Western-style diet

Although not particularly well-defined, a Western-style dietary pattern generally consists of high intake of refined, energy-dense, nutrient-poor food sources^{(Reference Odermatt14)}.

Evaluation of Western-style diet across sex is relatively extensive in preclinical models compared to other dietary patterns/constituents (Table 1). Overall, this evidence appears to suggest that young male rodents have more negative changes in the body composition profile, as well as a higher susceptibility to diet-induced obesity when exposed to a high-fat diet (HFD)^{(Reference de Souza, Wasinski and Donato15,Reference Maric, Krieger and van der Velden16,Reference Oraha, Enriquez and Herzog17,Reference Dhanraj, van Heerden and Pepper18,Reference Guerra-Cantera, Frago and Collado-Pérez19)} . It must however be mentioned that some discrepancies exist and may relate to species and/or diet differences^{(Reference Maric, Krieger and van der Velden16,Reference Gannon, Robison and Salinero20)} . Young female mice appear to have a greater ability to utilise fat in the diet as a source of fuel^{(Reference Oraha, Enriquez and Herzog17)}, increase energy expenditure^{(Reference Maric, Krieger and van der Velden16)} and increase AQP7/Aqp7 glycerol channel abundance (regulation influences glycerol release by adipocytes and reduced function is associated with obesity)^{(Reference Iena, Jul and Vegger21)}. Additionally, a more favourable immune response is observed in young female rodents exposed to an HFD^{(Reference Binenbaum, Atamni and Fotakis22,Reference Braga Tibaes, Azarcoya-Barrera and Wollin23)} .

Table 1. Preclinical evidence for Western-style diet-related sex differences

FGF21, fibroblast growth factor 21; HFD, high-fat diet; IGF2, insulin-like growth factor 2.

Although a consensus appears to be emerging in young animals, for aged animals the picture is less clear. It appears that the protection from HFD observed in young female mice diminishes with age, with females having greater weight gain and impairment in glucose tolerance compared to males^{(Reference Guerra-Cantera, Frago and Collado-Pérez19,Reference Salinero, Anderson and Zuloaga24)} . This may in part relate to changes in sex hormones. Indeed, ovariectomy of HFD-fed female mice enhanced adipose tissue inflammation leading to moderate changes in metabolism. However, gonadectomised HFD-fed males had improved metabolic outcomes that were associated with increased CD11c+ adipose tissue macrophages and increased proinflammatory cytokines^{(Reference Varghese, Griffin and Abrishami25)}. It should be noted that many diets utilised to model high-fat/Western-style diets in rodents (as outlined earlier) are refined (i.e. made from individual purified component rather than whole food). As such they lack many components of a complete control ‘chow’ diet. Indeed, dietary fibre source (e.g. soluble v. insoluble) and even amount are often overlooked in these studies, compromising the validity of these experiments. Morrison et al. utilised a refined diet with matched fibre source/content in their experimentation of low-fat diet v. HFD across both sexes, reporting that the lack of soluble fibre and not fat content primarily drives gut microbiota alterations previously associated with a refined HFD. In contrast to the aforementioned results, they report that male body weight increase is independent of dietary fat. However, when the amount of dietary fibre is comparable in all dietary groups, aged females do still appear to display increased weight gain in response to HFD^{(Reference Morrison, Jašarević and Howard26)}. This is in line with recent reports that the prebiotic effects of dietary fibres are sex-specific^{(Reference Kadyan, Park and Wang27)}, although the mechanisms responsible for such differences remain to be elucidated.

