PROJECT THESIS
Vegetarian and vegan diets and the risk of cardiovascular disease, ischemic heart disease and stroke: a systematic review and meta-analysis of prospective cohort studies
**************************** front page Project Thesis *************************
Abstract
Background: Plant-based dietary patterns have been recommended for cardiovascular disease (CVD) prevention due to their beneficial effect on cardiovascular risk factors.
Vegetarian diets have been shown to improve cardiovascular risk factors and have
consistently shown to reduce the risk of ischemic heart disease (IHD), however, uncertainties remain regarding the impact of vegetarian diets on risk of cardiovascular disease (CVD) and stroke.
Objective: To clarify the association between vegetarian or vegan diets and cardiovascular disease risk we conducted a systematic review and meta-analysis of prospective cohort studies on incidence or mortality from CVD, IHD and stroke among vegetarians versus nonvegetarians.
Methods: PubMed and Embase databases were searched up to 14th of February 2020. We included prospective cohort studies with adjusted relative risk (RR) estimates and 95%
confidence intervals (CIs) for incidence or mortality from CVD, IHD and stroke (overall and subtypes), comparing vegetarians and vegans to non-vegetarians. Summary RRs (95% CIs) were calculated using a random effects model.
Results: Ten prospective cohort studies with a total of 621 282 participants and 115 392 CVD cases, 30 377 IHD cases and 8 178 stroke cases were included in the meta-analysis. The summary RR was 0.85 (95% CI: 0.79–0.92, I2=68.0%, n=8) for CVD, 0.79 (95% CI: 0.71–
0.89, I2=67.3%, n=8) for IHD and RR 0.88 (95% CI: 0.73–1.07, I2=70.8%, n=9) for stroke when comparing vegetarians vs nonvegetarians. There was no indication of publication bias with Egger's test, Begg's test or by inspection of the funnel plots. The inverse association between vegetarian diets and CVD and IHD risk persisted across multiple subgroup and sensitivity analyses but was slightly stronger in the US studies than among European studies, while for stroke a significant inverse association was observed in Asian, but not in US and European studies.
Conclusion: This meta-analysis suggests vegetarian diets may reduce the risk of
cardiovascular disease overall as well as ischemic heart disease. Further studies are needed to clarify the association between vegetarian diets and stroke risk.
Introduction
Cardiovascular disease (CVD), mainly due to ischemic heart disease (IHD) and stroke is the leading cause of leading cause of death and disability globally(1). In 2017, the global age- standardized death rate from IHD was 117 per deaths per 100 000 while the age-standardized incidence was 136 new cases per 100 000, according to the Global Burden of disease
Study(2). For stroke, the 2017 global standardized death rate was 80 deaths per 100 000 while the age-standardized incidence rate was 150 per 100 000 In 2019, an estimated 9.14 million IHD deaths and 6.55 million deaths due to stroke occurred according to the Global Burden of Disease Study (3). To handle this burden there is a continued need for primary and secondary prevention of CVD. Reducing the prevalence of known cardiovascular risk factors is essential in this work and diet and other lifestyle factors like smoking cessation and
physical activity are important factors (4). Plant-based diets have shown to reduce known cardiovascular risk factors (5, 6). In fact, vegetarians who predominantly exclude meat, fish and poultry from their diet, exhibit a more favorable cardiovascular risk profile compared to people consuming meat (7-9). This includes lower prevalence of hypertension (9-11), high serum cholesterol (12), and type 2 diabetes mellitus (13). Studies further report less
overweight and obesity in vegetarians; a finding consistent with that fact that vegetarians indeed do have lower body mass index (BMI) (13, 14). Plant-based diets are recommended in several guidelines as a measure to prevent cardiovascular disease (4, 15, 16).
There are several subgroups of vegetarian dietary patterns which all have in common the exclusion of meat and meat products, fish and poultry. Lacto-ovo-vegetarians are the most common variant, allowing the consumption of milk products and eggs. Some vegetarians further exclude eggs (lacto-vegetarians) and others do the opposite by excluding milk
products (ovo-vegetarians). The strictest form of vegetarians are vegans, excluding all animal products, including honey and gelatin. Although not strictly considered vegetarian diets, pescetarians (fish eaters) and semi-vegetarians (restricting their meat intake to a minimum) are close relatives to the vegetarian diet (17).
Vegetarian diets have consistently shown to reduce the risk of IHD based on evidence from mainly prospective cohort studies. A pooled analysis of five cohort studies by Key et al found a 24% lower IHD mortality among vegetarians vs. non-vegetarians (18). Other studies that have been subsequently published also reported associations in the direction of reduced risk (19-21), although not all studies reported statistically significant associations (19, 21). In
contrast, results on risk of stroke or cerebrovascular disease (CBVD) have been less consistent with null results reported in a pooled analysis (22), and one separate study (19), but a positive association reported in the EPIX-Oxford study (20) and inverse associations reported in two Taiwanese studies (23). Studies on CVD overall have also shown mixed results, with some studies showing an inverse association (19, 20, 22), and others showing no significant association (24, 25). Therefore, to clarify these findings we conducted a
systematic review and meta-analysis of prospective cohort studies on vegetarian and vegan diets and risk of CVD, IHD and stroke risk.
METHODS Study design
We conducted a systematic review and meta-analysis on studies reporting risk estimates for vegetarian or vegan diets and risk of incidence and mortality from CVD, IHD and stroke, both overall and subtypes.
