Intake levels of dietary long-chain PUFAs modify the association between genetic variation in FADS and LDL-C.

Polymorphisms of the FA desaturase (FADS) gene cluster have been associated with LDL, HDL, and triglyceride concentrations. Because FADS converts α-linolenic acid (ALA) and linoleic acid into PUFAs, we investigated the interaction between different PUFA intakes and the FADS polymorphism rs174547 (T>C) on fasting blood lipid and lipoprotein concentrations. We included 4,635 individuals (60% females, 45-68 years) from the Swedish population-based Malmö Diet and Cancer cohort. Dietary intakes were assessed by a modified diet history method including 7-day registration of cooked meals. The C-allele of rs174547 was associated with lower LDL concentration (P = 0.03). We observed significant interaction between rs174547 and long-chain ω-3 PUFA intakes on LDL (P = 0.01); the C-allele was only associated with lower LDL among individuals in the lowest tertile of long-chain ω-3 PUFA intakes (P < 0.001). In addition, significant interaction was observed between rs174547 and the ratio of ALA and linoleic FA intakes on HDL (P = 0.03). However, no significant associations between the C-allele and HDL were detected within the intake tertiles of the ratio. Our findings suggest that dietary intake levels of different PUFAs modify the associated effect of genetic variation in FADS on LDL and HDL.

The blood concentrations of LDL-cholesterol, HDL-cholesterol, and triglycerides have a strong genetic influence (1, 2). In recent genome-wide association studies, single-nucleotide polymorphisms (SNPs) in the fatty acid desaturase (FADS) gene cluster that includes FADS1, FADS2, and FADS3 genes, were associated with LDL, HDL, triglyceride (3–5) and fasting glucose concentrations (6–8). First, an association was observed between the C-allele of rs174547 in FADS1 and decreased HDL and increased triglyceride concentrations in a study comprising almost 40,000 Europeans (4). The same allele was associated with lower mRNA levels of FADS1 and FADS3 in liver (4). More recently, a meta-analysis of >100,000 individuals reported genome-wide significant associations of the FADS locus and LDL, HDL, as well as triglyceride concentrations (5).

FADSs are key enzymes in the endogenous desaturation of α-linolenic acid (ALA, C18:3ω-3) and linoleic acid (LA, C18:2ω-6) into long-chain PUFAs, where FADS1 is a Δ-5 desaturase and FADS2 a Δ-6 desaturase (9, 10). Further, SNPs in the FADS locus have been associated with blood concentrations of long-chain PUFAs as well as with cholesterol concentrations (9, 11–16). Long-chain PUFAs regulate the fluidity of cell membrane, act as second messengers in intracellular signaling pathways, and regulate transcription (17). Dietary intakes of the long-chain ω-3 PUFAs docosahexanoic acid (DHA, C22:6ω-3) and eicosapentanoic acid (EPA, C20:5ω-3) have been reported to lower serum triglyceride levels (18), and higher dietary intake of ω-3 was associated with higher HDL and LDL in the Malmö Diet and Cancer (MDC) cohort (19). Additionally, long-chain PUFAs, such as arachidonic acid (AA, C20:4ω-6) and EPA, are precursors for inflammatory molecules such as eicosanoids (9, 20), and high long-chain PUFA concentration has been associated with lower prevalence of both metabolic syndrome and cardiovascular disease (9, 10). Further, previous studies have suggested that dietary intake levels of different PUFAs interact with FADS1 variation to affect blood lipids (13, 21).

Because FADS are key regulators of ALA and LA desaturation, we examined whether different dietary intake levels of PUFAs modify the association between the rs174547 polymorphism in the FADS locus and blood concentrations of LDL, HDL, and triglycerides.

SUBJECTS AND METHODS

Study population

The MDC cohort is a population-based prospective cohort including 28,449 participants, with baseline data collection conducted throughout the years 1991–96 (22). The study population includes individuals born during 1923–50 (23) and living in Malmö, the third-largest city in Sweden, with about 295,000 citizens. The participation rate was approximately 40% (24). Among MDC participants recruited from November 1991 to February 1994 (n = 12,445), a random 50% was invited to further participate in a carotid artery disease study, the MDC Cardiovascular Cohort. In total, 6,103 individuals underwent a review of their medical history, a physical examination and a laboratory assessment of cardiovascular risk factors (22, 24, 25). Information on LDL, HDL, and triglyceride fasting blood concentrations was available in 5,363 individuals. Totally, information on diet, FADS genotype, and blood lipids was available in 4,943 individuals. After excluding individuals with diabetes mellitus (self-reported diagnosis or using anti-diabetes medication, n = 123), users of lipid-lowering medication (n = 117), and those with a history of cardiovascular event (coronary event or stroke, n = 117), the study sample included 4,635 individuals (45–68 years, 60.3% female). Information about the history of cardiovascular event (coronary event or stroke) was taken from the national Swedish Hospital Discharge register and the local register of stroke (24). All individuals provided written, informed consent, and the ethics committee of Lund University approved the MDC study protocols.

