CVD remains the leading cause of morbidity and mortality worldwide. The WHO estimated that 17·3 million people in the world died from CVD in 2008, including 7·3 million from CHD, and 6·2 million from strokes( 1 ). There are a number of modifiable risk factors for CVD, such as high levels of serum LDL-cholesterol (LDL-C), hypertension, diabetes, overweight/obesity, smoking, low physical activity and diet. Indeed, diets that are rich in SFA and trans fatty acids (TFA) are associated with an increased risk of CVD, and it is largely agreed that this is due, in the most part, to increased serum LDL-C( Reference Mensink, Zock and Kester 2 ). Furthermore, evidence from pharmacological studies show that lowering LDL-C by an average of 1·8 mmol/l (by use of statins) reduces risk of IHD and stroke by 60 and 17 %, respectively( Reference Law, Wald and Rudnicka 3 ). Despite this, the relationship between SFA and CVD risk remains controversial( Reference Chowdhury, Warnakula and Kunutsor 4 ).
The UK dietary guidelines recommend <10 % of total energy intake from SFA, but according to the most recent National Diet and Nutrition Survey consumption of SFA is above these recommendations, at 11·9 % of total energy intake( Reference Bates, Lennox and Prentice 5 ). Milk and dairy products contribute about 27 % of SFA intake in the UK diet( Reference Bates, Lennox and Prentice 5 ). However, evidence from a number of prospective cohort studies show that consumption of milk and other dairy products (excluding butter) are not consistently associated with an increased risk of CVD. Milk is a unique and complex food that is nutritionally complete for the sustenance of young mammals. Milk consumption in most mammals ceases soon after weaning, this coincides with down-regulation of the gene for lactase, leading to a severe compromise in the ability to digest lactose, the sugar contained within milk. However, human subjects are unique within the animal kingdom being the only mammal that continues to consume another animals’ milk past infancy and throughout our lifespan. This is made possible in the majority of the population by possession of one of a number of SNP in the lactase gene, which results in persistence of lactase throughout life. The majority of individuals of European origin possess a version of the gene that remains active, which results in about 90 % of Europeans being able to digest lactose( Reference Sahi 6 ). Selection of these mutant SNP in the lactase gene throughout human development suggests that there may be some advantage to the ability to consume milk.
The present paper will provide a brief overview of the consumption of milk and dairy products in the diets of UK adults, and will summarise the evidence on the effects of milk and dairy consumption on CVD mortality and biomarkers.
Trends in milk and dairy consumption
In the UK, current milk consumption is about 1·5 litres of milk per person per week, with the majority of this consumed as semi-skimmed milk (70 %) followed by whole milk (20 %) and skimmed milk (10 %). Over the past few decades, trends in milk and dairy product consumption have shown considerable variation (Fig. 1)( 7 ). For example, consumption of whole milk has shown a dramatic decline since the 1970s from about 2·7 litres per person per week in 1974 to 0·3 litres per person per week in 2012 (Fig. 1(a)). In the early 1990s, the decline in whole milk consumption was partially replaced by semi-skimmed milk, consumption of which has remained fairly constant at about 1 litre per person per week over the last decade, while the intake of whole milk continues to decline. Fig. 1(b) shows the UK trends in consumption of other dairy products such as cheese, yoghurt and fromage frais, cream and butter. The trend for cheese and cream consumption has remained fairly constant at about 100 g and 20 ml per person per week for cheese and cream, respectively since the 1970s. In contrast, the trend for yoghurt and fromage frais consumption has increased significantly from the early 1970s with 33 ml per person per week to about 200 ml per person per week in 2012. The consumption of butter in the UK shows a similar downward trend as for whole milk, due, in part, to recommendations to reduce the amount of total and saturated fat in the diet, but also because of the increasing availability of other spreads.
The contribution of milk and dairy foods to nutrient intakes
Milk and dairy products are complex foods containing a number of different components. Table 1 shows the contribution of the dairy food group, which includes milk, cheese, yoghurt, butter, cream and fromage frais to energy and nutrient intakes in UK adults (age 19–64 years)( Reference Bates, Lennox and Prentice 5 ). It is clear that milk and other dairy products are important sources of a number of nutrients in the UK diet. Indeed, the dairy food group alone contributes more than the daily reference nutrient intake for vitamin B12 and provides about 50 % of the recommended nutrient intake for calcium and phosphorus. Milk and dairy products are also the main source of iodine in the UK diet, contributing about 40 % of the recommended nutrient intake. Adequate iodine levels are important for both men and women throughout life, but particularly in women of childbearing age as iodine levels below recommendations during pregnancy have been associated with reduced cognitive outcome in their children( Reference Bath, Steer and Golding 8 ). Although dairy is an important contribution to iodine intake in the UK diet, there have been inconsistent reports of iodine concentrations in milk with a recent study showing a 30 % lower iodine concentration in organic compared with conventional milk( Reference Payling, Juniper and Drake 9 ). Therefore, although milk and dairy products are not an essential dietary component, they make an important contribution to the provision of key nutrients.
