In Cochrane's 2020 analysis of saturated fat, the authors assessed various cut-offs for saturated fat intake.
As they explained (1):
To explore the WHO NUGAG specific question about the effect of the population consuming < 10%E as SFA relative to > 10%E SFA, we assessed effects of all studies where the mean assessed intervention SFA intake was < 10%E and the mean control SFA intake was > 10%E. We explored the effect of reduction of %E from SFA in gradual increments by using cut-offs of 7%E (where all studies with a mean intervention SFA intake < 7%E and mean control SFA intake > 7%E were pooled), 8%, 9%, 10%, 11%, 12% and 13%.
Based on the results, they provided the following figure:
(Image: Taken from reference 1. Each cut-off consists of studies where the intervention group had saturated fat intakes less than the cut-off and the control group had saturated fat intakes greater than the cut-off)
According to the Cochrane authors:
The figure suggests reductions in cardiovascular outcomes in studies where saturated fat intake was greater than 10%E in control groups, and less than 10%E in intervention groups.
But the graph is only as valid as the underlying data.
Therefore, let us take a closer look.
The 10% Cut-off
The "suggestion" of benefit that Cochrane claims for lowering saturated fat below 10% is based on the following pooled analysis of five trials:
To use Cochrane's own words, however, this analysis is "difficult to interpret." The data are limited, there is large heterogeneity, and the pooled estimate is imprecise.
Second, the "suggestion" of benefit is driven by the STARS trial — a trial that should have never been included in the first place.
Here is the analysis without STARS:
A very different result, which no longer "suggests" a benefit (to use Cochrane's interpretation).
As explained in a previous post, STARS involved much more than saturated fat reductions. The intervention also involved reductions in junk foods (e.g., cookies, pastry, cakes), reductions in trans fat, increases in omega-3 (doubled EPA+DHA), and a lower caloric intake (with weight loss).
Thus, we have no way of separating the possible effects of saturated fat from all these changes.
The Cochrane investigators, in fact, excluded STARS from their polyunsaturated fat (PUFA) reviews because it was "multifactorial."
But if STARS could not assess the specific effects of PUFA because it was multifactorial, then it also cannot assess the specific effects of saturated fat.
Furthermore, STARS had other major limitations, including:
A highly selected group of male patients (many with familial hypercholesterolemia — a condition irrelevant to the general population)
A drug run-in period prior to randomization
Lack of adequate blinding
Not designed to test clinical events
No evidence of proper randomization or allocation concealment
Systematic differences in care between the groups
Potential conflicts of interest (funded in part by Unilever and Bristol-Myers Ltd)
Thus, not only was STARS incapable of testing saturated fat due to its clear multifactorial intervention, but it was also a poor-quality trial that could not tell us anything reliable about clinical events.
Harder Endpoints
The alleged cardiovascular benefit in Cochrane's analysis for lowering saturated fat is dependent on very soft, subjective cardiovascular endpoints.
Worse still, Cochrane's analysis included six trials that were not designed to evaluate cardiovascular events or did not have evidence of blinding, including for outcome assessment (Black 1994, Ley 2004, Moy 2001, Oslo Diet-Heart 1966, and STARS 1992).
In the STARS trial, for example, about 71% of the cardiovascular events — a post-hoc endpoint — were very soft outcomes susceptible to bias.
As Ramsden and colleagues stated in 2016 (2):
Unfortunately, several critical deficiencies in the collection and reporting of non-fatal CHD events in these RCTs [randomized controlled trials] make it unfeasible to reliably estimate risk of composite CHD or non-fatal events.
Thus, a more reliable analysis would look at the harder outcomes.
To this end, I removed the very soft cardiovascular endpoints from Cochrane's 10% cut-off analysis and found no suggestion of benefit for lowering saturated fat:
(Image: The 10% cut-off analysis after excluding the more subjective cardiovascular events like angina. The endpoint here mostly consists of cardiovascular deaths, non-fatal heart attacks, and non-fatal strokes)
Indeed, taken at face value, Cochrane's analysis for the 10% cut-off did not suggest a benefit for all-cause mortality, cardiovascular mortality, myocardial infarctions, and CHD mortality.
For example, here is the forest plot for CHD mortality:
(Image: Incidentally, this is identical to Cochrane's "low summary risk of bias" analysis)
And for all-cause mortality:
(Image: Similar to their overall analysis, there was no evidence of a mortality benefit for lowering saturated fat below 10%E. Note the lack of heterogeneity and improved precision for this objective endpoint)
A Look at the Individual Trials
Of the five trials included in Cochrane's 10% cut-off analysis, only three were designed to test dietary changes in relation to cardiovascular outcomes or mortality: Sydney Diet-Heart 1978, Veterans Admin 1969, and WHI 2006.
The Sydney Diet-Heart Study evaluated the effect of lowering saturated fat and increasing polyunsaturated fat on survival, but there was no evidence of benefit for the lower saturated fat group, and the results even indicated potential harm (3):
In this cohort, substituting dietary linoleic acid in place of saturated fats increased the rates of death from all causes, coronary heart disease, and cardiovascular disease.
The Veterans Admin trial, on the other hand, is often cited as evidence for lowering saturated fat. But this is based on a post-hoc endpoint. There were no statistically significant effects on the primary endpoint (sudden death or definite MI) or total mortality (4):
We consider our own trial, with or without the support of other published data, to have fallen short of providing a definitive and final answer concerning dietary prevention of heart disease.
The third trial, WHI, also failed to find a benefit, this time for a dietary pattern lower in total fat combined with increases in vegetables, fruits, and grains (5):
Over a mean of 8.1 years, a dietary intervention that reduced total fat intake and increased intakes of vegetables, fruits, and grains did not significantly reduce the risk of CHD, stroke, or CVD in postmenopausal women.
