Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Fat myths and cardiovascular disease
- Uffe Ravnskov (31 March 2001)
-
Are refined carbohydrates more important?
- Ruth Livingstone (31 March 2001)
-
Painted pill
- J A Consumer (1 April 2001)
-
Alternative to reducing dietary fat
- Jeffrey J Segall (2 April 2001)
-
Let's revisit fat intakes achieved...
- Allan Sinclair (2 April 2001)
-
Nice Review, Wrong Conclusion
- Richard E Collins (3 April 2001)
-
A litany of shortcomings - both of the study and of the BMJ's presentation of it
- Jim Mann (4 April 2001)
-
Unfair treatment for an exciting result
- Federico Relimpio (4 April 2001)
-
Meta-analysis of dietary trials: Potential perils of combining "oranges" and "apples"
- Frank B Hu (10 April 2001)
-
Re: Meta-analysis of dietary trials: Potential perils of combining "oranges" and "apples"
- George Davey Smith (20 April 2001)
-
Dietary fat intake and prevention of cardiovascular disease
- Rosely Sichieri (30 April 2001)
-
Re: Re: Meta-analysis of dietary trials: Potential perils of combining "oranges" and "apples"
- Frank M Sacks (1 May 2001)
-
Inaccurate précis
- Ivan J Perry (12 May 2001)
-
Authors reply
- Lee Hooper, Carolyn Summerbell, Nigel Capps, Rudolph Riemersma, Rachel Thompson, Shah Ebrahim, Ge (6 June 2001)
-
Re: Meta-analysis of dietary trials: Potential perils of combining "oranges" and "apples"
- Eddie Vos (24 July 2001)
|
|
|||
|
Uffe Ravnskov Magle nStora Kyrkogata 9, S-22350, Lund
Send response to journal:
|
In spite of the trivial findings Dr Hooper and his coworkers concluded that their meta-analysis of the cholesterol-lowering dietary trials supports ”a central role of dietary fat intake in the causation of cardiovascular disease”.1 This surprising conclusion is based on a borderline significant improvement of the combined cardiovascular events, and a subgroup analysis showing a significant improvement of the same outcome in five trials of more than two years duration. These effects may rather be due to bias. Most of the trials reviewed by Dr. Hooper et al were open or single-blind. It is widely acknowledged that such trials are prey to bias, and had they been drug trials they would never have been accepted by any major medical journal to-day. The only double-blind dietary trial with a favourable outcome for cardiovascular events was the Veterans Administration Trial, but that trial was biased by a significantly larger number of heavy smokers in the control group.2 Another disquieting finding was the higher degree of atherosclerosis at autopsy in the treatment group, certainly not in support of a beneficial effect of dietary intervention. The only outcome free of bias is total mortality. The mean weighted rate ratio for that event was close to one in all subgroups, indicating that dietary fat intake has no importance, and this conclusion is supported by numerous epidemiological studies. Dynamic population studies of more than hundred time periods in 35 countries found no association between change of animal fat consumption and change of coronary mortality; twenty-seven cohort and case-control studies including more than 150,000 participants did not find any difference between the character of dietary fat in coronary patients and control individuals; and four cohort studies found no supportive associations between dietary fat and degree of atherosclerosis at autopsy.3 Add to that the numerous cross-sectional studies, most of which found no or even contradictory associations between fat consumption and the morbidity and mortality of cardiovascular disease.3 Trans fat, a product of the industrial hardening of vegetable oils with many unfortunate effects on experimental animals, is the only type of fat that consistently has been found to be consumed more often by coronary patients.4 It is time to tell the populations that the dietary advices they have received since many decades are not only ineffective, they may even have adverse health effects.5 Uffe Ravnskov. MD, PhD
1. Hooper L, Summerbell CD, Higgins JPT, Thompson RL, Capps NE, Davey Smith G, Riemersma RA, Ebrahim S. Dietary fat intake and prevention of cardiovascular disease: systematic review. BMJ 2001;322:757-63 (31 March) 2. Dayton S, Pearce ML, Hashimoto S, Dixon WJ, Tomayasu U. A controlled clinical trial of a diet high in unsaturated fat in preventing complications of atherosclerosis. Circulation 1969;40,suppl II:1-63. 3. Ravnskov U. The questionable role of saturated and polyunsaturated fatty acids in cardiovascular disease. J Clin Epidemiol 1998;51:443-60. 4. Oomen C, Ocké MC, Feskens EJM, van Erp-Baart MAJ, Kok FJ, Kromhout D. Association between trans fatty acid intake and 10-year risk of coronary heart disease in the Zutphen Elderly Study: a prospective population-based study. Lancet 2001;357:746-51. 5. Ravnskov U. The Cholesterol Myths. Washington: New Trends Publishing, 2000. |
|||
|
|
|||
|
Ruth Livingstone, General Practitioner Little Surgery, Stamford, Lincs
Send response to journal:
|
