Interaction between dietary fat intake and genetic variants on CVD-related traits was examined by five Brazilian studies (62–66). In a cross-sectional study of 567 participants (62), a significant interaction was reported between olive oil intake and Apolipoprotein E (APOE) genotype on low-density lipoprotein cholesterol (LDL) (P_interaction = 0.028), where a high intake of olive oil (≥ once a week) was associated with lower LDL levels in men carrying the “ε2” allele but had no effect in men without the “ε2” allele. In this study (62), a high polyunsaturated fatty acid (PUFA) intake (> twice a week) was associated with increased LDL levels in carriers of the “ε4” allele, but this was not observed in participants without the “ε4” allele (P_interaction = 0.04). A reduction in triglyceride levels in response to a high PUFA intake was also observed in carriers of the “ε2” allele but not in participants without the “ε2” allele (P_interaction = 0.04). A high PUFA intake was also associated with increased high-density lipoprotein cholesterol (HDL) concentration in participants without the “ε4” allele and reduced HDL levels in carriers of the “ε4” allele (P_interaction = 0.018) (62). In contrast, a cross-sectional study of 252 Brazilian women (63) observed increased triglyceride and very-low density lipoprotein cholesterol (VLDL) in response to a low PUFA or a high fat diet intake in carriers of the “ε4” allele of APOE, but not in non-carriers (P_interaction < 0.05 for both). The findings of the first study (62) indicate that, PUFA intake might be beneficial in increasing HDL levels in individuals without the “ε4” allele, while in those with the “ε4” allele, PUFA intake might contribute to a rise in triglyceride and LDL levels which is associated with higher risk of CVDs (67). Nonetheless, the findings of the second study (63) suggest a detrimental effect of low PUFA intake in carriers of the “ε4” allele. The differences in the findings could be attributed to the small sample sizes and the fact that, the second study (63) was conducted in women unlike the first study (62). PUFA is a ligand for peroxisome proliferator-activated receptors (PPARs) which are involved in regulating several lipid-pathway genes and it has been suggested that, increased consumption of PUFA might promote the expression of APOE and hepatic uptake of “ε4”-containing VLDL particles (68, 69).

Furthermore, a cross-sectional study of 228 Brazilian participants from the Health Survey of São Paulo (HS-SP) (64) observed significant interactions between a GRS based on seven SNPs (Table 1) and the Brazilian Healthy Eating Index Revised (BHEI-R) on the risk of dyslipidaemia. Participants with a higher GRS (5–8) had a lower odds ratio for dyslipidaemia with an intake of BHEI-R oil component above the median (P_interaction = 0.019); while those with a GRS > 9 had a lower odds ratio for dyslipidaemia with an intake of BHEI-R solid fats, alcoholic beverages and added sugars (SoFAAS) component below the median (P_interaction < 0.001). Similarly, a cross-sectional study involving 250 pregnant women (65) observed significant interactions between fatty acid desaturase (FADS) SNPs (rs174561 and rs3834458) and dietary α-linolenic acid (ALA) and linoleic/α-linolenic acid ratio (LA/ALA) on plasma concentrations of omega-3 (n-3) PUFAs. It was reported that, in women with high ALA intake, plasma ALA concentrations were higher in homozygotes for the minor allele (p < 0.05), compared to carriers of the major allele (MM and Mm) of rs174561 and rs3834458. However, the P-values given in the study (p = 0.004 for rs174561 and p = 0.028 for rs3834458) seem to represent associations stratified by genotype, instead of interactions. FADS are involved in the synthesis of PUFA and their activation is linked to inflammation and coronary artery disease (70, 71), and these findings suggest that SNPs which alter the activation of FADS might affect plasma concentration of PUFA. In another cross-sectional study of 113 adolescents from the Obesity, Lifestyle and Diabetes in Brazil (BOLD) study (66), no significant interactions were reported between seven genes involved in the one-carbon metabolism pathway (Table 1) and fat intake on lipid-related traits.

Interaction between dietary fat intake and genetic variants on glycemic traits was investigated by two cross-sectional studies (72, 73) using data from the BOLD study. In the first study which consisted of 200 participants (72), a high total fat intake [37.98% of total energy intake (TEI)/day] was shown to interact with a 10-SNP metabolic-GRS (Table 1), where individuals with 5 or more risk alleles had increased homeostasis model assessment estimate of insulin secretion (HOMA-B) (P_interaction = 0.016), fasting insulin (P_interaction = 0.017), body fat mass (P_interaction = 0.009), and decreased insulin:glucose ratio (P_interaction = 0.01), but the interaction did not influence homeostasis model assessment estimate of insulin resistance (HOMA-IR), glycated hemoglobin (HbA1c), or waist circumference (WC). Similarly, the second BOLD study (73) which also examined the interaction between dietary fat intake and a 10-SNP metabolic-GRS did not find significant interactions between the GRS and dietary fat intake on fasting glucose, fasting insulin or HbA1c (Table 1). The mechanisms through which dietary fat intake influence glycemic traits are unclear, although a sustained increase in blood glucose levels following a high fat meal has been reported (74).

