) was inversely associated with insulin resistance (− 16.7%, P < 0.001) | ↓ |
| Fourth and highest quintiles vs. lowest quintile of trans-palmitoleic acid (trans-16:1n-7) (quintile 4 HR = 0.41; 95% CI 0.27, 0.64; quintile 5 HR = 0.38; 95% CI 0.24, 0.62) were inversely associated with incidence of diabetes | |||||||
| Djoussé et al. (2011) [24] | USA | Investigate the association between plasma phospholipid n-3 FA and incident diabetes | 3088 American men and women, 75 years of age on average, free from diabetes at baseline, from the Cardiovascular Health Study | Age, race, sex, clinic site, BMI, alcohol consumption, physical activity, smoking, linoleic acid, LDL-cholesterol | 10.6 years | Highest quartile vs. lowest quartile of plasma α-linolenic acid (RR = 0.57; 95% CI 0.36, 0.90) was inversely associated with T2D | ↓ |
| Mozaffarian et al. (2013) [25] | USA | Investigate the associations between trans-palmitoleic acid (trans-16:1n-7) and metabolic risk and incident diabetes | 2617 multi-ethnic men and women, 45–84 years of age, free from diabetes at baseline, from the Multi-Ethnic Study of Atherosclerosis cohort | Age, race/ethnicity, education, clinic, smoking status, alcohol use, physical activity, waist circumference, BMI | 5 years | trans-palmitoleic acid (trans-16:1n-7) was associated with lower fasting insulin (− 9.1%, P = 0.002) | ↓ |
| Highest quintile vs. lowest quintile of trans-palmitoleic acid (trans-16:1n-7) (HR = 0.52; 95% CI 0.32, 0.85) was inversely associated with incident diabetes | |||||||
| Mahendran et al. (2013) [26] | Finland | Cross-sectional and prospective studies (prospective pool only reported here) to investigate the associations between fasting serum glycerol and FA and predictors for worsening hyperglycemia and T2D | 4335 Finnish men (prospective pool only reported here), 57 years of age on average, free from diabetes at baseline from the Metabolic Syndrome in Men Study | Age, BMI, current smoking, physical activity | 4.5 years | Elevated glycerol (OR = 1.18, 95% CI 1.12, 1.24), FFA (OR = 1.19, 95% CI 1.10, 1.29), MUFA (OR = 1.09, 95% CI 1.06, 1.12), SFA, and monounsaturated n-7 and n-9 (OR = 1.09, 95% CI 1.06, 1.12) predicted worsening of hyperglycemia and development of incident T2D | ↑ |
| n-6 FA (OR = 0.92, 95% CI 0.89, 0.95) were associated with reduced risk for the worsening of hyperglycemia and conversion to T2D | ↓ | ||||||
| Santaren et al. (2014) [27] | USA | Prospective and cross-sectional studies (prospective pool only reported here) to investigate the associations between pentadecanoic acid (15:0) and trans-palmitoleic acid (trans-16:1n-7) and T2D | 659 multi-ethnic men and women, 40–60 years of age, free from diabetes at baseline, from the Insulin Resistance Atherosclerosis Study | Fully adjusted model: age, sex, ethnicity, physical activity, total energy intake, total dairy intake, total hydrogenated food intake, BMI | 5 years | Serum pentadecanoic acid (15:0) (OR = 0.73; 95% CI 0.56, 0.95) was inversely associated with incident diabetes risk | ↓ |
| Serum trans-palmitoleic acid (trans-16:1n-7) was not associated with T2D | ↔ | ||||||
| Mahendran et al. (2014) [28] | Finland | Investigate erythrocyte membrane fatty acids as predictors of worsening hyperglycemia and incident T2D | 1346 Finnish men 45–73 years of age free from diabetes at baseline from the Metabolic Syndrome in Men Study | Age, BMI, current smoking, physical activity | 5 years | Palmitoleic acid (16:1n-7) (2.8 × 10−7) dihomo-ϒ-linolenic acid (20:3n-6) (2.3 × 10−4), the ratio of 16:1n-7 to 16:0 (as a marker of desaturase activity) (1.6 × 10−8), and the ratio of 20:3n-6 to 18:2n-6 (as a marker of desaturase activity) (9.4 × 10−7) predicted the worsening of hyperglycemia | ↑ |
| Linoleic acid (18:2n-6) (P = 0.0015) and the ratio of 18:1n-7 to 16:1n-7 (as a marker of elongase activity) (P = 1.5 × 10−9) predicted a decrease in glucose AUC | ↓ | ||||||