Dear Editor It is disappointing to see how little real health benefit is gained by the efforts of medical professionals to persuade their patients to reduce their dietary intake of fat. Coronary heart disease was almost unknown in the 19th century. The emergence of coronary heart disease as a major cause of death in developed nations in the 20th century may have many causes. However, it could be argued that the rise in incidence of coronary heart disease follows some 20-30 years after the rise in the consumption of refined carbohydrates, particularly sugar. We know that many of the conditions associated with coronary artery disease (diabetes, hypertension and obestity)are associated with insulin resistance. Perhaps we are looking in the wrong direction. Could the dietary villian be refined carbohydrates? |
|||
|
|
|||
|
J A Consumer NASA
Send response to journal:
|
Put it like this: The pills are labeled thusly, "This pill will produce DEATH in exactly 5 years and 5 minutes". Then the pill label is removed and each of a population swallows a pill and goes merrily on his way. The population is monitored each week for a long while,--over a two year period, say, or three years, or four. All is well! Alas, not one has expired. The pills are harmless! Tobacco, asbestos, liberal laws and courts, etc. The problem is one of degradation, and only when the critical zone is reached do the visible results begin to appear. 5 yr, 6 min.? Health is in shades of gray, and symtoms often lag--- Just a Consumer |
|||
|
|
|||
|
Jeffrey J Segall Home
Send response to journal:
|
EDITOR- The findings of the systematic review by Hooper et al [1] do not justify their support for a “central role of dietary fat intake in the causation of cardiovascualar disease”, and implicitly in its prevention. Instead, they should have referred to the alternative dietary hypothesis on milk and lactose. [2-5] The challenge for researchers was well expressed by Artaud-Wild et al: “The strong association of milk and many of its components with CHD mortality, both before and after adjustment for CSI [Cholesterol-Saturated Fat Index], suggested that something about milk must be strongly related to CHD mortality”.[4] That challenge should now be met by a randomised comparison of a lactose (and galactose) free diet with a diet of reduced and modified fat intake. Jeffrey J Segall 1. Hooper L, Summerbell CD, Higgins JPT, Thompson RL, Capps NE, et al. Dietary intake and prevention of cardiovascular disease: systematic review. BMJ 2001;322:757-63. 2. Wells WW, Anderson SC. The increased severity of atherosclerosis in rabbits on a lactose-containing diet. J Nutr 1959;68:541-49. 3. Segall JJ. Is milk a coronary health hazard? Br J Prev Soc Med 1977;31:81-5. 4. Artaud-Wild SM, Connor SL, Sexton G, Connor WE. Differences in coronary mortality can be explained by differences in cholesterol and saturated fat intakes in 40 countries but not in France and Finland. A paradox. Circulation 1993;88:2771-79. 5. Segall JJ. Epidemiological evidence for the link between dietary lactose and atherosclerosis. In: Colaco CALS, ed. The glycation hypothesis of atherosclerosis. Austin TX: Landes Bioscience, 1997:185-209. |
|||
|
|
|||
|
Allan Sinclair, Manager of Public Health Lakeland Regional Health Authority, AB, Canada
Send response to journal:
|
I have reviewed study and the extra Table describing the papers included in the Hooper et al. review and I feel I must make some comments. To me, the studies included seem to fall into at least three different categories... Firstly, some of the studies appear to have achieved only very modest fat intake changes (ie. reduction of 3-5% total fat and 2-3% sat fat). Secondly, some studies do not record dietary results achieved at all. Thirdly, a few studies actually increased fat intake quite substantially with additions of corn/olive oil supplements. As noted in the article, follow-up times were indeed disapointingly short... and this cannot be helped, but inclusion of studies with no intake data and/or very modest or increased intake effects can be helped. I would like to see the review analysed on the basis of achieved dietary change. I suspect that including studies with missing intake data, short follow-up times, and overall very modest diet effects all conspire to attenuate the results. I am also of the belief that the baseline fat intakes are, in general, far too high a starting point. At the turn of the century in North America, the diets of affluent people provided approximately 20% of calories in the form of fat...by todays standards, these (formerly)average people would have a diet comparable to the intervention group in your included study #27 with intakes of 18% total/6% sat fat. A study by the way that did show a positive effect even with only a 1.7 year follow-up. Allan Sinclair, M.H.Sc.,
|
|||
|
|
|||
|
Richard E Collins, Director of Heart Disease Reversal Program Alegent Health, Heart Institute, Omaha, NE. USA
Send response to journal:
|
