Fat and Heart Disease:
New Research or Dubious Science?

by Fiona Chew, Ph.D., Professor, S.I. Newhouse School of Public Communications, Syracuse University, N.Y.

“Butter Is Back” and “Study Questions Saturated Fat and Heart Disease Link” were among the many eye-grabbing headlines published across mainstream media following a March 2014 report that questioned the detrimental health effects of saturated fats.

Reprinted with permission, Center for Science in the Public Interest. ©Dennis Cox/fotolia.com

But, tempting as it may be, is it prudent to push aside prevailing counsel and order that steak with buttered potatoes? Science not only asks the question; it provides the answer.

The Controversy
Public health officials and health-promoting organizations have long categorized saturated fats as “bad” fats because of their documented harmful effects on the heart, and unsaturated fats as “good” fats because of their positive effects on the same organ. It was, therefore, little surprise that when Rajiv Chowdhury, M.D., Ph.D., and his cohort questioned conventional wisdom and long-standing dietary recommendations to avoid foods containing saturated fats (e.g., beef, cheese and butter) and replace them with more foods containing unsaturated fats (e.g., nuts, vegetable oils and salmon), a faction ensued. Meanwhile, as pundits, physicians and policymakers pontificated, the American public watched, read and waited, wondering whom to believe and what, if anything, to do.

Conclusions from the New Research
Chowdhury et al.’s controversial research appeared as an article entitled “Association of Dietary, Circulating, and Supplement Fatty Acids With Coronary Risk: A Systematic Review and Meta-analysis” in the March 18, 2014 issue of the Annals of Internal Medicine.

Comprising three meta-analyses, the review by the group of international researchers concluded that saturated fats were not as detrimental as previously believed, indicating that, overall, the evidence did not support higher consumption of polyunsaturated fats or monounsaturated fats, and lower consumption of saturated fats to reduce heart disease risk. The review did confirm, however, that trans fats were linked to higher heart disease risk.

While Chowdhury et al.’s conclusions received significant media attention and were consistent with Siri-Tarino et al.’s 2010 research in American Society of Nutrition suggesting a lack of evidence linking saturated fat to an increased risk of heart disease, the science behind the controversy finds no substantive data to support the meta-analysis. In fact, it repudiates it as a potential new paradigm.

New Analysis of Old Research
To arrive at their conclusions, Chowdhury et al. aggregated the results of 76 completed scientific research studies, including: 32 observational studies with 512,420 participants that examined the relationship between dietary fatty acids intake and coronary risk; 17 observational studies with 25,727 participants looking at fatty acid biomarkers and coronary risk; and 27 randomized controlled trials with 105,085 participants investigating the effect of fatty acid supplementation on coronary outcomes.

In the observational studies, the researchers compared relative risk of coronary heart disease among participants consuming different types of fat – saturated, polyunsaturated, monounsaturated and trans fats – and different subtypes of fatty acids within each of these fat types. They found no statistically significant differences in high versus low consumption of saturated fats, monounsaturated fats and polyunsaturated fats or fatty acid subtypes and coronary risk. Higher consumption of trans fats (and total trans fatty acids), however, was associated with a statistically significant higher risk.

The meta-analysis of randomized controlled trials examined supplementation with linoleic acid and omega-3 and omega-6 fatty acids, and found no statistically significant reduction in heart disease risks. Since the results were based on the aggregation of existing studies and limited to the observed participants, who may or may not have been optimal candidates for testing specific associations, the researchers called for further trials in the area of supplementation.

As mentioned above, the results of this Chowdhury et al. meta-analysis of clinical trials were consistent with Siri-Tarino et al.’s 2010 analysis, which aggregated 21 studies with 347,747 participants who were tracked over a period of 5 to 23 years and found no significant evidence supporting the association of dietary fat consumption to an increased risk of coronary heart disease, stroke or cardiovascular disease (CVD).

Reactions from the Scientific Community
Just as it did in 2010 in the wake of Siri-Tarino et al., the scientific community strongly critiqued and commented on the Chowdhury et al. meta-analysis and its conclusions about fat consumption, with many scientists publishing comments in the Annals of Internal Medicine “Letters and Comments” section.

Drs. Walter Willett, Frank Sacks and Meir Stampfer of Harvard University’s School of Public Health pointed out discrepancies in the number of studies included in the review’s fatty acids biomarker analysis and the varying results (non-significant in the former versus significant in the latter). From the supplement table data, they concluded that the results for "both intake and biomarkers for long-chain ω-3 fatty acids support benefit." They also reported that the varied findings in the randomized controlled trials were expected, because many in the study population had relatively high ω-3 fatty acids consumption, which would therefore show little benefit (Annals of Internal Medicine. 161:6:453, 2014).