EAR, estimated average requirement; DRV, daily recommended value; RNI, reference nutrient intake.
Saturated fat from milk and CVD
Higher dietary SFA consumption is associated with increased risk of CVD, due primarily to the serum total cholesterol and LDL-C raising effects of SFA( Reference Mensink, Zock and Kester 2 ). The association between SFA and increased serum LDL-C has led to dietary recommendations worldwide for the restriction of SFA intake. Dietary recommendations by the FAO/WHO( 10 ), UK dietary recommendations (Department of Health, 1991)( 11 ) and, dietary guidelines for Americans( 12 ) recommend intake of dietary SFA to <10 % of total energy intake. Despite these recommendations current SFA intakes in the UK are about 11·9 % of total energy( Reference Bates, Lennox and Prentice 5 ).
Milk and dairy products are the greatest contributor to dietary SFA in the UK diet, contributing about 27 % of SFA intake. As a result, guidance to reduce or eliminate dairy from the diet has been a common practice for CVD risk reduction. However, the evidence for the relationship between dairy consumption and CVD mortality does not support dairy restriction as an effective strategy for CVD reduction. It is important to recognise that we do not consume individual nutrients, but complex foods and diets that contain specific nutrients within various matrices. This can give rise to disparity between the biological effects of nutrients in different foods that may have contributed to the inconsistencies in the relationships of different SFA-rich foods and CVD mortality. Clear evidence for this comes from the Multi Ethnic Study of Atherosclerosis, in which different SFA-rich foods were shown to produce differential effects on CVD risk( Reference de Oliveira Otto, Mozaffarian and Kromhout 13 ). In 5209 subjects after a 10-year period (from 2000 to 2010). A lower hazard ratio (HR) for CVD was reported for every 5 g/d (HR 0·79; 95 % CI 0·68, 0·92) or 5 % of energy from dairy SFA (HR 0·62; 95 % CI 0·47, 0·82), whereas the equivalent intake of SFA from meat sources was associated with greater HR for CVD (HR for +5 g/d and a +5 % of energy from meat sources of SFA: 1·26; 95 % CI 1·02, 1·54 and 1·48; 95 % CI 0·98, 2·23, respectively)( Reference de Oliveira Otto, Mozaffarian and Kromhout 13 ). Furthermore, the substitution of 2 % of energy from meat sources of SFA with energy from dairy SFA was associated with a 25 % lower CVD risk (HR 0·75; 95 % CI 0·63, 0·91), suggesting that dairy foods containing SFA attenuated the detrimental association of SFA with CVD mortality. While this finding was attributed to the effects of other components within dairy foods, such as calcium, magnesium, bioactive peptides and proteins, it may also have been due to a difference in the relative proportions of different SFA between meat and dairy.
Further evidence for the beneficial association between dairy and CVD comes from an investigation of the association between plasma phospholipid fatty acids and incidence of CHD( Reference Khaw, Friesen and Riboli 14 ). In the present study, the enrichment of plasma phospholipid with even chain SFA: palmitic acids (C16:0) and stearic (C18:0), but not myristic (C12:0), was associated with significantly higher risk of CHD, while the odd chain SFA indicative of dairy consumption: pentadecanoic C15:0 and hexadecanoic acid C17:0 were associated with lower risk. These finding were corroborated by associations between similar circulating biomarkers of dairy fat and the incidence of stroke in US men (Health Professionals Follow-up Study n 51 529) and women (Nurses’ Health Study n 121 700)( Reference Yakoob, Shi and Hu 15 ). Odd chain fatty acids are found in milk and dairy products and result from bovine biohydrogenation( Reference Vlaeminck, Fievez and Cabrita 16 ). Their appearance in human plasma or tissue samples is now recognised as a specific biomarker of dairy intake, as man is unable to synthesise these fatty acids endogenously( Reference Smedman, Gustafsson and Berglund 17 ). These data support the prospective cohort data that suggest that milk and dairy products are not associated with detrimental effects on CVD risk.