My complements on the meta-analysis, but I am amazed at the the conclusion. Your headlines state that low fat diets are not much help, but mortality is decreased by 9% and events by 16%. If you are in the 9% group, it is 100% death. To say that this is only "marginal" is sending the wrong message to America. We have observed in the Ornish Program that survival is even more pronounced at the five year mark. It takes months to remodel the coronaries and to stablize plaque. Progression of coronary heart disease does not occur over night. The strongest case occurs in societies that have a lifetime of fat restriction, events and survival are significantly better in comparison to countries with high fat diets. No need to change the study, just the headlines. Low fat diets eventually create long term benefits! Just give the diet time! In addition, one must remember that studies have shown that the most beneficial effect is the change in the protein base from animal to plant coupled with a low fat diet. Richard E. Collins, MD |
|||
|
|
|||
|
Jim Mann, Professor in Human Nutrition and Medicine,Head of Department of Human Nutrition Department of Human Nutrition, University of Otago, P O Box 56, Dunedin, New Zealand
Send response to journal:
|
The BMJ website of the week (31 March) makes a claim which has already reached this part of the world, and will no doubt trouble those involved with health promotion worldwide: "So it’s official: low fat diets don’t do much to improve cardiovascular morbidity and mortality." This statement is derived from the paper in the same issue by Hooper et al (pages 757-763), the title of which declares it to be a systematic review relating to "Dietary fat intake and prevention of cardiovascular disease". The opening sentence of the website review is unfortunate since it does not accurately represent the conclusions of the article. The article itself is disappointing since while it might fulfil the Cochrane review criteria for a metaanalysis of clinical trials, it certainly does not provide a systematic review of the topic. It is surely important to consider the overall appropriateness of each trial before its inclusion in a metaanalysis. Several of the trials included are small or of short duration. It is difficult to imagine how the risk of coronary heart disease which has accrued over a prolonged period, perhaps a lifetime, may be reversed by dietary modification in less than two years. The DART trial (fat modification arm) contributed the greatest number of cardiovascular endpoints, yet there are at least two reasons why this trial was most unlikely to be able to show a beneficial effect. Limited dietary instruction was given and not surprisingly there was no appreciable reduction in cholesterol. Since modification of fat quality results in a predictable average change in cholesterol level, there was clearly limited compliance with the dietary advice. Furthermore the study was of too short a duration (2 years) to have expected a reduction in clinical events and mortality. The Veteran’s Administration Study which contributed the second largest number of deaths fulfilled all the criteria for a good randomised, controlled, double-blind clinical trial. Yet the experimental diet involved a ratio of polyunsaturated to saturated fatty acids of 1.5, far greater than would today be considered desirable. Such a diet would be expected to reduce cholesterol and coronary heart disease morbidity and mortality – as indeed it did, but perhaps not total mortality, as was also the case. Is it appropriate to compare such a dietary intervention with that used in the Oslo Study where saturated fats were replaced by a range of whole grain cereals, vegetables, fruit and some unsaturated fatty acids and in which study cardiovascular events and total mortality were reduced. Meta- regression analyses were employed to disentangle these issues but one might question whether the statistical power was sufficient to achieve this. One might also question why the authors chose to examine total cardiovascular rather than coronary artery endpoints since there has never been any serious understanding that changing dietary fat does much to prevent strokes. The link is with coronary heart disease. Finally one might expect that in a systematic review it would be appropriate to interpret the results of the clinical trials in the context of the enormous body of descriptive epidemiology which supports the dietary fat–cholesterol–coronary heart disease link. While evidence based medicine now relies almost exclusively on clinical trials as far as drug treatments are concerned, when considering evidence based nutrition with nutrient-disease links accumulating over a lifetime it is imperative to consider epidemiological and experimental evidence in conjunction with trials. Having described a litany of shortcomings of the published metaanalysis, it is nevertheless important to note the tentative conclusions offered by the authors of the paper : "There is a small but potentially important reduction in cardiovascular risk with reduction or modification of dietary fat intake, seen particularly in trials of longer duration." The conclusion appears to have been largely ignored in the website annotation and to be underplayed in the "This week in the BMJ" paragraph referring to this paper. It is most disappointing to find that a Journal as distinguished and influential as the BMJ has utilised a rather sensational and inaccurate approach. Jim Mann, Professor in Human Nutrition and Medicine, University of Otago, P O Box 56, Dunedin, New Zealand Murray Skeaff, Senior Lecturer in Human Nutrition, University of Otago, P O Box 56, Dunedin, New Zealand Stewart Truswell, Professor of Human Nutrition, University of Sydney, New South Wales, Australia |
