At the recent Swiss Re/The BMJ Food for Thought conference, the closing speakers tried to find some agreement on dietary fat guidelines. You can listen to the exchange between Professor Dariush Mozaffarian, Professor Salim Yusuf and Dr Fiona Godlee here.
Fiona started with: “The point about saturated fat is: the evidence is now looking pretty good, but the guidance hasn’t shifted… there doesn’t seem to have been an enormous ‘mea culpa’ from the scientific community that we got it so wrong. That does surprise me.”
Salim replied: “We got brainwashed by a very questionable study, called The Seven Countries Study, many years ago and it was ingrained in our DNA and generations of us were brought up with that… Somebody said that you need to wait for guidelines committees to die before you can change the guidelines committees”!
Fiona then said: “Maybe one outcome of this meeting would be for this meeting to say ‘that’s gone now’, the science has changed. Am I right Salim? Am I right Dariush? It seems to be that should be an outcome of some sort from this meeting.”
Alas, the UK guidelines committee shows no signs of such change, let alone the ‘mea culpa’ that Fiona suggests might be in order. On May 8th 2018, a draft report was issued by the Scientific Advisory Committee on Nutrition (SACN) for saturated fat. I wrote about it in my ‘Monday note’ that week. Comments on the draft document were required to be submitted back to SACN by 5pm on July 3rd 2018. I submitted mine on 26th June with George Henderson as a co-author (the fabulous @puddleg on twitter). I was most grateful for his input – in the area of population studies and SFA/PUFA substitution especially.
As I said in the May post, I focused on the cardiovascular disease (CVD) and coronary heart disease (CHD) area of the report for two reasons: i) because that’s my area of expertise and ii) more importantly, because CHD is why the saturated fat guidelines were introduced in the first place. I can add a third reason, following the May post, as that revealed that there was no evidence and/or evidence of no effect for saturated fat and all non-CVD conditions reviewed by the panel (type 2 diabetes, cancer, cognitive impairment and dementias). Hence, if CVD and CHD don’t withstand scrutiny, there’s nothing left for actual outcomes.
The SACN report also reviewed surrogate end points, which are pointless when data are available on actual health outcomes. The most common accusation levelled at saturated fat in the world of surrogate end points is that it raises LDL cholesterol (and that is seen as a bad thing per se). Given that every food that contains fat contains all three fats (saturated, monounsaturated and polyunsaturated), if the proportion of saturated fat in the diet is higher, unsaturated fat is necessarily lower and vice versa. In these two posts (here and here), I explain that vegetable oils, which we are encouraged to eat instead of things like butter, contain plant sterols. Plant sterols lower human cholesterol (by replacing it with plant cholesterol in essence), but this is not good for health according to the evidence. Hence it may not be the case that saturated fat raises cholesterol, but that unsaturated fat (especially such fats rich in plant sterols) lowers cholesterol. The deleterious health outcome from plant sterols must be taken into account and this is not currently happening.
You can see the draft report here. I started with the helpful summary table, Table 8.1 on p94, so that I could see where it was claimed that saturated fat had a harmful impact on actual health outcomes. There were two types of evidence reviewed: Randomised Controlled Trials (RCTs) and Prospective Cohort Studies (PCSs) (there are definitions of these in the glossary here). There were two circumstances examined: a reduction in saturated fat intake and swapping saturated fat out and polyunsaturated fat in. There were six health outcomes reviewed: CVD mortality; CVD events; CHD mortality; CHD events; strokes and peripheral vascular disease. That makes for a lot of permutations! Thankfully much of the summary table found no evidence or evidence of no effect (they are different, but can both be ignored for the purposes of reviewing claims against saturated fat).
I went through the claims systematically and the paragraphs in the report to which they related and completed the table, which was provided by SACN for response submissions. The response that I submitted is exactly as in the table below.
The five things that SACN did
I found the SACN report to be biased. I can’t put it any other way. I recently shared the declarations of interest of the committee members. When the first (and highly impactful) article from my PhD was published (Ref 18), one of the SACN saturated fat committee members wrote a letter to the editor defending the guidelines and closing with “Given the improvements in both levels of cholesterol and rates of heart disease world-wide in those countries which have introduced measures to reduce intakes of saturated fats, it would seem that an RCT, with corresponding ethical approval for its conduct, would surely be required to reverse existing recommendations?” Does that seem like someone open to impartially reviewing the dietary guidelines on saturated fat?
When reviewing the draft report, I found that SACN had done five things to try to make a case against saturated fat:
1) SACN presented duplicated studies and studies that simply reported other studies, which gave the appearance of there being more evidence supporting their claims than there was;
2) They left out some studies – all of which found evidence of no effect (one of mine (Ref 1); Schwingshackl & Hoffman (Ref 2) and Hamley (Ref 3));
3) They presented non-significant results as if significant;
4) They looked for and reported fixed effects methodology when the random effects methodology concluded that there were no significant findings;
(This is a technical aspect of meta-analysis, which is the methodology for pooling data from RCTs or PCSs. There are virtually no circumstances when using fixed effects methodology is the correct methodology. The pooled trials and cohort studies need to be virtually identical for that to be appropriate and this rarely, if ever, is the case. There were a number of occasions when the correct methodology – the random effects method – was reported in a paper as showing that there was no effect and yet the committee searched the paper for mentions of fixed effects tests, usually done as sensitivity tests, and reported the fixed effects outcome if it showed that there was an effect. I reported this as bias when I saw it.)