Search strategy
A search was conducted using PubMed and OvidEmbase databases from their inception in 1958 and 1947 respectively and up to 14th of February 2020. The search was later updated to the 3014 and three additional studies were identified by the updated search. We searched for prospective cohort studies reporting the association between a vegetarian or vegan diet and incidence or mortality from IHD or stroke. Terms like vegetarian(s) or vegan(s) or
vegetarian/vegan diet or plant-based diet were searched in combination with the following terms: cardiovascular disease, ischaemic heart disease, coronary heart disease, myocardial infarction, cerebrovascular disease, stroke, cerebral infarction, cerebral haemorrhage, and subarachnoid haemorrhage. We also applied a term for ‘Seventh-day Adventist’ as many of the relevant cohorts have been done in this religious group. We used relevant ‘subject headings’ for both PubMed and Embase databases but also added a free-text search to
retrieve any articles in press. A detailed search strategy is available for both the PubMed- and Embase databases in the supplementary text.
Inclusion criteria
Prospective cohort studies were included if they reported adjusted relative risk (RR)
estimates and 95% confidence intervals (95% CIs) for the association between vegetarian or
vegan vs. non-vegetarian diets and incidence and/or mortality from CVD, IHD or stroke. A full list of the excluded studies is shown in Supplementary Table 1.
Data extraction and study quality
Data were extracted to tables and included information regarding results and study characteristics. More specifically the information extracted was as follows: study name, author, year of publication, location by country or region, age and gender, years of follow-up, sample size and number of cases or deaths from CVD, IHD, total stroke, ischemic stroke and hemorrhagic stroke. We also extracted RRs with 95% CIs and information regarding
confounders adjusted for in each study´s statistical analysis. All studies were critically appraised according to the Newcastle Ottawa Scale (NOS) rewarding 0-9 stars to each study.
We assessed ‘selection’ (of the exposed and non-exposed group including whether or not outcome of interest was excluded from start), ‘comparability’ (points given if they had adjusted for age and one more factor) and ‘outcome’ (i.e., outcome assessed by record linkage, follow-up > 5 years, follow-up complete or nearly complete). Study quality were considered low (0-3 stars), medium (4-6 stars) or high (7-9 stars) (26).
Data extraction was performed by one author (JSD) and checked for accuracy by second author (DA). The PRISMA guidelines were followed throughout the process (27).
Statistical methods
We calculated summary RRs and 95% CIs comparing vegetarians and vegans to non- vegetarians in relation to the risk of CVD, IHD and stroke (overall, ischemic and
hemorrhagic) using the random effects model by DerSimonian (28). For studies reporting both on incidence and mortality from CVD, IHD or stroke, only risk estimates for incidence were used in the main analysis. When studies reported on both risk of IHD and stroke/CBVD separately, but not for CVD overall, we pooled results for the two outcomes using a fixed effects model, before inclusion in the main analysis of CVD. Since vast majority of CBVD deaths are due to stroke we included studies reporting on CBVD mortality together with studies on stroke (each outcome was analyzed separately in subgroup analyses).
The random-effects method allowed detection of heterogeneity within subgroups (Cochran’s Q-statistics, Chi-squared-statistics) (29) and quantification of such heterogeneity (I2-
statistics) (30). We considered the p-value for Chi-squared statistics significant if <0.05 and I2-percentage as low (0-33%), moderate (34-66%) or high (67-100%). Subgroup analyses
were conducted stratified by sex (men or women), years of follow-up (less or more than 10 years), exclusion of early follow-up years or not, outcome of interest (incidence or mortality), number of cases or deaths (< 250, 250-499 and ³ 500), geographic location (Europe, North America or Asia), study quality (low, moderate, high), and adjustment for confounding factors (age, education, smoking, alcohol, BMI, physical activity) to investigate potential sources of heterogeneity in the main analysis. For stroke, we also conducted separate subgroup analyses distinguishing between ischemic- and hemorrhagic stroke. An additional sensitivity analysis including only those studies reporting both BMI-adjusted and BMI- unadjusted risk estimates was conducted to assess more directly the impact of BMI on the results, since BMI could be an intermediary variable through which vegetarian/vegan diets can influence cardiovascular risk.
Egger´s test (31), Begg´s test (32), and inspection of the funnel plots (33) was used to investigate publication bias and we considered a p-value<0.10 or funnel plot asymmetry to indicate possible publication bias. To assess the robustness of the summary RR and to ensure that the summary RR was not driven by one very large study or a study with an extreme RR, sensitivity analyses were conducted re-calculating the meta-analysis summary RR when omitting one study at a time from the analysis. The statistical analyses were conducted using Stata version 15 (StataCorp, TX, USA).
RESULTS
Our search yielded a total of 2709 publications (details are visible in the flow-chart in figure 1). After screening and full-text assessment, 10 publications from 10 unique prospective cohort studies were included in the analysis (19-25, 34-36); there were 5 studies from Europe, 3 from North America, and 2 from Asia (Table 1). Two publications reported RRs from two studies in the same paper (23, 25). For 3 studies we used data from a pooled re- analysis of Key et al. 1999 (22) as the original publications did not contain data relevant for our main analysis (35-37). Two of these original publications (35, 36) did, on the other hand report risk estimates for men and women separately and we later included those estimates in a sex-stratified subgroup analysis on ischemic heart disease. As the original publication from the Heidelberg study (24) did not contain data on CBVD, we also used data from the pooled analysis of Key et al 1999 (22) for our stroke analysis. To summarize, this yielded 8 studies for the analysis on CVD (after an additional pooling of data from three studies reporting on
IHD and stroke separately) (19), 8 studies for our analysis on IHD (19-22, 24, 25), and 9 studies for our analysis on stroke (after including studies reporting RRs for CBVD) (19, 20, 22, 23, 25).