|||
|
|
|||
|
Federico Relimpio, Marqués de Paradas s/n, 41001 Seville, Spain Centro de Especialidades Virgen de los Reyes
Send response to journal:
|
I think that data presented by Hooper et al in the BMJ are not analyzed properly. Taking into account the prolonged time that it takes to make an atherosclerotic cardiovascular tree, it is not fair to demand a dietary intervention to prove clear-cut statistically significant benefits in a so short period. For this reason, I think that only studies with long follow-up should be considered at this regard. If one makes so (even though such a period remains considerably short), a significant degree of protection is obtained. In addition, if one compares that very degree of protection with that achieved by well-established protecting practices like low-dose aspirin, blood pressure lowering or lipid lowering therapy, one could find that dietary intervention to lower fat is not much less effective and considerably cheaper. Taking into account that we deal with healthy people (primary prevention), offering a protective RR of 0.76 in only two years is not only a relevant result, but an encouraging one. Just compare that with RRs obtained from other well-documented risk-lowering interventions in the primary prevention setting. Given the current epidemic of metabolic and cardiovascular disease, the message that authors give to the scientific community is openly misleading. I would strongly favour a correction of data interpretation. |
|||
|
|
|||
|
Frank B Hu, Assistant Professor of Nutrition Harvard School of Public Health
Send response to journal:
|
Dear Sir, Despite its popular use in combining findings from multiple studies, meta-analysis has many well-known pitfalls (1). These include lack of homogeneity of the studies, failure to consider important covariates, inadequate understanding of the scientific subject in question, failure to consider quality of the studies, and biases in including or excluding certain studies. Some of these pitfalls are exemplified in the paper by Hooper et al. (BMJ 2001;322:757-63), who reviewed 27 intervention trials of fat reduction or modification. There are several problems with this meta-analysis. First, it mixed conceptually different dietary intervention approaches, i.e., total fat reduction vs. using unsaturated fats to replace saturated fat. Neither epidemiology nor clinical trials support a benefit of low-fat diets on either serum cholesterol or risk of coronary heart disease (CHD). Substitution of vegetable oils rich in unsaturated fats for saturated or trans fats, on the other hands, lowers LDL and reduces cardiovascular endpoints, as indicated by several previous systematic reviews of this topic. A second problem is that the meta- analysis mixed studies specifically on cardiovascular disease with those designed for other purposes (such as cancer prevention, weight loss, and so on). It is questionable whether cardiovascular endpoints are adequately ascertained in studies designed for other purposes. Another problem of the meta-analysis is that it did not adequately consider compliance of the subjects. Poor compliance to a low-fat diet is a well-known problem in dietary trials. For example, the DART study (2) was unable to achieve anywhere close to the goal set for total fat. This and other trials showed minimal reduction in serum cholesterol with the dietary intervention, demonstrating poor adherence. In two earlier trials which showed cardiovascular benefits of unsaturated fats (3, 4), adipose tissue fatty acid levels were used to monitor compliance. Although the Finnish Mental Hospital Study (4) was not included in the meta-analysis as it did not meet the subjective criteria, it did provide important evidence for an effect of fat modification. This meta-analysis highlights potential perils in combining “oranges” and “apples” in aggregating dietary studies. The included trials are a veritable hodge-podge of aims, methodologies, populations, and quality – This same kind of approach was applied by one of the authors to cholesterol lowering drug trials (5). The results from that meta-analysis would predict that statin drugs would increase mortality in the 4S, WOSCOPS, LIPID, and CARE trials, whereas the opposite in fact occurred. Thus, similar degree of caution is needed to interpret results from either a meta-analysis or an individual study. Frank B. Hu, MD PhD
Dept. of Nutrition, Harvard School of Public Health, Boston, MA 02115 Reference: 1. Bailar JC, 3rd. Passive smoking, coronary heart disease, and meta- analysis. N Engl J Med 1999;340(12):958-9. 2. Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM, et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). Lancet 1989;2:757-761. 3. Dayton S, Pearce ML, Hashimoto S, Dixon WJ, Tomiyasu U. A controlled clinical trial of a diet high in unsaturated fat in preventing complications of atherosclerosis. Circulation 1969;40:(Suppl II):1-63. 4. Turpeinen O, Karvonen MJ, Pekkarinen M, Miettinen M, Elosuo R, Paavilainen E. Dietary prevention of coronary heart disease: The Finnish Mental Hospital Study. Int J Epidemiol 1979;8:99-118. 5. Smith GD, Song F, Sheldon TA. Cholesterol lowering and mortality: the importance of considering initial level of risk. Bmj 1993;306(6889):1367- 73. |
