Abstract

INTRODUCTION: In the past, few systematic reviews were created examining the relation between meat consumption and CVD. A recent systematic review found insufficient evidence for an association between meat consumption and CHD. And another one concluded that processed meat - but not red meat - was associated with a higher risk of CHD. Both reviews did not examine the relation with total CVD, and did not present data about white meat/poultry, or specific types of meat.
OBJECTIVES:

• To review all prospective studies which published information about total meat, and meat products, in relationship with cardiovascular disease.
• To define the amount of consumption found to be related with possible effects on CVD.
• To define possible effect modification by confounders.

DATA SOURCE: The Pubmed database was searched (No start date - May 25th, 2010) for relevant articles using the keywords "meat, meats, vegetarian, vegetarianism, or vegetarians", and a fair amount of other keywords" combined with "prospective, cohort, follow-up, or longitudinal". The exact search term is described Here.
Prospective studies published in the English language were included. Reference lists were searched for additional articles.
RESULTS: 44 articles were found which provided information about 23 different cohorts. Of these, 0 articles were excluded.

Total CVD.

• Total meats: Data was provided by 6 cohorts, including 3,289, and 2,380 cases for CVD risk, and mortality, respectively. No evidence was found for an association with CVD risk (RR = 0.97), or CVD mortality (RR = 1.11).
• Red meat: Data was provided by 2 cohorts, including 170, and 19,577 cases for CVD risk, and mortality, respectively. Inconclusive evidence was found for an association (RR = 1.33).
• Processed meat: Data was provided by 1 cohort, including 19,577 cases for CVD mortality. Inconclusive evidence was found for an association (RR = 1.17).

CHD.

• Total meats: Data was provided by 9 cohorts, including 1,303, and 2,928 cases for CHD risk, and mortality, respectively. Suggestive evidence was found that total meats increase CHD mortality (RR = 1.47). This effect may be restricted to subjects being a vegetarian for > 5 years, and for deaths before age 80. Inconclusive evidence was found for an association with CHD risk (RR = 1.25).
• Red meat: Data was provided by 4 cohorts, including 1,610, and 1,119 cases for CHD risk, and mortality, respectively. Suggestive evidence was found that red meat increases CHD mortality among women (RR = 1.54).
• Processed meat: Data was provided by 4 cohorts, including 1,481, and 970 cases for CHD risk, and mortality, respectively. Suggestive evidence was found that processed meat increases CHD risk (RR = 1.51). No evidence was found for an association with CHD mortality (RR = 0.75).

Stroke.

• Total meats: Data about stroke mortality was provided by 7 cohorts, including 12,322 cases. Inconclusive evidence was found for an association (RR = 0.80).
• Red meat: Data was provided by 3 cohorts, including 1,130, and 1,224 cases for stroke risk, and mortality, respectively. Inconclusive evidence was found for an association.
• Processed meat: Data about stroke mortality was provided by 2 cohorts, including 2,189 cases. No evidence was found for an association (RR = 0.87).

CONCLUSION: Suggestive evidence was found that total meats consumption increases CHD mortality, but this effect may be restricted to subjects being a vegetarian for > 5 years, and for deaths before age 80. Suggestive evidence was found that red meat consumption increases CHD mortality among women, and that processed meat consumption increases CHD risk. No/inconclusive evidence was found for any association with stroke or total CVD. No evidence was found for an association between consumption of poultry, pork, beef, or specific meat items and any end point. No clear evidence was found for a difference in effects between processed and red meats, and no data is available about the modifying effect of meat fats.
LIMITATIONS: All analysis' were based on data from a small amount of cohorts, or included results from primarily vegetarian cohorts, which were poorly adjusted for dietary life style variables. None of the articles presented data about the relation between lean meats vs fatty meats for any type of meat. And none of the articles mentioned any effect from meat fats. In addition, practically no data is available about specific meat items in relation to any CVD end point.