Study characteristics
Table 1 entails a summary of characteristics and RRs of the cohort studies included in our main analysis. For a more extensive list of all data included in both our main and subgroup analysis we refer to supplementary table 3. The included studies were as follows: Adventist Mortality study (22), Adventist Health Study 1 (22), Heidelberg study (24), Oxford
Vegetarian study (25), Health Food Shoppers study (25), Adventist Health study 2 (21), EPIC-Oxford study (20), Tzu Chi Health study (23), Tzu Chi Vegetarian study (23) and UK Biobank study (19). Participants were from North America (USA), Europe (United Kingdom and Germany) and Asia (Taiwan). All studies used dietary questionnaires to ascertain
vegetarian status. Only 4 studies used validated questionnaires (20, 21, 23, 37). For the rest of the studies, information on validation of questionnaires was not available. Participants were all regarded as health-conscious people. The exposed groups (vegetarians) were often recruited from i.e., Health Food Shops, vegetarian magazines, Tzu Chi volunteer societies (Buddhist monks), from Seventh-day Adventists communities, although participants from UK Biobank study(19) were from the general population. Vegetarian status was ascertained in slightly different ways: most studies used questions explicitly asking i.e., how often participants ate meat, fish or poultry (19, 21, 22, 24, 25) or a yes or no question like “Do you eat any meat?” (EPIC-Oxford, Tzu Chi Health study, Tzu Chi Vegetarian study) (20, 23).
One other study (Health Food Shoppers study) asked participants whether they considered themselves vegetarian or not without further elaboration on frequency of consumption of meat, fish and poultry etc. In the current analysis the main analysis of vegetarians included vegans in most studies, except for the UK Biobank study(19), but a separate analysis was conducted for vegans when possible. Age ranged from 10 to 90 years. Follow-up ranged from 5.14 to 21 years (for data used both in main and subgroup analyses). The number of incident cases or deaths ranged from 219 to 106690 for cardiovascular disease, 60 to 24794 for ischemic heart disease, 31 to 330 for cerebrovascular disease, 54 to 5946 for total stroke, 31 to 519 for ischemic stroke and 28 to 300 for hemorrhagic stroke. Sample size ranged from 1724 to 422791 across all outcomes. All studies used record linkage for outcome
ascertainment except the Heidelberg study (24) who retrieved copies of death certificates from the Registrar’s office.
Cardiovascular disease
Eight prospective cohort studies (6 publications) (19-22, 24, 25) with 621 282 participants and 115 392 CVD cases were included in the analysis of vegetarian diets and CVD. The summary RR for vegetarians compared to non-vegetarians was 0.85 (95% CI: 0.79–0.92, I2=68.0%, pheterogeneity=0.003 (figure 2, supplementary figure 1). There was no indication of publication bias with Egger's test (p=0.28), Begg's test (p=0.39), or by inspection of the funnel plot (Supplementary Figure 7). The summary RRs (95% CI) ranged from 0.84 (0.77- 0.91) when excluding the Health Food Shoppers study to 0.88 (0.83-0.94) when excluding the Adventist Health study 1 (Supplementary Figure 12).
Ischemic heart disease
Eight prospective cohort studies (6 publications) (19) including 621282 participants and 30 377 IHD cases were included in the analysis of vegetarian diets and ischemic heart disease.
The summary RR for vegetarians compared to non-vegetarians was 0.79 (95% CI: 0.71–0.89, I2 = 67.3%, pheterogeneity=0.003) (figure 3, supplementary figure 2). There was no indication of publication bias with Egger's test (p=0.61), Begg's test (p=0.90), or by inspection of the funnel plot (Supplementary Figure 8). The summary RR (95% CI) ranged from 0.76 (0.70- 0.83) when excluding UK Biobank study
to 0.83 (0.76-0.90) when excluding Adventist Health study 1 (Supplementary Figure 13).
2 studies (2 publications) (20, 21) including 121 496 participants and 3192 IHD cases were included in the analysis of vegan diets and IHD and the summary RR (95% CI) was 0.84 (0.68-1.04) (Supplementary Figure 6).
Stroke
9 prospective cohort studies (5 publications) (19, 20, 22, 23, 25) including 561 359
participants and 8 178 stroke cases were included in the analysis of vegetarian diets and total stroke. The summary RR for vegetarians compared to non-vegetarians was 0.88 (95% CI:
0.73–1.07, I2=70.8%, pheterogeneity=0.001) (figure 4, supplementary figure 3). There was no indication of publication bias with Egger's test (p=0.19), Begg's test (p=0. 0.47), or by inspection of the funnel plot (Supplementary Figure 9). The summary RR (95% CI) ranged from 0.83 (0.69-1.00) when excluding EPIC-Oxford study to 0.93 (0.78-1.11) when
excluding Tzu Chi Vegetarian study (Supplementary Figure 14).
When subtypes of stroke were analyzed, the summary RR for vegetarians vs. non-vegetarians was 0.56 (95% CI: 0.22-1.42, I2=81.9%, pheterogeneity=0.004) for ischemic stroke (Figure 5, Supplementary Figure 4) and 0.77 (95% CI: 0.19-3.09, I2=84.8%, pheterogeneity=0.01) for hemorrhagic stroke (Figure 6, Supplementary Figure 5).
Subgroup analysis
Table 2 and 3, and supplementary table 4 and 5 show results from all subgroup analyses.
Although, there were high heterogeneity in our main analysis as measured by I2, this was for CVD and IHD mainly driven by differences in the strength of the association, rather than due to differences in the direction of the association, as all studies reported risk estimates in the direction of an inverse association. For stroke the results were less consistent as there was heterogeneity also with regard to the direction of the association.