Sex differences in response to other components of the Western-style dietary pattern (e.g. high fructose, high sugar, low fibre) have been less extensively covered, and the results are generally mixed. Greater metabolic abnormalities have been reported in female animals receiving 10 % fructose supplementation^{(Reference Roglans, Baena and Sangüesa28)}. Similarly, a sweet-fat diet (standard laboratory control diet supplemented with sweet cookies, sunflower seeds and lard) resulted in more intense fat accumulation and weight gain in females as a result of suppressed carbohydrate and fat metabolism^{(Reference Bazhan, Iakovleva and Dubinina29)}. Furthermore, female mice maintained on a cafeteria diet had more extensive liver steatosis, higher non-alcoholic fatty liver disease scores and elevated triglyceride (TAG) content compared to males, with no difference in body weight gain or adiposity index observed^{(Reference Gasparin, Carreño and Mewes30)}. High-sucrose consumption in mice led to more extensive dysregulation of the oxylipin profile (oxidation products of PUFA) in the brains of female mice^{(Reference Norman, Nuthikattu and Milenkovic31)}. The mechanistic basis for which remains unclear. Intriguingly, others have reported the complete opposite with males displaying greater weight gain, glucose intolerance and hepatic inflammation on either high-fat, high-sugar or high-fat, high-fructose diets in agreement with the HFD studies outlined earlier^{(Reference Daly, Saxena and Singh32,Reference De Souza, Barros and De Souza33)} .

Ethical consideration and lack of stratification by sex mean that clinical evaluation of Western-style diets across sex is scarce. However, in one such study conducted in young healthy adults 7-d exposure to a high-fat, high-energy diet did not result in any metabolic outcomes in either males or females^{(Reference Whytock, Shepherd and Cocks34)}, indicating that young healthy individuals can tolerate acute exposure to a Western-style diet. Consumption of a high-fructose meal however led to increased postprandial hepatic de novo lipogenesis in females only^{(Reference Low, Cornfield and Charlton35)}, suggesting that women may be more responsive to higher levels of fructose in the diet. It should however be noted that the opposite has also been reported^{(Reference Tran, Jacot-Descombes and Lecoultre36)} and in line with this, Couchepin et al. observed that healthy young female mice were more resistant to fructose overfeeding compared to their male counterparts^{(Reference Couchepin, Lê and Bortolotti37)}.

Observational studies evaluating sex differences are similarly lacking and can be difficult to discern whether differences relate to biological/metabolic effects or merely food preference/portion size. Indeed, in a cross-sectional multi-ethnic study of middle-aged individuals (45–57 years) it was reported that women have a higher diet quality (as assessed by the HEI-2010)^{(Reference Maskarinec, Namatame and Kang38)}. Diet quality reduced adiposity across both sexes but intriguingly females displayed a stronger association than men^{(Reference Maskarinec, Namatame and Kang38)}. In line with this, Ruiz-Canela et al. reported (in a study population of 55–80-year-olds with CVD risk) that significant differences in BMI relating to consumption of a pro-inflammatory diet were restricted to females, although other indices of general and abdominal obesity were consistent across both sexes^{(Reference Ruiz-Canela, Zazpe and Shivappa39)}. Furthermore, in a Japanese cohort, increased SFA intake was associated with increased all-cause mortality in females only^{(Reference Nagata, Nakamura and Wada40)}. Similarly, UK Biobank analysis (of 40–69 years) suggested that higher sugar, SFA and dietary fibre intake may subtly modulate all-cause mortality and/or dementia risk to a greater extent in females^{(Reference McKenzie, Harris and Peters41)}. In contrast to the aforementioned studies, a cross-sectional study of a Taiwanese population with dyslipidaemia described that greater consumption of a Western-dietary pattern (highest quartile) increased general obesity, central obesity and high body fat regardless of sex^{(Reference Lin, Hsu and Lee42)}.

It has been posited that these sex difference may relate to changes in the microbiota. Indeed, in human subjects, a Western-style diet (high-fat/low-fibre) reportedly leads to an altered microbial profile across males and females, with higher levels of Campylobacter, Blautia, Flavonifractor and Erysipelatoclostridium in females^{(Reference Bailén, Bressa and Martínez-López43)}. However, a functional understanding of these changes requires further elucidation. HFD feeding in rats induces sex-related alterations in gut microbiome composition and metabolome^{(Reference Bailén, Bressa and Martínez-López43,Reference Shi, Wei and Xing44,Reference Hases, Stepanauskaite and Birgersson45)} which correlate to metabolic measures such as insulin resistance^{(Reference Hases, Stepanauskaite and Birgersson45)}. Kim et al. suggest that the microbial impact may be mediated via the pregnane X receptor (a xenobiotic-sensing nuclear receptor) which reportedly primes the gut microbiome towards an obesity-prone microbial configuration in a sex (male) specific manner^{(Reference Kim, Choi and Dutta46)} (Table 2).