|||
|
|
|||
|
George Davey Smith, Professor of Clinical Epidemiology Bristol University
Send response to journal:
|
The predictive value of meta analysis Hu et al criticise our meta-analysis of fat lowering trials on the basis that a meta-analysis I was involved with of cholesterol lowering trials (1) 'would predict that statin drugs would increase mortality in the 4S WOSCOPS, LIPID, and CARE trials, whereas the opposite in fact occurred'. This is not the case. Our meta-analysis was a meta-regression which demonstrated that the greater the cholesterol reduction, the greater the reduction in coronary heart disease (CHD) mortality. Trials up until the time of our meta-analysis (1993) had generally only produced small cholesterol reductions (e.g. 9% in the WHO clofibrate study and the Lipid Research Clinics cholestyramine study), the exception being the partial ileal bypass surgery trial which produced a 23% reduction (2). The meta- regression analysis relating percent cholesterol reduction to CHD mortality would have predicted an odds ratio of 0.51 for CHD deaths on treatment with simvastatin in the 4S trial, which produced a 25% net reduction in cholesterol level; the observed result was 0.58. This would clearly have produced a benefit in terms of all-cause mortality. At the time our analysis was carried out there was insufficient direct data to be certain about whether statins would result in any adverse influence on other causes of death. We showed that in trials up to that date a small increase in non-CHD mortality was observed. This was only seen in drug trials (not in diet or surgical studies) and the increase in non-CHD mortality was not related to the degree of cholesterol reduction, demonstrating that it was not the fall in cholesterol which produced this effect, but it was a side-effect of the drug therapies used (with fibrate drugs being the class contributing most data to the meta-analysis). The question of whether the statin agents would have such side effects was an open question in 1993; since then large-scale trials have ruled out any major effect in this regard. The purpose of meta-analysis and meta- regression is to provide a quantitative overview of current trial findings and highlight areas of uncertainty. The power of trials and meta-analyses of such trials over observational data, which Hu et al seem to value more highly in their response (with respect to trans fatty acids in relation to CHD events, for example) is highlighted by the work of this group on vitamin E supplement use and CHD, where (probably due to uncontrolled confounding) highly misleading suggestions of major benefits were claimed (3), which failed to be confirmed by large-scale RCTs (4). (1) Davey Smith G, Song F, Sheldon TA. Cholesterol lowering and mortality: the importance of considering initial level of risk. BMJ 1993;306:1367-1373. (2) Ebrahim S, Davey Smith G, McCabe C, Payne N, Pickin M et al. What role for statins? A review and economic model. Health Technology Assessment 1999;3(19):1-91. (3) Stamfer MJ et al. Vitamin E consumption and the risk of CHD in women. NEJM 1993;328:1444-9 (4) Hooper L, Ness A, Davey Smith G. Vitamin E and CHD. Lancet, in press. |
|||
|
|
|||
|
Rosely Sichieri, Assistant Professor in Epidemiology State University of Rio de Janeiro
Send response to journal:
|
EDITOR: The paper by Hooper at al. (BMJ 2001;322: 757-63) on dietary fat intake and prevention of cardiovascular disease is timely and well conducted. Reduction of dietary fat is one of the main points included on the dietary recommendations of many countries, however research has not proved that eating a low fat diet increase the life expectancy. Therefore, and based on the results of the systematic review conducted by the authors, I consider that their conclusions are misleading. The authors concluded, " alteration of dietary fat intake had small effects on total mortality" in the abstract-results. However, the overall relative risk and the relative risks of the detailed analysis conducted are highly consistent with lack of association between fat intake and mortality (rate ratio 0.98; CI 0.86-1.12). There is also a lack of association for cardiovascular mortality (rate ratio 0.91; CI 0.77-1.07). The only association found was with cardiovascular events, however this association was vanished when the authors excluded a trail that supplemented with fish oil. The total mortality with 1430 events allows a powerful message of lack of association for a very hard outcome, but the authors concluded, " we are left with a suggestion that less total fat or less of any individual fatty acid fraction in the diet is beneficial". How the authors were left with this suggestion is not easily grasp from the data. Rosely Sichieri
|
|||
|
|
|||
|
Frank M Sacks, Professor of Cardiovascular Disease Prevention Harvard School of Public Health
Send response to journal:
|