For CVD and IHD the inverse associations persisted in subgroup analyses stratified by sex, follow-up years, exclusion of early follow-up years outcome subtype, exclusion of prevalent disease at baseline geographic location number of cases study quality and adjustment for confounding factors. In the analysis of CVD there was little evidence of heterogeneity between subgroups, with the exception of the subgroup analysis stratified by geographic location, which showed a stronger association in North American studies than in European studies (Table 2, Supplementary Table 4). In the analysis of IHD there was also little
evidence of heterogeneity between subgroups, with the exception of the subgroup stratified by whether early follow-up was included or not which showed a weaker
association when early follow-up was excluded compared to when it was included (Table 3, Supplementary Table 5). The null results for stroke also persisted in most subgroup
analyses, however, there was significant heterogeneity between subgroups when analyses were stratified by duration of follow-up and geographic location with significant inverse associations among studies with shorter vs. longer follow-up and among two Asian
(Taiwanese) studies, but not among European or American studies (Table 3, Supplementary Table 5).
DISCUSSION
The present meta-analysis and systematic review showed a statistically significant 15 % reduction in risk of CVD, and a significant 21% reduction in risk of IHD for vegetarians compared to non-vegetarians, but no significant association was observed for total stroke or subtypes of stroke. Although, there was high heterogeneity in the main analyses for CVD and IHD, this was largely explained by differences in the strength of the associations as all
studies reported risk estimates in the direction of an inverse association. For CVD the association was stronger among North American studies than among the European studies.
For stroke, studies were less consistent, and an inverse association was observed only among Asian studies and not in European or American studies. There were no obvious indications for publication bias in neither of the three main analyses. The findings regarding IHD are consistent with previous meta-analyses (7, 38-40), however, to our knowledge this is the first meta-analysis to report a significant reduction in CVD risk overall as well, while previous meta-analyses found no significant association (7, 38-40). This difference is likely due to the larger sample size and statistical power in the current analysis. Our finding of no association for stroke is consistent with previous meta-analyses on vegetarian diets and stroke (7, 38-40), however, we cannot entirely rule out a weak to moderate association, but further studies are needed to clarify this.
There are several strengths and limitations to our analysis. Heterogeneity was an apparent issue across all outcomes, questioning the ability to allow generalization. However, in performing subgroup analysis, most CVD and IHD results remained unchanged across subgroups, suggesting a wide recommendation could be possible. Although stroke results were non-significant and highly inconsistent, our subgroup analysis showed a possible protective effect of vegetarian diets among Asian populations, although the opposite was suggested in the UK population by the findings in the EPIC Oxford study (20). Furthermore, we performed sensitivity allowing a direct comparison between only studies performing both BMI-adjustment and no BMI-adjustment. Although, no significant subgroup effect was detected for BMI, our analysis showed a possible mediating effect of BMI in reducing IHD risk among vegetarians as results from studies with BMI-adjustment showed a lower risk- reduction in vegetarians. Although, our subgroup analysis controlled for multiple factors, one can never completely rule out residual confounding because of measurement errors or
unmeasured confounders, however in the EPIC-Oxford study results for IHD persisted across strata of smoking and presence of other risk factors, which might potentially suggest
independent effect (20). We also performed several tests to assess for publication bias,
although one can never fully exclude missing data. However, our meta-analysis had increased statistical strength due to the inclusion of more studies than previous meta-analyses (7, 38- 40), reflecting the largely consistent findings for CVD and IHD.
One other strength with this present meta-analysis was a detailed search-strategy which enabled a wider search; however, one publication was recognized as missing, indicating a possible lack of certain terms. However, we applied multiple subject headings and identical terms were used in a free text search. As exposure assessment in most cohorts was only performed at baseline (except in the EPIC-Oxford study), participants may have changed their diet during the course of follow-up. This could lead to misclassification of dietary habits and offset our results. However, follow-up dietary data in EPIC-Oxford study, compared to baseline data, showed a greater dietary change among vegetarians, which would only underestimate the inverse association observed for IHD and CVD in our analysis (20). The use of self-reported dietary questionnaires may further allow misclassification of dietary habits, as self-identified vegetarians may in fact eat small amounts of meat, fish or poultry.
However, this would likely only underestimate our findings. Last, no studies were of low quality, reducing the risk of bias, although only two studies were of high quality, suggesting moderate risk of bias.
Our results are consistent with studies on food groups and CVD and IHD. Studies have shown increased risk of both CVD, IHD (41-43), but also stroke associated with consumption of red and processed meat (44). Evidence have shown that red meat increases atherosclerosis in mice through increased production of trimethylamine-N-oxide (TMAO) through a gut- dependent pathway. In fact, vegetarians exhibit lower levels of TMAO, suggesting a link between meat- consumption and CVD risk (17). However, there are other dietary
explanations for reduced CVD risk among vegetarians. Compared to meat-eaters, vegetarians tend to have a higher intake of fruit and vegetables, nuts and whole grains (45), and such food groups have generally been shown to reduce the risk of CVD, IHD and stroke (46-48). Also, vegetarians have a higher intake of plant-based proteins (45), i.e. legumes, which have been shown to reduce the risk of IHD (49). Last, vegetarian diets are also associated with lower BMI (7) and reduced weight gain, lower serum total and LDL-cholesterol (17), lower blood pressure or prevalence of hypertension (9-11), and lower risk of type 2 diabetes (13), all of which are important CVD risk factors.
The present meta-analysis suggests a vegetarian diet offer important health benefits by reducing the risk of both CVD and IHD, although not stroke. These are findings with important public health implications given that CVDs still are the main causes of death globally and suggest that adoption of plant-based dietary patterns such as vegetarian diets can be useful for reducing the CVD burden.
Future research should focus on additional large-scale and high-quality studies as they are needed to clarify results for stroke and stroke subtypes, as well as to provide results stratified by other risk factors to better rule out potential residual confounding. Further studies from other geographic regions also needed to improve generalization of the data. Detailed and repeated dietary assessment are also important because diet can change during follow-up.
Further studies should also focus on recruiting vegans as there were few studies with sufficient number of vegans to allow for statistical strength in meta-analysis.