| Palmitoleic acid (16:1n-7) (OR = 0.54, CI 95%: 0.35, 0.82) and linoleic acid (OR = 0.54, 95% CI 0.35, 0.82) were inversely associated with T2D | ↓ | ||||||
| n-3 PUFA did not show any associations with worsening hyperglycemia or T2D | ↔ | ||||||
| Virtanen et al. (2014) [29] | Finland | Investigate the associations between serum n-3 PUFA, EPA, DPA, DHA, α-linolenic acid, hair mercury and risk of incident T2D | 2212 Finnish men, 42–60 years of age, free from T2D at baseline, from the Kuopio Ischaemic Heart Disease Risk Factor study | Fully adjusted model: age, examination year, BMI, family history of T2D, smoking, years of education, leisure-time physical activity, alcohol intake, serum linoleic acid | 19.3 years | Highest vs. lowest quartile of EPA + DPA + DHA (HR = 0.67; 95% CI 0.51, 0.87) had inverse association with risk for T2D | ↓ |
| Lemaitre et al. (2015) [30] | USA | Investigate the association between plasma phospholipid very long-chain SFA (VLSFA) at baseline with subsequent incident diabetes | 3179 men and women, 75 years of age on average, free from diabetes at baseline, from the Cardiovascular Health Study | Age, sex, race, clinic, education, smoking, alcohol use, physical activity, treated hypertension, ischemic heart disease, self-reported health status, BMI, waist circumference | 18–19 years | Highest vs. lowest quartile of plasma concentration of arachidic acid (20:0) was associated with 32% lower risk for diabetes | ↓ |
| Ma et al. (2015) [31] | USA | Investigate the association of circulating palmitic acid (16:0), stearic acid (18:0), oleic acid (18:1n-9) and metabolic risk factors and incident diabetes | 3004 men and women, 74 years of age on average, free from diabetes at baseline, from the Cardiovascular Health Study | Age, sex, race, education, clinic, smoking status, alcohol consumption, leisure-time physical activity, prevalence of ischemic heart disease, hypertension at baseline, consumption of per cent energy from protein, per cent energy from carbohydrate, total energy | 18 years | Palmitic acid (16:0) (HR = 1.89. 95% CI 1.27, 2.83) and stearic acid (18:0) (HR = 1.62, 95% CI 1.08, 2.41 were associated with risk for diabetes | ↑ |
| Oleic acid (18:1n-9) was not associated with risk for diabetes | ↔ | ||||||
| Lankinen et al. (2015) [32] | Finland | Investigate fasting proportions of plasma fatty acids, estimated desaturases, and elongases as predictors for worsening glycasemia and incidence of T2D | 1364 Finnish men, 45–68 years of age, free from diabetes at baselines, from the Metabolic Syndrome in Men cohort | Age, BMI, smoking, physical activity at baseline, baseline fasting glucose | 5.9 years | Total SFA (P = 2.3 × 10−4), palmitoleic acid (16:1n-7) (P = 2.3 × 10−5), dihomo-ϒ-linolenic acid (20:3n-6) (P = 1.6 × 10−5). estimated stearoyl-CoA desaturase 1 (P = 2.3 × 10−5), and Δ6-desaturase (D6D) enzyme (P = 9.2 × 10−8) activities predicted the worsening of glycaemia | ↑ |
| PUFA, linoleic acid (18:2n-6) (P = 2.2 × 10−4), and elongase activity (P = 3.3 × 10−8) predicted a decrease in glucose AUC | ↓ | ||||||
| Estimated D6D activity (HR = 1.52; 95% CI 1.21, 1.92) and dihomo-ϒ-linolenic acid (20:3n-6) (HR = 1.46; 95% CI 1.16, 1.84) were associated with risk of incident T2D | ↑ | ||||||
| Steffan et al. (2015) [33] | USA | Investigate the association between serum levels of non-esterified FA and risk of T2D as well as any interaction by n-3 FA | 5697 multi-ethnic men and women, 45–84 years of age, free from diabetes at baseline, from the multi-Ethnic Study of Atherosclerosis | Fully adjusted model: age, sex, race, education, field center, current smoking, current alcohol intake, plasma n-3 FA, waist circumference, C-reactive protein | 11.4 years | Highest quartile vs. lowest quartile of non-esterified FA (HR = 1.86; 95% CI 1.45, 2.38) was associated with incidence diabetes. Higher diabetes incidence was found across successive quartiles | ↑ |