Trans fatty acids and CVD
Other important fatty acids present in milk and dairy foods are TFA, which are synthesised via bacterial metabolism of unsaturated fatty acids in the rumen of cows( Reference Lock and Bauman 18 ). The intake of TFA from industrially hydrogenated vegetable oils (iTFA) has a negative impact on cardiovascular health( Reference Mozaffarian, Katan and Ascherio 19 , Reference Brouwer, Wanders and Katan 20 ). However, the association between ruminant TFA (rTFA) and CVD remains inconclusive( Reference Gebauer, Chardigny and Jakobsen 21 , Reference Brouwer, Wanders and Katan 22 ), with some studies showing a cardioprotective association( Reference Mozaffarian, Katan and Ascherio 19 , Reference Jakobsen, Overvad and Dyerberg 23 ). In an attempt to resolve conflicting reports, a systematic review and meta-analysis was undertaken by Bendsen et al. ( Reference Bendsen, Christensen and Bartels 24 ) who reported that the relative risk (RR) for high v. low quintiles of total TFA intake (2·8 to approximately 10 g/d) was 1·22 (95 % CI 1·08, 1·38; P = 0·002) for CHD events and 1·24 (95 % CI 1·07, 1·43; P = 0·003) for fatal CHD. In addition, rTFA intake (0·5–1·9 g/d) was not significantly associated with CHD risk (RR 0·92; 95 % CI 0·76, 1·11; P = 0·36), although there was a trend towards a positive association for iTFA (RR 1·21; 95 % CI 0·97, 1·50; P = 0·09). The authors concluded that while iTFA may be positively related to CHD, rTFA were not, but the limited number of studies available prevented a firm conclusion on the critical importance of the source of TFA. In contrast to previous findings, a recent prospective cohort study by Kleber et al. ( Reference Kleber, Delgado and Lorkowski 25 ) showed that total TFA content in erythrocyte membranes of 3259 participants of the Ludwigshafen Risk and Cardiovascular Health Study was inversely associated with adverse cardiac outcomes, while rTFA (trans-palmitoleic acid) was associated with reduced risk. In addition, erythrocyte membrane iTFA was associated with no increased risk of adverse cardiac outcomes( Reference Kleber, Delgado and Lorkowski 25 ). However, it is important to highlight that total TFA concentration in erythrocyte membranes in the present study population was relatively low compared with levels in other studies, and this may have been too low to observe an effect of TFA on CVD mortality. Furthermore, it has been suggested that the lack of an association between rTFA and CHD risk may be due to a lower intake from ruminant sources compared with iTFA( Reference Bendsen, Christensen and Bartels 24 ). Despite this controversy, there is little doubt that dietary iTFA are associated with increased CVD mortality( Reference de Souza, Mente and Maroleanu 26 ). In response there has been a substantial decrease in iTFA over the past 10–15 years, due to the voluntary action by the UK food industry( 27 ). This has led to an increase in the relative proportion of rTFA in the UK diet, although the absolute intake of ruminant fat is unchanged, with the current mean population intake of total TFA (0·7 % food energy in adults)( Reference Bates, Lennox and Prentice 5 ) below the recommended population maximum of 2 % of food energy intake( 11 ). Although milk and milk products (including butter) contribute to 32 % of this intake( Reference Bates, Lennox and Prentice 5 ), current rTFA intake is not considered to be a major cause for concern with respect to cardiovascular health at a population level( Reference Brouwer, Wanders and Katan 22 , Reference Tardy, Morio and Chardigny 28 ). However, the impact of any increase in dietary TFA would need to be monitored.
Effects of milk and dairy foods on CVD risk: evidence from observational studies
The potential effects of milk and dairy consumption on CVD mortality would best be determined using adequately powered randomised control intervention studies, which have CVD events and CVD-related deaths as outcomes. However, for obvious financial and logistical reasons such studies have not been performed. The most informative data on the relationship between milk and dairy consumption and CVD is provided by long-term prospective cohort studies( Reference Elwood, Pickering and Givens 29 ).
Several influential reviews have focused on the impact milk and dairy food consumption and CVD risk, some of which have conducted meta-analyses of available cohort data (Table 2)( Reference Elwood, Pickering and Givens 29 – Reference Elwood, Pickering and Hughes 32 ). Elwood et al. ( Reference Elwood, Pickering and Hughes 32 ) conducted a meta-analyses of ten prospective cohort studies that examined the associations between milk and risk of IHD and stroke. Using a pooled estimate of the relative odds for IHD and stroke, the meta-analysis revealed no association with IHD (RR 0·87; 95 % CI 0·74, 1·03) and a significant inverse association for stroke (RR 0·83; 95 % CI 0·77, 0·90) in the subjects with the highest milk compared with those with the lowest intakes. These findings, together with a combined estimate of risk for both IHD and stroke (ten studies RR 0·84; 95 % CI 0·78, 0·90), suggested that consumption of milk was associated with a modest reduction in CVD risk. This work was extended by Elwood et al. ( Reference Elwood, Pickering and Givens 29 ) to include nine prospective cohort studies of milk and dairy products and IHD and eleven studies for stroke. The meta-analysis indicated a 15 % lower RR for all-cause mortality (RR 0·85; 95 % CI 0·77, 0·98) and an 8 % lower overall RR of IHD (RR 0·92; 95 % CI 0·80, 0·99) in the subjects with the highest dairy consumption compared with those with the lowest intakes. Furthermore, a significant inverse association was observed for the risk of stroke (RR 0·79; 95 % CI 0·68, 0·91) in the subjects with the highest dairy consumption compared with those with the lowest intakes. These findings supported previous meta-analyses by Elwood et al. ( Reference Elwood, Pickering and Hughes 32 ), and support a reduction in IHD and stroke in subjects consuming the highest amount of milk and dairy products compared with the lowest intakes.