These findings have important public health implications given that CVDs still are the main causes of death globally and suggest that adoption of plant-based dietary patterns such as vegetarian diets could be useful for reducing the CVD burden. This paper supports a stronger emphasizes on vegetarian dietary patterns in public health recommendations as a measure for CVD prevention.
In conclusion, we found a significant inverse association for CVD and IHD in comparing vegetarians to non-vegetarians. For stroke and subtypes of stroke, uncertainty remains, as there was only a non-significantly reduced risk for these outcomes. Among vegans, we did not find a significantly reduced risk of CVD, IHD or stroke, but there were an insufficient number of studies in our analysis. These findings are in agreement with already existing guidelines recommending plant-based dietary patterns for CVD prevention (4, 15, 16), however more emphasis may be put on vegetarian diets. Further studies are needed to clarify the association between vegetarian diets and stroke risk, as well as the association between vegan diets and CVD, IHD and stroke.
Conflict of interest
No conflict of interest
Author contributions
Literature search and screening: JSD
Data extraction and checking data extractions: JSD, DA Statistical analyses: JSD
Drafting of manuscript: JSD
Critical revision of manuscript: JSD, MS, DA Study supervision: MS, DA
Funding
This project has not received funding.
References
1. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al.
Global Burden of Cardiovascular Diseases and Risk Factors, 1990–2019: Update From the GBD 2019 Study. J Am Coll Cardiol. 2020;76(25):2982-3021.
2. Dai H, Much AA, Maor E, Asher E, Younis A, Xu Y, et al. Global, regional, and national burden of ischaemic heart disease and its attributable risk factors, 1990–2017: results from the Global Burden of Disease Study 2017. American Heart Journal - Quality of Care and Clinical Outcomes. 2020;0:1-11.
3. Krishnamurthi RVI, Takayoshi ; Feigin, Valery L. Global, Regional and Country- Specific Burden of Ischaemic Stroke, Intracerebral Haemorrhage and Subarachnoid Haemorrhage: A Systematic Analysis of the Global Burden of Disease Study 2017.
2020;54(2):171-9.
4. Arnett D, Blumenthal R, Albert M, Buroker A, Goldberger Z, Hahn E, et al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2019;140:e596-e646.
5. Najjar RS, Moore CE, Montgomery BD. A defined, plant-based diet utilized in an outpatient cardiovascular clinic effectively treats hypercholesterolemia and hypertension and reduces medications. 2018;41(3):307-13.
6. Yokoyama Y, Levin SM, Barnard ND. Association between plant-based diets and plasma lipids: a systematic review and meta-analysis. Nutr Rev. 2017;75(9):683-98.
7. Dinu M, Abbate R, Gensini GF, Casini A, Sofi F. Vegetarian, vegan diets and multiple health outcomes: A systematic review with meta-analysis of observational studies.
Crit Rev Food Sci Nutr. 2017;57(17):3640-9.
8. Wang F, Zheng J, Yang B, Jiang J, Fu Y, Li D. Effects of Vegetarian Diets on Blood Lipids: A Systematic Review and Meta‐Analysis of Randomized Controlled Trials. J Am Heart Assoc. 2015;4(10):e002408-e.
9. Yokoyama Y, Nishimura K, Barnard ND, Takegami M, Watanabe M, Sekikawa A, et al. Vegetarian Diets and Blood Pressure
A Meta-analysis. JAMA Intern Med. 2014;174(4):577-87.
10. Pettersen BJ, Anousheh R, Fan J, Jaceldo-Siegl K, Fraser GE. Vegetarian diets and blood pressure among white subjects: results from the Adventist Health Study-2 (AHS-2).
Public Health Nutr. 2012;15(10):1909-16.
11. Appleby PN, Davey GK, Key TJ. Hypertension and blood pressure among meat eaters, fish eaters, vegetarians and vegans in EPIC-Oxford. Public Health Nutr.
2002;5(5):645-54.
12. Thorogood M, Carter R, Benfield L, McPherson K, Mann JI. Plasma lipids and lipoprotein cholesterol concentrations in people with different diets in Britain. Br Med J (Clin Res Ed). 1987;295(6594):351-3.
13. Serena T, Butler T, Yan R, Fraser GE. Type of Vegetarian Diet, Body Weight, and Prevalence of Type 2 Diabetes. Diabetes Care. 2009;32(5):791-6.
14. Davey GK, Spencer EA, Appleby PN, Allen NE, Knox KH, Key TJ. EPIC-Oxford:
lifestyle characteristics and nutrient intakes in a cohort of 33 883 meat-eaters and 31 546 non meat-eaters in the UK. Publi Health Nutrition. 2003;6(3):259-69.
15. Willett W, Rockström J, Loken B, Springmann M, Lang T, Vermeulen S, et al. Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems. The Lancet. 2019;393(10170):447-92.
16. Helsedirektoratet. Kostråd ved primær- og sekundærforebygging av hjerte- og karsykdom (internet document). Oslo: Helsedirektoratet (last updated 5th of march 2018, read Feb 20, 2021). Available from
https://www.helsedirektoratet.no/retningslinjer/forebygging-av-hjerte-og-
karsykdom/kartlegging-av-levevaner-og-rad-om-livsstilstiltak-som-forebygging-av-hjerte-og- karsykdom/kostrad-ved-primaer-og-sekundaerforebygging-av-hjerte-og-karsykdommer.
17. Oussalah A, Levy J, Berthezène C, Alpers DH, Guéant J-L. Health outcomes associated with vegetarian diets: An umbrella review of systematic reviews and meta- analyses. Clinical nutrition (Edinburgh, Scotland). 2020;39(11):3283-307.
18. Key TJ, Fraser GE, Thorogood M, Appleby PN, Beral V, Reeves G, et al. Mortality in vegetarians and nonvegetarians: Detailed findings from a collaborative analysis of 5
prospective studies. Am J Clin Nutr. 1999;70(3 suppl):516S-24S.