Table 2. Human evidence for Western-style diet-related sex differences

Mediterranean diet

A Mediterranean diet (MedDiet) pattern appears to be highly beneficial, with adherence associated with a reduction in all-cause mortality. The MedDiet consists of a proportionally higher intake of unprocessed cereals, legumes, olive oil, fruits, nuts and vegetables, along with moderate consumption of fish, dairy and meat products^{(Reference Sikalidis, Kelleher and Kristo47)}.

In contrast to Western diet, preclinical studies investigating MedDiet across sexes are limited. This predominantly relates to the fact that preclinical studies tend to focus on aspects/constituents of the MedDiet rather than MedDiet in its entirety. Some of these constituents such as dietary fibres, lipids (e.g. MUFA and PUFA) and plant bioactives will be discussed in later sections.

Human evidence evaluating MedDiet across sexes is surprisingly limited with many studies failing to provide stratification of results/analysis by sex, despite inclusion of both sexes in the experiment. This is quite a significant issue, presumably relating to a lack of power that needs to be resolved imminently. From the available evidence, studies in younger (24–53 years, premenopausal) adults suggest that the MedDiet confers more favourable changes in glucose/insulin homoeostasis in men than in women^{(Reference Bédard, Corneau and Lamarche48,Reference Bédard, Riverin and Dodin49)} . In line with this, improvements in TAG levels, HDL-cholesterol ratios and waist circumference are more pronounced in men than in women^{(Reference Leblanc, Bégin and Hudon50)}. Furthermore, MedDiet adherence leads to a significant decrease in adiponectin concentration in men only^{(Reference Bédard, Tchernof and Lamarche51)}, as well as a more favourable redistribution of LDL subclasses from smaller to larger LDL^{(Reference Bédard, Corneau and Lamarche52)}. This appears to be independent of circulating NEFA concentrations (believed to be an important factor in insulin resistance). Similar results were reported after 3-year MedDiet adherence in older (~66 years) overweight/obese individuals with metabolic syndrome, in which a reduction in weight, waist circumference, fasting glucose, insulin and TAG were more pronounced in men than in women^{(Reference Soldevila-Domenech, Pastor and Sala-Vila53)}. This appears to be consistent with CVD, in which an association between MedDiet adherence appears to be stronger^{(Reference GómezSánchez, Gómez Sánchez and Patino-Alonso54,Reference Bilenko, Fraser and Vardi55)} , although no difference has also been reported^{(Reference Bédard, Dodin and Corneau56)}. In contrast to this, 1-year Mediterranean-like diet intervention in elderly healthy subjects led to female-specific (but also country-specific) reduction in epigenetic ageing score^{(Reference Gensous, Garagnani and Santoro57)}. Also, from a neurological disease perspective, women appear to have more favourable outcomes in response to MedDiet adherence. Indeed, an inverse association between MedDiet and dementia risk was established among women, but not among men^{(Reference Andreu-Reinón, Chirlaque and Gavrila58)}. Similarly, in a cross-sectional analysis adherence to both MedDiet and Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay diet was significantly associated with a higher age of Parkinson's disease onset^{(Reference Metcalfe-Roach, Yu and Golz59)}, especially in women. However, for colorectal cancer no disease-modifying effect was observed as a result of MedDiet^{(Reference Schulpen and van den Brandt60)}. This was also true for all cancer risk, which despite displaying an inverse association in females only, failed to reach significance after full adjustment of confounding factors^{(Reference Schulpen and van den Brandt60,Reference Schulpen and van den Brandt61)} (Table 3).

