Right on the heels of the updated Cochrane review on “Reduced dietary salt for the prevention of cardiovascular disease,” (and the terrible reporting on it), another one is updated: “Reduced or modified dietary fat for preventing cardiovascular disease.” One of the authors (Lee Hooper) worked on the salt review as well.
Given that this is a Cochrane review, it uses only RCTs (and only studies that tracked morbidity and mortality, not just intermediate risk factors). Of course as always, the caveats of basing overall conclusions on topics only on RCTs should be noted, but given that this issue has so much data from different study designs, I was curious if the conclusions from the new Cochrane would match recent publications:
- The “consensus” article as blogged about here
- The 2 meta-analyses last year: one on prospective cohorts on saturated fats and CVD (Siri-Tarino et al.), and the other on RCTs that increase PUFA in place of saturated fats (Mozaffarian et al.) as discussed here
- The 2009 paper by Skeaff & Miller on prospective cohorts, RCTs, and coronary heart disease here (PDF)
Note: If you want a short summary directly from Lee Hooper, the Cochrane site has a short podcast here.
This is an updated review to the 2001 report that included 27 trials that were at least 6 months duration with over 18000 subjects total. This yielded only over 1400 deaths, over 800 cardiovascular deaths, and over 1200 cardiovascular events, and found no clear relationship between altering fat content in the diet and cardiovascular mortality, but reducing saturated fat did reduce cardiovascular events by 16%.
In this new review, 48 intervention trials (that provided dietary advice, advice plus supplementation, or provided meals) of duration of at least 6 months were included with a total of 60 comparisons (some trials have more than 1 arm). 25 comparisons (61958 subjects) were between a reduced fat diet and a usual or control diet, 15 (13004 subjects) were between a modified fat diet and usual or control diet, 10 (4931 subjects) were between a reduced and modified fat diet and usual or control, and 9 (1290 subjects) between a low fat and modified fat diet. 21 of the comparisons gave data on morality, 16 on cardiovascular mortality (65978 subjects and 1407 deaths) and 23 on combined cardiovascular events (65508 subjects and 4887 events).
Trials either 1) reduced total fat intake (<30% energy from fat), 2) modified fat intake (while aiming to maintain total fat intake at 30% or more)- e.g. increased MUFA or PUFA intake and reduced saturated fat intake, or 3) reduce and modify fat intake compared to a usual or control diet.
Here are some of the results (highlighted red if significant and notable):
- No significant effect of any dietary fat intervention (and dietary fat subgroups of modified fat diet vs usual, reduced fat vs usual, reduced and modified vs usual) compared to usual or control diet on total mortality (RR 0.98; n=71790, 4292 deaths). This held up with further subgrouping by trial duration, cardiovascular risk, mode of intervention, total fat in control, saturated fat in control, gender, and setting. After meta-regression of difference between control and intervention of: duration, total fat intake, saturated fat intake, trans fat intake, weight, serum LDL-C, there were not statistically significant differences between each of these and mortality (statistical power was low though).
- No significant effect of any dietary fat intervention (and no effect with the same subgrouping by dietary fat as above) compared to usual or control diet on cardiovascular mortality (RR 0.94; n=65978, 1407 deaths). Similar subgrouping as above did not reveal any significance, nor did a mega-regression (similar as above) here also did not show any statistical significance (again limited statistical power).
- There was a reduction in combined cardiovascular events (cardiovascular deaths, (cardiovascular morbidity (non-fatal myocardial infarction, angina, stroke, heart failure, peripheral vascular events, artrial fibrillation), and unplanned cardiovascular interventions (coronary artery bypass surger or angioplasty) with dietary fat interventions compared to usual diet (RR 0.86; n=65508, 4887 events). This was maintained in a sensitivity analysis (RR 0.83) (and Peto odds ratio meta-analysis*). Subgrouping by dietary fat change revealed that modification of fat vs usual fat reduced events (RR 0.82) as well as reduced and modified fat vs usual diet (RR 0.77). Importantly, there was no significant effect from trials comparing reduced fat vs usual intake (RR 0.97). A fixed effects meta-analysis on reduced and modified fat yielded a significant (p=0.05) reduction in cardiovascular event risk (RR 0.84). Subgrouping here found differences in study duration (studies more than 2 years showed an effect but not less), an effect from trials giving dietary advice plus supplementation (vs only small effect with advice alone and no effect from trials where meals were provided). Saturated fat intake (not provided by many trials) did not have an effect. Gender was important- cardiovascular events were reduced in men but not women (trials limited in women) or in combined studies. Setting also was important- community settings reduced events but residential institutions did not. Oddly, studies that were published in the 1960s and 90s reduced events, but not studies from the other decades. Meta-regression results were similar to that of total and cardiovascular mortality above (authors suggest in discussion that a lack of significant relationship between study duration and events may be because the diet is less rigorously adhered to over long periods but they have no evidence to support it). *for explanations of each analysis see this link
Secondary Outcomes (individual cardiovascular events)
- No significant effect of altering dietary fat on myocardial infarction (RR 0.93; n=64891, 2068 infarcts). The modified fat intake group, reduced fat vs usual fat, and reduced and modified fat vs usual intake was not statistically significant either (RR 0.91, 0.97, 0.90 respectively).
- No significant effect of altering dietary fat on stroke (RR 0.99; n=59853, 1140 strokes). Subgrouping by dietary fat change did not show significance.
- No significant effect of altering dietary fat on cancer deaths (RR 0.98; n=65724, 2818 cancer deaths), and subgrouping revealed a barely significant effect in the modified fat intake group of increasing cancer deaths, but there were only 91 deaths.