Lastly, we could cite another relevant trial that Cochrane did not include in their analysis: The Minnesota Coronary Experiment.
The Minnesota Coronary Experiment was designed to test the effects of replacing saturated fat with omega-6 linoleic acid on cardiovascular events and deaths.
According to the data, the experimental group reduced their saturated fat intake to less than 10%E, but there was no evidence of benefit (2):
(Image: Taken from reference 2. Like other trials, the Minnesota Coronary Experiment did not find evidence of an all-cause mortality or CHD mortality benefit for lowering saturated fat to 9.2%E)
Therefore, I agree with Dr. Gershuni, who stated (6):
The existing evidence does not support the notion that dietary SFA causes heart disease. There is no demonstrable benefit for reducing SFA to < 10%.
The Lower Cut-offs
In a 2015 presentation, the lead author of Cochrane's review admitted the following (7):
We also did something about testing cut-offs for saturated fat intake for WHO . . . I don't think it's hugely interesting. It's very weak evidence for looking at how you might set a cut-off for saturated fat intake.
This "very weak evidence," I will add, becomes meaningless in the cut-offs below 10%.
To illustrate, let us look at the 7% and 8% cut-offs:
These bottom cut-offs are based on a single trial conducted by Black et al. — a trial with only two cardiovascular events.
(Image: As highlighted, the Black 1994 trial is based on two cardiovascular events)
Furthermore, the Black trial was at "moderate to high" risk of bias, was designed to evaluate skin cancer (not cardiovascular events), and did not actually test saturated fat.
The intervention, for instance, included not only a reduction in saturated fat but also a reduction in vegetable/seed oils (8).
Of note, intakes of omega-6 linoleic acid dropped to under 4%E, and Black himself attributed the apparent benefit on skin cancer mainly to the lower omega-6 (9):
Thus, the influence of fat on MNSC [non-melanoma skin cancer] occurrence was primarily that resulting from lowering fat intake, primarily omega-6 FA.
Conclusions
Similar to their main analyses, Cochrane's "cut-off" analyses suffer from many issues, including flawed trial selection and reliance on subjective endpoints.
The persistent biases in Cochrane's reviews against saturated fat could, in part, be explained by the pre-conceived biases of the main author, Dr. Lee Hooper.
Take the following statement from Dr. Hooper, commenting on Cochrane's first analysis in 2001 (10):
I was disappointed that we didn't find something more definitive . . . If this exhaustive analysis didn't provide evidence of the dangers of saturated fat, it was probably because the studies reviewed didn't last long enough, or perhaps because the participants didn't lower their saturated-fat intake enough.
Thus, Dr. Hooper — who is "primarily responsible" for planning and carrying out Cochrane's saturated fat review — seems to have a fundamental problem in accepting the lack of evidence against saturated fat.
We should also note that Dr. Hooper is a member of and funded by the World Health Organization to carry out systematic reviews. And these reviews are commissioned "specifically to underpin WHO nutrition guidance.”
The WHO, as some readers may already know, recommends lowering saturated fat intake to under 10% of total energy.
As described here:
In the end, the data show that we can disregard the recommendations of the WHO. There is simply no credible evidence for lowering saturated fat consumption below 10% of total calories.
References
1) Hooper, L., Martin, N., Jimoh, O. F., Kirk, C., Foster, E., & Abdelhamid, A. S. (2020). Reduction in saturated fat intake for cardiovascular disease. Cochrane database of systematic reviews, (8). (https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD011737.pub3/epdf/full)
2) Ramsden, C. E., Zamora, D., Majchrzak-Hong, S., Faurot, K. R., Broste, S. K., Frantz, R. P., ... & Hibbeln, J. R. (2016). Re-evaluation of the traditional diet-heart hypothesis: analysis of recovered data from Minnesota Coronary Experiment (1968-73). bmj, 353.
3) Ramsden, C. E., Zamora, D., Leelarthaepin, B., Majchrzak-Hong, S. F., Faurot, K. R., Suchindran, C. M., ... & Hibbeln, J. R. (2013). Use of dietary linoleic acid for secondary prevention of coronary heart disease and death: evaluation of recovered data from the Sydney Diet Heart Study and updated meta-analysis. Bmj, 346, e8707.
4) Dayton, S., Pearce, M. L., Hashimoto, S., Dixon, W. J., & Tomiyasu, U. (1969). A controlled clinical trial of a diet high in unsaturated fat in preventing complications of atherosclerosis. Circulation, 40(1s2), II-1.
5) Howard, B. V., Van Horn, L., Hsia, J., Manson, J. E., Stefanick, M. L., Wassertheil-Smoller, S., ... & Kotchen, J. M. (2006). Low-fat dietary pattern and risk of cardiovascular disease: the Women's Health Initiative Randomized Controlled Dietary Modification Trial. Jama, 295(6), 655-666.
6) Gershuni, V. M. (2018). Saturated fat: part of a healthy diet. Current Nutrition Reports, 7, 85-96.
7) https://youtu.be/Zwe_JF7Aqb8?t=3320
8) Jaax, S., Scott, L. W., Wolf Jr, J. E., Thornby, J. I., & Black, H. S. (1997). General guidelines for a low‐fat diet effective in the management and prevention of nonmelanoma skin cancer.
9) Black, H. S., & Rhodes, L. E. (2016). Potential benefits of omega-3 fatty acids in non-melanoma skin cancer. Journal of clinical medicine, 5(2), 23.
10) https://www.nbcnews.com/health/health-news/what-if-bad-fat-isn-t-so-bad-flna1c9472217