Willett et al. additionally identified several drawbacks to the analyses sampled, including: erroneous data in an important study that was included in the pooled analysis; inclusion of participants with prevalent CVD at baseline, instead of only healthy participants; and lack of acknowledgement by the meta-analysts of earlier pooled analyses that allowed direct comparisons between different fats, indicating that replacing saturated fatty acids (SFAs) with polyunsaturated fatty acids (PUFAs) was associated with lower coronary risks. Likewise, Willett et al. argued that the Chowdhury et al. review did not acknowledge data demonstrating the reduction of low-density lipoprotein cholesterol (LDL) when SFAs were replaced with monounsaturated fatty acids (MUFAs) or PUFAs. They maintain that a comprehensive review of the scientific literature would support the replacement of SFAs with PUFAs to reduce CVD risk.

Other scientists stated that Chowdhury et al. misinterpreted the results of some of the studies, including Drs. Christine Dawczynski, Marcus E. Kleber, Winfried März, Gerhard Jahreis and Stefan Lorkowski, who said data from their own meta-analysis using some of the same studies showed a positive association between saturated fats and coronary outcomes (Annals of Internal Medicine. 161:6:453-454, 2014). Other researchers, such as Drs. Frank Davidoff and Irwin H. Rosenberg, accused the meta-analysts of ecological fallacy and extrapolating results based on individuals or subgroups in a study population to the entire study population (Annals of Internal Medicine. 161:6:454, 2014). Still other critiques drew attention to differential results based on the inclusion or exclusion of studies, and raised the question of omission bias (Liebman, Katan & Jacobson, Annals of Internal Medicine. 161:6:454- 455, 2014; Morenga, Mann & Skeaff, Annals of Internal Medicine. 161:6:455, 2014).

Finally, there were calls by experts to: further examine and differentiate the dietary sources of fatty acids (McCaulley, Annals of Internal Medicine. 161:6:456, 2014; Schwingshackl & Hoffman, Annals of Internal Medicine. 161:6:455-456, 2014); account for the effect of the replacement food, such as sugar and carbohydrates, on coronary risk (Diekman et al., Annals of Internal Medicine. 161:6:456-457, 2014); develop dietary guidelines based on the totality of evidence (Geleijnse, Brouwer & Kromhout, Annals of Internal Medicine. 161:6:457- 458, 2014); and focus on a “whole-diet” approach (O’Neil & Itsiopoulos, Annals of Internal Medicine. 161:6:458, 2014).

Media Storm
News organizations and medical/health journalists were also swift to round up various perspectives. The New York Times quoted Dr. Frank Hu, nutrition and epidemiology professor at Harvard’s School of Public Health, as cautioning that the Chowdhury et al. findings did not represent “a green light” for more consumption of foods high in saturated fats, such as steak and butter. Advising the public to eat Mediterranean diet foods such as fish, high-fiber grains, olive oil, nuts and avocados, Hu emphasized that it was misleading to analyze individual nutrients in isolation because, when saturated fats are reduced, they are often replaced by refined carbohydrates, which can also increase heart disease risk. The focus on a whole diet approach was echoed by Yale University nutrition science professor Dr. David Katz.

New York University nutrition professor Dr. Marion Nestle indicated in The Boston Globe that the Chowdhury et al. study created confusion for everyone, and advised the public to eat all things in moderation.
Meanwhile, Dr. Dariush Mozaffarian, a co-author of the 2014 meta-analytic study, defended his and his cohort’s conclusions, elaborating in The Boston Globe that, in their review, saturated fat was found to increase the size of the LDL, or artery-clogging cholesterol, particle, but not the number of LDL particles, which scientists now consider the main culprit for higher heart disease risk.

Scientific Evidence
While Chowdhury et al.’s review questioned the link between saturated fats and CVD, the cumulative scientific data, in conjunction with the scientific community, continue to support the connection.

The Nutrition Evidence Library (NEL) in the U.S. Department of Agriculture’s Center for Nutrition Policy and Promotion houses the complete body of available scientific evidence to address food and nutrition questions, and it has become the nexus of facts – and the link between factions – in this larger-than-life health discussion. Conducting systematic reviews of the findings under its purview, the library fully documents available studies and provides evidence on the association of various types of fats and CVD risk, among other things. In fact, many of the findings establishing the relationship between fatty acids and coronary heart disease were predicated on the library’s systematic reviews.

