Dietary Fat and Heart Disease: Evaluating the Quality of Individual “Core” Trials

Dear Editor:

I recently read Lichtenstein's compelling supplement, “Dietary Fat and Cardiovascular Disease: Ebb and Flow Over the Last Half Century,” that traces the evolution of US dietary guidelines from the 1970s in the context of dietary fat and reviews the controversies and current evidence of dietary fat and cardiovascular disease (CVD) (1). However, the author's characterization of the evidence from randomized controlled trials (RCTs) in the section titled “Dietary Fat and Cardiovascular Disease: Present” starkly contrasts with that of a recent comprehensive review of the meta-analytical evidence (2). In opposition to the AHA Presidential Advisory on Dietary Fats and Cardiovascular Disease (3), the review of meta-analyses of RCTs found that the Advisory's core trials were not representative of the best available evidence and, because their recommendations hinge on the “core” trials, the strength of their recommendations was overstated (2). Although predefined inclusion criteria are commendable and transparent, it does not preclude researchers from evaluating the quality of the 4 “core” trials included.

Brief Review of the “Core” Trials

The dichotomy of “core” and “noncore” trials implies a hierarchy of evidence, or “good” and “bad” categories, and simplifies a complex topic to assist with public messaging. As such, the “core” trials are presented as the definitive studies on dietary fat and heart disease, especially the replacement of SFA with PUFA, mostly soybean oil. Unfortunately, in this case, the dichotomy is not an indication of the quality of the studies included. The “core” trials contain similar methodological and design flaws as the “noncore” trials (2).

Overall, major confounding variables erroneously favored the intervention (PUFA) groups. In the Finnish Mental Hospital Study, the control (SFA) group consumed ∼9 times more trans-fat (TFA), 15–49% more sugar, took 2 times more cardiotoxic medication, and <50% of the patient population completed both periods of the trial (4). For this and other reasons, the Finnish Mental Hospital Study was not included in 8 of the last 10 meta-analyses of RCTs (2). The Oslo Diet-Heart Study was at a high risk of bias owing to an unclear blinding protocol, because the study author knew the patient allocations and determined the diagnosis for soft endpoints such as angina; the intervention group received extensive nutrition counseling; and the intervention group ate 5 g/d more ω-3 fatty acids and restricted their TFA intake, whereas the control group consumed nearly 10% of energy as TFA (5). Similarly, the Los Angeles Veterans Administration Trial control group consumed ≥2% more trans-fat, had more heavy smokers, and only consumed ∼16% of the RDA for vitamin E per day, whereas the intervention group consumed nearly 150% of the RDA of vitamin E daily and exhibited poor adherence to the diet (6).

Given the limitations of each individual trial, it is unclear whether the effects on heart disease can be attributed to the replacement of SFA with PUFA. The substantial dietary differences, poor adherence and follow-up, high risk of bias, and excessive TFA intake in the control (SFA) groups of the “core” trials seem to contradict Lichtenstein's conclusion that, “The majority of the data indicate that replacing food sources of saturated with unsaturated fat, particularly polyunsaturated fat, is associated with lower risk” (1).

An Avenue to a Consensus?

A recent meta-analysis accounted for the aforementioned flaws by differentiating trials as adequately or inadequately controlled (7). Based on 5 adequately controlled trials, the replacement of SFA with PUFA (mostly ω-6 PUFAs) was unlikely to decrease coronary heart disease (CHD) events, CHD mortality, or total mortality.

Ramsden et al. (8) reanalyzed unpublished data from a diet-heart trial and found that, despite a 13.3% drop in total cholesterol, all-cause mortality (HR: 1.62; 95% CI: 1.00, 2.64; P = 0.051), CVD mortality (HR: 1.70; 95% CI: 1.03, 2.80; P = 0.04), and CHD mortality (HR: 1.74; 95% CI: 1.04, 2.91; P = 0.04) were higher in the PUFA group. Although controversy exists regarding the unknown TFA intake of each group, the drop in cholesterol and restriction of other forms of trans-fat (biscuits, cakes, and pastries) indicate that the intervention group most likely consumed either similar amounts of, or less, TFA. Interestingly, Ramsden et al. (8) also reported that higher risk of CVD mortality was related to the magnitude of increase of ω-6 intake. In agreement, a recent Cochrane systematic review found that consuming high amounts of ω-6 compared with low intake led to a significant increase in CVD mortality (RR: 1.33; 95% CI: 0.86, 2.05; P = 0.04) (9).

Based on the best available evidence, it may be more prudent to recommend a greater emphasis on ω-3 intake; reducing concentrated sources of ω-6 such as seed oils; consuming adquate dairy products; and switching to unprocessed sources of lean meats. The relation between dietary fat and heart disease is a highly contentious topic that tends to spark professional debate across the spectrum of health care and academia that can unintentionally lead to public confusion and distrust. It is apparent that we share the same concern for appropriate messaging to the public and I sincerely hope that our dialogue strengthens the clarity of this message and subsequent nutrition recommendations.

Notes

The author reported no funding received for this work.

Author disclosures: The author reports no conflicts of interest.

Abbreviations used: CVD, cardiovascular disease; RCT, randomized controlled trial; TFA, trans-fatty acid.