Five Brazilian cross-sectional studies (75–79) investigated the interaction between plasma fatty acids and genetic variants on systemic inflammation, using data from the HS-SP. The first study (75) consisted of 262 adults, and significant interactions were identified between plasma n-3 and adiponectin (ADIPOQ) SNP rs2241766 (P_interaction = 0.019); arachidonic acid and ADIPOQ rs16861209 (P_interaction = 0.044); docosapentaenoic acid and ADIPOQ rs16861209 (P_interaction = 0.037); and SFA and ADIPOQ rs17300539 (P_interaction = 0.019) on the risk of systemic inflammation. Carriers of the “G” allele of rs2241766 had a reduced odds ratio of having inflammatory biomarkers when plasma n-3 levels were above the median, while participants with the “CC” genotype of rs16861209 had a lower odds ratio of having inflammatory biomarkers in the 50th percentile of plasma arachidonic acid and docosapentaenoic acid. Moreover, carriers of the “A” allele of rs17300539 had a higher odds ratio of having inflammatory biomarkers in the upper 50th percentile of plasma SFA compared to those with the “GG” genotype (75). In the second study (76), which consisted of 262 participants, an interaction was also observed between plasma arachidonic acid/eicosapentaenoic acid ratio and toll-like receptor 4 (TLR4) SNP rs11536889, in which individuals with the “C” allele had an increased odds ratio of having inflammatory biomarkers at the higher percentile of arachidonic acid/eicosapentaenoic acid ratio (P_interaction = 0.034). Similarly, the third study consisting of 262 participants (77) identified a significant interaction between plasma palmitoleic acid and C-reactive protein (CRP) SNP rs1417938, where individuals with the “AA” genotype had a higher odds ratio of having inflammatory biomarkers with a plasma palmitoleic acid above the median (P_interaction = 0.047).

In line with these findings, an increasing risk of having inflammatory biomarkers in response to increasing plasma SFA was observed in carriers of the “A” allele of tumor necrosis factor-α (TNF-α) SNP rs180062 (−308G/A) (P_interaction = 0.041); while a decreasing risk with increasing plasma stearic acid was found in participants with the “GG” genotype (P_interaction = 0.046), in a sample of 281 participants from the HS-SP (78). Furthermore, a decreasing risk of metabolic syndrome (MetS) was observed in response to increasing plasma stearic acid levels in “A” allele carriers of interleukin 1 beta (IL1B) SNP rs16944 (P_interaction = 0.043), and in response to increasing plasma arachidonic acid levels in those with the “GG” genotype of interleukin 10 (IL10) SNP rs1800896 (P_interaction = 0.007), in a sample of 301 participants from the HS-SP (79). However, no significant interactions were identified between total SFA, myristic acid, palmitic acid, stearic acid and ADIPOQ SNPs rs1501299 and rs266729; TLR4 SNPs rs11536889 and rs5030728; and CRP SNP rs1205 on inflammatory biomarkers in three of the studies (75, 78, 79). Plasma fatty acid profile is considered an indicator of dietary fatty acid intake (75) and these findings suggest that plasma fatty acid profile can interact with SNPs of several genes and modify the risk of systemic inflammation which is linked to cardiometabolic diseases such as type 2 diabetes and CVDs (75).

Three Brazilian cross-sectional studies (66, 72, 73) investigated the interactions between carbohydrate intake and genetic variants on cardiometabolic traits, using data from the BOLD study. In the first study which consisted of 113 participants (66), a total carbohydrate intake of 47.7% TEI was associated with a significantly increased homocysteine concentration (P_interaction = 0.031) in carriers of the “AA” genotype of fucosyltransferase 2 (FUT2) SNP rs602662. Carbohydrate intake also interacted with Catechol-O-Methyltransferase (COMT) SNP rs4680, increasing oxidized-LDL more in carriers of “AA” than “GG” genotype (P_interaction = 0.005) (66). Notwithstanding, after applying Bonferroni correction for multiple testing, none of the interactions were considered significant (66). Moreover, the other two studies (72) which consisted of 200 participants and (73) which consisted of 187 participants, from the BOLD study, did not identify significant interactions between carbohydrate intake and a metabolic-GRS based on 10 SNPs (Table 1) on cardiometabolic traits.

Three studies (66, 73, 80) investigated the interaction between protein intake and genetic variants on cardiometabolic traits, two of which (66, 73) used data from the BOLD study. A cross-sectional study of 1191 overweight and normal weight children (80) observed a significantly increased BMI (p = 0.01) among participants carrying the “T” allele of FTO SNP rs79149291 with a protein intake above 12.7% TEI/day (80). Similarly, in the BOLD study discussed above (66), those with a protein intake of 16.99% TEI who were carriers of the “AA” genotype of FUT2 SNP rs602662 (P_interaction = 0.007) had increased homocysteine levels (66). However, in the other BOLD study (73), there were no interactions between protein intake and a GRS based on 10 SNPs (Table 1) on obesity or diabetes traits.

The interaction between micronutrients and genetic variants on cardiometabolic traits was examined by five Brazilian studies (81–85). A cross-sectional study of 335 healthy young adults (81), observed a pronounced increase in systolic blood pressure (SBP) (P_interaction = 0.016) among carriers of the “G” allele of Angiotensinogen (AGT) SNP rs699 with a higher plasma magnesium (209.3 mg). Similarly, among those with a high calcium intake (573.3 mg), carriers of the “T” allele of Bradykinin Receptor B2 (BDKRB2) SNP rs1799722 had significantly higher SBP (P_interaction = 0.015) and diastolic BP (DBP) (P_interaction = 0.014) than carriers of the “CC” genotype (81). In line with these findings, a case-control study of 234 elderly people (82) reported an interaction between sodium intake and angiotensin-converting enzyme (ACE) SNP rs4646994 on the risk of hypertension, where carriers of the “I/I” genotype with a high sodium intake (>2 g/day) had an increased risk of hypertension (P_interaction = 0.007). Furthermore, in a cross-sectional study of 1298 healthy adults (83), those carrying the “T” allele of Cytochrome B-245 Alpha Chain (CYBA) (p22phox) with more than 86.5 mEq sodium per 12 h of urine collection, had increased SBP (P_interaction < 0.001) and DBP (P_interaction = 0.011). Sodium is known to increase BP by reducing vasodilation (86), while dietary calcium is believed to stabilize intracellular calcium in smooth muscles, thereby reducing vasoconstriction and BP (87). Additionally, the “A” allele of AGT SNP rs699 is thought to be a risk factor for elevated SBP, possibly due to its association with a rise in plasma AGT levels (60, 81), and the findings of the study discussed above (81) indicate that, the protective effect of the “G” allele might be lost in the presence of higher plasma magnesium.