RR; relative risk.
In another meta-analysis of seventeen prospective cohort studies Soedamah-Muthu et al. ( Reference Soedamah-Muthu, Ding and Al-Delaimy 30 ) showed a modest inverse association between milk intake and risk of overall CVD (four studies; RR 0·94 per 200 ml/d; 95 % CI 0·89, 0·99). However, milk intake was not associated with risk of CHD (six studies; RR 1·00 per 200 ml/d; 95 % CI 0·96, 1·04), stroke (six studies; RR 0·87; 95 % CI 0·72, 1·05) or total mortality (eight studies; RR per 200 ml/d 0·99; 95 % CI 0·95, 1·03). A recent meta-analysis by Qin et al.( Reference Qin, Xu and Han 31 ), which included a total of twenty-two prospective cohort studies, showed an inverse association between dairy consumption and overall risk of CVD (nine studies; RR 0·88; 95 % CI 0·81, 0·96) and stroke (twelve studies; RR 0·87; 95 % CI 0·77, 0·99). However, no association was found between dairy consumption and CHD risk (twelve studies; RR 0·94; 95 % CI 0·82, 1·07), which supports previous findings( Reference Soedamah-Muthu, Ding and Al-Delaimy 30 ). Qin et al.( Reference Qin, Xu and Han 31 ) also investigated the association between individual dairy foods on risk of CVD, including stroke and CHD. Interestingly, cheese consumption was associated with a significantly reduced risk of stroke (four studies; RR 0·91; 95 % CI 0·84, 0·98) and CHD (seven studies; RR 0·84; 95 % CI 0·71, 1·0). Recently, Praagman et al.( Reference Praagman, Dalmeijer and van der Schouw 33 ) also reported a significant association between cheese consumption and stroke mortality, although no impact on CHD mortality was found( Reference Praagman, Dalmeijer and van der Schouw 33 ). One possible explanation for the observed beneficial effects of cheese consumption on stroke and CHD risk may be the relatively high calcium content that increases saponification of SFA in the gut, rendering them resistant to digestion leading to increased faecal fat excretion( Reference Nestel, Chronopulos and Cehun 34 ). This mechanism is supported by the results from a prospective cohort study in which the observed inverse association between cheese consumption and CHD was attenuated when calcium content was used as a confounder in the analysis( Reference Louie, Flood and Burlutsky 35 ). Furthermore, a meta-analysis of randomised control trials (RCT) investigating the impact of calcium from dairy and dietary supplements estimated that increasing dairy calcium intake by 1241 mg/d resulted in an increase in faecal fat of 5·2 (1·6–8·8) g/d( Reference Christensen, Lorenzen and Svith 36 ).
Since publication of the meta-analyses above (Table 2), additional prospective cohort studies have been published. For example, the Rotterdam Study consisting of 4235 men and women aged 55 years and above showed that total dairy, milk, low-fat dairy, and fermented dairy were not significantly related to incident stroke or fatal stroke after a 17·3-year follow-up period( Reference Praagman, Franco and Ikram 37 ). In addition, the authors reported a significant inverse relationship between high-fat dairy consumption and fatal stroke (HR 0·88 per 100 g/d; 95 % CI 0·79, 0·99), but not incident stroke (HR 0·96 per 100 g/d; 95 % CI 0·90, 1·02). Total dairy or individual dairy foods were not associated with incident CHD or fatal CHD.
Contrary to these data and since these meta-analyses a study conducted by Michaelsson et al. ( Reference Michaelsson, Wolk and Langenskiold 38 ), reported that the milk intake was significantly associated with markedly higher total and CVD mortality and fracture risk in 61 433 Swedish women from the mammography cohort. This relationship was also observed in a cohort of 45 339 Swedish men, although the relationship was considerably weaker( Reference Michaelsson, Wolk and Langenskiold 38 ). However, the authors concluded that the study should be ‘interpreted with caution, due to the inherent possibility of residual confounding and reverse causation phenomena, which is often associated with observational study designs’. In addition, when this data were reanalysed, an inverse association was observed for the number of CVD deaths against milk consumption( Reference Hellstrand and Michaëlsson 39 ). These inconsistent findings between milk intake and CVD mortality observed with the same dataset require further investigation. Furthermore, since the study by Michaelsson et al. ( Reference Michaelsson, Wolk and Langenskiold 38 ), two further large prospective cohort studies have been published on the relationship between milk and myocardial infarction and IHD mortality in Japanese (n 94 980; 19 years follow-up)( Reference Wang, Yatsuya and Tamakoshi 40 ) and Danish (n 98 529; 5·4 years follow-up)( Reference Bergholdt, Nordestgaard and Varbo 41 ) populations, both of which reported no association with myocardial infarction or IHD.