Genotyping

The genotyping of the FADS rs174547 (T/C) was performed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry on the Sequenom Mass-ARRAY platform (San Diego, CA). Genotyping was successful in 5,806 (96%) of the 6,055 individuals with DNA available and the rs174547 was in Hardy-Weinberg equilibrium (P = 0.92). In addition, 5,490 of the 6,055 individuals, of whom we still have DNA available, were additionally genotyped by the TaqMan allelic discrimination on ABI 7900 with a concordance rate for the two methods of 99.2%.

Dietary information

Dietary intake was measured by a modified diet history methodology combining a 168 item dietary questionnaire, a 7 day menu book and a 1 h diet history interview, especially designed for the MDC study (26). The 168 item dietary questionnaire covered food items regularly consumed during the past year. The participants were asked to fill in the frequency of food intake and estimate the usual portion sizes using a booklet with photographic aids. The 7 day menu book covered cooked lunch and dinner meals, cold beverages (including alcoholic beverages), medications, natural remedies, and dietary supplements used by the participants. To complete the dietary data, the participants were interviewed, for 1 h about their food choices, food preparation practices, and portion sizes (by using a more-extensive booklet of photos) of the food reported in the menu book. The trained interviewers checked the menu book and questionnaire for very high reported intakes and overlapping information. The average daily food intake (grams per day) was calculated based on the information from the menu book, interview, and questionnaire, and converted into nutrient and energy intakes by using the MDC Food and Nutrient Database, developed from the PC KOST-93 of the Swedish National Food Administration (26).

The different PUFA intakes were converted into the percentage contributed by the specific PUFAs to total energy intake (E%). The PUFA dietary variables included in this study were: ALA (C18:3ω-3); long-chain ω-3 PUFA (EPA [C20:5ω-3], docosapentanoic acid [DPA, C22:5ω-3], and DHA [C22:6ω-3]); total ω-3 PUFA (ALA, EPA, DPA, and DHA); total ω-6 PUFA (LA [C18:2ω-6], γ-linolenic acid [C18:3ω-6], and AA [C20:4ω-6]); ALA to LA ratio (ALA/LA); total ω-3 to total ω-6 PUFA ratio (ω-3/ω-6 PUFA).

The relative validity of the modified diet history method was examined in 105 women and 101 men. As the reference method, a total of 18 days of weighed food records was collected during 3 consecutive days, every second month for 1 year. The energy-adjusted correlation coefficients between the modified diet history method and the reference method were in men: 0.22 for ALA, 0.23 for LA, 0.55 for AA, 0.24 for EPA, 0.37 for DPA, and 0.20 for DHA. In women, the coefficients were 0.58 for ALA, 0.68 for LA, 0.44 for AA, 0.38 for EPA, 0.40 for DPA, and 0.27 for DHA (27).

Other variables

HDL, triglyceride, and total cholesterol concentrations were determined in overnight fasting blood samples by standard methods at the department of Clinical Chemistry, University Hospital of Malmö. LDL concentration was calculated by the Friedewald formula: LDL = total cholesterol – HDL – (triglycerides/2.2) (24). For individuals with a triglyceride concentration of more than 400 mg per deciliter (4.5 mmol/l), LDL was defined as missing (25). Body mass index (BMI) was calculated as weight in kilograms divided by square of height in meters (kg/m2).