Similarly, in a longitudinal study of 1088 children with a follow up of 4.6 years (84), those with a deficit of plasma vitamin D (<75 nmol/L) and carriers of the risk allele (“A”) of FTO SNP rs9939609 had increased BMI (P_interaction = 0.033). However, a cross-sectional study examining folate intake in 5914 healthy adults (85) did not identify interactions between folate intake and MTHFR SNP rs1801133 on homocysteine concentrations.

Three Brazilian studies (85, 88, 89) examined the interaction between alcohol intake and genetic variants on cardiometabolic traits. In a cross-sectional study of 113 participants (88), a significant interaction was observed between alcohol intake and endothelial nitric oxide synthase (eNOS) SNP rs2070744 (−786 T/C) on plasma nitrite levels. Individuals carrying the “C” allele who consumed alcohol had lower plasma nitrite levels (P_interaction = 0.033). However, there were no significant interactions between alcohol intake and rs2070744 on BP (88). Similarly, in a cross-sectional study of 3,803 participants from the Pelotas Birth Cohort (85), an interaction was identified between alcohol intake and MTHFR SNP rs1801133 (C677T), in which men with the “677TT” genotype who consumed ≥ 15 g of alcohol per day had the highest homocysteine concentration (P_interaction = 0.002); but the interaction was not observed in women. Moreover, a prospective cohort study of 964 postmenopausal women (89), reported no interactions between alcohol intake and APOE genotype on lipid traits. A rise in homocysteine concentration is attributed to a deficiency in B vitamins and folate, and SNPs of MTHFR might affect homocysteine concentration by impairing folate metabolism (90). However, it is unclear how alcohol intake modifies the activity of MTHFR, and the finding of the study (85) suggests a sex-specific response.

Two studies (91, 92) investigated the interaction between smoking and genetic variants on cardiometabolic traits in Brazilians. In a cross-sectional study of 391 participants (91), smoking interacted with APOA-IV SNPs rs693 (XbaI), rs675 (Thr347Ser) and rs5110 (Gln360His), increasing BMI in individuals with the “X*2” (P_interaction = 0.007) and “347Ser” (P_interaction = 0.02) alleles. However, men with the “360His” allele who were non-smokers had a larger WC than homozygotes for the “Gln” allele (P_interaction = 0.018) (91). Similarly, in a cross-sectional study of 673 overweight adults (403 women and 270 men) (92), carriers of the “S2” allele of APOC3 SNP rs5128 had increased triglycerides and the effect was more pronounced in women who smoked than in non-smokers (P_interaction = 0.009). Serum APOC3 concentration has been shown to be positively associated with triglyceride levels, and smoking has been reported to lower the concentration of APOC3 but only in women without central obesity (93), indicating a sex-specific response which is influenced by obesity traits.

Interactions between physical activity and genetic variants on cardiometabolic traits were investigated by nine Brazilian studies (66, 85, 91, 94–99). In a longitudinal study of 197 overweight or obese children (94), a physical exercise program (3 sessions/week for 12 weeks) interacted with adrenoceptor beta 2 (ADRB2) SNP rs1042714, decreasing triglyceride levels and triglyceride-glucose index (P_interaction = 0.001 for both) more in carriers of the “Glu27Glu” genotype than those carrying the “Gln27” allele. A cross-sectional study of 1701 children and adolescents (95) also reported higher BMI and WC in individuals with the “TT” genotype of fibronectin type III domain containing 5 (FNDC5) SNP rs16835198 compared to carriers of the “G” allele only in those with lower levels of cardiorespiratory fitness (CRF) (P_interaction = 0.038 and P_interaction = 0.007 for WC and BMI, respectively); and lower limb strength (P_interaction = 0.040 and P_interaction = 0.044 for WC and BMI, respectively). Physical activity has been proposed to alter the expression of certain genes (100), and the findings of these studies indicate that, the effect of physical activity on lipid, glycemic and anthropometric traits might be influenced by SNPs of ADRB2 and FNDC5 genes.

Similarly, a sedentary behavior (a screen time of > 378 min/day) was shown to increase cardiometabolic risk score in carriers of “AA” genotype of FTO SNP rs9939609 with a low CRF but not in those with a high CRF in a cross-sectional study of 1,215 children and adolescents (P_interaction = 0.047) (96). Along this line, a randomized controlled trial of 34 participants (97) reported that, a 45-min walk on a treadmill at moderate intensity resulted in a reduction in SBP (P_interaction = 0.02) and DBP (P_interaction < 0.01) in carriers of the “I” allele of ACE SNP rs4646994 compared with a non-exercise control session, but the reduction was not observed in participants with “DD” genotype. However, five studies (66, 85, 91, 98, 99) did not identify significant interactions between physical activity and genetic variants on cardiometabolic traits as shown in Table 1.