The balance of current evidence, including meta-analyses of prospective cohort studies, indicates that milk and dairy products are associated with no detrimental effect on risk of CVD, with some evidence of a moderately protective effect of milk consumption. However, a further meta-analysis that includes all of the current prospective cohort data is required to confirm this and more studies are required to determine the effects of individual dairy products on CVD risk.
Effects of dairy on blood lipids and lipoproteins
In the absence of randomised control dairy intervention studies with clinical endpoints, the bulk of evidence for cause and effect relationships between dairy foods and CVD has relied heavily upon validated CVD biomarkers as outcome measures in RCT. There is consistent evidence that consumption of dietary SFA increases total and LDL-C concentrations, a robust biomarker of CHD risk( Reference Mensink, Zock and Kester 2 ). Replacement of SFA with unsaturated fatty acids has a beneficial reduction on serum LDL-C and the clinically relevant total cholesterol: HDL-cholesterol (total-C:HDL-C) ratio( Reference Mensink, Zock and Kester 2 , Reference Jebb, Lovegrove and Griffin 42 , Reference Vafeiadou, Weech and Altowaijri 43 ). However, not all classes of SFA have the same effects on blood lipids. High dietary intakes of lauric (C12:0), myristic (C14:0) and palmitic (C16:0) acids have been shown to elevate serum total and LDL-C, whereas stearic acid (C18:0) has minimal impact, due, in part, to its more limited absorption( Reference Kris-Etherton, Griel and Psota 44 ). These SFA are also associated with a concomitant increases in HDL-C concentrations, a lipoprotein that is generally considered to be anti-atherogenic( Reference Mensink 45 ). This differential effect of dietary fats on different lipoprotein fractions highlights the importance of expressing dietary effects on the clinically relevant total-C:HDL-C ratio( Reference Lemieux, Lamarche and Couillard 46 ). Given that a high proportion of the C12:0, C14:0 and C16:0 in the human diet is derived from milk fat, it would be predicted that the consumption of milk and dairy foods would be associated with adverse effects on serum LDL-C and total-C:HDL-C. However, evidence indicates that dairy food consumption, with the exception of butter( Reference Engel and Tholstrup 47 ), is associated with limited or no significant detriment to serum lipids. In support of this, a cross-sectional analysis of 2 512 Welsh men from the Caerphilly cohort study showed no significant difference in serum total cholesterol or LDL-C concentrations, and a significant negative association between the highest compared with the lowest quartile of dairy consumers( Reference Livingstone, Lovegrove and Cockcroft 48 ). Negative associations between dairy consumption, confirmed by dietary assessment and biomarkers of dairy intake: plasma phospholipid levels of C15:0 and C17:0, and the proportion of the pro-atherogenic small dense LDLIII particles, was reported in another cross-sectional study in 291 healthy men( Reference Sjogren, Rosell and Skoglund-Andersson 49 ). Stronger evidence from a meta-analysis of twenty RCT with a total of 1 677 subjects showed that there was no significant change in LDL-C with either low and full-fat dairy consumption( Reference Benatar, Sidhu and Stewart 50 ). In contrast, studies that used butter, invariably produced the predicted increases in LDL-C( Reference Engel and Tholstrup 47 ). This again suggests that the other components of dairy foods, such as proteins, bioactive peptides and calcium may be involved with the amelioration of the detrimental effects of dairy SFA.
Differential impact of high- and low-fat dairy foods
There is no established nutritional benefit of whole-fat dairy consumption, except in children under 2 years, compared with lower fat alternatives. With respect to the latter, skimming milk fat to produce low-fat milk and dairy products is a common and an effective way of lowering SFA intake. However, there is currently no consensus on whether fat-reduced dairy foods are associated with a reduced risk of CVD( Reference Benatar, Sidhu and Stewart 50 ) and studies in this area give inconsistent data, with few RCT that directly compare whole with low fat alternatives. Minimal benefit has been reported in a prospective study of 33 636 women, which suggested no significant differences between consumption of high- v. low-fat dairy products on risk of myocardial infarction( Reference Patterson, Larsson and Wolk 51 ). Furthermore, findings from a 12-month RCT concluded that in overweight adults inclusion of reduced-fat dairy foods had no impact on blood lipids, blood pressure (BP) and arterial compliance( Reference Crichton, Howe and Buckley 52 ). Moreover, a meta-analysis conducted by Soedamah-Muthu et al.( Reference Soedamah-Muthu, Ding and Al-Delaimy 30 ), showed that there was no significant difference between consumption of high-fat (RR 1·04; 95 % CI 0·89, 1·21) or low-fat dairy (RR 0·93; 95 % CI 0·74, 1·17) on CHD risk.