| Higher diabetes incidence was observed for individuals with n-3 levels below the 75th percentile | ↑ | ||||||
| No associations were observed in those with n-3 FA ≥ 75th percentile | ↔ | ||||||
| Takkunen et al. (2016) [34] | Finland | Investigate the associations between serum fatty acid composition and T2D, insulin secretion, and insulin sensitivity | 407 overweight men and women, 40–65 years of age, with impaired glucose tolerance at baseline, from the Finnish Diabetes Prevention Study | Age, sex, study group, study center, smoking, alcohol intake, waist circumference, physical activity at leisure | 11 years | 20:4n-6 (HR = 0.78; 95% CI 0.67, 0.95), 20:5n-3 (HR = 0.72; 95% CI 0.58, 0.88), 22:5n-3(HR = 0.74; 95% CI 0.69, 0.90), 22:6n-3 (HR = 0.73; 95% CI 0.59, 0.90), Δ5-desaturase (D5D) (HR = 0.78; 95% CI 0.64, 0.94), and total n-3 FA (HR = 0.75; 95% CI 0.57,0.86) were inversely associated with T2D | ↓ |
| Yakoob et al. (2016) [35] | USA | Investigate the associations between pentadecanoic acid (15:0), heptadecanoic acid (17:0), trans-palmitoleic acid (trans-16:1n-7), and incident diabetes | 3333 men and women, 30–75 years of age, free from diabetes at baseline, from the Nurses’ Health Study and Health Professionals Follow-Up Study | Age, race, smoking status, physical activity, alcohol, family history of diabetes, parental history of MI, hypercholesterolemia, hypertension, menopausal status, postmenopausal hormone use, fruits, vegetables, fish, meats, whole grains, sugar-sweetened beverages, polyunsaturated fat, calcium, glycemic load, biomarker levels of trans-18:1, trans-18:2, 16:0, 18:0, BMI | 15.2 years | Highest quartile vs. lowest quartile of plasma pentadecanoic acid (15:0) (HR = 0.56; 95% CI 0.39, 0.83), heptadecanoic acid (17:0) (HR = 0.57; 95% CI 0.38, 0.83), and trans-palmitoleic acid (trans-16:1n-7) (HR = 0.48; 95% CI 0.33, 0.70) were inversely associated with risk for diabetes | ↓ |
| Yary et al. (2016) [36] | Finland | Investigate the associations between serum n-6PUFA, Δ5-desaturase (D5D), Δ6-desaturase (D6D), and T2D risk | 2189 men, 42–60 years of age, free from T2D at baseline, from the Kuopio Ischaemic Heart Disease Risk Factor Study | Fully adjusted model: age, examination year, family history of T2D, BMI, smoking, education, leisure-time physical activity, alcohol intake, energy, serum long chain n-3 PUFA concentrations | 19.3 years | Highest quartile vs. lowest quartile of estimated D5D activity (HR = 0.55; 95% CI 0.41, 0.74), total n-6 PUFA (HR = 0.54; 95% CI 0.41, 0.73), linoleic acid (HR = 0.52; 95% CI 0.39, 0.70), and arachidonic acid (HR = 0.62; 95% CI 0.46, 0.85) were inversely associated with T2D | ↓ |
| Higher concentrations of dihomo-ϒ-linolenic acid (20:3n-6) (HR = 1.38; 95% CI 1.04, 1.84) and D6D activity (HR = 1.50; 95% CI 1.14, 1.97) were associated with risk for T2D | ↑ | ||||||
| Howard et al. (2018) [37] | USA | Randomized, parallel design: decreased-fat, increased vegetable, fruit, and grain vs. comparison diet | 48,835 postmenopausal women from the Women’s Health Initiative dietary intervention | 8.1 years | Decreased-fat, increased vegetable, fruit, and grain group had lower rates of initiation of insulin therapy during the intervention (HR = 0.74; 95% CI 0.59, 0.94) and follow-up (HR = 0.88; 95% CI 0.78, 0.99) | ↓ | |
| In subgroup analysis of biomarkers, the intervention reduced the risk of developing glucose ≥ 100 mg/dL (OR = 0.75; 95% CI 0.61, 0.93) |
BMI body mass index, CI confidence interval, DHA docosahexaenoic acid, DPA docopentaenoic acid, EPA eicosapentaenoic acid, FA fatty acid, FFQ food frequency questionnaire, HR hazard ratio, OR odds ratio, PUFA polyunsaturated fatty acid, RR relative risk, SCD stearoul-CoA desaturase, SFA saturated fatty acid, T2D type 2 diabetes