A self-administered questionnaire was used to determine lifestyle factors including cigarette smoking, alcohol intake, and physical activity habits. Three categories of smoking status were used: current (including irregular smoking), former, and never smokers. Alcohol habits were divided into five categories. Individuals reporting no alcohol consumption during the last year in the questionnaire, which also were zero reporters of alcohol in the 7 day menu book, were categorized as zero consumers of alcohol. We divided the other study participants’ alcohol consumption (grams per day) into categories with different cutoffs according to gender. The cutoff levels for females were 5, 10, and 20 g of alcohol per day and the cutoff levels for males were 10, 20, and 40 g of alcohol per day. The leisure-time physical activity level (PAL) was calculated from a list of 17 different activities in the questionnaire. The time spent on each activity was multiplied by an intensity factor, creating a leisure-time physical activity score. The leisure-time physical activity score was then divided into quintiles, with the same cutoffs for both genders. Separate categories for smoking, alcohol intake, and leisure-time physical activity were constructed for the subjects with missing data.

Statistical analysis

SPSS, Inc. PASW Statistics 18.0 was used for statistical analysis. Statistical significance was set at P < 0.05, and all P values are 2-sided. All of the covariates, except season, were differentially distributed between tertiles of total ω-3 and ω-6 PUFA intake. Assuming an additive model, associations with the FADS rs174547 (T/C) genotype categories were investigated using the General Linear Model adjusted for age and sex (basic analysis), and thereafter, age, sex, and BMI. Interaction between FADS genotype and tertiles of dietary intake levels on serum lipid concentrations were studied by introducing a multiplicative factor of genotypes and diet tertiles as continuous variables in addition to these main factors as separate variables. The interaction analyses were adjusted for potential confounders: age, sex, BMI, season of diet collection (four categories), cigarette smoking, leisure-time physical activity, alcohol intake, and total energy intake. All continuous variables except age were Ln-transformed to achieve normal distribution when testing for trend across FADS genotype categories and interaction between FADS genotype categories and tertiles of dietary intake levels of PUFA on LDL, HDL, and triglycerides; before transformation, a very small amount (0.001 g) was added to ω-3 PUFA intake to handle zero intakes.

In sensitivity analyses, potential misreporters of energy were excluded. Misreporters of energy intake were identified by comparing the individually estimated PAL expressed as the energy expenditure divided by the basal metabolic rate (BMR), with energy intake divided by BMR, further explained elsewhere (28). Individuals were defined as misreporters when the ratio of the reported energy intake to BMR was outside the 95% confidence limits of the calculated PAL (i.e., under- and over-reporters).

RESULTS

Each C-allele of rs174547 associated with 0.05 mmol/l lower LDL concentration (P trend = 0.03, ), but not with HDL or triglyceride concentrations (P trend = 1.00 and P trend = 0.10, respectively) in the basic analysis. BMI was significantly associated with genotype, and when BMI was included as a covariate, a significant association of 0.02 mmol/l higher triglyceride concentration per C-allele was observed (P trend = 0.04). Similar to the basic analysis, the association with LDL concentration remained significant (P trend = 0.047), and no association with HDL concentration was detected (P trend = 0.52) after adjusting for BMI.

TABLE 1.

Characteristics of the Malmö Diet and Cancer cardiovascular cohort individuals by FADS1 rs174547 genotype

Characteristics	All(N = 4,635)	T/T(N = 2,054)	T/C(N = 2,056)	C/C(N = 525)	P trenda
Women, N (%)	2,795 (60.3)	1,227 (59.7)	1,240 (60.3)	328 (62.5)	0.52
Age (y)	57.7 (52.3–62.6)	57.6 (52.2–62.3)	57.7 (52.2–62.6)	58.0 (52.2–62.7)	0.54
BMI (kg/m )	25.1 (22.8–27.7)	25.3 (23.0–27.8)	25.0 (22.9–27.3)	24.8 (22.7–27.3)	0.047
Fasting glucose (mmol/l)	4.9 (4.6–5.3)	4.9 (4.6–5.3)	4.9 (4.6–5.3)	4.9 (4.6–5.2)	0.44
LDL cholesterol (mmol/l)	4.10 (3.5–4.8)	4.10 (3.5–4.8)	4.10 (3.5–4.8)	4.00 (3.4–4.7)	0.03
HDL cholesterol (mmol/l)	1.35 (1.1–1.6)	1.36 (1.1–1.6)	1.36 (1.1–1.6)	1.36 (1.1–1.6)	1.00
Triglycerides (mmol/l)	1.14 (0.9–1.6)	1.13 (0.9–1.5)	1.13 (0.9–1.6)	1.18 (0.9–1.6)	0.10
Dietary intake
ALA (E%)	0.73 (0.6–0.9)	0.73 (0.6–0.9)	0.72 (0.6–0.9)	0.71 (0.6–0.8)	0.06
Long-chain ω-3 PUFA (E%)	0.20 (0.1–0.3)	0.20 (0.1–0.3)	0.20 (0.1–0.3)	0.21 (0.1–0.4)	0.41
Total ω-3 PUFA (E%)	0.97 (0.8–1.2)	0.98 (0.8–1.2)	0.97 (0.8–1.2)	0.96 (0.8–1.2)	0.47
Total ω-6 PUFA (E%)	4.89 (4.0–5.9)	4.93 (4.1–5.9)	4.87 (4.0–5.9)	4.83 (4.0–5.7)	0.27
ALA/LA (E%)	0.15 (0.1–0.2)	0.15 (0.1–0.2)	0.15 (0.1–0.2)	0.15 (0.1–0.2)	0.44
Total ω-3/ω-6 PUFA (E%)	0.19 (0.2–0.2)	0.19 (0.2–0.2)	0.19 (0.2–0.2)	0.19 (0.2–0.2)	0.72