5) They ignored evidence that didn’t suit them (paragraphs 8.40-8.47 were extraordinary).
Those are the headlines. If you want the Full Monty, it is below…
|Comments by paragraph||Comments |
Please insert each new comment in a new row
|Section 8||This section only includes comments on clauses that led to claims of evidence in Table 8.1 (p94).|
|8.2||This paragraph contains a number of errors: |
1) One of the four Harcombe et al papers has been omitted (PCSs at the time the guidelines were set) (Ref 1).
2) Schwingshackl & Hoffman (2014) has been omitted (Ref 2).
3) Hamley (2017) has been omitted (Ref 3). (This was published during the committee deliberations and should also have been included – it can be now at draft report stage.)
4) Schwab et al (2014) (Ref 4) reported findings from Hooper et al (Ref 5) and Jakobsen et al (Ref 6) , rather than doing any meta-analysis of their own and so should be deleted.
5) Van Horn et al (2008) (Ref 7) is about assumed risk factors, not disease, and so should not appear in this section. Additionally, as with Schwab et al (2014) (Ref 4), Van Horn et al (2008) (Ref 7) reported passages from other articles, rather than doing any meta-analysis of their own and so should not be included.
6) For the purposes of meta-analysis, Micha & Mozaffarian (2010) (Ref 8) is a duplication of Mozaffarian et al 2010 (Ref 9) and should be deleted.
7) This paragraph reported: “The Hooper et al. (2015) review included virtually all randomised controlled trials (RCTs) included in other studies.” This is correct. This paragraph should clarify the differences between Hooper and other studies, not just the similarities: Hooper et al (Ref 10) was the only meta-analysis to include four small studies for which CVD data were not published or peer reviewed (See 8.6).
|8.3-8.7 & 8.12||Saturated fat intake and CVD: Table 8.1 claimed “Adequate evidence for reduced saturated fat intake on reduced CVD events” from RCTs (Nothing claimed from PCSs).|
|8.3||This paragraph omitted Schwingshackl & Hoffman (2014) (Ref 2). This examined evidence for all-cause mortality, CVD mortality, CVD events and MIs for both reduced fat intake and modified fat intake with a systematic review and meta-analysis of RCTs. This study found “The present systematic review provides no evidence (moderate quality evidence) for the beneficial effects of reduced/modified fat diets in the secondary prevention of coronary heart disease. Recommending higher intakes of polyunsaturated fatty acids in replacement of saturated fatty acids was not associated with risk reduction.” |
As explained in 8.2, Schwab et al (2014) (Ref 4) and Van Horn et al (2008) (Ref 7)were included and should not have been.
|8.4||As explained in 8.2, Van Horn et al (2008) (Ref 7) should be deleted. Schwingshackl & Hoffman (2014) (Ref 2) should be included.|
|8.6||This paragraph should have noted the limitations of the Hooper et al (Ref 10) report (not least as it is almost entirely relied upon for the case against saturated fat) (Ref 11): |
– SACN should have questioned why the Hooper et al review found something that the other meta-analyses didn’t. The Hooper et al finding included 4 small studies (646 people in total), not included in any other meta-analysis, which were primarily studies of: diabetes (Ref 12); skin cancer (Ref 13); hypercholesterolemia (Ref 14); and glucose intolerance (Ref 15), but for which unpublished, non-peer-reviewed CVD event information was obtained by Hooper et al in personal correspondence.
Houtsmuller (Ref 12) was the most striking outlier in Hooper et al’s meta-analysis for CVD events. This paper claimed “One group of patients was put on a diet (I), consisting of carbohydrates 50 cal%, saturated fats 35 cal% and proteins 15 cal%.” This is nutritionally impossible. All foods that contain fat contain all three fats. There is no food comprising 100% saturated fat. No diet can equate total fat to saturated fat with no intake of unsaturated fat. Such data are not robust and should not have been included by Hooper et al. As every other researcher has shown, when only the dietary trials that had CVD/CHD as measured outcomes and peer reviewed data, are included, there are no findings to report.
Additionally, when a sensitivity test was undertaken on the RCTs that actually significantly reduced SFA intake (as opposed to having the aim of reducing SFA intake), the CVD events finding (for >52,000 participants) reduced from 17% to 9% and was no longer statistically significant (Hooper et al 2015 Table 8, p121) (Ref 10).