In contrast, data from the Nurses’ Health Study cohort illustrated that the associated RR of CHD varied according to the fat content of dairy foods with an estimated 20 % lower RR with low-fat dairy consumption (RR 0·80; 95 % CI 0·73, 0·87) compared with a 12 % higher RR with high-fat dairy consumption (RR 1·12; 95 % CI 1·05, 1·20)( Reference Hu, Stampfer and Manson 53 ). Furthermore, observational studies investigating the relationship between consumption of different types of dairy on cardio-metabolic risk factors have indicated that low-fat dairy consumption is an effective strategy to promote lower BP levels( Reference Engberink, Hendriksen and Schouten 54 – Reference van Meijl and Mensink 56 ), circulating markers of inflammation( Reference Esmaillzadeh and Azadbakht 57 ), the ratio of total-C:HDL-C( Reference Mensink, Zock and Kester 2 ) and LDL-C concentration( Reference Kai, Bongard and Simon 58 ), as well as aid in weight maintenance or reduction( Reference Abargouei, Janghorbani and Salehi-Marzijarani 59 ). Further evidence from well-controlled dietary intervention studies is required before a definitive conclusion can be drawn on the benefits of low- and high-fat dairy.
There have also been a number of studies suggesting that the specific milk proteins have differential effects on lipids. Whey (60 g/d for 12 weeks) has been shown to produce significant reductions in serum TAG and total and LDL-C in comparison with a casein control group( Reference Pal, Ellis and Dhaliwal 60 ). Furthermore, a significant decrease in the postprandial appearance of TAG after consuming a whey meal of 21 % compared to control and 27 % compared with the casein meals were reported( Reference Pal, Ellis and Ho 61 ). In addition to the specific dairy proteins, different dairy foods have been shown to have a range of lipid effects( Reference Astrup 62 ). It has been reported that cheese may have a differential effect on blood lipids compared with other dairy foods( Reference Nestel, Chronopulos and Cehun 34 , Reference Tholstrup, Hoy and Andersen 63 , Reference Hjerpsted, Leedo and Tholstrup 64 ), with prolonged ripening of cheddar cheese resulting in more pronounced lipid-lowering effects in a pig model( Reference Thorning, Bendsen and Jensen 65 ). A meta-analysis that included five of these RCT showed that when compared with butter intake, cheese consumption reduced LDL-C by 6·5 % (−0·22 mm/l; 95 % CI −0·2, −0·14) and HDL-C by 3·9 % (−0·05 mm/l 95 % CI −0·09, −0·02) but had no effect on TAG( Reference de Goede, Geleijnse and Ding 66 ). In addition, a recent RCT reported that consumption of 80 g/d of a high-fat cheese (27 % fat) compared with no cheese or 50 g/d of fat and salt-free cheese for 8 weeks resulted in no changes in total or LDL-C. Those in the high-fat cheese group with metabolic syndrome at baseline had significant reductions in total cholesterol (−0·70 mm/l) compared with control and a significantly higher reduction in TAG( Reference Nilsen, Hostmark and Haug 67 ). These data show that dairy products do not exert the negative effects on blood lipids which would be predicted solely from their SFA content, and highlights a need for additional studies before firmer conclusions can be made on the differential effects of dairy products on serum lipid and lipoprotein concentrations.
Overall, the current evidence presented in this section suggests that the fatty acid profile of milk may not be as detrimental for lipid risk factors as previously thought, and supports differential effects of dairy foods, particularly cheese.