Data are median (inter-quartile range) or number (%), if not otherwise indicated. ALA, α-linolenic acid; BMI, body mass index; E%, energy percentage of total energy intake; LA, linoleic acid; PUFA, polyunsaturated fatty acid. Sex, age, diet variables, and LDL, HDL, and triglycerides (n = 4,635). BMI (n = 4,633). Fasting glucose (n = 4,624).

Ln transformed. Adjusted for age and sex.

PUFA intakes did not differ according to FADS genotypes (Table 1). No significant interaction was observed between the FADS genotype categories and intake levels of total ω-3 PUFAs on LDL concentration (). However, we observed a significant interaction between the FADS genotype categories and long-chain ω-3 PUFA intake on LDL concentration (P = 0.01). The C-allele was significantly associated with lower LDL among individuals within the lowest tertile of long-chain ω-3 PUFA intake (≤0.14 E%, P < 0.001), but not among those in the mid- (0.14–0.28 E%) or highest tertiles (>0.28 E%). When examining within-genotype categories, the high long-chain ω-3 PUFA intake was associated significantly with higher LDL concentration among the CC-genotype (P < 0.001) and TC-genotype carriers (P = 0.04) but not among TT-genotype carriers (P = 0.17) (). In addition, there was a significant interaction between the FADS genotype categories and ALA/LA intakes on HDL concentration (P = 0.03) despite lack of significant associations between the FADS genotypes and HDL concentration in any of the ALA/LA tertiles. However, we observed significant associations between ALA/LA and HDL concentration among CC- (P = 0.046) and TC-genotype carriers (P = 0.02) but not among those with the TT-genotype (). No significant interactions were observed between the FADS genotype categories and any of the different PUFA intake levels on triglyceride concentration (Table 2).

TABLE 2.

Association between rs174547 (T/C) for each additional C-allele and blood lipids in strata of diet intakes among 4,635 individuals

	LDL Cholesterol			HDL Cholesterol			Triglycerides
Diet variables (E%)	Effect size	P-trenda	P-intb	Effect size	P-trenda	P-intb	Effect size	P-trenda	P-intb
ALA			0.94			0.55			0.47
Low (≤0.65)	−0.034	0.39		−0.003	0.81		0.024	0.37
Medium (0.65–0.80)	−0.078	0.03		0.003	0.98		0.028	0.29
High (≥0.80)	−0.021	0.55		0.009	0.71		0.013	0.89
Long-chain ω-3 PUFA			0.01			0.53			0.70
Low (≤0.14)	−0.138	<0.001		−0.006	0.43		0.025	0.30
Medium (0.14–0.28)	0.001	0.98		0.007	0.69		0.020	0.53
High (≥0.28)	−0.007	0.86		0.004	0.95		0.025	0.49
Total ω-3 PUFA			0.38			0.78			0.87
Low (≤0.86)	−0.071	0.05		0.004	0.67		0.006	0.91
Medium (0.86–1.09)	−0.049	0.24		0.005	0.93		0.037	0.17
High (≥1.09)	−0.013	0.62		0.001	0.78		0.024	0.55
Total ω-6 PUFA			0.35			0.90			0.90
Low (≤4.35)	−0.052	0.16		0.002	0.89		0.013	0.53
Medium (4.35–5.48)	−0.074	0.04		0.004	0.97		0.025	0.33
High (≥5.48)	−0.007	0.83		0.002	0.86		0.030	0.53
ALA/LA			0.49			0.03			0.15
Low (≤0.14)	−0.022	0.51		−0.020	0.07		0.041	0.17
Medium (0.14–0.16)	−0.055	0.12		0.015	0.35		0.040	0.25
High (≥0.16)	−0.059	0.12		0.014	0.35		−0.013	0.78
Total ω-3/ω-6 PUFA			0.73			0.26			0.16
Low (≤0.17)	−0.056	0.11		−0.002	0.58		0.035	0.30
Medium (0.17–0.22)	−0.018	0.77		−0.007	0.67		0.043	0.10
High (≥0.22)	−0.064	0.05		0.015	0.38		−0.007	0.64

