Put down the apple and have some cheddar

    Although both cheese and meat are lumped into the “watch out!” category in heart health recommendations, dairy products often show neutral or positive associations with cardiovascular health. But how do cheese-rich diets fare in randomized trials when compared to other diets? This trial tested three diets against each other in a highly controlled fashion: a cheese diet, meat diet, and high-carb diet.

    Study under review: Diets with high-fat cheese, high-fat meat, or carbohydrate on cardiovascular risk markers in overweight postmenopausal women: a randomized crossover trial

    Over the past several decades, we have made slow progress toward recognizing that not all fatty acids are the same. Saturated fat now has subcategories of short-, medium-, and long-chained fatty acids that determine how they are metabolically handled in the body. Even within the sub-categories there are differences in the blood lipid effects of fatty acids. For example, palmitic acid is more atherogenic than stearic acid, even though both are long-chained saturated fatty acids.

    Recently, there has been an increased interest in the kind of food supplying the fat. Dairy products are an excellent example because the impact of dairy fat on blood lipids is very inconsistent. For example, most research shows no association between cardiovascular disease risk and cheese intake. Intervention trials suggest that there isn’t a similarly harmful effect on blood lipids when increasing fat intake from cheese compared to fat intake from butter.

    Despite this emerging evidence, health authorities like the American Heart Association continue to recommend limiting saturated fat intake primarily through reducing the intake of fatty meats, butter, and cheese. Therefore, it seems unlikely that an individual would replace butter with cheese in their diet in an attempt to improve their health. A more likely scenario would be replacing fat with carbohydrates, which is currently advocated by many health authorities around the world. Additionally, others may consider replacing fatty meat in the diet with low-fat cheese and lean meats. The study under review aimed to explore the effect of cheese and meat as sources of saturated fat, and their replacement with carbohydrates, on cardiovascular disease risk markers.

    While some official recommendations suggest limiting saturated fat intake to lower the risk of cardiovascular disease, recent evidence suggests that the type of saturated fat and the food source from which it is ingested matter. This study took a look at how fat from cheese or meat affected markers for cardiovascular risk markers, compared to carbohydrates.

    Who and what was studied?

    Overweight but otherwise healthy postmenopausal women were recruited from the University of Copenhagen area to undergo a randomized, crossover, open-label (both the researchers and participants know which treatment is being administered) intervention consisting of three diet phases lasting two weeks each, separated by two-week washout periods. The study protocol is summarized in Figure 1.

    All foods were provided to the participants for every intervention phase. The amount of food was based on estimated energy requirements to encourage weight maintenance. Two of the diets had a high saturated fat content supplied from different foods (cheese or meat) and were designed to match the macronutrient composition of the average Danish diet (15% protein, 35% fat [15% saturated fat], and 50% carbohydrates). The third diet (CHO diet) served as a realistic control diet that was higher in carbohydrate (60%) and lower in fat (25%) than the cheese or meat diet, since current global recommendations for a healthy diet promote consuming less fat from dairy and meat in favor of carbohydrate-containing foods.

    All three diets were matched for their ratio of saturated (SFAs), monounsaturated (MUFAs), and polyunsaturated (PUFAs) fatty acids, as well as dietary protein (15%), fiber, sugar, cholesterol, and sodium. Controlling these factors allowed for a comparison of food-specific SFAs (cheese vs meat) and differences in micronutrients (e.g. calcium).


    Postmenopausal women consumed three highly controlled weight-maintenance diets for two weeks each that differed only in their macronutrient composition (cheese and meat diets vs. CHO diet) or in their primary source of dietary fat (cheese vs. meat).

    What were the findings?

    As intended by experimental design, bodyweight remained stable throughout all three interventions. Waist and hip circumference, blood pressure, fasting glucose and insulin, and HOMA-IR (an estimate of insulin resistance) also did not differ significantly.

    The main findings of this study (summarized in Figure 2) were that the cheese diet and meat diet, respectively, resulted in a 5% and 8% greater HDL-c concentration and an 8% and 4% greater ApoA-I concentration, compared to the CHO diet, with no significant difference between the two. However, the cheese diet also resulted in a significantly lower apoB:apoA-I ratio (-5%) than the CHO diet, whereas the meat diet did not affect the apoB:apoA-I ratio. There were no significant differences in other blood lipids (total cholesterol, LDL-c, triglycerides, apoB, total cholesterol to HDL-c ratio, or LDL-c to HDL-c ratio) between the three diets.