Even had the one Hooper et al finding (among seven other non-findings) retained significance following the sensitivity test (and it didn’t), any finding would still have lacked generalisability. The Hooper et al review (Ref 10) did not include a single study of healthy people of both genders and thus any findings would have lacked generalisability and could not be extrapolated to populations.
|8.7||This paragraph should be deleted. Van Horn et al (Ref 7) is about assumed risk factors, not disease/events. Additionally, the statement “a systematic review of 83 primary studies and 19 review articles concluded that low intake of saturated fats (<7% of total energy) resulted in reduced risk of CVD” is disingenuous. |
i) As above, this was related to assumed risk factors and not CVD events (which is the focus of this section) and ii) 82 studies and 19 review articles did not conclude that low intake of saturated fats (<7% of total energy) resulted in reduced risk of CVD. Van Horn et al (Ref 7) referenced one paper that claimed reducing SFA to <7% energy and dietary cholesterol of <200 mg/day reduced LDL. Three RCTs were similarly reported as having lowered LDL with a <7% SFA and <200mg/day cholesterol diet.
|8.12||This paragraph should be amended to report that there is adequate evidence of no effect for reduced saturated fat intake on CVD events (Refs, 2, 10).|
|8.13-8.17||Substitution of saturated fats with polyunsaturated fats and CVD: Table 8.1 claimed “Adequate evidence for substitution of saturated fats with polyunsaturated fats on reduced CVD events” from RCTs.|
|8.13||This paragraph omitted Schwingshackl & Hoffman (2014) (Ref 2). This examined evidence for all-cause mortality, CVD mortality, CVD events and MIs for both reduced fat intake and modified fat intake with a systematic review and meta-analysis of RCTs. This study found “The present systematic review provides no evidence (moderate quality evidence) for the beneficial effects of reduced/modified fat diets in the secondary prevention of coronary heart disease. Recommending higher intakes of polyunsaturated fatty acids in replacement of saturated fatty acids was not associated with risk reduction.” |
This paragraph should delete Schwab et al (2014) (Ref 4) (See 8.16).
This paragraph should delete Van Horn et al (2008) (Ref 7) (See 8.17).
|8.14||This paragraph claimed that there was a 27% lower risk of CVD events – “p<0.05” – with SFA replaced by PUFA. The original Hooper et al paper (Table 9. p121) (Ref 10) reported the P value as 0.14, which makes the finding non-significant.|
|8.15||This paragraph should clarify that Ramsden et al (2013) (Ref 16) contributed nothing to the claim for replacing SFA with PUFA and the impact on CVD events. Ramsden et al (2013) focused on CVD mortality. The word “events” did not appear in the Ramsden paper or appendix. |
This paragraph was also not reflective of the conclusions of Ramsden et al’s paper. Ramsden et al’s conclusion was: “In this cohort, substituting dietary linoleic acid [a PUFA] in place of saturated fats increased the rates of death from all causes, coronary heart disease, and cardiovascular disease” [my emphasis].
|8.16||This paragraph claimed that “Schwab et al (2014) (Ref 4)…reported on the effect of saturated fat substitution with unsaturated fats (PUFA or MUFA) on CVD events. The authors reported a 14% reduction in RR of CVD events (RR 0.86, 95% CI 0.77 to 0.96; p=0.07; I² = 50%; 24 RCTs; 65,508 participants, 4586 CVD events.” |
This is not correct. Schwab et al reported “A SR [A Systematic Review – my emphasis] concluded that there is moderate evidence that substitution of unsaturated fatty acids (MUFA or PUFA) for SFA can reduce CVD events by 14% (reference 67).” Reference 67 in their paper was for the Hooper et al 2011 review (Ref 5). Schwab et al did not conduct an SR or meta-analysis of their own; they merely referenced another.
Similarly, this paragraph in the SACN report continued “There was no effect of saturated fat change on CVD mortality (RR 0.94, 95% CI 0.85 to 1.04; p=0.23; I2 =0%; 16 RCTs; 65,978 participants, 1407 CVD deaths).” All of this is from the Hooper et al 2011 review (which has been superseded by the 2015 review) and not from any work undertaken by Schwab et al.
This paragraph should be deleted.
|8.17||This paragraph claimed “A systematic review by Van Horn et al. (2008) reported a reduced risk of CVD when saturated fats were substituted with unsaturated fats including MUFA (<20% of energy) and PUFA (<10% of energy). However, it was unclear on which of the included RCTs this statement was based and no meta-analysis was performed.” |
The only reference to such data in the Van Horn et al paper is “The American Heart Association recommends a diet … SFA and TFA should be replaced isocalorically with complex carbohydrates and/or UFA, including both MUFA (not to exceed 20% of energy) and PUFA (not to exceed 10% of energy).” (p292) As with 8.16, Van Horn et al have not done their own systematic review. SACN have reported Van Horn et al’s reporting of general advice. No RCTs have been studied by Van Horn et al (and this SACN paragraph is in the RCT/CVD event evidence section). There are a number of references to CVD events in the Van Horn et al paper (related to alcohol, exercise, vitamins, obesity, for example) but none relate to SFA/PUFA replacement.
This paragraph should be deleted.
|8.18-8.19||Substitution of saturated fats with polyunsaturated fats and CVD: Table 8.1 claimed “Limited evidence for substitution of saturated fats with polyunsaturated fats on reduced CVD mortality” from PCSs.|
|8.18||This paragraph claimed “Evidence from systematic reviews of PCS indicate a reduction in CVD mortality when saturated fats were substituted with PUFA (Schwab et al., 2014; Van Horn et al., 2008) or a combination of MUFA and PUFA (Schwab et al., 2014), however there was no formal meta-analysis of these data which limits their quality.” |
Schwab et al (2014) (Ref 4) reported no evidence from systematic reviews of PCS for CVD mortality and substitution of saturated fats with PUFAs, or a combination of MUFA and PUFA.