Effect size (β) = difference in lipid concentration for each additional C-allele. ALA, α-linolenic acid; BMI, body mass index; E%, energy percentage of total energy intake; LA, linoleic acid; PUFA, polyunsaturated fatty acid.

Adjusted for age and sex.

Adjusted for age, sex, season, alcohol intake, cigarette smoking, leisure time physical activity, BMI, and total energy intake.

Fig. 1.

Association between rs174547 and LDL in strata of long-chain ω-3 PUFA among 4,635 individuals. The CC-genotype was associated with 0.14 mmol/l lower LDL concentration among individuals with low intakes of long-chain ω-3 PUFA (≤0.14 E%, P = 2.3 × 10−4), but not among those with medium (0.14–0.28 E%) or high (>0.28 E%) intakes (P = 0.98 and P = 0.86, respectively). A low intake of long-chain ω-3 PUFA was associated with 0.20 mmol/l and 0.06 mmol/l higher HDL concentration among individuals with CC-genotype (P = 2.7 × 10−3) and TC-genotype (P = 0.04) but not among those with TT-genotype (P = 0.17).

Fig. 2.

Association between ALA/LA and HDL in strata of rs174547 among 4,635 individuals. A high ALA/LA was associated with 0.04 mmol/l and 0.02 mmol/l higher HDL concentration among individuals with CC-genotype (P = 0.046) and TC-genotype (P = 0.02) but not among those with TT-genotype (P = 0.96). There was no association between the CC genotype and HDL in individuals with low, medium, or high ALA/LA (P = 0.07, P = 0.35, and P = 0.35, respectively).

We also examined interactions separately for men and women. We observed a significant interaction of the FADS genotype categories and long-chain ω-3 PUFA on LDL concentrations in men (P = 0.01), but not in women (P = 0.39). However, these interactions did not significantly differ between the genders (P = 0.08), and there was no significant difference in gender distribution across the tertiles of long-chain ω-3 PUFA. No significant interactions of FADS genotype categories and ALA/LA on the levels of HDL were detected, or any of the different PUFA intakes on triglycerides levels, either in men or in women.

There were significantly more males than females in the highest tertiles of total ω-3 and ω-6 PUFA intake compared with the mid- and lowest tertiles (P < 0.001). Because the gender distribution across the tertiles of ω-3 and ω-6 PUFA intakes was not equal, we repeated all of the interaction analyses using gender-specific cutoffs for the long-chain ω-3 and ω-6 PUFA tertiles. However, the above-reported results remained essentially the same when using gender-specific cutoffs.

In sensitivity analyses, when excluding suspected misreporters of energy (19.3% of the study sample), the interaction between the FADS genotype categories and long-chain ω-3 PUFA intakes on LDL remained significant (P = 0.04). The interaction between the FADS genotype categories and ALA/LA intake levels on HDL cholesterol was slightly attenuated after excluding misreporters (P = 0.06). Interactions between FADS genotype categories and the different PUFA intake levels on triglyceride remained nonsignificant.

DISCUSSION

This study indicates that the dietary intake levels of long-chain ω-3 PUFAs may modify the association between FADS rs174547 and LDL concentration. Additionally, the intake levels of ALA/LA may modify the associated effect of FADS genotype categories on HDL concentrations.