- No significant effect of altering dietary fat on cancer diagnosis (RR 0.96; n=59082, 6115 diagnoses). Most of this data was from studies that reduced dietary fat.
- No significant effect of altering dietary fat on diabetes diagnosis (RR 0.96; n=49859, 3367 diagnoses). Most of this data was from one study.
- No significant effect of altering dietary fat on non-fatal myocardial infarction diagnosis (RR 0.95; n=54883, 1403 people with at least 1 infarct).
- There wasn’t really enough data to analyze altering dietary fat on quality of life.
These outcomes are from studies that report deaths and/or cardiovascular events as to give hints to what the mechanisms behind altering dietary fat may be effecting those outcomes.
- There was no effect of modified fat intake on weight (but only 2 studies with 99 subjects). 16 studies (n=11058) showed reducing fat compared to usual diet reduces weight slightly (-0.83kg). 10 comparisons of reduced fat to usual intake found a reduction in BMI (-0.47kg/m^2).
- There was no effect of modified fat intake on LDL-C (but only 2 studies with 116 subjects). 14 comparisons (n=6971) found a significant effect from reducing fat on LDL-C compared to usual diet (-0.10mmol/L). The reduced and modified fat intake group also reduced LDL-C (but only 4 comparisons, n=627).
- None of the dietary fat changes altered HDL-C significantly.
- All 3 dietary fat interventions reduced total cholesterol: modified fat (-0.44mmol/L, 8 comparisons, n=2280), reduced fat intake (-0.10mmol/L, 15 comparisons, n=7602), and reduced and modified fat intake (-0.26mmol/L, 5 comparisons, n=2131).
- Modified fat intake and reduced and modified fat intake reduced fasting triglycerides (-0.11mmol/L, 5 comparisons, n=706) and (-.27mmol/L, 3 comparisons, n=218) respectively. Reducing total fat only compared to usual diet did not.
- Reducing fat had no clear effect on systolic or diastolic blood pressure, and data was too limited for other changes.
So increasing MUFA and/or PUFA and reducing saturated fat may reduce cardiovascular events (by 18% overall), but since saturated fat intake (or even biomarkers of it such as LDL or HDL) wasn’t reported in many studies it is difficult to get a sense of a real effect; the authors acknowledge that it is difficult to get a sense of true relationships between dietary fat changes and mortality (and morbidity) because differences in intakes between the intervention and control arms may be too small. And it is important to note that subgroup analysis found that to reduce cardiovascular events, the dietary intervention had to last more than 2 years, and the effect was only seen in men (trials are still limited in women subjects).
Since relative risks/percentages are difficult to think about if you don’t have the disease incidence at hand, I am thrilled that they authors included the following figure that puts the results into perspective (I put red boxes over the cardiovascular event data that is significant):
So at the population level, increasing MUFA and/or PUFA and reducing saturated fat might lead to a meaningful reduction in cardiovascular events according to these trials and analyses.
The evidence doesn’t show that altering dietary fat (either by reducing it or increasing unsaturated fats and reducing saturated fat) on average significantly alters total or cardiovascular mortality, or individual cardiovascular events: infarctions, stroke, cancer deaths, cancer diagnosis, non-fatal myocardial infarction diagnosis, quality of life, HDL, or systolic and diastolic blood pressure (though data is too limited on most). Some cardiovascular risk factors are positively altered: reducing fat had a small effect on weight and BMI reduction and LDL-C, and all interventions reduced total cholesterol. Modified fat intake and the reduced and modified intake groups has reduced fasting triglycerides. Overall, the clinical significance of the small changes of one, some, or all of these may represent the mechanism(s) of the reduced cardiovascular events, or perhaps it was something else.
It should be noted that nearly all the studies included here were in industrialized countries, so it may not necessarily pertain to developing and transition countries. Because it included multiple countries, however, it doesn’t exclusively represent a particular cultural diet.
To return to the other recent research discussed at the beginning of this post- the conclusions are different with the “consensus” article (coronary heart disease isn’t evident in this new one). They also contrast with Siri-Tarino et al. that found no association between saturated fat intake and cardiovascular disease in 21 cohorts, but also found no relationship between saturated fat intake and coronary heart disease or stroke as this new one also did. And Mozaffarian et al. suggested that substituting PUFA in place of saturated fat reduces coronary heart disease, but as the authors of this Cochrane put it this evidence was “limited and circumstantial” because they inferred it from the combined 19% reduction in myocardial infarction or coronary heart disease disease death- this was similar to the previous version of this Cochrane review (the authors note that they used similar studies). Both of those links (Siri-Tarino et al. & Mozaffarian et al.) can be found here. Skeaff and Miller analyzed 28 cohorts (n=280,000) and found that total fat intake is not associated with coronary heart disease mortality or coronary heart events, similar to this Cochrane. They also had “partly updated the previous version of this [Cochrane] review” and found similar results to these, but also a “marginally significant relationship between the ratio of polyunsaturated/saturated fat and cardiovascular events but none with mortality. This review was unable to find a significant relationship from PUFA or MUFA to cardiovascular events.
Based on multiple, large meta-analyses from different study designs, saturated fat is still in a bit of hot water. But with all of this research we now know it is much less of a threat than originally thought. Increasing unsaturated fats (especially PUFA) still may have some benefits in the cardiovascular system.
And we can say with pretty good confidence that total fat (in the variation of most industrialized consumption) doesn’t need lowering.
Hooper L, Summerbell CD, Thompson R, Sills D, Roberts FG, Moore H, & Davey Smith G (2011). Reduced or modified dietary
fat for preventing cardiovascular disease Cochrane Database of Systematic Reviews (7) : 10.1002/14651858.CD002137.pub2.