Table 3. Human evidence for Mediterranean style diet-related sex differences

ApoA-2, apolipoprotein A2; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol; MedDiet, Mediterranean diet; MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; PD, Parkinson's disease; sdLDL, small-dense LDL.

Vegetarian/vegan diets

Food constituents derived from animal sources are limited/absent from vegetarian/vegan diets. Despite an abundance of studies investigating such diets in the context of metabolism and disease, sub-analysis by sex is consistently missing. As such, the existence of any sex differences in response to vegetarian diet is not entirely clear. Blood sampling of healthy age-matched vegetarians and non-vegetarians revealed a purportedly beneficial increase in adiponectin levels in female vegetarians, which was not present in males^{(Reference Vučić Lovrenčić, Gerić and Košuta63)}. However, no diet-dependent or sex-dependent differences were found in insulin, Homeostatic model assessment for insulin resistance index (HOMA2-IRI), inflammatory and metabolic biomarkers^{(Reference Vučić Lovrenčić, Gerić and Košuta63)}. Reviewed extensively by Adams and Sabaté, there is evidence to suggest that the cardio-protective effects of a vegetarian diet may be sex-specific^{(Reference Adams and Sabaté64)}. The available evidence suggests that a vegetarian dietary pattern is associated with a reduction in CVD outcomes for vegetarian men relative to omnivorous men, whilst for women this association is less strong/non-existent^{(Reference Adams and Sabaté64)}. In line with this a 4-year longitudinal study reported that low intake of vegetables was significantly associated with type 2 diabetes risk in men, but not in women. Although this may relate to the lack of study power (fewer women in the low-intake category), it could also relate to differences in vegetable preference^{(Reference Whytock, Shepherd and Cocks34)}. Conversely, Kim et al. did not find any association between plant-based diet and CVD, nor any apparent sex differences within a US population^{(Reference Kim, Caulfield and Rebholz65)}. They did however report that those with a high plant-based diet index (i.e. above median) had a 5 % lower risk in all-cause mortality in the overall study population which was not influenced by sex^{(Reference Kim, Caulfield and Rebholz65)}. Together there is some evidence to support vegetarian diet-related sex differences, particularly in the context of CVD however further investigation is clearly warranted to gain a greater understanding metabolically and for other diseases.

Ketogenic diet

Ketogenic diets are low in carbohydrate content and high in fat, shifting energy reliance from glucose to ketone bodies. Twenty-five day ketogenic diet adherence in individuals with severe obesity resulted in significantly larger excess body weight loss and a greater reduction in γ-glutamyl transferase in males^{(Reference D'Abbondanza, Ministrini and Pucci66)}. This greater benefit in males has been reported by others^{(Reference Volek, Sharman and Gómez67,Reference Lyngstad, Nymo and Coutinho68)} . Interestingly, this difference does not appear to be present when considering post-menopausal females^{(Reference D'Abbondanza, Ministrini and Pucci66)}, again emphasising the importance of the menopause (and likely sex hormones) in metabolism and response to diet. In rats maintained on a high-fat, high-sugar-diet, the beneficial effects of ketogenic diet intervention were largely similar across both sexes although these benefits correlated significantly with plasma β-hydroxybutyrate in females only^{(Reference Sahagun, Bachman and Kinzig69)}. In the context of pancreatic cancer, strict ketogenic diet in combination with gemcitabine (chemotherapy medication) prolonged survival. Intrudingly, when stratified by sex this result remained significant for males only^{(Reference Cortez, Rodriguez Lanzi and Hong70)}.