Introduction

Previous reviews: Systematic reviews examining the relation between meat consumption and CVD are practically restricted to publications in the last few years. In 2009, Mente A. examined the possibility of a causal relationship between meat consumption and CHD according to the Bradford Hill Criteria (Mente A). He found insufficient evidence for an association. And in 2010, Micha R. examined the relation between red- and/or processed meat consumption and CHD/stroke (Micha R). She found that processed meats - but not red meats - were associated with higher CHD incidence.
Limitations of these reviews: Mente A. did not stratify effects of meats according to their specific types, and only examined the relation with CHD.
Micha R. chose to present results as a continuous analysis. This means relative risks were calculated per increase or 50-100 g/day. Such an analysis suggests that results from prospective studies show a straight line, and that each additional, identical unit of consumption would show an identical effect.
But results from prospective studies rarely show a straight line. For example, in 2009 Sinha R. presented results from a cohort of very large size in which the effect of processed meat consumption on total CVD mortality were examined (Sinha R). The tables below show the difference between the actual effects of categorized analysis as presented by the authors (Sinha R) vs the effects as they are suggested to be from continuous analysis (Micha R). RR's for both analysis were similar among women, but differed among men. Among men, the actual effect sizes were weaker than would be suggested from continuous analysis, and where a small increase in risk (RR = 1.02) would be expected for someone increasing his daily intake from the 1st to 2nd quintile of consumption (4 to 11 g), the actual results showed a small, but significant protective effect for such an increase (RR = 0.92; 95% CI = 0.87-0.98).
In short, continuous analysis seldom reflects true effect sizes for particular levels of consumption, and eliminates the possibility of examining differend effects at intermediate levels of consumption.

Red meats vs processed meats: Micha R. presented results for red meats (unprocessed red meat), processed meat (primarily processed red meats), and total meats (the total of these 2 categories). She concluded that red meats consumption was not associated with CHD, whereas processed meats and total meats increased CHD risk. However, the analysis for red meats included results from 3 cohorts of very small-moderate size, including only 598 cases from prospective data (769 cases after including case-control data). Data about within-cohort comparisons for effects of red meats vs processed meats was available from 2 of these prospective studies.
Even though effects from another cohort were not clearly stratified into red meats vs processed meats (according to the criteria used by Micha R.), within-cohort comparisons showed that RR's for red meats were significantly higher among men, and nonsignificantly higher among women, compared to effects for processed meats (see following table). Since this cohort was of very large size, and included a total of 19,577 cases, it seems preliminary to use the effects from Micha R. as evidence for a difference in effect between red meats vs processed meats.

My systematic review: Results are presented for highest vs lowest units of consumption, taking into account significant effects at intermediate levels of consumption. Data about vegetarian cohorts is included, and if different analysis were performed in these cohorts, the RR was chosen from the analysis most likely to reflect meat consumption - instead of vegetarianism - as the only variable.
Total meats includes the combined effect of any red meat (lamb-, beef-, and/or pork) with any white meat (poultry and/or fish).

|References:
Mente A. A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med. 2009 Apr 13;169(7):659-69. Full text.
Micha R. Red and processed meat consumption and risk of incident coronary heart disease, stroke, and diabetes mellitus: a systematic review and meta-analysis. Circulation. 2010 Jun 1;121(21):2271-83. Abstract.
Sinha R. Meat intake and mortality: a prospective study of over half a million people. Arch Intern Med. 2009 Mar 23;169(6):562-71. Full text. |

Total meats and total cardiovascular disease.

CVD risk: Data was provided by 2 cohorts, including 3,289 cases. A significant protective effect was found in one cohort of very small size (Jamrozik K [13]). No other association was found. The average RR = 0.97.
Effect modification: No data was found.
CVD mortality: Data was provided by 5 cohorts, including 2,380 cases. A significant protective effect was found in one cohort of very small size (Jamrozik K [13]), but a nonsignificantly increased risk was found in another cohort of very small size (Chang-Claude J [4]). No other associations were found. The average RR = 1.11 for meat consumption.
Effect modification: No effect modification was found by gender (Key TJ [16]).

Conclusion: Inconsistent findings were done relating meat consumption to total CVD mortality. Inconclusive evidence was found for an association between total meats consumption and total CVD mortality. No associations were found with total CVD risk.