In the BOLD study consisting of 113 participants (66), a total fat intake of 25.36% TEI interacted with Betaine-Homocysteine S-Methyltransferase (BHMT) SNP rs492842, increasing vitamin B12 concentrations (P_interaction = 0.034) in participants with the “TT” genotype. A case-control interventional study of 126 obese women (101) also reported that, a hypocaloric diet (< 600 kcal/day) for 7 weeks was associated with a decreased abdominal circumference (P_interaction = 0.04) among carriers of the “A” allele of FTO SNP rs9939609. Furthermore, in a prospective cohort study of 3,701 women, breastfeeding (> 6 months duration) interacted with FTO SNP rs9939609, decreasing BMI (P_interaction = 0.03), fat mass (P_interactin = 0.03), and WC (P_interaction = 0.04) in carriers of the “A” allele (102).

In summary, research in Brazil stands out in comparison to the rest of the gene-lifestyle research in LACP for being the most abundant; twenty-nine studies investigated gene x lifestyle interactions in the Brazilian population as shown in Table 1, covering a wide range of cardiometabolic traits. Dietary fat intake and plasma fatty acid profile were the most frequently investigated dietary factors examined by seven and five studies, respectively, although all the studies examining plasma fatty acid profile used data from the HS-SP. Carbohydrate intake was examined by only three studies, all of which used data from the BOLD study. Similarly, protein intake was investigated by only three studies, two of which used data from the BOLD study. Physical activity was the most frequently examined lifestyle factor, followed by smoking and alcohol intake. Breastfeeding was examined by only one study (102), and lifestyle factors such as socioeconomic status, level of education, and the effect of rural and urban environments were not investigated. Only one study was conducted in rural settings (88), but it was not focused on interaction of the rural environment with genetic variants. The FTO SNP rs9939609 was the most studied, being explored by five studies (84, 85, 96, 98, 99). Overall, the findings call for further research into lifestyle factors such as socioeconomic status, level of education and the effect of rural and urban environments as well as other dietary factors such as fruit and vegetable intake.

The interaction between dietary fat intake and genetic variants on CVD-related traits was examined by five Mexican studies (103–107). In a cross-sectional study of 224 participants with T2D (103), interactions between monounsaturated fatty acid (MUFA) intake and APOE genotype on blood lipid concentrations were reported. A low MUFA intake (< 10–15% TEI) was found to be associated with higher total cholesterol (TC) (P_interaction = 0.016), non-HDL (P_interaction = 0.024) and LDL (P_interaction = 0.030) only in carriers of the “ε2” allele of APOE SNP rs7412. Similarly, interactions between MUFA intake (P_interaction = 0.001), total fat intake (P_interaction = 0.001), dietary cholesterol intake (P_interaction = 0.019) and Dopamine Receptor D2/Ankyrin Repeat and Kinase Domain Containing 1 (DRD2/ANKK1) SNP rs1800497, increasing triglyceride levels in carriers of the “A2A2” genotype were observed in a cross-sectional study of 175 Mexican adults with T2D (104). MUFA intake has been linked to decreased triglyceride concentration (108) which is consistent with the findings of the first study (103). However, the findings of the second study (104) imply that MUFA intake might not be beneficial for individuals with the “A2A2” genotype of rs1800497. Both studies were conducted in participants with T2D which is known to affect lipid metabolism (109). Moreover, as highlighted by the authors of the second study (104), the effect of dietary fat intake on triglycerides concentration may be influenced by other factors including physical activity and the level of insulin resistance.

A Mexican case-control study consisting of 100 participants with normal weight and 100 participants with obesity (105) also found significant interactions between SFA intake and leptin receptor (LEPR) SNP rs1137101 on TC (P_interaction = 0.002) and triglyceride (P_interaction = 0.02) levels. It was reported that, a SFA intake of ≥ 12 g/day was associated with a 3.8 times higher risk of hypercholesteroleamia and a 2.4 times higher risk of hypertriglycerideamia compared to an intake of < 12 g/day in participants carrying the “G” allele of rs1137101 (105). An interaction between total fat intake with LEPR SNP rs1137101 on TC (P_interaction = 0.001) was also reported in this study (105), where a high intake of total fat (≥83 g/d) was associated with a 4.1 times higher risk of hypercholesteroleamia in carriers of the “G” allele of rs1137101. Similarly, in a prospective cohort study involving a dietary intervention in 41 participants with hypercholesterolemia (106), interactions were observed between consumption of a diet low in SFA (<6% TEI/day) in addition to another diet containing 15 g of soluble fiber and 25 g of soy protein for 2 months and Glucose-Fructose Oxidoreductase Domain Containing 2 (GFOD2) SNP rs12449157 on TC (P_interaction = 0.006) and LDL (P_interaction = 0.025). Participants carrying the “G” allele had a larger decrease in TC and LDL in response to the dietary intervention compared to subjects with the “AA” genotype of rs12449157 (106). In this study (106), baseline LDL and TC levels were higher in carriers of the “G” allele, but they responded better to the dietary intervention, which indicates that the genetic risk of dyslipidaemia can be modified by a dietary intervention. However, in another study of 31 Mexican participants with dyslipidaemia (107) from the same cohort as above (106), using the same dietary intervention, no significant interactions were identified between the diet and Calpain 10 (CAPN10) SNPs rs5030952, rs2975762, and rs3792267 on lipid traits. It has been reported that SFA of different types and from different food sources might have different effects on cardiometabolic traits (110, 111), however, both studies (106, 107) used the same dietary intervention. Nonetheless, factors such as physical activity have also been reported to influence the effect of dietary fat intake on cardiometabolic traits (104), which could explain the differences in the findings.

