Epidemiological evidence is suggestive, but inconclusive, for an association between consumption of red and processed meat and risk of stroke. We aimed to assess this association by conducting a meta-analysis of prospective cohort studies.
We performed a literature search on PubMed database through June 2012 to indentify prospective cohort studies of red and processed meat intake in relation to risk of stroke. Reference lists of the retrieved articles were also reviewed. Both fixed-effects and random-effects model were assumed to compute the summary risk estimates.
Five large independent prospective cohort studies were identified. These studies contained a total of 2 39 251 subjects and 9593 stroke events. Comparing the highest category of consumption with lowest category, the pooled relative risks (RRs) of total stroke were 1.15 (95% confidence interval (CI), 1.05–1.25) for total meat (red and processed meat combined) (n=4), 1.09 (95% CI, 1.01–1.18) for red meat (n=5) and 1.14 (95% CI, 1.05–1.25) for processed meat (n=5); the corresponding RRs of ischemic stroke (highest vs lowest quintile) were 1.15 (95% CI, 1.04–1.28), 1.13(95% CI, 1.01–1.25) and 1.19 (95% CI, 1.08–1.31). Consumption of red and/or processed meat was not associated with hemorrhagic stroke. In the dose–response analysis, the risk of stroke increased significantly by 10% and 13% for each 100 g per day increment in total and red meat consumption, respectively, and by 11% for each 50 g per day increment in processed meat consumption.
Findings from this meta-analysis indicate that consumption of red and/or processed meat increase risk of stroke, in particular, ischemic stroke.
In most developed countries, stroke remains an important cause of mortality and the most common cause of disability, and is becoming a great health burden in future decades.1 For instance, in the United States, the direct and indirect cost of stroke for 2010 is $73.7 billion,2 whereas, across the European countries, the total annual cost of stroke was estimated to be €27 billion.3
Meat is a major source of protein and fat for humans. Consumption of meat, in particular processed meat, has been suggested to be associated with an increased risk of some diseases, including cancers,4, 5 type 2 diabetes6 and cardiovascular disease.4, 6 But the association between red and processed meat consumption and risk of stroke remains unknown. Although a previous meta-analysis found a null association, the pooled result could be limited because only three studies containing less than 2300 cases of stroke were included, and no two studies evaluated the same meat and stroke subtypes.6 In fact, several subsequent prospective cohort studies have reported a positive association between red and processed meat intake and risk of stroke.7, 8, 9 We therefore chose to conduct a meta-analysis of prospective studies with the following objectives: to update the evidence concerning the relationship between red and processed meat consumption and risk of stroke; to investigate the effects of red and processed meat on stroke subtypes; and to further evaluate whether there is a dose–response relationship. We attempted to plan, conduct and report this meta-analysis in adherence to the guidelines of the ‘Meta-analysis of Observational Studies in Epidemiology group’.10
Materials and methods
We performed a literature search through June 2012 on PubMed database (www.ncbi. nlm.nih.gov/pubmed) without language restrictions. We combined text terms and, where appropriate, MeSH (Medical Subject Headings) terms for meat (meat, meat products, beef, pork, veal, mutton, lamb, ham, other specific unprocessed red and processed meat subtypes) and stroke (cerebrovascular disorders, cerebrovascular disease, stroke, intracranial hemorrhage and brain hemorrhage).
In addition, the reference lists of the retrieved articles were comprehensively reviewed to identify additional studies. We also attempted to contact authors for additional information of the retrieved articles, if necessary.
Inclusion and exclusion criteria
Studies were included if they met the following criteria: (i) prospective design; (ii) the exposure of interest was consumption of red meat (unprocessed) and/or processed meat; (iii) the outcome of interest was stroke; and (iv) the relative risk (RR) with corresponding 95% confidence interval (CI) were provided.
In the present study, ‘red meat’ was defined as unprocessed meat from beef, veal, pork, mutton and lamb, and excluding poultry, fish or eggs; ‘processed meat’ was defined as any meat preserved by smoking, curing or salting or addition of chemical preservatives, such as bacon, salami, sausages, hot dogs or processed deli or luncheon meats, and excluding fish or eggs; and ‘total meat’ was defined as food item that includes both ‘red meat’ and ‘processed meat’ into a single item in the studies identified in the search. We only included the studies that described clear definitions of red and/or processed meat consumption, and excluded those in which the ‘total meat’ item contained unprocessed poultry. If the same population was used in more than one study, we included the study with the longest follow-up duration.
The following data were extracted from each included eligible study: the first author’s last name, publication year, the name of the cohort study, years of follow-up, the sex and age of participants, number of cases and participants, red and processed meat consumption categories, the RRs or hazard ratios of stroke and corresponding 95% CIs for each category of red and processed meat consumption, and variables adjusted for in the analysis. We extracted the maximally adjusted RR with corresponding 95% CI for the highest vs the lowest category of meat intake for use in the main analyses. Study selection and data extraction were conducted independently by two authors (G-CC and Q-FL), with any disagreements resolved by consensus.