In the published article, Prof GD Smith was abundantly clear about the goals of his meta-analysis, but now he obfuscates! The objective was stated in the title: "Cholesterol lowering and mortality: the importance of considering initial level of risk", and all the figures and tables concerned this relationship. The central finding was an equation that predicted odds ratio in the treated compared to placebo group as a function of the inverse of the initial CHD risk. Stated plainly, the higher the initial risk, the greater the relative risk reduction in mortality. The authors used this equation to predict the results of cholesterol lowering treatment, and even to give clinical guidelines. This equation yields increases in mortality when the rates in the statin trials are used. Moreover, the premise of the equation, that odds ratio is a function of risk, has not generally proven true in the statin trials where relative risk reduction for mortality and other endpoints has been largely independent of level of risk. |
|||
|
|
|||
|
Ivan J Perry, Professor of Public Health Department of Epidemiology & Public Health, University College Cork, Republic of Ireland
Send response to journal:
|
Hooper and colleagues may be interested to note that their paper "Dietary fat intake and prevention of cardiovascular disease: systematic review" (BMJ 2001:322:757-63) has been cited in "Medicine Weekly - the voice of Irish medicine", in support of the argument that the link between dietary fat and heart disease is largely "spurious and confounded". I assume that this spin on their systematic review can be traced in part to the "This Week in the BMJ" headline "Reducing dietary fat has little effect on cardiovascular disease". Clearly for many readers of the BMJ this headline conveys the essential message of the paper. Few will have read the paper in its entirety and even fewer will have taken the trouble to download and peruse the unweildly table giving details of the studies included in the meta analysis. The references to "40 intervention arms" in the "This Week in the BMJ" summary and to 27 studies in the abstract of the paper are frankly misleading. The reader should be alerted to the fact that the conclusions on cardiovascular disease events are based on 16 studies of which only 5 had a mean follow-up of greater than 2 years. The mortality data are even more sparse, based on 11 studies with interventions of variable intensity. The protective effect of reduction or modification of dietary fat on CVD events observed in the 5 studies with at least 2 years follow-up is the major finding from this review and should have been highlighted by the BMJ. Both the authors of the paper and the journal have a responsibility not to add to the confusion on the role of dietary fat in the causation of cardiovascular disease. |
|||
|
|
|||
|
Lee Hooper, Lee Hooper: Research Associate in Evidence Based Care and Systematic Review (please see original paper for these), Carolyn Summerbell, Nigel Capps, Rudolph Riemersma, Rachel Thompson, Shah Ebrahim, Ge
Send response to journal:
|
Thank you to all who have taken the trouble to comment on our recently published systematic review1. We are interested in the contrasting interpretations of the findings: Drs Ravenskov, Livingstone, Segall and Sichiery feel that the results point to lack of efficacy of dietary fat modification, whilst Drs Collins, Mann and Relimpo emphasise its effectiveness. We agree with Drs Collins, Relimpo and JA Consumer that the important factor in interpretation of the results is time. As pointed out, lifestyles take time to harm. We don’t expect to see excess deaths a couple of years after a young person starts smoking, but we know that, in a cohort, effects on survival will be evident by retirement age. In the meantime we expect to see an increase in smoking related morbidity, minor at first (coughs, chest infections) with more serious illness gradually appearing. It seems likely that this is the type of pattern (but healthful rather than harmful) we are seeing when we look for effects of a fat modified diet. Up to about two years we see very little outward effect, although clearly many trialists were hoping to find effects in shorter periods of time than this. From two to five years we see an improvement in cardiovascular events in the modified fat group (but no effects on death). It may be that in ten years or so there is a clear and strong effect on cardiovascular deaths and even total mortality, but we do not know this as there are no trials long enough to confirm or refute this. Meta-regression strengthens this possibility, showing a greater difference between cardiovascular events in modified fat and control groups over longer times. This was the reason for our tentative conclusion that there is ‘a small but potentially important reduction in cardiovascular risk with reduction or modification of dietary fat intake’. To address some of the criticisms of the methodology: Professor Hu and colleagues are concerned over our inclusion of both fat lowering and fat modifying interventions in the review. Both types of diet reduce serum LDL cholesterol. It is interesting to note that almost all of the event data come from fat modification trials: of 40 intervention