There are a number of references to CVD mortality in the Van Horn et al paper (related to alcohol, exercise, vitamins, obesity, for example) but none relate to SFA/PUFA replacement.
This paragraph should be deleted.
|This paragraph reiterated the claim that there was a 27% lower risk of CVD events. This was not a significant finding. The SACN report states: “These findings were consistent with the results of other systematic reviews of RCTs and the evidence was considered adequate.” This sentence needs to be deleted as per the comments on Ramsden et al (2013) (Ref 16), Schwab et al, 2014 (Ref 4) and Van Horn et al, 2008 (Ref 7), which left no other findings. |
The conclusion of this section should be:
i) There is adequate evidence from two meta-analyses of RCTs (Refs 2, 10) that replacing SFA with PUFA has no effect on CVD events.
ii) There is no evidence from systematic reviews or meta-analyses of PCSs to draw any conclusion about the impact of substitution of saturated fats with polyunsaturated fats on CVD mortality.
|I have ignored all sections related to MUFAs, carbohydrates and protein as no claims were made against saturated fat. Table 8.1 claimed that there was limited evidence that substituting saturated fats with MUFAs would increase CHD events and that there was adequate evidence that substituting saturated fats with carbohydrates would increase CHD events. I have not addressed these claims.|
|8.31-8.38||Saturated fat intake and CHD: Table 8.1 claimed “Moderate evidence for reduced saturated fat intake on reduced CHD events” from RCTs.|
|8.31||8.31 Contains a number of errors: |
1) A Harcombe et al paper has been omitted (Ref 1). This should be included in the list of systematic reviews without meta-analysis.
2) The SACN report reference “Harcombe 2016a” (Ref 17) is incorrectly reported as a systematic review without meta-analysis. It is a systematic review with meta-analysis.
3) Seven, not six, papers evaluated the results from PCS. There are two Harcombe et al papers published for cohort evidence (Refs 1, 17).
NB for convenience – the three Harcombe et al papers referenced as 2016a, 2016b and 2015 in the SACN report are (with the references in this document in brackets):
Harcombe Z, Baker JS, Cooper SM, Davies B, Sculthorpe N, DiNicolantonio JJ & Grace F (2015) Evidence from randomised controlled trials did not support the introduction of dietary fat guidelines in 1977 and 1983: a systematic review and meta-analysis. Open Heart 2, e000196 (Ref 18).
Harcombe Z, Baker JS & Davies B (2016a) Evidence from prospective cohort studies does not support current dietary fat guidelines: a systematic review and meta-analysis. British Journal of Sports Medicine (Ref 17).
Harcombe Z, Baker JS, DiNicolantonio JJ, Grace F & Davies B (2016b) Evidence from randomised controlled trials does not support current dietary fat guidelines: a systematic review and meta-analysis. Open Heart 3 (Ref 19).
|8.33||This paragraph should report for Harcombe et al (2015) (Ref 18) that “the included papers deliberately only examined papers published before 1983 to examine the evidence base for the dietary guidelines at the time they were introduced. This, the only paper to examine this question for RCTs, found no evidence for the guidelines introduced.” |
It is important to document the fact that there was no evidence to support the introduction of the dietary fat guidelines at the time they were introduced. Any evidence being sought is retrospective.
|8.38||This paragraph rightly reported that Hooper et al (2015) (Ref 10) reported no effect of reduced saturated fat intakes on CHD events. |
The paragraph then wrongly tried to claim that a fixed effects model would have given a significant result. A fixed effects model cannot be used given the heterogeneity of dietary fat trials. The I² of 66% confirms this. Hooper et al calculated the fixed effects model as a sensitivity test, not as a finding that SACN can opt for in preference to the random effects (correct) conclusion. At no time has Hooper et al made a claim for a significant finding for CHD events.
This paragraph confirms the confirmation bias of the SACN panel.
|8.39-8.47||Saturated fat intake and CHD: Table 8.1 claimed “Moderate evidence for reduced saturated fat intake on reduced CHD mortality” “Moderate evidence for reduced saturated fat intake on reduced CHD events” both from PCSs.|
|8.39||A Harcombe et al paper is missing (Ref 1).|
|8.40||This paragraph confirmed de Souza et al (Ref 20) – the most recent systematic review – “…reported no association between the highest and lowest intakes of saturated fats and CHD mortality” AND “Furthermore no association was reported between the intake of saturated fats and total CHD.” |
The reason for restating the conclusion of de Souza et al (Ref 20) will become clear.
|8.41||This paragraph confirmed Chowdhury et al (Ref 21) – the most comprehensive systematic review and meta-analysis – found “No association was found with CHD outcomes when comparing the top tertile of saturated fat intakes with the bottom tertile.” This was the conclusion reported by Chowdhury. This finding came from Random Effects methodology, which is the correct methodology that every researcher in this field has used. It is the only acceptable methodology to use given the heterogeneity of the dietary trials that have been undertaken since 1965. |
Again, this paragraph tries to present results from a Fixed Effects methodology, when the correct Random effects methodology has been presented by Chowdhury et al. This paragraph confirms the confirmation bias of the SACN panel.