    An apolipoprotein (apo) is a protein that binds lipids such as fat and cholesterol to form lipoproteins. ApoA-I is the major protein component of HDL, while apoB is the major protein component of VLDL and LDL. There is also a subclass of apoB that forms chylomicrons. It has been argued that apoB is superior to LDL-c in estimating the risk of cardiovascular disease, and that apoA-I is a stronger predictor of heart attack and stroke risk than HDL-c. Moreover, the ratio between apoB and apoA-I is considered an excellent as well as simple predictor of adverse cardiovascular events – the lower the ratio, the lower the risk. In addition to changes in some blood lipids, the cheese and meat diets resulted in a 40 and 21% greater fecal fat excretion and 28% and 39% greater fecal bile acid excretion, respectively, than the CHO diet. The fecal fat excretion was significantly different between the cheese and meat diet, but this difference amounted to only 0.9 grams per day (eight to nine kcal) and there were no differences among the three diets in fecal energy excretion.

    The cheese and meat diets resulted in a more beneficial blood lipid profile than the CHO diet through their effects on HDL-c and Apo A-I, and resulted in a greater fecal excretion of fat and bile. The cheese diet also showed a small advantage over the meat diet through its reduction of the apoB:apo A-I ratio.

    What does the study really tell us?

    This study’s main goal was to compare the effects of a high-SFA diet on blood lipids when the primary source of saturated fat was obtained from cheese or meat. The results indicate that cheese may have small but significant benefits over meat. Additionally, the study showed that reducing total fat intake from 35% to 25% of calories and replacing those calories with starchy carbohydrates results in unfavorable changes in blood lipids.

    The study had a strong crossover design with highly controlled dietary interventions. All diets were matched for calories, the SFA:MUFA:PUFA ratio, dietary protein, fiber, sugar, cholesterol, and sodium, which made it possible to compare the food-specific effects of cheese and meat. The cheese diet’s primary source of fat was cheese, which averaged 96-120 grams or about 3.5 to four ounces per day and provided 38% of the total fat intake. Another 11% of total fat intake was consumed as lean cuts of beef and pork. In comparison, the meat diet contained an average of 164 grams or about 5.8 ounces per day of fatty beef and pork, which provided 52% of the total fat intake. It is possible that the different sources of protein may have influenced the results, but no human studies have compared the blood lipid effects of casein and meat protein.

    The major micronutrient difference between the cheese and meat diet was calcium intake, which was purposely not matched because it is unlikely that an individual who replaces cheese in the diet with meat would supplement calcium to compensate. Thus, the three-fold difference in calcium intake between the cheese and meat diets (1278 vs. 401 milligrams) is more representative of a real-world outcome. Previous research has shown that high calcium intake independently attenuates SFA-induced increases in total and LDL-cholesterol without affecting HDL-c. It was suggested that these benefits were attributable to calcium’s ability to interfere with fat absorption. Although the current study failed to find a significant difference in total and LDL-cholesterol between the cheese and meat diets, the apoB:apo A-I ratio was significantly less with the cheese diet. However, the difference in fecal fat excretion amounted to less than one gram per day and there was no difference in fecal energy excretion, suggesting that something other than calcium was responsible for the superiority of the cheese diet.

    One limitation that must be acknowledged is the study sample of postmenopausal women, which makes generalization to more heterogeneous populations difficult. Additionally, the sample size was very small, with only 14 people completing all three interventions, and the intervention length was relatively short, lasting only two weeks per diet. This may have caused the study to be underpowered for detecting some differences in other cardiovascular risk markers. For instance, if we look at the non-significant findings of this study, it appears that the cheese diet may have prevented increases in total and LDL-cholesterol seen with the meat diet, reduced the total cholesterol to HDL-c ratio and LDL-c to HDL-c ratio, and reduced apoB concentrations.

    Finally, it’s important to note that this study was 100% funded by several international dairy institutes, and one of the six authors is a member of the U.S. Global Dairy Platform advisory board and is currently the principal investigator of research projects supported by grants from several of the funders of this study. However, this researcher only supplied scientific consultation and the funders had no say in the study design, analysis and interpretation of data, or writing of the manuscript.

    This study tells us that a diet higher in fat (35% vs. 25%) and lower in carbohydrates (50% vs. 60%) has a favorable impact on blood lipids when the SFA:MUFA:PUFA ratio, and amount of dietary protein, fiber, sugar, cholesterol, and sodium are held constant. Eating a diet that replaces some meat with cheese may have additional benefits mediated through factors such as the increased calcium intake.

    The big picture

    This study builds upon previous work that suggests the food matrix in which nutrients are obtained plays an important role in health. In volume one of the ninth NERD we discussed a study that suggested the milk-fat globule membrane (MFGM) mediated the inconsistent effects of dairy fat on blood lipids. The study compared heavy cream that maintains an intact MFGM to butter that has the MFGM destroyed during the churning process. The results showed that consuming dairy fat with the MFGM intact led to significant reductions in total and LDL-cholesterol, apoB, and the apoB:apo A-I ratio, but that consuming butter had the opposite effects. Both types of dairy fat led to similar increases in HDL-c and apo A-I.