arms included, 2/15 that aimed to lower total fat intake had 80 CV events and 46 deaths; 10/14 that aimed to modify the type of fat eaten had 663 CV events and 1003 deaths recorded, other arms with events aimed to make both changes. We share Dr Ravenskov’s concern that very few of the included trials were double blinded. Blinding participants to their own dietary intake is costly and complicated, and may not even be feasible. The most important type of blinding is that of the physician or assessor who decides whether or not a cardiovascular (CV) event has taken place. Counting combined CV events in trials with adequate physician blinding we find 423 events/8521 person years in the control group and 365/9205 in the intervention groups (relative risk ~0.80). This compares with 220/2442 and 208/2469 (relative risk ~0.93) from trials where physician blinding is unclear or inadequate. Lack of physician blinding does not appear to be exagerating the effect of dietary fat on cardiovascular events, rather the reverse. Rules for inclusion of studies into the review were decided before we were aware that some studies would aim to modify fat intake, while increasing overall fat intake. It is inappropriate to change inclusion criteria in response to unwelcome results of the studies found. The two trials that resulted in increased fat intake (London corn/olive2 and MRC soya3) almost certainly did result in dramatic changes in type of fat ingested, and in serum total cholesterol levels (people in the fat modification arm of MRC soya had mean falls in serum cholesterol of 0.64 mmol/L more than those in the control arm at 4 years, in the corn oil arm of London corn/olive cholesterol fell by 0.58 mmol/L more than the control at 12 to 18 months, however cholesterol rose in the olive oil arm, including only 26 participants, by 0.3 mmol/L more than control). Figure 1 illustrates the relationship between mortality rate ratio and the observed change in % energy taken as fat achieved in the intervention group (compared with the control group). When the diet increased total fat intake, the treatment effect tended to favour the control group. This result did not reach statistical significance.
Professor Mann is right in his comments that some trials included were small and of short duration. When we began the review we did not know whether there might be short term effects of dietary fat on CV events or mortality (via thrombotic mechanisms for example) that might be visible early after dietary modification. Results from the review in trials of 6 to 24 months duration make this unlikely, which is useful to know. In meta-analysis small trials have an important role to play in augmenting larger trial data, despite their not having sufficient power individually to provide useful data on death or events. However, there are too few small trials to provide much help here. Again, this was not known when we began the review. As stroke and peripheral vascular events were (sadly) poorly recorded, the numbers of cardiovascular events were little different from coronary heart disease events (25 peripheral vascular events and 32 strokes, 57 events, in the control group, and 27 and 18 respectively, 45 events, in the intervention group - as there are fewer peripheral vascular and stroke events in the intervention group adding these to the coronary events will tend to enhance the apparent effect of fat intervention compared to control, not reduce it). Compliance is always an issue in lifestyle trials, and for this reason we expected to see a differential effect of subgrouping trials according to food provided or dietary advice given. However, if anything diet advice trials seem to have a slightly better effect on cardiovascular events (original paper, fig 5). Despite trials not achieving their stated dietary goals, the mean fall in serum total cholesterol was 0.64 mmol/L (11.1% of the intitial mean 5.8 mmol/L) in the experimental compared to control groups. We did, of course, try to separate out the effects of reducing fat intake, from those of modifying type of fats consumed, through meta- regression (using dietary fat intakes actually achieved, as in figure 1). Unfortunately there is insufficient information from adequately documented trials to generate robust estimates of the effects of total fat, saturated fat, poly-unsaturated fat or even change in serum cholesterol on CV events (original paper, table 2). However, despite some epidemiological observations not supporting a benefit of low fat diets on coronary heart disease4, all of the meta-regressions using trial data suggest that less total fat, or less of any fat fraction, is protective of cardiovascular events. That the association for mono-unsaturated fats is the most robust (so that the greater the reduction achieved in mono-unsaturated fats in the intervention group compared to the control group, the greater the reduction in cardiovascular events) should make us pause in our interpretation of the observational data. We feel that it is important to realise that half a century on from the genesis of the diet-heart hypothesis there have been so few people enrolled in dietary fat trials, and for such very