The reason for restating the conclusion of Chowdhury et al (Ref 21) will become clear
|8.42||This paragraph confirmed Siri-Tarino et al (Ref 22) – “No association was found between upper and lower quartiles of saturated fats intake and CHD.” |
The reason for restating the conclusion of Siri-Tarino et al (Ref 22) will become clear.
|8.43||This paragraph confirmed Skeaff & Miller (Ref 23) – “There was no association with CHD mortality at 5 to 16 years follow-up … or CHD events at 5 to 20 years follow-up…” AND |
“Analysis of 5% total energy increments in saturated fats also showed no association for either CHD mortality… or CHD events.”
The reason for restating the conclusion of Skeaff & Miller (Ref 23) will become clear.
|8.44||This paragraph is referring to the wrong Harcombe et al paper and it has reported the wrong conclusion. This should read: Harcombe et al (Ref 1) included data from 6 PCS, all published before 1982 (for the reason stated in 8.33), involving 31,445 participants and 360 CHD deaths with a mean follow-up of 6.2 to 7.5 years. The data were not conducive to meta-analysis. It was reported that one of the six studies found an association between CHD deaths and intakes of saturated fats across countries; none found a relationship between CHD deaths and saturated dietary fat in the same population.” |
The reason for restating the conclusion of Harcombe et al (Ref 1) will become clear.
|8.44B New para needed||Harcombe et al (Ref 17) concluded: “Across 7 studies, involving 89,801 participants (94% male), there were 2,024 deaths from CHD during the mean follow-up of 11.9 ± 5.6 years. The death rate from CHD was 2.25%. Eight data sets were suitable for inclusion in meta-analysis; all excluded participants with previous heart disease. Risk ratios (RR) from meta-analysis were not statistically significant for CHD deaths and total or saturated fat consumption. The risk ratio (RR) from meta-analysis for total fat intake and CHD deaths was 1.04 (95% CI 0.98 to 1.10). The RR from meta-analysis for saturated fat intake and CHD deaths was 1.08 (95% CI 0.94 to 1.25).” |
The reason for stating the conclusion of Harcombe et al (Ref 17) will become clear.
|8.45||This paragraph confirmed Mente et al (Ref 24) – “when the highest intakes of saturated fats were compared with the lowest, no association between saturated fats and coronary outcomes were identified.” |
The reason for restating the conclusion of Mente et al (Ref 24) will become clear.
|As per 8.38, this paragraph reported “Hooper et al (2015) (Ref 10) also found no effect on CHD events when using a random-effects model.” This was the correct reporting of the correct conclusion from Hooper et al. However, the SACN committee again tried to use fixed effects methodology and sensitivity tests undertaken by Hooper et al to claim a different conclusion, one that was not made by the original researchers. At no time has Hooper et al made a claim for a significant finding for CHD events. |
This paragraph confirms the confirmation bias of the SACN panel.
This paragraph should be amended to reflect the correct conclusion: that there is adequate evidence of no effect of saturated fat intake on CHD events (Hooper et al (2015) (Ref 10)).
|The conclusions from paragraphs 8.40-8.45 inclusive have been reiterated in this document to show that NONE of: de Souza et al (Ref 20); Chowdhury et al (Ref 21); Siri-Tarino et al (Ref 22); Skeaff & Miller (Ref 23); Harcombe et al (Ref 1); Harcombe et al (Ref 17); OR Mente et al (Ref 24) found an association between saturated fat intake and CHD mortality or CHD outcomes. The conclusion of this section should have been a categorical statement “The Committee found adequate evidence of no effect.” |
Instead, this paragraph reported one fixed effects test from just one of these studies (Chowdhury et al (Ref 21)) and ignored all other evidence: “The committee, on balance, therefore considered these data to be moderate evidence” for reduced saturated fat intake on CHD mortality and CHD events.
This paragraph is an extraordinary example of the confirmation bias of the SACN panel. An independent panel could not have concluded as this paragraph did from the conclusions presented in paragraphs 8.40-8.45 inclusive.
|8.48-8.54||Substitution of saturated fats with polyunsaturated fats and CHD: Table 8.1 claimed “Limited evidence for substitution of saturated fats with polyunsaturated fats on reduced CHD events” from RCTs.|
|8.48||“Five systematic reviews analysed the results from RCTs (Ramsden et al., 2016 (Ref 25); Hooper et al., 2015 (Ref 10); Micha & Mozaffarian, 2010 (Ref 8); Mozaffarian et al., 2010 (Ref 9); Skeaff & Miller, 2009 (Ref 23)” is not correct. Micha & Mozaffaria 2010 merely reported the finding of Mozaffarian et al., 2010, which was in print at the time, so this was duplication and should be removed. |
8.48 Hamley’s paper “The effect of replacing saturated fat with mostly n-6 polyunsaturated fat on coronary heart disease: a meta-analysis of randomised controlled trials” (Ref 3) was published during the committee deliberations and should have been included.