The summary risk estimates were calculated by using both fixed- and random-effects model.11 The two models yielded very similar estimated. We therefore present the results based on the random-effects model, which considers both within- and between-study variation.
For the dose–response analysis, we used the method proposed by Greenland and Longnecker12 to estimate study-specific slopes from the natural logarithm of the RR across categories of exposure. For each study, the median or mean level of meat intake for each category was assigned to each corresponding RR estimate. When the median or mean intake per category was not provided, we assigned the midpoint of the upper and lower boundaries in each category as average intake. If the highest or lowest category was open-ended, we assumed the width of the interval to be the same as in the closest category. If the person-years by levels of intake were not provided in the primary studies, we approximated them from years of follow-up and number of subjects. We used 100 g as the approximate average serving size for total and red meat, and 50 g for processed meat.
Heterogeneity test was performed by use of Q and I2 statistics.13 For the Q statistic, a P-value <0.1 was considered statistically significant heterogeneity. Publication bias was investigated by use of both Begg rank correlation test and Egger linear regression test.14, 15 All statistical analyses were done using STATA software, version 11.0 (StataCorp, College Station, TX, USA). All P-values are two-sided, and P<0.05 was considered statistically significant, unless explicitly stated.
A total of 576 citations were identified through the primary search. After screening the titles and the abstracts, nine articles appeared to be relevant to this meta-analysis and were selected for full review. Of these full texts, two16, 17 were excluded for lacking of a clear description of types of meat; one18 study was excluded because unprocessed poultry was included in the total meat item; we further excluded two19, 20 texts in which the same population was used in the latter two studies (published results in one7 article). At last, four7, 8, 9, 21 articles (one7 article contained two large independent cohort studies) were included in the meta-analysis.
The five large cohort studies were carried out between 2003 and 2012, and contained a total of 2 39 251 subjects and 9593 stroke events. Characteristics of the selected studies are shown in Table 1. Of the five studies, two were conducted in the United States, two were carried out in Sweden and one in Japan. One study included both men and women, two studies included men only and two consisted of women only. The mean follow-up duration ranged from 10.1 to 26 years. The number of participants included in the primary studies ranged from 34 670 to 84 010, and the number of cases ranged from 1379 to 2663. Four studies separately presented results on total, red and processed meat in relation to risk of stroke for the highest vs the lowest quintiles of consumption, and one study provided results on both red and processed meat for the highest vs the lowest quartiles of consumption. Age, body mass index, smoking, alcohol intake and history of diabetes and hypertension were adjusted for all included studies, and four studies additionally adjusted for physical activity, aspirin use, family history of myocardial infarction and intake of total energy, fish, fruits and vegetables.
High vs low analysis
Figure 1 shows the multivariable-adjusted RRs for individual studies and all studies combined for the highest vs the lowest categories of total (red and processed meat combined), red and processed meat consumption. Overall, the pooled analyses showed that individuals in the highest categories of total, red or processed meat consumption had a statistically significant increased risk of total stroke, compared with those in the lowest categories. The summary RRs were 1.15 (95% CI, 1.05–1.25) for total meat (red and processed meat combined) (n=4), 1.09 (95% CI, 1.01–1.18) for red meat (n=5) and 1.14 (95% CI, 1.05–1.25) for processed meat (n=5). No significant heterogeneity was found (Figure 1).
Four7, 8, 9 studies presented results for stroke subtypes. All RRs in these studies were estimated based on the comparison of the highest with the lowest quintile of red and processed meat consumption. For these studies, the pooled RRs of total meat consumption for ischemic stroke and hemorrhagic stroke were 1.15 (95% CI, 1.04–1.28) and 1.16 (95% CI, 0.81–1.66), respectively. The pooled RRs of red meat consumption for ischemic stroke and hemorrhagic stroke were 1.13 (95% CI, 1.01–1.25) and 0.99 (95% CI, 0.77–1.28), respectively, and the corresponding RRs of processed meat consumption for ischemic stroke and hemorrhagic stroke were 1.19 (95% CI, 1.08–1.31) and 1.23 (95% CI, 0.96–1.58), respectively. Little evidence of heterogeneity was observed.
The dose–response analysis of the primary studies showed that the risk of stroke increased significantly by 10% for each 100 g per day increment in total meat consumption (n=4; RR=1.10; 95% CI, 1.05–1.15), by 13% for each 100 g per day increment in red meat consumption (n=5; RR=1.13; 95% CI, 1.03–1.23) and by 11% for each 50 g per day increment in processed meat consumption (n=5; RR=1.11; 95% CI, 1.02–1.20), with low study heterogeneity.