Although several earlier studies, similar to our results, have found an association between the minor allele of rs174547 (or an SNP in high allelic association) and decreased LDL concentrations (6, 13, 29–31), some inconsistency between the reported associations with lipids and lipoproteins remains. Kathiresan et al. (4) demonstrated that the minor C-allele of rs174547 associated with lower HDL and higher triglyceride concentrations in Caucasians and similar associations were observed in a Japanese population (21). However, because the same C-allele was found to associate significantly with decreased LDL, but not with HDL or triglycerides in a Mongolian population, it was suggested that differences in dietary fish intake between Japanese and Mongolians could provide an explanation of the dissimilar results (21). The Mongolian diet is mainly based on livestock products, with a very low intake of fish compared with the Japanese diet, which, in general, includes a very large amount of fish products. Consistently, Japanese individuals had higher plasma concentrations of ω-3 PUFA and a higher ω-3/ω-6 PUFA compared with Mongolians (32). In our study, the rs174547 C-allele was associated with lower LDL concentration only in the lowest intake tertile of long-chain ω-3 PUFA. Because the dietary source of long-chain fatty acids mainly comes from fatty fish, our finding is in agreement with the suggestion of Nakayama et al. (21) that among individuals with low fish intake, in our study represented by those in the lowest intake tertile of long-chain PUFAs, the association between the FADS genotype categories and LDL concentrations may be accentuated.

In contrast to Lu et al. (13), who reported a significant association between the minor allele of rs174546 (corresponding to the C-allele of rs174547 of our study) and lower HDL concentration among individuals with high consumption of ω-6 PUFAs (≥5.26 E%), we observed no significant interaction between ω-6 PUFA intakes and the FADS genotype on HDL concentration. However, because low intake of ω-6 PUFA could indirectly reflect a high ALA/LA intake, our observation of interaction between the FADS genotype categories and ALA/LA intakes on HDL cholesterol concentration, as well as association between high ALA/LA intake and higher HDL concentration among the C-allele carriers, could be in line with findings by Lu et al.

FADS1 is responsible for the desaturation of ALA and LA into EPA and AA (9, 10). Kathiresan et al. (4) demonstrated that the minor allele of rs174547 (in our study associated with lower LDL concentration) was associated with lower gene expression of the FADS1 and FADS3 transcripts. Tanaka et al. (31) showed that individuals homozygous for the minor allele of an SNP in high allelic association with FADS rs174547 had lower plasma concentrations of AA compared with individuals homozygote for the major allele. The SNP accounted for 18.6% of the additive variance in AA concentration. The potential mechanism for association between high ALA/LA intake and higher HDL among the minor allele carriers could be that the minor allele, which has been associated with a lower desaturase activity (14), may affect the availability of long-chain ω-3 PUFAs differently between the genotypes. The metabolic derivatives, eicosanoids, of these long-chain ω-3 PUFAs are potent activators of peroxisome proliferator-activated receptors (9, 20), which can regulate the expression of genes directly involved in HDL production (13). Therefore, our result may reflect a limited long-chain PUFA availability among rs174547 minor allele carriers and thereby association between HDL concentrations and ALA/LA intake.

The strengths of our study include a relatively large sample size and detailed information on dietary intakes based on a 168 item dietary questionnaire, a 7 day menu book, and a 1 h interview. However, the estimated dietary intakes were based on self-reports, and the limitation of a short-term diet measurement to reflect the “habitual” intake may have introduced misclassification of dietary intakes and attenuation of the associations. Although dietary data from the MDC is in general of very high relative validity (33), the relative validity of some PUFAs is still rather low, especially that of long-chain ω-3 PUFAs in men, which is a weakness of the current study. One reason for the low validity may be the infrequent consumption of fatty fish (major source of long-chain ω-3 PUFAs) among many of the study participants. Because the fish consumed at main meals was only registered during a limited number of 7 days, misclassification may be a problem. Furthermore, this is a cross-sectional study, which limits our possibility to investigate causality. Finally, we performed multiple tests, and thus some of the observed significant associations and interactions could be due to chance and need to be replicated.

It is generally accepted that genetic and environmental factors influence the blood lipid and lipoprotein concentrations, but very little is known about such interactions. Our findings suggest that the dietary intakes of different PUFAs may modify the associated effect of the genetic variation in FADS on LDL and HDL concentration. Our results emphasize the importance of dietary fat composition in modifying the effect of genetic susceptibility on blood lipid and lipoprotein concentrations and highlight the potential of developing individualized prevention strategies for dyslipidemia and cardiovascular disease in the future. However, for such strategies to become possible, a major challenge now is to learn more about the importance and mechanisms of interactions between genetic variants, dietary intakes, and blood lipid concentrations. It is therefore important to replicate our results in well-powered studies with good-quality dietary data.