This specifically addressed the confounding variables in diet heart trials and sought to focus on the results from trials that most accurately tested the effect of replacing SFA with mostly n-6 PUFA. This found “When pooling results from only the adequately controlled trials there was no effect for major CHD events (RR = 1.06, CI = 0.86–1.31), total CHD events (RR = 1.02, CI = 0.84–1.23), CHD mortality (RR = 1.13, CI = 0.91–1.40) and total mortality (RR = 1.07, CI = 0.90–1.26). Whereas, the pooled results from all trials, including the inadequately controlled trials, suggested that replacing SFA with mostly n-6 PUFA would significantly reduce the risk of total CHD events (RR = 0.80, CI = 0.65–0.98, P = 0.03), but not major CHD events (RR = 0.87, CI = 0.70–1.07), CHD mortality (RR = 0.90, CI = 0.70–1.17) and total mortality (RR = 1.00, CI = 0.90–1.10).”
|8.49||As 8.48 – the duplicated reference to the finding in Mozaffarian et al, 2010 needs to be removed i.e. Micha & Mozaffarian, 2010 needs to be removed for the review of evidence from RCTs.|
|8.50||Hooper et al (2015) (Ref 10) has been wrongly and disingenuously reported. The claim that “there was a 24% reduction in CHD events (RR 0.76, 95% CI 0.57 to 1.00; >3000 participants, 737 events)” is misleading. This is not statistically significant, as it includes the line of no effect. Hooper et al have never made such a claim for CHD events.|
|8.51||Skeaff & Miller (2009) (Ref 23) has been wrongly and disingenuously reported. The SACN report claimed that high PUFA and lower saturated fats “reduced the risk for CHD events (RR 0.83, 95% CI 0.69 to 1.00; p=0.05; I2 =44.2%; 8 RCTs; 4528 participants, 284 events…” This is not statistically significant, as it includes the line of no effect.|
|8.52||Mozaffarian et al (2010) (Ref 9) should be excluded for its study selection, as was explained in this peer reviewed critique paper (Ref 26). Mozaffarian et al (2010) (Ref 9) omitted two studies that cautioned about the potential harm/toxicity of PUFAs (Refs 27, 28) and it included the non-randomised, non-controlled Finnish Mental Hospital cross-over trial (Ref 29), which all other respectable researchers, including Cochrane, omitted .|
|8.54||The Micha & Mozaffarian 2010 (Ref 8) article did not do a different meta-analysis. The finding quoted in 8.54 is from the abstract of Mozafarrian (2010) (Ref 9). The non robust study (Mozaffarian 2010) should not have been included, let alone duplicated. |
This paragraph should be deleted.
|8.55-8.60||Substitution of saturated fats with polyunsaturated fats and CHD: Table 8.1 claimed “Adequate evidence for substitution of saturated fats with polyunsaturated fats on reduced CHD mortality” and “Adequate evidence for substitution of saturated fats with polyunsaturated fats on reduced CHD events” both from PCSs.|
|8.55||Reference to Schwab should be deleted (2014) (Ref 4). As reported in paragraph 8.56, Schwab et al. (2014) (Ref 4) limited their analysis to a summary of the findings of Jakobsen et al (2009) (Ref 6) and thus this duplication is misleading.|
|8.56||As paragraph 8.55, this paragraph should be deleted.|
|8.57||This paragraph reported “Overall, a 5% lower energy intake from saturated fats and a concomitant higher energy intake from PUFA was significantly associated with a decrease in CHD deaths (HR 0.74, 95% CI 0.61 to 0.89; p-value not reported) and CHD events (HR 0.87, 95% CI 0.77 to 0.97; p-value not reported).” |
The two p-values were reported by Jakobsen et al (2009) (Ref 6) as 0.40 and 0.70 respectively, which are not significant. This non-significance was confirmed by the fact that significance could not be reached for women or men for coronary events or for coronary deaths when reviewed separately (three out of four confidence intervals including 1.0 and all p values for between-study heterogeneity and for effect modification by sex being substantially higher than 0.05 – as high as 0.81 for example).
|8.58||This paragraph reported “Farvid et al. (2014), in a systematic review with meta-analysis of 13 PCS with 310,602 participants, reported on the substitution of saturated fats with dietary linoleic acid (n-6 PUFA). Increasing percent of energy from linoleic acid (by 5%) instead of saturated fats was associated with a 13% lower risk of CHD deaths using a fixed-effect (Mantel-Haenszel) model (RR 0.87, 95% CI 0.82 to 0.94; p<0.05; I2=0.0; 10 PCS). This finding was similar using a random-effects model (RR 0.86, 95% CI 0.76 to 0.97). There was a 9% lower risk of CHD events using a fixed-effect model (RR 0.91, 95% CI 0.87 to 0.96; p=0.012; I2 =55.9%; 8 PCS), which was non-significant using a random-effects model (RR 0.90, 95% CI 0.80 to 1.01).” |
As above, fixed effects methodology is not appropriate for studies with such heterogeneity (there are virtually no cirumstances when a fixed effects model is appropriate). The reporting of the CHD events as non-significant using a random effects model is correct and thus this finding can be ignored.
The reporting of the CHD deaths as significant using a random-effects model is incorrect. The sentence “This finding was similar using a random-effects model (RR 0.86, 95% CI 0.76 to 0.97)” is from P1572 of Farvid et al . The full extract is “Across 10 cohort studies that examined the association between LA and total CHD events (14 estimates), LA consumption was inversely associated with risk of total CHD events. The fixed effect summary of RR for comparing the highest with lowest category was 0.85 (95% CI, 0.78–0.92; Figure 2) with medium heterogeneity (I2=35.5%). This finding was similar using a random-effects model (RR, 0.86; 95% CI, 0.76–0.97).”