Dietary fibre

Often overlooked as a key contributor in health and disease, dietary fibre is the undigestible part of the plant, typically obtained from wholegrain cereals, fruits and vegetables. European and US guidelines suggest an intake of 30–35 g daily for men and 25–32 g daily for women (discrepancy between males and females relates to fact that many countries calculate recommendation based upon for total energy intake) but actual dietary fibre intake is significantly lower^{(Reference Barber, Kabisch and Pfeiffer71)}. In adolescents, increasing dietary fibre to recommendation levels decreased predicted fasting glucose, fasting insulin, Homeostatic model assessment for insulin resistance (HOMA-IR), Systolic blood pressure (Hg SBP), and diastolic blood pressure (Hg DBP) regardless of sex^{(Reference Dong, Chen and Gutin72)}. In line with this a cross-sectional analysis found that higher daily dietary fibre consumption was associated with beneficial effects on cholesterol in both males and females^{(Reference Zhou, Wu and Tang73)}. Analysis of the European prospective investigation into cancer and nutrition cohort revealed that total dietary fibre was inversely associated with colorectal cancer (Hazard Ratio per 10 g daily increase in fibre 0⋅87, 95 % CI: 0⋅79, 0⋅96) which did not differ by sex^{(Reference Murphy, Norat and Ferrari74)}. Similarly, an inverse relationship between dietary fibre and multiple sclerosis has been reported in a case control study with the trends similar across males and females^{(Reference Ghadirian, Jain and Ducic75)}. In the context of depression, dietary fibre may be more favourable in females with an inverse association between depression and dietary fibre consumption established in females only^{(Reference Chrzastek, Guligowska and Piglowska76)}. Discrepancies in the impact of dietary fibres across sex may relate to changes in the gut microbiota, indeed oligofructose supplementation in mice led to broad changes in faecal community structure (increasing Bacteroidetes at the expense of Lachnospiraceae) in females but not males. How dietary fibre type (e.g. soluble or insoluble) influences metabolism and health outcomes across sex is yet to be explored and represents a major gap in our current knowledge.

PUFA

As alluded to in the Western-diet section of this review, lipid metabolism appears to be sexually dimorphic. Indeed, vast differences in lipid species have been identified across sex, particularly when considering age × sex interaction, with the most prevalent of these differences found across phosphatidylcholine, sphingomyelin and TAG species^{(Reference Slade, Irvin and Xie77)}. It is therefore not surprising that specific dietary lipid types e.g. PUFA exert different effects across the sexes. Females have significantly higher peripheral DHA than males^{(Reference West, Michaelson and Miles78)}. In rats this higher DHA concentrations is found in the liver, plasma, erythrocytes and heart (53, 75, 36 and 25 % higher, respectively, compared with males) but not the brain^{(Reference Kitson, Smith and Marks79)}. This may be linked to higher Δ6-desaturase expression in females relative to males, which appears to be limited to the liver^{(Reference Kitson, Smith and Marks79)}. Women show a greater increase in circulating EPA in response to α-linolenic acid consumption^{(Reference Childs, Kew and Finnegan80)}. Similarly, EPA and DHA supplementation increases plasma TAG EPA to a greater extent in females^{(Reference Walker, Browning and Mander81)}. The source of PUFA (e.g. krill oil v. fish oil) may also alter these sex differences adding further complexity to the interaction^{(Reference Ghasemifard, Hermon and Turchini82)}.

Sex × diet interactions may influence brain PUFA concentrations^{(Reference Chen, Haven and Lecaj83)}. Higher n-3 PUFA concentrations appear to benefit different cognitive domains in a sex-specific manner^{(Reference Duchaine, Fiocco and Carmichael84)}. In mice receiving DHA supplementation, a reduction in anxiety and depressive-like behaviours was observed in male mice only and coincided with sex-specific gut microbiota interactions in response to DHA which correlated with behavioural finding^{(Reference Davis, Hecht and Jasarevic85)}. This in contrasts with a report in human subjects which showed n-3 fatty acid intake to be negatively associated with depressive symptoms in only women^{(Reference Beydoun, Fanelli Kuczmarski and Beydoun86)}. From a diabetes perspective n-3 PUFA status was inversely associated with diabetes in overweight/obese females but not in males^{(Reference Abbott, Burrows and Thota87)}. This is supported by a systematic review and meta-analysis of randomised controlled trials which found that n-3 PUFA intervention improved insulin resistance in women but not in men^{(Reference Abbott, Burrows and Thota88)}. Furthermore, PUFA appear to be more protective against hypertriglyceridaemia in females, compared to males^{(Reference Kaviani, Taylor and Stevenson89)}. Interestingly, the ability of n-3 PUFA to reduce platelet aggregation (a factor in CVD) is sex-specific. In men, only EPA treatment reduces aggregation, whilst in women, only DHA treatment reduced platelet aggregation^{(Reference Phang, Lincz and Garg90)}. Both increased n-3 and n-6 PUFA intake were found to be inversely associated with non-alcoholic fatty liver disease risk, irrespective of sex^{(Reference Cui, Li and Ren91)}. There is growing evidence suggesting that oxylipin (bioactive oxidation products of PUFA) production and profile is differentially altered across sexes in response to the intake of various n-3 and n-6 PUFA^{(Reference Gabbs, Zahradka and Taylor92,Reference Ferdouse, Leng and Winter93,Reference Pauls, Ragheb and Winter94,Reference Leng, Winter and Aukema95)} , although this seems to be less extensive in the brain^{(Reference Ferdouse, Leng and Winter96)}. As mediators of PUFA, such differences in the oxylipin profile may provide in part some explanation for the varying disease-modifying influences observed across sexes.