Total meats and heart disease.

Background:

• Dr. Hirayama examined the effects of a small amount of food groups in relation to a large amount of mortality end points in a Japanese cohort of very large size. An extended review of his work was published as a book in 1990 (Hirayama T [5]). Data about this cohort is seldom included in current systematic reviews about the relations mentioned. Dr. Hirayama published a lot of articles stating that vegetables and meats were related to several disease end points, adjusted for age and sex. However, the book included one page showing effects after multivariate analysis including cigarette smoking, meat, green-yellow vegetables, and alcohol. And this analysis showed that a large amount of previously published effects completely changed when these variables were taken into account.
  Since a) Dr. Hirayama himself only published sex, and age-adjusted results in the English language, while results following multivariate analysis were completely different b) the results were published as a book and not in a peer-reviewed journal, and c) Dr. Hirayama was the only researcher examining this cohort, results from his cohort are debatable. Results will be presented including effects from his work, but his work will not be included in the evidence for a possible effect.
• In 1999, Key TJ published a collaborative analysis of combined data from 4 prospective studies, including primarily vegetarian subjects, and including 2,264 IHD deaths. However, when vegans, fish eaters and occasional meat eaters were filtered out of the results, the final analysis included 1,433 deaths for the association between meat eaters vs vegetarians. Findings about 3 of these cohorts were published following this publication, including a longer follow-up period. Cohort-specific results were chosen to be included in the systematic review, instead of results from the collaborative analysis, because this allowed for a larger amount of cases to be included in the analysis.
• In 2002, Appleby PN published results about "The Oxford Vegetarian Study". Results were published for a) vegetarians vs nonvegetarians, and b) meat eaters vs non-meat eaters. Though it seems obvious to include the results from the meat eaters vs non-meat eaters, the non-meat eating group in this analysis included both low consumers of meats, and fish eaters. Therefore, results for vegetarians vs nonvegetarians were included.

CHD risk: Data was provided about 3 cohorts, including 1,303 cases. In one cohort, a significantly increased risk was found among women, but not men (Buckland G [16]); while in another cohort, a significantly increased risk was found among men, but not women (Fraser GE [1]). No other associations were found. The average RR = 1.25.
Effect modification: No data was found.
CHD mortality: Data was provided about 8 cohorts, including 2,928 cases. Significantly increased risks were found in 2 cohorts, including 1,659 cases (57% of all cases). No other associations were found, but all RR's for meat consumption were > 1. The average RR for meat consumption = 1.47 (excluding incomplete data from Knekt P [7], Kinjo Y [5], and Tanaka H [3]).
Effect modification: No effect modification was found by gender (Key TJ [11 + 16]), or a history of CVD/diabetes (Key TJ [11]). An analysis of 4 cohorts showed that vegetarianism was protective among subjects being a vegetarian for > 5 years, and who died before age 80 only (Key TJ [11]).

Conclusion: Inconsistent findings were done among both men and women, relating meat consumption to CHD risk. Inconclusive evidence was found for an association between total meat consumption and CHD risk.
Significantly increased risks of meat consumption were found in 2 cohorts, one of which was of very small size. Suggestive evidence was found that total meat consumption increases CHD mortality (+ 47%). This effect may be restricted to subjects being a vegetarian for > 5 years, and who died before age 80.
Limitations: Most significant effects were found among cohorts consisting of primarily vegetarian subjects, and few results were adjusted for other dietary variables. But RR's for meat consumption were consistently > 1.

Total meats and stroke.

Stroke risk: No data was found.
Stroke mortality: Data was provided by 7 cohorts, including 12,322 cases. A significant protective effect was found in one cohort of very large size (Kinjo Y [5]). No other associations were found. The average RR = 0.80 for meat consumption.
Effect modification: No significant effect modification was found by gender (Kinjo Y [5], Key TJ [11 + 16]), smoking status (Kinjo Y [5]), or age (Key TJ [11]).