A random effects finding from another part of the paper has thus been wrongly connected to the CHD deaths fixed effects model. No random effects result was presented alongside the fixed effect RR for CHD deaths and thus there is no significant random effects finding to be taken into consideration.
Notwithstanding that Schwab (2014) (Ref 4) should not have been mentioned and neither Jakobsen et al (2009) (Ref 6) nor Farvid et al (2014) (Ref 30) survived scrutiny, there are a number of factors that should be taken into consideration in this section:
1) Jakobsen et al (2009) (Ref 6) and Farvid et al (2014) are mathematical modelling exercises. No PUFAs were swapped in for SFAs. The Hamley paper (Ref 3) is the important one, as this did test replacement of SFAs with PUFAs and it also addressed the issue of the quality of trials. It concluded: “When pooling results from only the adequately controlled trials there was no effect for major CHD events (RR = 1.06, CI = 0.86–1.31), total CHD events (RR = 1.02, CI = 0.84–1.23), CHD mortality (RR = 1.13, CI = 0.91–1.40) and total mortality (RR = 1.07, CI = 0.90–1.26).”
2) There have been a number of PCSs since Jakobsen et al (2009) (Ref 6) and Farvid et al (2014) (Ref 30) that would counter any findings had they been significant (Refs 31-33).
3) There have been a number of studies warning about the potential harm from administration of polyunsaturated fats (Refs 16, 25, 27, 28, 31, 34). Public health advice thus needs to be extremely cautious in this area.
|8.59||The conclusion of this section should be – as reported – that there is adequate evidence of no effect for CHD mortality. |
The conclusion for CHD events should also be that there is adequate evidence from meta-analyses of RCTs (Refs 3, 10, 23) that replacing SFA with PUFAs has no effect on CHD events.
|8.60||The conclusion of this section should be – there is adequate evidence of no effect for both CHD mortality and CHD events from meta-analyses of PCSs.|
|Table 8.1||Needs to change to reflect all of the above.|
|15.16||There was adequate evidence from RCTs that reducing intake of saturated fats had no effect on CVD events.|
|15.19||There was adequate evidence from RCTs that substituting saturated fats with PUFA had no effect on CVD events.|
|15.20||There was no evidence from PCSs that substituting saturated fats with unsaturated fats was associated with a lower risk of CVD mortality. There was no evidence for CVD events.|
|15.26||There was adequate evidence from RCTs that reducing intake of saturated fats had no effect on CHD events.|
|15.27||There was adequate evidence from PCS that reducing intake of saturated fats had no effect on CHD mortality or CHD events.|
|15.29||There was adequate evidence from RCTs that replacing SFA with PUFAs had no effect on CHD events.|
|15.30||There was adequate evidence from PCSs that replacing SFA with PUFAs had no effect on both CHD mortality and CHD events.|
- Harcombe Z, Baker JS, Davies B. Evidence from prospective cohort studies did not support the introduction of dietary fat guidelines in 1977 and 1983: a systematic review. Br J Sports Med. 2016.
- Schwingshackl L, Hoffmann G. Dietary fatty acids in the secondary prevention of coronary heart disease: a systematic review, meta-analysis and meta-regression. BMJ Open. 2014.
- Hamley S. The effect of replacing saturated fat with mostly n-6 polyunsaturated fat on coronary heart disease: a meta-analysis of randomised controlled trials. Nutrition journal. 2017.
- Schwab U, Lauritzen L, Tholstrup T, Haldorsson TI, Riserus U, Uusitupa M, et al. Effect of the amount and type of dietary fat on cardiometabolic risk factors and risk of developing type 2 diabetes, cardiovascular diseases, and cancer: a systematic review. Food & Nutrition Research. 2014.
- Hooper L, Summerbell CD, Thompson R, Sills D, Roberts FG, Moore H, et al. Reduced or modified dietary fat for preventing cardiovascular disease. Cochrane database of systematic reviews (Online). 2011.
- Jakobsen MU, O’Reilly EJ, Heitmann BL, Pereira MA, Bälter K, Fraser GE, et al. Major types of dietary fat and risk of coronary heart disease: a pooled analysis of 11 cohort studies. The American journal of clinical nutrition. 2009.
- Van Horn L, McCoin M, Kris-Etherton PM, Burke F, Carson JA, Champagne CM, et al. The evidence for dietary prevention and treatment of cardiovascular disease. Journal of the American Dietetic Association. 2008.
- Micha R, Mozaffarian D. Saturated fat and cardiometabolic risk factors, coronary heart disease, stroke, and diabetes: a fresh look at the evidence. Lipids. 2010.
- Mozaffarian D, Micha R, Wallace S. Effects on coronary heart disease of increasing polyunsaturated fat in place of saturated fat: a systematic review and meta-analysis of randomized controlled trials. PLoS Med. 2010.
- Hooper L, Martin N, Abdelhamid A, Davey Smith G. Reduction in saturated fat intake for cardiovascular disease. Cochrane Database of Systematic Reviews. 2015(6).
- Harcombe Z. Dietary fat guidelines have no evidence base: where next for public health nutritional advice? Br J Sports Med. 2016.