short periods (in marked contrast to the numbers of people and years in the statin trials). Over 50 years, dietary fat trials have amassed only 30,000 person years of observation in trials over 6 months long, and only 8,300 person years of observation in trials longer than two years for which useable data on mortality are available. Much more data is available already for statins - the 4S trial5 alone amassed over 19,000 person-years of observation, a mean of over 4 years for each person. We feel that some of the comments on our systematic review are actually justifications as to why a larger effect was not seen. Given the paucity of the data available, we feel it is remarkable that an effect was seen at all. We are criticised for including studies that provided interventions other ‘than would today be considered desirable’. However, what is clear to us following the review is that, whatever views may be fashionable, due to a lack of good long term trial data we still do not know what types of dietary fat changes (fat lowering or fat modification) are protective against cardiovascular events or death. Data from cohort studies (that may conflict with trial data due to higher levels of inherent bias) and trials with intermediate outcomes (such as serum cholesterol) do not give us the whole picture. Lee Hooper, Carolyn Summerbell, Nigel Capps, Rudolph Riemersma, Rachel Thompson, Shah Ebrahim and George Davey Smith. Reference List 1. Hooper L, Summerbell CD, Higgins JPT, Thompson RL, Capps N, Davey Smith G et al. Dietary fat intake and prevention of cardiovascular disease: systematic review. BMJ 2001;322:757-63. 2. Rose GA, Thomson WB, Williams RT. Corn oil in treatment of ischaemic heart disease. British medical Journal 1965;1:1531-3. 3. MRC. Controlled trial of soya-bean oil in myocardial infarction. Lancet 1968;2:693-9. 4. Hu FB, Stampfer MJ, Manson JE, Rimm EB, Colditz GA, Rosner BA et al. Dietary fat intake and the risk of coronary heart disease in women. N.Engl.J.Med. 1997;337:1491-9. 5. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383-9. |
|||
|
|
|||
|
Eddie Vos, maintains http://health-heart.org Sutton Qc Canada
Send response to journal:
|
The Hu, Sacks and Willett rapid response in eBMJ to the Hooper et al meta-analysis regarding the lack of conclusive evidence of cardiovascular benefit by dietary fat changes suggests that the Finnish Mental Hospital Study (1) gives important evidence of benefit of unsaturates. This Finnish study would have been rightfully excluded as no fish oil or other omega-3 directed studies were included like, for example, the Lyon Diet Heart Study. The referred to male-subject part of the Finnish study demonstrates the effects of simultaneous increases in omega-6 linoleic and in omega-3 alpha-linolenic fatty acids in hospital diets. While the abstract and the paper itself stress the benefits of polyunsaturated over saturated (dairy) fats via an observed 16% serum cholesterol reduction, there was no consideration by the authors of the fact that the soybean oil used was both high in omega-3 and in omega-6 polyunsaturates. The Finnish study could be considered the world's first large omega-3 trial as the benefits found would be in line with the Lyon study (and possibly in spite an excessively high omega-6 supply). It is now well accepted that linoleic and alpha-linolenic acids have fundamentally different -and naturally opposing- roles in health ... with high omega-3 but low omega-6 fats (fish, canola and flax) demonstrating benefits in cardiovascular diseases and in arrhythmia prevention. For those not current with this Finnish cross-over design trial, the linolenic intake was 5.6 g/d, vs. 1.9 g/d in the control group (before processing losses) while similarly linoleic was 34 g/d vs. 11 g/d [coronary heart disease death or major ECG change: 4.2 vs. 12.7 / 1000 man -years; RR=0.33; P=0.001]. To put these amounts of polyunsaturates in perspective, the International Society for the Study of Fatty Acids and Lipids in their 1999 Adult Adequate Intake recommendations (http://www.issfal.org.uk/adequateintakes.htm) suggests that this linolenic intake would be just below "adequate" in the control group and became "adequate" in the treatment group. On the other hand, linoleic was about twice the "upper limit" in the control and was boosted to 5 times the ISSFAL recommended upper limit in the treatment group. With these confounding and possibly opposing effects in health, it is no wonder that the past 5 decades of fat-changes studies (apart from omega -3 studies) have not come up with clear answers and as long as polyunsaturate studies and statin drug studies, continue to be presented as being "cholesterol lowering studies" without considering their multiple positive and negative effects, confusion will reign. Hooper et al provided an important contribution by pointing out some of the weaknesses in the studies and in the popular fat recommendations. (1) Turpeinen O, Karvonen MJ, Pekkarinen M, Miettinen M, Elosuo R, Paavilainen E. Dietary prevention of coronary heart disease: The Finnish Mental Hospital Study. Int J Epidemiol 1979 Jun;8(2):99-118. Eddie Vos Sutton Qc Canada J0E 2K0 vos@health-heart.org |
|||