Manipulating the fatty acid profile of milk
Modification of the fatty acid profile of bovine milk offers an alternate strategy for lowering the population's intake of SFA, by removing SFA from the food chain, while preserving the beneficial contributions that dairy products make to the protein and micronutrient content of the human diet( Reference Shingfield, Bonnet and Scollan 68 ). Partial replacement of milk SFA with cis-MUFA or cis-PUFA through supplementation of the cows’ diet with plant oils or oilseeds reduces synthesis of short- and medium-chain SFA by the bovine mammary gland, and increases the long-chain (>C18) unsaturated fatty acid concentration in the milk( Reference Givens, Shingfield, Williams and Buttriss 69 , Reference Glasser, Ferlay and Chilliard 70 ). Inclusion of 49 g/kg of dry matter of rapeseed oil in the ruminant diet for a 28-d period increased cis-MUFA from 20 to 33 g/100 g fatty acids, while reducing SFA from 70 to 55 to 60 g/d fatty acids( Reference Givens, Kliem and Humphries 71 ). This feeding regimen inadvertently leads to increased concentrations of naturally produced rTFA, predominantly trans-linoleic acid (trans-18:1) and trans-MUFA, in the milk. However, despite this increase in rTFA, the consumption of the modified dairy products is not thought to have a major impact on CVD risk( Reference Kleber, Delgado and Lorkowski 25 ). In addition, cell culture studies have shown that rTFA has minimal impact on endothelial function and gene expression( Reference Livingstone, Givens and Jackson 72 ), although whether rTFA intake, through manipulation of the fatty acid profile of milk and dairy products to decrease SFA content, impacts on cardiovascular health, has yet to be determined.
Consumption of SFA-reduced milk and milk products, by feed modification has been shown to be beneficial to CVD risk, in healthy and hypercholesterolaemic populations when compared with conventional dairy products( Reference Livingstone, Lovegrove and Givens 73 ). Poppitt et al. demonstrated that consumption of 20 % energy daily as conventional or feed-modified SFA-reduced butter for a 3-week period resulted in significant reduction in both total cholesterol and LDL-C during the modified butter feeding( Reference Poppitt, Keogh and Mulvey 74 ). However, the evidence is relatively limited and the majority of studies have used butter only and relied on serum lipid levels as biomarkers of CVD risk. This knowledge gap is being addressed at the University of Reading with the RESET (Replacement of Saturated fat in dairy on total cholesterol) study investigating the impact of reducing SFA intake by using modified milk and dairy products on vascular function and CVD risk biomarkers, without limiting dairy product consumption. Milk that has a substantial proportion of SFA replaced with cis-MUFA will be collected from cows fed a diet supplemented with 1 kg/d of high-oleic sunflower oil. Cheddar cheese and butter will be produced from this milk and these dairy foods will be supplied to volunteers with increased CVD risk for a 12-week period in a randomised, crossover, double-blind, controlled study. The impact of these modified dairy products on fasted and postprandial vascular function, BP, lipids, insulin sensitivity and inflammatory biomarkers will be determined relative to typical commercially available products. The project, which started in late 2013, will provide unique evidence of the physiological effects of modified SFA-reduced dairy products, which could contribute to food-based dietary recommendations for CVD risk reduction.
Effects of milk and dairy products on blood pressure and arterial stiffness
Hypertension, defined as systolic BP ⩾ 140 mm Hg and/or diastolic BP of ⩾ 90 mm Hg, is one of the leading risk factors in the development of stroke, CHD, heart failure and end stage renal disease( Reference Lawes, Vander Hoorn and Rodgers 75 ). BP is modifiable by environmental and lifestyle factors, with diet as one of the most important mediators( Reference Appel, Brands and Daniels 76 ). The Dietary Approaches to Stop Hypertension trial demonstrated that a diet consisting of reduced total and SFA fats, high intakes of low-fat dairy products, and fruits and vegetables significantly lowered BP in normotensive and hypertensive individuals( Reference Appel, Moore and Obarzanek 77 ). Moreover, the magnitude of BP reduction was of greater magnitude after the diet rich in low-fat dairy products compared with the fruit and vegetable-rich diet, which omitted dairy products altogether( Reference Appel, Moore and Obarzanek 77 ). The findings from cross-sectional and prospective studies have shown an inverse association between consumption of dairy products, particularly low-fat varieties and risk of hypertension( Reference Livingstone, Lovegrove and Cockcroft 48 , Reference Toledo, Delgado-Rodriguez and Estruch 55 , Reference van Meijl and Mensink 56 , Reference Pereira, Jacobs and Van Horn 78 – Reference Soedamah-Muthu, Verberne and Ding 83 ). These findings were confirmed in a meta-analysis by Soedamah-Muthu et al. ( Reference Soedamah-Muthu, Verberne and Ding 83 ), in which nine prospective cohort studies and a total of 57 256 participants, showed a reduced RR for hypertension (pooled RR 0·97; 95 % CI, 0·95, 0·99 per 200 g/d) and intake of total dairy( Reference Soedamah-Muthu, Verberne and Ding 83 ). A few RCT have examined the effects of dairy products on BP( Reference van Meijl and Mensink 56 , Reference Stancliffe, Thorpe and Zemel 84 , Reference Maki, Rains and Schild 85 ). For example, a randomised cross-over trial by Van Meijl and Mensink( Reference van Meijl and Mensink 56 ) in thirty-five healthy overweight and obese men and women indicated that daily consumption of low-fat dairy products compared with carbohydrate-rich products for 8 weeks, significantly reduced systolic BP by 2·9 mm Hg. However, a recent study by Maki et al. ( Reference Maki, Rains and Schild 85 ) observed no significant effects of consuming low-fat dairy products, compared with low-fat non-dairy products, on BP in sixty-two men and women with prehypertension or stage 1 hypertension( Reference Maki, Rains and Schild 85 ).