- Houtsmuller AJ, Zahn KJ, Henkes HE. Unsaturated fats and progression of diabetic retinopathy. Doc Ophthalmol. 1979.
- Black HS, Herd JA, Goldberg LH, Wolf JE, Thornby JI, Rosen T, et al. Effect of a Low-Fat Diet on the Incidence of Actinic Keratosis. New England Journal of Medicine. 1994.
- Moy TF, Yanek LR, Raqueño JV, Bezirdjian PJ, Blumenthal RS, Wilder LB, et al. Dietary Counseling for High Blood Cholesterol in Families at Risk of Coronary Disease. Prev Cardiol. 2001.
- Ley SJ, Metcalf PA, Scragg RK, Swinburn BA. Long-term effects of a reduced fat diet intervention on cardiovascular disease risk factors in individuals with glucose intolerance. Diabetes Res Clin Pract. 2004.
- Ramsden CE, Zamora D, Leelarthaepin B, Majchrzak-Hong SF, Faurot KR, Suchindran CM, et al. 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. 2013.
- Harcombe Z, Baker J, Davies B. Evidence from prospective cohort studies does not support current dietary fat guidelines: A systematic review and meta-analysis. Br J Sports Med. 2016.
- Harcombe Z, Baker JS, Cooper SM, Davies B, Sculthorpe N, DiNicolantonio JJ, et al. Evidence from randomised controlled trials did not support the introduction of dietary fat guidelines in 1977 and 1983: a systematic review and meta-analysis. Open Heart. 2015.
- Harcombe Z, Baker JS, DiNicolantonio JJ, Grace F, Davies B. Evidence from randomised controlled trials does not support current dietary fat guidelines: a systematic review and meta-analysis. Open Heart. 2016.
- de Souza RJ, Mente A, Maroleanu A, Cozma AI, Ha V, Kishibe T, et al. Intake of saturated and trans unsaturated fatty acids and risk of all cause mortality, cardiovascular disease, and type 2 diabetes: systematic review and meta-analysis of observational studies. BMJ. 2015.
- Chowdhury R, Warnakula S, Kunutsor S, Crowe F, Ward HA, Johnson L, et al. Association of Dietary, Circulating, and Supplement Fatty Acids With Coronary Risk: A Systematic Review and Meta-analysis. Ann Intern Med. 2014.
- Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease. The American journal of clinical nutrition. 2010.
- Skeaff CM, Miller J. Dietary fat and coronary heart disease: summary of evidence from prospective cohort and randomised controlled trials. Ann Nutr Metab. 2009.
- Mente A, de Koning L, Shannon HS, Anand SS. A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med. 2009.
- Ramsden CE, Zamora D, Majchrzak-Hong S, Faurot KR, Broste SK, Frantz RP, et al. Re-evaluation of the traditional diet-heart hypothesis: analysis of recovered data from Minnesota Coronary Experiment (1968-73). BMJ. 2016.
- Ravnskov U, DiNicolantonio JJ, Harcombe Z, Kummerow FA, Okuyama H, Worm N. The Questionable Benefits of Exchanging Saturated Fat With Polyunsaturated Fat. Mayo Clinic proceedings Mayo Clinic. 2014.
- Rose GA, Thomson WB, Williams RT. Corn Oil in Treatment of Ischaemic Heart Disease. BMJ. 1965.
- Woodhill JM, Palmer AJ, Leelarthaepin B, McGilchrist C, Blacket RB. Low fat, low cholesterol diet in secondary prevention of coronary heart disease. Advances in experimental medicine and biology. 1978.
- Miettinen M, Karvonen M, Turpeinen O, Elosuo R, Paavilainen E. Effect of cholesterol-lowering diet on mortality from Coronary Heart-Disease and other causes: A Twelve-year Clinical Trial in Men and Women. The Lancet. 1972.
- Farvid MS, Ding M, Pan A, Sun Q, Chiuve SE, Steffen LM, et al. Dietary linoleic acid and risk of coronary heart disease: a systematic review and meta-analysis of prospective cohort studies. Circulation. 2014.
- Praagman J, Beulens JW, Alssema M, Zock PL, Wanders AJ, Sluijs I, et al. The association between dietary saturated fatty acids and ischemic heart disease depends on the type and source of fatty acid in the European Prospective Investigation into Cancer and Nutrition–Netherlands cohort. The American journal of clinical nutrition. 2016.
- Praagman J, de Jonge EA, Kiefte-de Jong JC, Beulens JW, Sluijs I, Schoufour JD, et al. Dietary Saturated Fatty Acids and Coronary Heart Disease Risk in a Dutch Middle-Aged and Elderly Population. Arterioscler Thromb Vasc Biol. 2016.
- Dehghan M, Mente A, Zhang X, Swaminathan S, Li W, Mohan V, et al. Associations of fats and carbohydrate intake with cardiovascular disease and mortality in 18 countries from five continents (PURE): a prospective cohort study. The Lancet. 2017.
- Ramsden CE, Hibbeln JR, Majchrzak SF, Davis JM. n-6 Fatty acid-specific and mixed polyunsaturate dietary interventions have different effects on CHD risk: a meta-analysis of randomised controlled trials. British Journal of Nutrition. 2010.