    Cheese is a type of dairy that maintains an intact MFGM, and in the current study, led to a significantly lower apoB:apo A-I ratio than meat, along with non-significant improvements in total and LDL-cholesterol and apoB. Additionally, the cheese and meat diets led to similar improvements in HDL-c and apo A-I, which were increased relative to the CHO diet. These results corroborate the findings of the MFGM study and suggest that the MFGM’s effects are limited to LDL cholesterol and that HDL cholesterol is more responsive to changes in total fat intake regardless of source.

    In the big picture of the MFGM study review, we also discussed a handful of animal studies, which suggested that reduced cholesterol absorption or phospholipid-induced alterations in liver gene expression might mediate the effects of the MFGM on blood lipids. Although the study didn’t measure fecal output, it did confirm that consuming dairy with an intact MFGM led to changes in gene expression in humans. The current study did measure fecal output and showed that there was a clinically irrelevant difference in fat excretion and no difference in bile excretion between the cheese and meat diets. Together, these findings suggest the benefits of the MFGM may be attributable to changes in how the body processes blood lipids rather than through interference with absorption.

    Although absorption interference appears less likely to be responsible for the effects of the MFGM and changes in “bad” cholesterol, it may explain the significant increase in HDL-c and apo A-I in the cheese and meat diets relative to the CHO diet. Increased fecal bile acid excretion by drugs that block the reabsorption of bile acid has been shown to reduce the stimulation of the farnesoid X receptor (FXR) in the intestines and liver, which in turn has been shown to increase apo A-I gene expression and HDL-particle remodeling. This is consistent with the current study findings that the cheese and meat diets had significantly greater fecal bile acid excretion, HDL-c, and apo A-I than the CHO diet.

    It has been suggested that greater fat intake leads to greater bile acid excretion, which is consistent with the current study results. It also helps to explain why other research has shown HDL-c to be increased with low-carbohydrate diets compared to low-fat diets, and diets enriched with SFAs to increase apo A-I gene expression, while diets enriched with glucose decreased its expression.

    The superiority of a cheese-enriched diet compared to a meat-based diet may be partly due to the presence of an intact MFGM in the cheese that acts to mediate regulation of LDL-c and apoB in the liver. On the other hand, the ability of a higher fat diet to increase HDL-c and apoA-I may be owed to an increased excretion of bile acids with increasing fat intake. In other words, the benefits of the MFGM may be related to changes in how the body processes LDL cholesterol while total fat intake mediates changes in HDL cholesterol.

    Frequently Asked Questions

    Q. What is the difference between HDL and apoA-I or LDL and apoB?

    Since cholesterol doesn’t dissolve in water, it’s carried around in blood by lipoproteins. A coating of apolipoproteins on these lipoproteins helps them with transport and cell interaction. The lipoprotein called apoA-I is mostly on HDL particles, while apoB is on both LDL and other particles (namely VLDL and IDL). Figure 3 shows targets for the apoB:apoA-I ratio for men and women, as this number can quite be a useful measure to track for cardiovascular disease prevention.


    LDL particles are the ones that typically get stuck within artery linings and through inflammatory processes can propagate heart disease. However, as has been discussed a few times in NERD, LDL is not categorically “bad” and HDL is not categorically “good.” HDL-increasing drugs have been some of the biggest failures in major heart-related clinical trials, while many people that suffer heart attacks have normal or even low LDL cholesterol. This may be because high concentrations of LDL are more likely to be causative in heart disease whereas low levels of HDL are more permissive of heart disease.

    Q. Do different meats act similarly, and do different cheeses act similarly, with regard to lipid impacts?

    Beef and chicken diets with similar saturated fat content appear to produce similar lipid changes. Cheeses can vary substantially in their nutritional content, with wide ranges of total fat and protein concentrations and some differences in micronutrients, as well as substantially different processing (e.g. raw vs. pasteurized cheese). Different cheeses can also have different levels of conjugated linoleic acid, which may have some impact on heart disease risk.

    This topic has not been investigated in many randomized trials. The fact that diets are typically mixed (e.g. not composed of all brie cheese or only beef products as protein/fat sources) makes applying trials to real life more difficult.

    What should I know?

    The food we eat matters, at least when it comes to blood lipids. Evidence continues to accumulate that certain forms of dairy have a neutral or beneficial effect on cardiovascular disease risk markers, and these benefits may be owed to the presence of an intact MFGM. The current study demonstrated that consuming one-third of total fat intake as cheese with 11% from lean beef and pork is more beneficial than consuming half of total fat intake from fatty beef and pork.

    Although the amount of cheese consumed daily to obtain this amount may be considered unrealistic for most (about four ounces per day), if the beneficial effects are mediated by the MFGM, then it could be expected that similar effects occur with consumption of dairy fat from non-homogenized cream, milk, and yogurt – basically dairy that has not been mechanically altered. No study has yet to determine if a threshold limit for these benefits exists, so it may be prudent to consume some of these dairy products daily.