Plant bioactives

Sex has been suggested to modulate both the metabolism^{(Reference Zamora-Ros, Achaintre and Rothwell97)} and physiological effects of plant bioactives such as (poly)phenols^{(Reference Gibney, Milenkovic and Combet98)}. HPLC-MS/MS analysis of acute doses of grape seed (poly)phenols established clear sex differences in the metabolism and distribution of flavanols throughout the bodies of rats, with quantitative differences found in the plasma and brain^{(Reference Margalef, Pons and Iglesias-Carres99)}. In line with this supplementation with an oral formulation of resveratrol, JOTROL™ in 3xTg-AD mice resulted in Alzheimers disease (AD)-related gene expression changes (Adam10, Bace1, Bdnf, Psen1) some of which were brain region-dependent and sex-specific^{(Reference Dennison, Volmar and Modarresi100)}. Analysis of the Primary prevention of cardiovascular disease with a mediterranean diet (PREDIMED) study revealed that catechins, proanthocyanidins, hydroxybenzoic acids and lignans were inversely associated with type 2 diabetes, with women displaying stronger inverse associations. Additionally, a cross-sectional analysis of a Korean population reported an inverse association between flavonoid intake and obesity in women, whilst for men a positive association was determined for some subclasses (namely, flavonols, flavanones and anthocyanidins)^{(Reference Kim, Kim and Jun101)}. In a randomised double-blind parallel trial, a combination of 548 mg daily of polyphenols and 2 g daily of l-citrulline reduced ambulatory systolic blood pressure in women, but not in men^{(Reference Vors, Rancourt-Bouchard and Couillard102)}. Furthermore, a systematic review and meta-analysis described an inverse association between (poly)phenol consumption and gastric cancer. Interestingly, the risk reduction was greater for females, which the authors suggest may be partly explained by the fact that (poly)phenols can regulate female hormones which play a protective role against cancer^{(Reference Fagundes, Silva and Fernandes103)}. These differences may relate to impact on the gut microbiota which may be modulated in a sex-specific manner, indeed microbial changes associated with 7,8-dihydroxyflavone predicted body weight changes in females but not in males^{(Reference Sharma, Wu and Kumaraswamy104)}. In contrast to the female-specific improvements outlined earlier, the Reasons for geographic and racial differences in stroke (REGARDS) prospective cohort study reported that the inverse association between flavanone intake and ischaemic stroke risk did not differ by sex^{(Reference Goetz, Judd and Hartman105)}. Additionally, in mouse models of CVD both blackberry and gallic acid supplementation reduced atherosclerosis in males only^{(Reference Clark, Centner and Ukhanov106)}. Consistent with this, nettle extract altered lipid metabolism differently across sexes, with the activation of transcription factors that control lipid metabolism, and subsequent increase in HDL-cholesterol, specific to male mice^{(Reference Domjanić Drozdek, Odeh and Đikić107)}.