19. Petermann-Rocha F, Parra-Soto S, Gray S, Anderson J, Welsh P, Gill J, et al.
Vegetarians, fish, poultry, and meat-eaters: who has higher risk of cardiovascular disease incidence and mortality? A prospective study from UK Biobank. Eur Heart J.
2020;ehaa939:1-8.
20. Tong TYN, Appleby PN, Bradbury KE, Perez-Cornago A, Travis RC, Clarke R, et al.
Risks of ischaemic heart disease and stroke in meat eaters, fish eaters, and vegetarians over 18 years of follow-up: Results from the prospective EPIC-Oxford study. BMJ. 2019;366:1- 10.
21. Orlich MJ, Singh PN, Sabaté J, Jaceldo-Siegl K, Fan J, Knutsen S, et al. Vegetarian dietary patterns and mortality in Adventist Health Study 2. JAMA Intern Med.
2013;173(13):1230-8.
22. Key TJ, Fraser GE, Thorogood M, Appleby PN, Beral V, Reeves G, et al. Mortality in vegetarians and nonvegetarians: detailed findings from a collaborative analysis of 5
prospective studies. The American journal of clinical nutrition. 1999;70(3 Suppl):516S-24S.
23. Chiu THT, Chang H-R, Wang L-Y, Chang C-C, Lin M-N, Lin C-L. Vegetarian diet and incidence of total, ischemic, and hemorrhagic stroke in 2 cohorts in Taiwan. Neurology.
2020;94(11):e1112-e21.
24. Chang-Claude J, Hermann S, Eilber U, Steindorf K. Lifestyle determinants and mortality in German vegetarians and health-conscious persons: results of a 21-year follow- up. Cancer Epidemiol Biomarkers Prev. 2005;14(4):963-8.
25. Appleby PN KT, Thorogood M, Burr ML, Mann J. Mortality in British vegetarians.
Public Health Nutr. 2002;5(1):29-36.
26. Wells G, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle- Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses.
Ottawa: Ottawa Hospital Research Institute Available online at:
http://wwwohrica/programs/clinical_epidemiology/oxfordasp (last updated 2019, assessed Feb 8 2021). 2019.
27. Moher D, Liberati A, Tetzlaff J, Altman DG, Group TP. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med.
2009;6(7):1-6.
28. DerSimonian R, Laird L. Meta-Analysis in Clinical Trials. Control Clin Trials.
1986;7(3):177-88.
29. Kulinskaya E, Dollinger MB. An accurate test for homogeneity of odds ratios based on Cochran’s Q-statistic. BMC Med Res Methodol. 2015;15(1):1-19.
30. Higgens JPT TS. Quantifying heterogeneity in a meta‐analysis. Satistics in Medicine.
2002;21(11):1539-58.
31. M. E, Smith G, C Minder M. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629-34.
32. CB. B, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50(4):1088-101.
33. Light RJ, Pillemer DB. Summing up : the science of reviewing research. Cambridge, Mass: Harvard University Press; 1984.
34. Appleby PN, Crowe FL, Bradbury KE, Travis RC, Key TJ. Mortality in vegetarians and comparable nonvegetarians in the United Kingdom. The American journal of clinical nutrition. 2016;103(1):218-30.
35. Fraser GE, Lindsted KD, Beeson WL. Effect of risk factor values on lifetime risk of and age at first coronary event. The Adventist Health Study. Am J Epidemiol.
1995;142(7):746-58.
36. Snowdon DA, Phillips RL, Fraser GE. Meat consumption and fatal ischemia heart disease. Prev Med. 1984;13(5):490-500.
37. Beeson WL, Mills PK, Phillips RL, Andress M, Fraser GE. Chronic disease among Seventh-day Adventists, a low-risk group. Rationale, methodology, and description of the population # Adventist health study 1 / AHS-1 OBS: er vel bare tverrsnittsdata dette? Cancer.
1989;64(3):570-81.
38. Glenn AJ, Viguiliouk E, Seider M, Boucher BA, Khan TA, Blanco Mejia S, et al.
Relation of Vegetarian Dietary Patterns With Major Cardiovascular Outcomes: A Systematic Review and Meta-Analysis of Prospective Cohort Studies. Front Nutr. 2019;6:80-.
39. Kwok CS, Umar S, Myint PK, Mamas MA, Loke YK. Vegetarian diet, Seventh Day Adventists and risk of cardiovascular mortality: A systematic review and meta-analysis. Int J Cardiol. 2014;176(3):680-6.
40. Huang T, Yang B, Zheng J, Li G, Wahlqvist ML, Li D. Cardiovascular disease mortality and cancer incidence in vegetarians: A meta-analysis and systematic review. Ann Nutr Metab. 2012;60(4):233-40.
41. Al-Shaar L, Satija A, Wang DD, Rimm EB, Smith-Warner, A. S, et al. Red meat intake and risk of coronary heart disease among US men: prospective cohort study. BMJ (Clinical research ed. 2020;371:m4141-m.
42. Micha R, Wallace SK, Mozaffarian D. Red and processed meat consumption and risk of incident coronary heart disease, stroke, and diabetes mellitus: a systematic review and meta-analysis. Circulation. 2010;121(21):2271-83.
43. Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major dietary protein sources and risk of coronary heart disease in women. 2010;122(9):876-83.
44. KALUZA J, WOLK A, LARSSON SC. Red meat consumption and risk of stroke: a meta-analysis of prospective studies. Stroke. 2012;43(10):2556-60.
45. Papier K, Tong TY, Appleby PN, Bradbury KE, Fensom GK, Knuppel A, et al.
Comparison of Major Protein-Source Foods and Other Food Groups in Meat-Eaters and Non- Meat-Eaters in the EPIC-Oxford Cohort. Nutrients. 2019;11(4):1-16.