The impact of dairy foods on BP and other more novel markers of vascular health are becoming increasingly relevant. Increased central arterial stiffening is a hallmark of the ageing process and the consequence of many diseases such as diabetes, atherosclerosis and chronic renal failure. Arterial stiffness is also a marker for increased CVD risk, including myocardial infarction( Reference Mitchell, Moye and Braunwald 86 ), heart failure( Reference Chae, Pfeffer and Glynn 87 ) and total mortality( Reference Benetos, Safar and Rudnichi 88 ), as well as stroke( Reference Vaccarino, Berger and Abramson 89 ) and renal disease( Reference Blacher, Guerin and Pannier 90 ). Arterial stiffness is measured by pulse wave velocity and augmentation index, both of which are predictive of heart attacks and stroke ( Reference Boutouyrie, Tropeano and Asmar 91 ) and all-cause mortality( Reference Janner, Godtfredsen and Ladelund 92 ). Pulse wave velocity measures the speed of propagation along the artery, whereas augmentation index is calculated from the BP wave form and is based on the degree of wave reflection. Significant relationships between dairy product intake and arterial pulse wave velocity have been shown in a cross-sectional( Reference Crichton, Elias and Dore 93 ) and longitudinal( Reference Livingstone, Lovegrove and Cockcroft 48 ) cohort studies. Livingstone et al. ( Reference Livingstone, Lovegrove and Cockcroft 48 ) used data from the Caerphilly Prospective Study, based on 2512 men followed for a mean of 15 years and showed a significant inverse relationship between dairy product intake and augmentation index. The subjects in the highest quartile of dairy product intake (mean 480 g/d), excluding butter, had 2 % (P = 0·02) lower augmentation index compared with subjects with the lowest dairy intake (mean 154 g/d), whereas across increasing quartiles of butter intake there was no impact on augmentation index, but a significant increase in insulin, serum TAG and total cholesterol concentrations, and diastolic BP( Reference Livingstone, Lovegrove and Cockcroft 48 ).
The mechanisms by which milk and dairy products may reduce BP and arterial stiffness are unclear. It has been hypothesised that bioactive peptides released during milk protein digestion, dairy fermentation or industrially by enzyme or chemical treatments, may be involved in the relationship between dairy consumption and BP( Reference FitzGerald, Murray and Walsh 94 , Reference Boelsma and Kloek 95 ). It has been proposed that these bioactive peptides may inhibit the action of angiotensin I converting enzyme, thereby reducing blood levels of angiotensin, preventing blood vessel constriction, and modulating endothelial integrity. Ballard et al. ( Reference Ballard, Bruno and Seip 96 ) showed that consumption of 5 g whey-derived peptide daily for a 2-week period significantly improved brachial artery flow-mediated dilation response( Reference Ballard, Bruno and Seip 96 ). A further study reported that although whey and casein exerted similar hypotensive effects, whey protein supplementation (60 g/d for 12 weeks) significantly reduced augmentation index compared with casein (60 g/d for 12 weeks)( Reference Pal and Ellis 97 ), an effect that requires confirmation. There is also evidence to suggest that certain peptides from milk proteins may modulate the release of vasoconstrictor endothelin-1 by endothelial cells, thus preventing an increase in BP( Reference Maes, Van Camp and Vermeirssen 98 ). Milk also contains a variety of other biologically active components such as calcium, potassium and magnesium that may exert impact on BP and arterial stiffness( Reference Fekete, Givens and Lovegrove 99 ).
The weight of existing evidence indicates that milk and dairy products (excluding butter) are not associated with detrimental effects on CVD risk factors and mortality, and may even exert favourable effects on CVD risk, by lowering BP and arterial stiffness. While the specific mechanisms that underpin these effects are not clear, the unique nutritional composition of milk and dairy foods has been implicated in improving vascular function and in attenuating the LDL-C-raising property of SFA. Our present dietary guideline to reduce intake of dietary SFA to 10 % of the total energy to lower CVD risk is still valid, but the elimination of milk and dairy from our diet is clearly not an evidence-based strategy for achieving this aim.
We would like to thank Professor Ian Givens for his collaboration on this work.
Conflicts of Interests
The authors have previously received funding for research from AHDB Dairy. J. A. L. has acted as an advisor to the Dairy Council. J. A. L. has received funding for research from Volac for BBSRC case studentship and ‘in kind’ foods from Arla for an MRC funded study.
J. A. L. and D. A. H. are sole authors of this manuscript.