Interactions between carbohydrate intake and genetic variants on cardiometabolic traits were examined by three Mexican studies (104, 112, 113). In a cross-sectional study of 3591 adults (112), carbohydrate intake was negatively associated with HDL concentrations in premenopausal women carrying the risk allele (“C”) of ATP Binding Cassette Subfamily A Member 1 (ABCA1) SNP rs9282541 (R230C), but not in those carrying the “R” allele (P_interaction = 0.037). In another cross-sectional study of 215 healthy adults (113), a high sucrose intake (>5% TEI) significantly increased TC (P_interaction = 0.034) and LDL (P_interaction = 0.037) more in participants with “B1B2/B2B2” genotype than those with “B1B1” genotype of cholesteryl ester transfer protein (CETP) SNP rs708272. However, the interaction did not influence triglycerides, HDL, BMI nor waist circumference (113). In contrast, the cross-sectional study discussed above (104), reported that the intake of maltose (0.68 ± 0.42 g/day) significantly decreased triglycerides (P_interaction = 0.023) in carriers of the “A1”allele of DRD2/ANKK1 SNP rs1800497. These findings indicate that carbohydrate intake might modulate lipid levels in Mexicans with certain genetic variants, but the mechanism through which carbohydrates affect lipid levels are unclear. Moreover, it has been reported that, the effect of carbohydrates on lipids might be dependent on glycemic index or glycemic load, and highly processed carbohydrates are linked to unfavorable lipid profiles (114).

Two cross-sectional studies examined the interaction between micronutrients and genetic variants on cardiometabolic traits (115, 116). In the first study which consisted of 231 healthy new-borns (115), a deficient maternal vitamin B12 (<2.0 mcg/d) was found to be associated with a smaller size baby at birth in mothers with the “TT” genotype of MTHFR SNP rs1801133 (P_interaction = 0.02) but a deficient maternal folate (<400 mcg/d) was not associated with anthropometric parameters (weight, length or BMI) of new-borns (115). A low vitamin B12 intake (<2.0 mcg/d) was also associated with increased homocysteine levels (P_interaction = 0.01) in carriers of the “TT” genotype of MTHFR SNP rs1801133 in a cross-sectional study of 130 healthy women (116). The “TT” genotype of MTHFR is associated with decreased enzymatic activity and increased homocysteine concentration (117) and the findings of these studies suggest that increasing the intake of vitamin B12 might improve fetal development in Mexican women with the “TT” genotype.

The cross-sectional study of 130 healthy women discussed above (116), was the only study which examined alcohol intake and no interaction was found between alcohol intake and MTHFR SNP rs1801133 on homocysteine levels which could be due to the fact that 80% of the studied population consumed less than 1 cup/week of alcohol (116).

Interactions between physical activity and genetic variants on cardiometabolic traits were investigated by four Mexican studies (113, 118–120). In the cross-sectional study discussed above (113), increased concentration of TC (P_interaction = 0.033) was observed in individuals carrying the “B2” allele of CETP SNP rs708272 who did not perform physical activity, compared to those with the “B1B1” genotype. However, there were no interactions on TG, HDL, TG:HDL ratio, LDL, BMI or WC (113). Similarly, interactions between physical fitness measured by muscular endurance (ME) and aerobic capacity with genetic variants were observed in a case-control study of 608 physically active adults (118), where higher levels of ME and aerobic capacity were associated with a lower WC in individuals with a high GRS based on 23 SNPs (Table 1) (P_interaction = 0.0001 for both). In this study (118), a higher risk of obesity was found in older participants (≥ 40 years) with the “AA” genotypes of APOB SNP rs512535 (P_interaction = 0.004) and tumor necrosis factor (TNFA) SNP rs361525 (P_interaction = 0.007) with low levels of ME. However, another cross-sectional study of 565 physically active participants (119) did not find significant interactions between physical fitness and six SNPs (ADIPOQ rs2241766, ACSL1 rs9997745, LIPC rs1800588, PPARA rs1800206, PPARG rs1801282 and PPARGC1A rs8192678) on glycemic traits. Moreover, the fourth cross-sectional study which consisted of 394 participants (120), did not identify interactions between physical activity and ADIPOQ SNP –11391G/A on adiponectin levels.

In a cross-sectional study of 206 Mexican women (121), an interaction between polycyclic aromatic hydrocarbons (PAHs) and Paraoxonase 1 (PON1) SNP rs661 (Q192R) on serum asymmetric dimethylarginine (ADMA) was observed, where individuals carrying the “R” allele had higher ADMA levels compared to those with the “QQ” genotype in response to higher levels of urinary 1-hydroxypyrene (P_interaction = 0.02). Increased levels of ADMA (p < 0.01) and fatty acid-binding protein 4 (p < 0.001) were also identified in individuals with the “RR” genotype of PON1 SNP rs661 with higher urinary arsenic levels (>45.0 μg/g of creatinine) in comparison with participants with the “QQ” genotype in a sample of 185 Mexican women (122). The mechanisms of the interaction may be shared in the case of exposure to PAHs as these are also involved in the generation of reactive oxygen species (123).

Overall, different cardiometabolic traits have been investigated in Mexico, where eleven out of fifteen studies found significant gene x lifestyle interactions (103–106, 112, 113, 115, 116, 118, 121, 122) as shown in Table 1. Dietary fat intake was the most frequently examined dietary factor, being investigated by five studies (103–107); followed by carbohydrate intake, which was examined by three studies (104, 112, 113). Physical activity was the most frequently examined lifestyle factor, while alcohol intake was investigated by only one study. Lifestyle factors such as smoking, socioeconomic status, level of education and the impact of rural and urban environments were not investigated. Moreover, dietary factors such as consumption of protein, complex carbohydrates, and fruits and vegetables have not been investigated, highlighting a need for further research.