Saturated fatty acids. A 2011 NEL review of 12 recent studies, including 10 randomized controlled trials, one non-randomized trial and a meta-analysis of 11 cohort studies, resulted in strong evidence showing that SFAs have a positive association with increased risk of CVD. The review also revealed that decreasing SFA consumption conversely led to CVD declines. Specifically, it found that when there was a 5 percent decrease in SFA energy which was replaced by MUFAs or PUFAs, CVD risk dropped. A total of more than 344,900 subjects participated in the studies examined.

Polyunsaturated fatty acids. A NEL review of 10 studies involving more than 425,000 subjects from one meta-analysis of 11 cohort studies, five randomized controlled trials and four cohort studies provided strong and consistent evidence that n-6 PUFAs were linked to improved blood lipids associated with CVD, particularly when PUFAs replaced dietary SFAs or trans fatty acids. Replacing SFAs with PUFAs resulted in decreased total cholesterol, LDL cholesterol and triglycerides, the evaluation found, with PUFA consumption significantly leading to declines in CVD risk.

Another NEL review of 28 studies comprising nine meta-analyses, four randomized controlled trials and 15 cohort studies provided moderate evidence that consuming two servings of seafood equivalent to 250 mg per day of long-chain n-3 fatty acids was linked to reduced deaths from coronary heart disease or sudden death in persons with CVD. Meanwhile, eight studies examining the relationship between consumption of plant n-3 PUFAs and CVD risk showed limited, but supportive, evidence based on NEL review that higher consumption of plant-based n-3 fatty acids may be associated with reduced mortality among persons with CVD.

Monounsaturated fatty acids. A NEL review of 13 studies, including one meta-analysis of 11 cohort studies, 11 randomized controlled trials and one cohort study with more than 350,500 subjects, yielded strong evidence that when MUFAs replaced SFAs, blood lipids related to CVD improved and CVD risk decreased.

Trans fatty acids. Results from a 2006 meta-analysis of four cohort studies with nearly 400,000 subjects by Chowdhury et al. co-author David Mozaffarian and other researchers in the New England Journal of Medicine showed a higher increase in CVD risk from a higher intake of energy from trans fatty acids.

Overall, the scientific record consistently has shown that saturated fats and trans fats have an adverse association with CVD, and that intake of these fats increases CVD risk, while replacing them with polyunsaturated or monounsaturated fats reduces that risk.

Higher vs. Lower Fat Intake
Chowdhury et al.’s conclusion that “[c]urrent evidence does not clearly support cardiovascular guidelines that encourage high consumption of polyunsaturated fatty acids and low consumption of total saturated fats,” was based on 49 observational studies. The authors aggregated the comparison of the top third and bottom third of participants’ dietary intake of saturated fats, polyunsaturated fats and monounsaturated fats among 538,141 participants. They did not report that one fat was better than another fat, only that higher intake versus lower intake of each type of fat was not associated with a higher or lower coronary risk. However, no absolute measures of the amount of fatty acids consumed were listed, since only the top and bottom thirds were assessed; therefore, these represent relative “highs” and “lows,” and may not denote actual high and low amounts of fats.

Additionally, Chowdhury et al. analyzed the pooled effects of 27 randomized controlled trials with 105,085 participants and compared the supplementation of linoleic acid, long chain omega-3 fatty acid and omega-6 fatty acid. They concluded that “supplementation with these nutrients does not statistically significantly reduce the risk for coronary outcomes,” but accepted data demonstrate otherwise.

SFAs are DOA
Despite the attention-grabbing headlines and heated debate surrounding the Chowdhury et al. analysis, butter is decidedly not back. It is squarely where it was before the controversy began. Perhaps the fact that there were polemics at all underlines, at least in part, that the permission to indulge has strong supporters.

In short, Chowdhury et al. gave us an “in.” It was a narrow entrée, with little support, but it opened the door to questions – and the possibility of a completely guiltless cheeseburger, pizza or steak. And, if nothing else, it reminded professionals and the public alike that our renouncement of SFAs is backed by reams of strong, clear and exacting research.

This scientific evidence stands up to scrutiny. In doing so, it has established the consistent association between the consumption of saturated fats and trans fats, and the increase in CVD risk. Likewise, it finds and substantiates that replacing saturated fats with polyunsaturated or monounsaturated fats is linked to lower CVD risk.

Questions, no matter how big and broad, beg answers, however, and further research related to Chowdhury et al. is recommended. In addition to macro studies examining the quality of the whole diet and the replacement of saturated fats with carbohydrates, micro studies examining the chain length of saturated fats and the way in which the body metabolizes these fats, as well as investigations into the quality of gut microbiomes, merit examination.

Read More:
Spotlight Article: Dietary Guidelines 2015 Offer New Advice
Also Noted: Dietary Guidelines 2015 for Individuals and Families
CECHE News: Freedom House Forum Honors Mark Palmer