Abstract

INTRODUCTION: Previous research suggested that vegetarianism may decrease heart disease mortality (Key TJ. 1999), but vegetarianism may also increase risk of death from other causes, which may lead to a lack of an association with mortality from all causes (Key TJ. 2003).
OBJECTIVES:

• To review all prospective studies which published information about total meat, and meat products in relationship with all-cause mortality/survival.
• To define the amount of consumption found to be related with possible effects on death.
• 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: 35 articles were found which provided information about 18 different cohorts. Of these, 0 articles were excluded.
Results are described when any evidence for an association - as defined in the Methods - was found. In addition, data about total meats, red meat, and processed meat is described.

• Total meats: Data was provided by 13 cohorts, including 17,142 deaths. Inconclusive evidence was found for an association (RR = 1.12). And no evidence was found for an association with meat fats.
• Red meat: Data was provided by 3 cohorts, including 75,698 deaths. Red meat consumption possibly increases all-cause mortality (RR = 1.31).
• Processed meat: Data was provided by 3 cohorts, including 85,634 deaths. Inconclusive evidence was found for an association (RR = 1.15).

CONCLUSION: No evidence was found for an association between total meats consumption and mortality. This evidence did not change when meat consumption and vegetarianism were considered 2 different variables. Red meat consumption possibly increases mortality, but this association could not be linked to any specific type of red meats. No evidence was found for an association between poultry, specific meat items, or meat fats, and mortality.
LIMITATIONS: Hardly any data is available about meat at different levels of consumption. And hardly any data is available about specific meat items, or effect modification by possible confounders.

|References:
Key TJ. Mortality in vegetarians and nonvegetarians: detailed findings from a collaborative analysis of 5 prospective studies. Am J Clin Nutr. 1999 Sep;70(3 Suppl):516S-524S. Full text
Key TJ. Mortality in British vegetarians: review and preliminary results from EPIC-Oxford. Am J Clin Nutr. 2003 Sep;78(3 Suppl):533S-538S. Full text |

Total meats and all-cause mortality.

Background/use of data:

• 7 publications were found about The EPIC Study (16). These included data about the same subpopulation with different length of follow-up (Trichopoulou A. 2003 vs Trichopoulou A. 2009, and Key TJ. 2003 vs Key TJ. 2009).
  The EPIC-Elderly study provided data about 74,607 elderly (aged ≥ 60) subjects - with or whithout diabetes - from 9 countries (including Greece & UK), and including 4,047 deaths (Trichopoulou A. 2005). But 3 other publications provided information about subgroups of this cohort from which the data partly overlapped with data from The EPIC-Elderly Study:
  a) Trichopoulou A (2006). 1,013 subjects from EPIC-Greece, and with diabetes, including 80 deaths.
  b) Trichopoulou A (2009). 23,349 subjects aged 20-86 from EPIC-Greece, and without diabetes, including 1,075 deaths.
  c) Key TJ (2009). 47,254 subjects aged 20-89 from EPIC-Oxford (UK), including 1,513 deaths.
  Data form The EPIC-Elderly Study was included in the systematic review, excluding data from the previous 3 publications, because the Greek and UK subcohorts included less deaths than the EPIC-Elderly Study (2,668 vs 4,047).
• 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 cohort studies, including primarily vegetarian subjects, and including 8,330 deaths. 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 deaths 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.
• In 1984, Kahn HA published results about "The Adventist Health Study". 3 different OR's were provided, following 3 different series of adjustments. The OR after adjustment for cigarettes was included in this systematic review, because cigarette smoking has been shown to influence mortality rates more convincingly than consumption of eggs/fish or eggs/salad.

Results: 37 articles providing information about 14 different cohorts were found, including a total of 72,642 cases. Survival was the end point in one article (Nube M [2]), and mortality was the end point in all remaining articles.
A significantly increased risk was found in one cohort (Kahn HA [1]). But and a significant protective effect was found in one other cohort (Jamrozik K [12]), and women had higher survival rates in another cohort (Nube M [2]). No other (non)significant associations were found. The average RR = 1.05 for total meats consumption, including data from all 14 cohorts. When one cohort with debatable results was excluded from the analysis (Hirayama T [4]), this left 17,142 cases, and the average RR became 1.12.
Effect modification: The significantly increased risk found by Kahn HA (1) was restricted to men. Another cohort showed a protective effect among women only (Nube M [2]). No other effect modification was found by gender (Key TJ [9], Key TJ [16]).
Meat consumption significantly increased mortality among elderly subjects without chronic diseases, but this association was found in a cohort of very small size (Fortes C [14]).
Meat fat: Data about meat fat was provided by 2 cohorts (Kahn HA [1], Jamrozik K [12]). No associations were found with mortality.