The interaction between dietary fat intake and genetic variants on CVD-related traits was examined by six Costa Rican studies (124–129). In a cross-sectional study of 420 participants (124), SFA intake interacted with APOE genotype and influenced blood lipid concentrations. A higher SFA intake (13.5% energy) was associated with higher levels of very-low density lipoprotein cholesterol (VLDL) (P_interaction = 0.03) and lower concentration of HDL (P_interaction = 0.02) in carriers of the “ε2” allele. However, no significant interactions were identified between SFA intake and APOE genotype on lipids in a case-control study involving 1,927 participants with myocardial infarction (MI) and 1,927 matched controls (125). In another cross-sectional study of 336 participants (126), SFA intake was found to interact with APOC3 genotype and impact on the concentration of TC (P_interaction = 0.0004) and LDL (P_interaction = 0.01). Homozygotes for the APOC3-455T-625T alleles had a 13% increase in TC and a 20% increase in LDL with a high SFA intake (>11% of energy intake), but the interaction was not significant in individuals with the APOC3-455C-625del allele (126). In the case-control study discussed above (125), a significant interaction between SFA intake and APOE genotype on the risk of MI (P_interaction = 0.0157) was also reported, in which carriers of the “ε4” allele had a 49% increased risk of MI compared to a 2.2 fold increased risk in those with the “ε2” allele in response to a high SFA intake (>11.8% of energy intake).

APOE plays a key role in lipid metabolism, being a main component of triglyceride-rich lipoproteins and HDL, and a ligand for LDL receptor (124, 130) and it is believed that the metabolism of fatty acids is impaired in carriers of the “ε4” allele which is considered a risk factor for CVDs (131). However, the above findings indicate that, a high SFA intake is more detrimental to carriers of the “ε2” allele than those carrying the “ε4” allele, highlighting the potential role of SFA intake in modifying genetic risk.

In accordance with the findings above, a case-control study of 1805 participants with a first non-fatal MI and 1,805 matched controls (127), reported an interaction between PUFA intake and PPARγ SNP rs1801282, influencing the risk of MI (P_interaction = 0.03). Individuals with the “Pro12/Pro12” genotype had a 34% reduced risk of MI per 5% increment in energy from PUFA compared to a 7% decreased risk in those carrying the “Ala12” allele (127). Similarly, a case-control study of 1932 participants with a first non-fatal MI and 2,055 matched controls (128), reported a significant interaction between long-chain omega-3 (LC n-3) PUFA intake and Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) SNP rs11206510 on the risk of MI (P_interaction = 0.012), where carriers of the “C” allele had an odds ratio for MI of 0.84 per 0.1% increase in total energy from LC n-3 PUFA, compared to an odds ratio of 1.02 in participants without the “C” allele (128). Along similar lines, a case-control study of 1936 participants with a first non-fatal MI and 2,035 matched controls (129) reported a significant interaction between omega-6 (n-6) PUFA intake and Phospholipase A2 Group IVA (PLA2G4A) SNP rs12746200 on the risk of MI (P_interaction = 0.005), in which participants with the “G” allele had a reduced risk of MI with an intake of n-6 PUFA above the median compared to those with the “AA” genotype. However, there were no significant interactions with n-3 PUFA intake (129).

These findings indicate that the beneficial effect of PUFA intake reported by some studies (101, 132) might be limited in individuals with certain genetic variants. PPARγ is a nuclear receptor which is involved in adipogenesis and plays a role in the metabolism of glucose and fatty acids (133, 134), and the “Ala12” allele of PPARγ SNP rs1801282 has been reported to slow down the release of PUFA from adipocytes, which could explain the smaller reduction in the risk of MI in comparison with carriers of the “Pro12/Pro12” genotype (127).

Interactions between other dietary factors and genetic variants on the risk of MI were examined by three Costa Rican studies (135–137). In a case-control study of 1,560 incident cases of non-fatal MI and 1,751 matched controls (135), sugar sweetened beverage (SSB) intake interacted with a GRS based on 3 SNPs of chromosome 9p21 (rs4977574, rs2383206 and rs1333049), increasing the risk of MI (P_interaction = 0.03). SSB intake also interacted with rs4977574, increasing the risk of MI in carriers of the “G” allele (P_interaction = 0.005), but there was no interaction with fruit juice intake (135). In another case-control study of 2,014 participants with a first acute non-fatal MI and 2,014 matched controls (136), an increased risk of MI with increasing coffee intake was observed in carriers of the “C” allele (also known as “slow metabolizers of caffeine”) of Cytochrome P450 Family 1 Subfamily A Member 2 (CYP1A2) SNP rs762551 compared to carriers of the “AA” genotype (P_interaction = 0.04). Similarly, in a case-control study consisting of 2,042 participants with a first non-fatal MI and 2042 control subjects (137), cruciferous vegetable intake (0.86 servings/day of half a cup) interacted with Glutathione S-Transferase Theta 1 (GSTT1) SNP rs17856199, lowering the risk of MI in carriers of the “*1” allele, but not in individuals with the “*0*0” genotype (P_interaction = 0.006). These findings indicate that, dietary factors other than fat intake, might also influence the risk of MI in Costa-Ricans with certain genetic variants.