Neither Begg rank correlation test nor Egger linear regression test suggested significant publication bias, with regard to consumption of total, red or processed meat in relation to total stroke risk (all P-values >0.3).
Findings from this meta-analysis of five large prospective cohort studies involving more than 9500 stroke cases indicate a statistically significant positive association between red and/or processed meat consumption and risk of stroke. In a comparison of the highest with the lowest category of meat intake, the risk of stroke increased significantly by 15% for red and processed meat intake, by 9% for red meat intake and by 14% for processed meat intake. With respect to stroke subtypes, the association was significant for ischemic stroke, but not for hemorrhagic stroke. In addition, the dose–response analysis also showed consistent associations between consumption of red and/or processed meat and increased risk of stroke.
Heterogeneity is often a concern in a meta-analysis. However, little indication of heterogeneity was found across our study, which may partially be explained by our strict inclusion criteria. We excluded the studies in which the types of meat has not been clearly described, or those who included poultry in the total meat item. Because consumption of poultry has been reported to reduce the risk of stroke,7 an inclusion of those studies could probably bias any true relationship between red and processed meat consumption in relation to risk of stroke. In addition, the RRs in all but one study were estimated based on the comparison of the highest with the lowest quintile of red and processed meat consumption.
Several potential mechanisms may explain the adverse effect of red meat consumption on risk of stroke. Red meat is a major source of heme iron. Intracellular iron chelator desferrioxamine has been experimentally suggested to inhibit inflammation and atherosclerosis in mice, indicating a potential role of iron in the progress of atherogenesis;22 heme iron has also been epidemiologically suggested to increase risks of atherosclerosis,23 type 2 diabetes24 and coronary heart disease,25 all of which may contribute to increase the risk of stroke. Furthermore, consumption of red meat was found to be positively associated with risk of blood pressure,26 and to increase the incidence of hypertension,27 the strongest risk factor for stroke.28
In this study, the adverse effect of processed meat consumption on risk of stroke was stronger than that of red meat. This additional harm of processed meat may be explained, at least in part, by other constituents in processed meat, such as sodium. Recent studies support a deleterious effect of high sodium diet on vascular structure.29 High sodium intake was also found to be associated with significantly increased risk of stroke and total cardiovascular disease;30 conversely, reduced sodium intake was suggested to significantly lower blood pressure in hypertensive individuals.31
The present study had several strengths. All studies included in this meta-analysis were of a prospective design, which eliminates the possibility of recall and selection biases. We have acknowledged that most of the primary studies suggested a nonsignificant positive association between red and/or processed meat intake and stroke. As the single prospective studies were mainly of limited power to prove statistical significance, this meta-analysis involving large number of stroke cases (>9500) enhanced the statistical power to assess the long-term effects of meat consumption on stroke risk.
Several limitations should also be acknowledged when interpreting the results from this meta-analysis. First, as a meta-analysis of epidemiological studies, it is not able to solve the problem of confounding that is universal in the included studies. We cannot entirely exclude the possibility of some confounders as a potential explanation for the observed findings. For instance, participants with a higher meat intake tended to be more likely to be smokers, have diabetes and hypertension, and have higher body mass index and higher intakes of alcohol.7, 9 However, all included studies have adjusted for these major potential confounders, including diabetes and hypertension, which may be intermediates of the association between meat consumption and risk of stroke. A Second limitation is that all studies assessed diet with a food-frequency questionnaire, and three studies of the included studies assessed meat intake only at baseline. Hence, some misclassification of exposure that could lead to an underestimation of the risk estimate is inevitable. Third, the number of hemorrhagic strokes included in this meta-analysis was relatively small, and thus, may limit our ability to detect a modest association between meat intake and hemorrhagic strokes. Finally, given that our meta-analysis was based on published studies, publication bias, which results from a tendency to publish only positive results, also merits consideration. In this meta-analysis, however, we detected little indication of such bias.
Carrying out controlled trials is, generally, an optimal approach to assess the effect of nutritional intervention on disease outcomes. However, such a trial concerning the adverse effect of red and processed meat is hardly feasible. Alternatively, as a part of a systematic evaluation of the role of meat, our study summarized the existing evidence from epidemiological studies in North America, Europe and Japan, and suggested an adverse effect of greater red and processed meat consumption on stroke, in particular ischemic stroke. Therefore, from a view of public health point, reducing red and processed meat consumption, and replacing it with other healthy dietary components, such as fruits, green leafy vegetables and fish, may bring appreciable benefits in stroke prevention, which would also reduce other cardiovascular diseases,32 type 2 diabetes33 and some cancers.34
This study was funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).