46. Aune D, Giovannucci E, Boffetta P, Fadnes LT, Keum N, Norat T, et al. Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all- cause mortality–a systematic review and dose- response meta-analysis of prospective studies. Int J Epidemiol.
2017;46(3):1029-56.
47. Aune D, Keum N, Giovannucci E, Fadnes LT, Boffetta P, Greenwood DC, et al.
Whole grain consumption and risk of cardiovascular disease, cancer, and all cause and cause specific mortality: systematic review and dose-response meta-analysis of prospective studies.
BMJ. 2016;353(2716).
48. Aune D, Keum N, Giovannucci E, Fadnes LT, Boffetta P, Greenwood DC, et al. Nut consumption and risk of cardiovascular disease, total cancer, all-cause and cause- specific mortality: a systematic review and dose-response meta-analysis of prospective studies. BMC Med. 2016;14(1).
49. Afshin A, Micha R, Khatibzadeh S, Mozaffarian D. Consumption of nuts and
legumes and risk of incident ischemic heart disease, stroke, and diabetes: a systematic review and meta-analysis. The American journal of clinical nutrition. 2014;100(1):278-88.
Figure 1. Flow chart
Figure 2. Forest plot vegetarian diet and CVD
CVD cardiovascular disease AMS Adventist Mortality; study AHS-1 Adventist Health study;
OVS Oxford Vegetarian study; HFSS Health Food Shoppers study; HBS Heidelberg study;
AHS-2 Adventist Health study 2; EPIC-Oxford Oxford component of the European Prospective Investigation into Cancer and Nutrition; UKB UK Biobank study Figure 3. Forest plot vegetarian diet and IHD
IHD ischemic heart disease AMS Adventist Mortality; study AHS-1 Adventist Health study;
OVS Oxford Vegetarian study; HFSS Health Food Shoppers study; HBS Heidelberg study;
AHS-2 Adventist Health study 2; EPIC-Oxford Oxford component of the European Prospective Investigation into Cancer and Nutrition; UKB UK Biobank study
Figure 4. Forest plot vegetarian diet and total stroke / cerebrovascular disease
HBS Heidelberg study AMS Adventist Mortality; study AHS-1 Adventist Health study; OVS Oxford Vegetarian study; HFSS Health Food Shoppers study; EPIC-Oxford Oxford
component of the European Prospective Investigation into Cancer and Nutrition; UKB UK Biobank study, TCVS Tzu Chi Vegetarian study; TCHS Tzu Chi Health study
Figure 5. Forest plot vegetarian diet and ischemic stroke
EPIC-Oxford, Oxford component of the European Prospective Investigation into Cancer and Nutrition; TCVS Tzu Vegetarian study; TCHS Tzu Chi Health study
Figure 6. Forest plot vegetarian diet and hemorrhagic stroke
EPIC-Oxford, Oxford component of the European Prospective Investigation into Cancer and Nutrition; TCVS Tzu Vegetarian study
Table 1. Characteristics of the included cohort publications
Author, year Country / region
Study name Sex Age (years)
F-up (years)
Cases / Sample size
Sex Outcome RR (95% CI) Adjustment Study
quality Snowden, 1984,
AMS
USA / North America
Adventist Mortality study (AMS)
30-84 21 1599 / 24172 M
W
IHD, mortality, M a IHD, mortality, W a
0.67 (0.56 - 0.77) 0.73 (0.63 - 0.83)
d Age 6
Fraser 1995, AHS-1
USA / North America
Adventist Health study (AHS-1)
> 24 5.14 448 / 27321 M W
IHD, incidence, M a IHD, incidence, W a
0.56 (0.4 - 0.78) 1.02 (0.74 - 1.41)
d Smoking, BMI, exercise, DM, hypertension, nut consumption
7
Key et al. 1999, AMS
USA / North America
Adventist Mortality study (AMS)
16–89 5.6 780 / 24538 598 / 24538 182 / 24538
M/W M/W M/W
CVD, mortality b IHD, mortality CBVD, mortality
0.72 (0.62 - 0.83) 0.74 (0.63 - 0.88) 0.65 (0.48 - 0.87)
d Age, sex and smoking 5
Key et al. 1999, AHS-1
USA / North America
Adventist Health study (AHS-1)
16–89 11.1 1238 / 28952 921 / 28952 317 / 28952
M/W M/W M/W
CVD, mortality b IHD, mortality CBVD, mortality
0.70 (0.61 - 0.80) 0.62 (0.53 - 0.73) 0.93 (0.73 - 1.19)
d Age, sex and smoking 6
Key et al. 1999, HBS
Germany / Europe
Heidelberg study (HBS)
16–89 9.9 31 / 1757 M/W CBVD, mortality 1.69 (0.69 - 4.15) d Age, sex and smoking 5
Appleby et al.
2002, OVS
UK / Europe
Oxford
Vegetarian study (OVS)
16 – 89 17.6 469 / 11045 250 / 11045 125 / 11045
M/W M/W M/W
CVD, mortality IHD, mortality CBVD, mortality
0.93 (0.77 - 1.12) 0.86 (0.67 - 1.12) 1.08 (0.75 - 1.54)
d Age, sex, smoking 5
Appleby et al.