Interaction between smoking and genetic variants on the risk of MI was investigated by three Costa Rican case-control studies (137–139), two of which found significant interactions (137, 138). In a case-control study of 492 participants with a first non-fatal MI and 518 matched controls (138), an interaction was observed between smoking status and Paraoxonase 1 (PON1₁₉₂) SNP rs661 on the risk of MI (P_interaction = 0.04), where the PON1_192Arg allele was associated with an increased risk of MI only in non-smokers. Similarly, in the case-control study discussed above (137), the combined intake of cruciferous vegetables (>5 servings/day) and smoking (1–10 cigarettes/day) in carriers of the “*1” allele of rs17856199, lowered the risk of MI (P_interaction = 0.008). However, there were no significant interactions with GSTM1 or GSTP1 genotype on the risk of MI (137). Moreover, in the third Costa Rican case-control study which involved 873 participants with a first non-fatal MI and 932 control subjects (139), no significant interactions were observed between smoking and CYP1A1 SNP rs1048943 or CYP1A2 SNP rs762551 on the risk of MI. Smoking has been linked to increased risk of MI (140, 141) although the mechanisms are unclear. Smoking is also believed to impair the activity of PON1, which is linked to increased risk of CVDs (142, 143), but this is not supported by the findings of the studies above, suggesting that Costa Ricans with certain genetic variants might respond differently to smoking.

One case-control study consisting of 1534 participants with a first non-fatal MI and 1,534 matched controls (144), investigated the interaction between a lifestyle cardiovascular risk score comprising of physical activity, smoking, alcohol consumption, waist-to-hip ratio, and socioeconomic status; and a GRS based on 14 SNPs (Table 1) on the risk of MI, and no significant interactions were identified.

The research in Costa Rica has mainly focused on CVD traits in adults, with an emphasis on the risk of MI, and dietary fat intake has been the most frequently examined exposure. Socioeconomic status was examined by one study (144), and lifestyle factors such as educational level, the effect of rural and urban environments as well as dietary factors such as consumption of protein, fiber and complex carbohydrates have not been explored, highlighting a need for further research.

Interaction between dietary fat intake and genetic variants on anthropometric traits were investigated by six studies (145–150), all of which used data from the Boston Puerto Rican Health Study (BPRHS). In a cross-sectional study of 930 Puerto Ricans from the BPRHS (145), a high intake of SFA (≥22 g/day) was associated with a 7.9% higher BMI in individuals with the “CC” genotype of APOA2 SNP rs5082 than those carrying the “T” allele (P_interaction = 0.003); but the SNP had no effect on BMI when SFA intake was low (<22 g/day). This study also observed that, among individuals with a high SFA intake (≥22 g/d), those with the “CC” genotype had a higher risk of obesity than participants carrying the “T” allele of the SNP rs5082 [Odds ratio (OR) = 1.84; 95% confidence interval (CI) = 1.38–2.47; P < 0.0001). A similar finding was reported in a prospective cohort study of 920 participants from the BPRHS (146), where a high intake of SFA (≥ 9.3% of total energy) was linked to higher BMI (P_interaction = 0.006), WC (P_interaction = 0.02), and hip circumference (HC) (P_interaction = 0.002) in participants carrying the minor allele (“T”) of LDL receptor related protein 1 (LRP1) SNP rs1799986 compared to individuals with the “CC” genotype; but the SNP had no effect on anthropometric traits when SFA intake was low (<9.3% of total energy). The “CC” genotype of APOA2 rs5082 is believed to affect body fat distribution by lowering plasma concentration of APOA2 and these findings indicate that, a low SFA intake might attenuate this genetic risk (145, 151).

An interaction of total fat intake with APOA1-75 on WC was also reported in a longitudinal study of 821 participants of the BPRHS (147), in which individuals carrying two copies of the major allele had a lower WC with a low total fat intake than those carrying the minor allele (P_interaction = 0.005). A longitudinal study performed in 1,171 participants (333 men and 838 women) of the BPRHS (148) also observed that, women with the “TT” genotype of lipoprotein lipase (LPL) SNP rs320 had lower BMI (P_interaction = 0.002) and WC (P_interaction = 0.001) with a high intake of PUFA but this was not observed in minor allele (“G”) carriers and there were no significant interactions in men. In contrast, another longitudinal study of 1,340 participants (395 men and 945 women) of the BPRHS (149) found that, men with the “GG” genotype of brain derived neurotrophic factor (BDNF) SNP rs6265 had higher BMI (P_interaction = 0.042), WC (P_interaction = 0.018), and HC (P_interaction = 0.009) with a low PUFA intake (<8.76% of energy) than those carrying the “A” allele but no difference was observed when PUFA intake was high (≥8.76% of energy) and the interaction was not observed in women. Interaction between Mediterranean diet with TCF7L2 SNP rs7903146 on obesity-related traits was also observed in a cross-section study of 1,120 Puerto Ricans of the BPRHS (150), where carriers of the “T” allele had lower WC (99.2 ± 0.9 vs. 102.2 ± 0.9 cm; P_interaction = 0.026) and weight (77.3 ± 1.0 vs. 80.9 ± 1.0 kg; P_interaction = 0.024) with a high Mediterranean diet score than individuals with “CC” genotype. However, there were no significant differences between the genotypes when the Mediterranean diet score was low. The findings suggest that a high intake of PUFA and Mediterranean diet might be beneficial in reducing the genetic risk of obesity-related traits in a sex-specific manner and call for further research into the mechanisms involved.