Conclusion: Few, and inconsistent effects were found. No evidence was found for an association between total meats consumption, or meat fats, and all-cause mortality.
Limitations: Few results were published at different levels of consumption (Kahn HA [1], Nube M [2], Mann JI [6], Fortes C [14]). Mostly, results were published for 2 units of consumption (e.g., vegetarians vs nonvegetarians), or as continuous data instead of categorized data. This makes it impossible to link any possible effect to a given level of consumption. Also, hardly any data is available about effect modification.

Total meats and all-cause mortality. Meat consumption vs vegetarianism.

Background: The previous analysis included data about both meat consumption and vegetarianism. But vegetarianism can modify a possibly effect of meat consumption by other differences in diet and lifestyle behaviors in vegetarians. Therefore, a second analysis was made stratifying the previous anlysis into data about total meats consumption vs vegetarianism.
Meat consumption: Data was provided about 12 cohorts, including 69,165 cases. Results were identical to the previous results including data about vegetarianism. When one cohort with debatable results was excluded from the analysis (Hirayama T [4]), this left 13,642 cases, and the average RR became 1.15.
Vegetarianism: In 1999, Key TJ (9) published an analysis of 4 different cohort studies, including 8,330 deaths. One of these cohorts (The Adventist Health Study [1]) was followed-up over 2 different time periods (see extended table). This analysis showed no effect of vegetarianism vs nonvegetarianism, though the DRR was slightly below 1 (DRR = 0.95; 95% CI = 0.82-1.11). Findings about 3 of these cohorts (Chang-Claude J [3], Appleby PN [6]) were published following this publication, including a longer follow-up period. RR's for vegetarians were > 1 in all 3 cohorts. The average RR = 1.04 for vegetarianism, when data about the EPIC-Oxford Study is included (Key TJ [16]).
Veganism: In 1999, Key TJ (9) published an analysis of 4 different cohort studies (see extended table). Occasionial meat eaters, fish eaters, and vegetarians all had lower mortality rates compared with regular meat eaters (≥ 1 time/wk), but mortality rates for vegans and regular meat eaters were identical.

Conclusion: Few effects were found. No evidence was found for an association between total meats consumption - or vegetarianism - and all-cause mortality.
Perspective: Key TJ (9), found a DRR of 0.95 for vegetarianism analysing results from 4 cohorts, and the RR for meat consumption was > 1 in my systematic review (RR = 1.15). Both effect sizes were driven by results from one cohort. Without results from "The Adventist Health Study", RR's would become 1.04, and 1.00 for vegetarianism, and meat consumption, respectively.

Red meat and all-cause mortality.

Data was provided by 3 cohorts, including 75.698 deaths. Significantly increased risks were found in 2 cohorts of moderate-very large size (Kelemen LE [17], Sinha R [20]), and including 75,230 cases (99.4% of all cases). But a significant protective effect was found in the remaining cohort, which was of small size (Whiteman D [10]). The average RR = 1.31.
Effect modification: No effect modification was found by smoking status (Sinha R [20]).

Conclusion: Findings were inconsistent, but the cohorts in which significantly increased risks were found included almost all cases. Red meat consumption possibly increases all-cause mortality (+ 31%). No clear level of consumption could be defined for this effect, but in the largest cohort, risk significantly increased at 21.4 g/1000 kcal/day, and gradually increased further at higher levels of consumption.

Processed meat and all-cause mortality.

Data was provided by 3 cohorts, including 85,634 cases. A significantly increased risk was found in a cohort of very large size among both men and women (Sinha R [20]), but a significant protective effect was found in a cohort of large size among men (Iso H [18]). No other associations were found of high vs low consumption. The average RR = 1.15.
Effect modification: No effect modification was found by smoking status (Sinha R [20]).

Conclusion: Inconsistent findings were done. Inconclusive evidence was found for an association between processed meat consumption and all-cause mortality.