Conclusion: A strong significant protective effect was found in one cohort of very large size, but no other associations were found. Inconclusive evidence was found for an association between total meats consumption and total stroke mortality. No data was found about stroke risk.

Red meat and total cardiovascular disease.

CVD risk: Data was provided by one cohort. No association was found.
Effect modification: No data was found.
CVD mortality: Data was provided by one cohort of very large size. A significantly increased risk was found among both men and women. And risk tended to increase with every increasing quintile of consumption. The average RR = 1.33.
Effect modification: Risk was increased among both never, and ever smokers. But risks were higher among never smokers for both sexes (see extended table).

Conclusion: Effects were restricted to findings from one cohort. Inconclusive evidence was found for an association between red meat consumption and total CVD mortality. No evidence was found for an association with total CVD risk.

Red meat and heart disease.

Background: In 2007, Qi L published data about a diabetic subgroup from "The Nurses' Health Study" relating red meat consumption to CHD. For the association with CHD risk, data about nonfatal MI - instead of total CHD - was included in the following analysis. Data about fatal CHD from this cohort was included in the analysis of CHD mortality, and therefore, data about total CHD would have led to a use of the CHD deaths in both following analysis'.
CHD risk: Data was provided by one cohort, including 1,610 cases. No association was found.
Effect modification: No data was found.
CHD mortality: Data was provided by 4 cohorts, including 1,119 cases. Significantly increased risks were found in 2 cohorts (Qi L [6]), Kelemen LE [17]). No associations were found in both remaining cohorts. The average RR = 1.45 (excluding incomplete data from Albert CM [10]).
Effect modification: Significantly increased risks were found in 2 cohort, including women only. No association was found in one cohort, including men only (Albert CM [10]). The average RR = 1.54 for women.

Conclusion: Significantly increased risks were found in 2 cohorts, one of which was of small size. These cohorts included women only. Suggestive evidence was found that red meat increases CHD mortality among women (+ 54%). No associations were found with CHD risk.

Red meat and stroke.

Stroke risk: No data was found about total stroke risk. Inconsistent findings were done with ischemic stroke (He K [6], Fung TT [6]), and haemorrhagic stroke (Iso H [6], He K [6]) in 2 cohorts (see extended table).
Effect modification: No data was found.
Stroke mortality: Data was provided by one cohort, including 1,224 cases (Sauvaget C [15]). No association was found.
Effect modification: No data was found.

Conclusion: Little data was found. Inconsistent findings were done relating red meat to risk of subtypes of stroke. No evidence was found for an association between red meat consumption and stroke risk or stroke mortality.

Processed meat and total cardiovascular disease.

CVD risk: No data was found.
CVD mortality: Data was provided by one cohort of very large size. Significantly increased risks were found among both men and women. Among men, a significant protective effect was found at the 2nd quintile of consumption, and the strong significance of the trend is driven by this heterogeneity in effects (see extended table). The average RR = 1.17.
Effect modification: Effects were similar among never vs ever smokers.

Conclusion: Effects were restricted to findings from one cohort. Inconclusive evidence was found for an association between processed meat consumption and total CVD mortality.

Processed meat and heart disease.

CHD risk: Data was provided by 2 cohorts, including 1,481 cases. Significantly increased risks were found in both cohorts. The average RR = 1.51.
Effect modification: No data was found.
CHD mortality: Data was provided by 2 cohorts, including 970 cases. A significant protective effect was found among men in one cohort (Iso H [19]). No other associations were found. The average RR = 0.75.
Effect modification: No data was found.

Conclusion: Significantly increased CHD risks of processed meat were found in 2 cohorts, one of which was of very small size. Suggestive evidence was found that processed meat consumption increases CHD risk (+ 51%). Associations with CHD mortality were not in line with the previous findings. No evidence was found for an association between processed meat consumption and CHD mortality.

Processed meat and stroke.

Stroke risk: No data was found.
Stroke mortality: Data was provided by 2 Asian cohorts, including 2,189 cases. No associations were found.
Effect modification: No data was found.

Conclusion: Little data was found. No evidence was found for an association between processed meat consumption and stroke risk or stroke mortality.