Interaction between total fat intake and genetic variants on CVD traits were reported by three studies (147, 152, 153). In a longitudinal study of 802 participants of the BPRHS (152), a significant interaction was observed between total fat intake and APOA5 SNP -1131T < C on plasma triglycerides (P_interaction = 0.032), where a high total fat intake (≥31% of total energy) was associated with a higher plasma triglyceride concentration in individuals with the “1131C” allele, although no difference between the genotypes was observed when total fat intake was low. This study (152) also observed an interaction between APOA5 SNP S19W with total fat intake on SBP (P_interaction = 0.002) and DBP (P_interaction = 0.007), where participants with the minor allele (“G”) had a higher SBP with a low total fat intake (< 31% of total energy), and a lower SBP with a high total fat intake in comparison with individuals with the “CC” genotype. The study on 821 participants of the BPRHS discussed above (147), also reported significant interactions between total fat intake and APOC3 -640 on DBP (P_interaction = 0.003), APOA4 N147S and APOA5 S19W on SBP (P_interaction = 0.001 and P_{interaction =} 0.002, respectively). It was observed that, homozygous for the major allele of APOA1-75, APOA4 N147S and APOA5 S19W had lower SBP with a low intake of total fat (< 31% of total energy) than those carrying the minor allele; while heterozygous for APOC3 -640 had lower DBP with a high total fat intake (≥ 31% from energy) (147). However, a randomized crossover trial involving 41 adults from Dominican, Puerto Rican and other Caribbean Hispanic origins (153), did not find significant interactions between a high fat diet and hepatic lipase (LIPC) SNP rs1800588 on HDL, LDL, TC or plasma glucose concentrations. A high intake of total fat has been associated with an unfavorable lipid profile and high blood pressure (154) and the above findings indicate that, this association might be influenced by variants of several genes.

Two studies investigated the interaction between carbohydrate intake and genetic variants on cardiometabolic traits (155, 156). In a longitudinal study involving 920 participants of the BPRHS (155), a significant interaction was observed between Perilipin 1 (PLIN 1) SNP 1,482 G > A and complex carbohydrate intake on WC (P_interaction = 0.002), where individuals carrying the “A” allele had a higher WC with a low intake of complex carbohydrate (<144 g/day) and a lower WC with a high intake of complex carbohydrate (≥144 g/day) than those with the “GG” genotype. Similarly, a cross-sectional study of 153 children descendent from Hispanic ancestry (156), identified significant interaction between carbohydrate intake (211.4 g/day) and total sugar intake (96.1 g/day), increasing hepatic fat fraction in carriers of the “GG” genotype of Patatin like phospholipase domain containing 3 (PNPLA3) SNP rs738409 (P_interaction = 0.04 and P_interaction = 0.01, respectively), but the interaction was not observed in individuals carrying the “C” allele. It has been reported that, body weight might be influenced by the type of carbohydrate consumed (157) which is supported by the findings of these studies, but the results also indicate that genetic variants might also play a role.

The interaction between micronutrient intake and genetic variants on cardiometabolic traits was investigated by two studies (158, 159). A cross-sectional study involving 1,734 Mexican Americans (158) reported a significant interaction between vitamin E and APOB SNP rs693 on LDL (P_interaction = 8.94 × 10-7), and between vitamin A and PCSK9 SNP rs11206510 on LDL (P_interaction = 7.65 × 10-5), but the direction of the interactions is unclear. Similarly, in the longitudinal study of 1,144 Puerto Ricans of the BPRHS discussed above (159), a significant interaction between vitamin D status and IRS1 rs2943641 on the risk of T2D was identified in women in which minor allele homozygotes (“TT”) had a lower risk of T2D compared with “C” allele carriers only when 25(OH)D was higher than the median [>17 ng/mL (42.4 nmol/L)] (P_interaction = 0.007), but the interaction was not observed in men. The findings of these studies indicate that micronutrients might modulate the association between genetic variants and lipid and glycemic traits, but further studies are needed to replicate and elucidate the mechanisms involved.

Only one study (160) examined the interaction between physical activity and genetic variants on cardiometabolic traits. This study (160) was a prospective cohort study of 9,645 adult Puerto Ricans, Mexicans, Dominicans, Cuban, Central American, and South American from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) cohort, USA, and a positive association was observed between a GRS based on 97 SNPs (Table 1) and BMI, but the effect of the GRS was stronger in the first tertile of moderate to vigorous physical activity compared to the third tertile (P_interaction = 0.005). Significant interactions following the same pattern were observed for fat mass (P_interaction = 0.003), fat percentage (P_interaction = 0.003) and fat mass index (P_interaction = 0.002) (160).

In summary, research in LACP diaspora has mainly focused on Puerto Ricans residing in USA and most of this evidence (10 out of 13 studies) comes from the same study (BPRHS). Dietary fat intake has been the most frequently studied, with carbohydrate intake being examined by only two studies. Similarly, physical activity was investigated by only one study and lifestyle factors such as socioeconomic status, level of education, and the effect of rural and urban environments have not been explored.

Two gene-diet interaction studies were reported in Chileans (161, 162). The first study (161) was a cross-sectional study of 2828 healthy Chilean adults, and a significant interaction was observed between consumption of SSB and a weighted genetic risk score (wGRS) based on 16 T2D risk SNPs (Table 1) on log-fasting glucose (P_interaction = 0.02), where the strongest effect was observed between the highest SSB intake (≥2 servings/day of 330 ml) and the highest wGRS. In this study (161), SSB intake also interacted with additive effects of Transcription Factor 7 Like 2 (TCF7L2) SNP rs7903146 (P_interaction = 0.002) and with the “G/G” genotype of Melatonin Receptor 1B (MTNR1B) SNP rs10830963 (P_interaction = 0.001), increasing log-fasting glucose levels. The second Chilean study (162) was a non-randomized controlled trial performed in 97 healthy women and 147 women with polycystic ovary syndrome, and there were no reported interactions between a high glycemic carbohydrate intake (75 g of glucose) during an oral glucose tolerance test and Insulin Receptor Substrate 1 (IRS-1) SNP rs1801278 on glycemic traits. In Chile, research has been limited to diabetes traits as outcomes and simple carbohydrates as exposure, reflecting a need for further research into other dietary and lifestyle factors such as socioeconomic status, level of education and the effect of rural and urban environments.