2002, HFSS
UK / Europe
Health Food Shoppers study (HFSS)
16 – 89 18.7 1117 / 10736 562 / 10736 330 / 10736
M/W M/W M/W
CVD, mortality IHD, mortality CBVD, mortality
0.95 (0.84 - 1.07) 0.85 (0.71 - 1.01) 0.99 (0.79 - 1.24)
d Age, sex, smoking 5
Chang-Claude, 2005, HBS
Germany / Europe
Heidelberg study (HBS)
≥ 10 21 219 / 1724
60 / 1724
M/W M/W
CVD, mortality IHD, mortality
0.83 (0.62 - 1.12) 0.70 (0.41 - 1.18)
d Age, gender, smoking, physical activity, alcohol, BMI, education
6
Orlich et al, 2013, AHS-2
USA / North America
Adventist Health study 2 (AHS-2)
≥ 25 5.79 987 / 73308 372 / 73308 987 / 73308 987 / 73308 372 / 73308 372 / 73308
M/W M/W M W M W
CVD, mortality IHD, mortality CVD, mortality, M CVD, mortality, W IHD, mortality, M IHD, mortality, W
0.87 (0.75 - 1.01) 0.81 (0.64 - 1.02) 0.71 (0.57 - 0.90) 0.99 (0.83 - 1.18) 0.71 (0.51 – 1.00) 0.88 (0.65 - 1.20)
d, e Age, sex, race, smoking, exercise, income, education, marital status, alcohol, region, sleep (plus menopause and hormone therapy in women)
7
Tong et al.
2019, EPIC- Oxford
UK / Europe
EPIC-Oxford study
20-90 18.1 3892 / 48188 2820 / 48188 2820 / 48188 1072 / 48188 1072 / 48188 519 / 48188 300 / 48188 2820 / 48188 2820 / 48188 1072 / 48188 1072 / 48188
M/W M/W M/W M/W M/W M/W M/W M W M W
CVD, incidence b IHD, incidence IHD, incidence c1 Total stroke, incidence Total stroke, incidence c1 Ischemic stroke, incidence Hemorrhagic stroke, incidence IHD, incidence, M
IHD, incidence, W Total stroke, incidence, M Total stroke, incidence, W
0.9 (0.82 - 0.98) b 0.78 (0.7 - 0.87) 0.83 (0.75 - 0.92) c1 1.2 (1.02 - 1.4) 1.21 (1.03 - 1.42) c1 1.12 (0.9 - 1.41) 1.43 (1.08 - 1.9) 0.77 (0.66 - 0.91) 0.78 (0.68 - 0.90) 0.99 (0.73 - 1.34) 1.28 (1.07 - 1.53)
d, e age, sex, method of recruitment, region, year of recruitment, education, deprivation, smoking, alcohol, physical activity, dietary supplements.
Pluss oral contraceptive and HRT use in women
c1 BMI-adjusted model
8
Chiu et al. 2020, TCHS
Taiwan / Asia
Tzu Chi Health study (TCHS, cohort 1)
≥ 20 6.1 54 / 5050 54/ 5050 31 / 5050
M/W M/W M/W
Total stroke, incidence Total stroke, incidence c2 Ischemic stroke, incidence
0.51 (0.25 - 1.06) 0.51 (0.25 - 1.06) c2 0.26 (0.08 - 0.88)
d age, sex, smoking, alcohol, betel nut, physical activity, education, hypertension, DM, dyslipidemia, IHD and BMI (≥24 vs <24)
c2 BMI-unadjusted model
8
Chiu et al. 2020, TCVS
Taiwan / Asia
Tzu Chi Vegetarian study (TCVS, cohort 2)
≥ 20 9.25 121 / 8302 54 / 8302 28 / 8302
M/W M/W M/W
Total stroke, incidence Ischemic stroke, incidence Hemorrhagic stroke, incidence
0.52 (0.33 - 0.82) 0.41 (0.19 - 0.88) 0.34 (0.12 - 1.00)
d age, sex, smoking, alcohol, betel nut, physical activity, education, hypertension, DM, dyslipidemia, IHD
7
AMS, Adventist Mortality; study AHS-1, Adventist Health study; OVS Oxford Vegetarian study; HFSS, Health Food Shoppers study; HBS Heidelberg study; AHS-2 Adventist Health study 2; EPIC-Oxford Oxford component of the European Prospective Investigation into Cancer and Nutrition; TCHS Tzu Chi Health study/cohort 1 TCVS Tzu Chi Vegetarian study/cohort 2 UKB UK Biobank study IHD ischemic heart disease DM diabetes mellitus BMI body mass index HRT hormone replacement therapy
Petermann- Rocha et al.
2020, UKB
UK / Europe
UK Biobank study (UKB)
37-73 8.5 106690 / 422791 6580 / 422791 24794 / 422791 24794 / 422791 2767 / 422791 5946 / 422791 5946 / 422791 1088 / 422791 106690 / 422791 106690 / 422791 24794 / 422791 24794 / 422791 5946 / 422102 5946 / 422102
M/W M/W M/W M/W M/W M/W M/W M/W M W M W M W
CVD, incidence CVD, mortality IHD, incidence IHD, incidence c3 IHD, mortality Total stroke, incidence Total stroke, incidence c3 Total stroke, mortality CVD, incidence, M CVD, incidence, W IHD, incidence, M IHD, incidence, W Total stroke, incidence, M Total stroke, incidence, W
0.91 (0.86 - 0.96) 0.91 (0.71 - 1.17) 0.96 (0.85 - 1.07) 0.90 (0.80 - 1.01) c3 0.88 (0.59 - 1.30) 0.84 (0.66 - 1.07) 0.82 (0.64 - 1.05) c3 0.87 (0.48 - 1.58) 0.88 (0.81 - 0.96) 0.93 (0.87 - 0.99) 0.91 (0.78 - 1.07) 1.02 (0.87 - 1.20) 0.81 (0.56 - 1.18) 0.88 (0.63 - 1.22)
d age, sex, deprivation, and ethnicity (white, mixed, south Asian, black, Chinese), morbidity (43 diseases), smoking, sedentary time(h/day), alcohol, physical activity, BMI
e age, deprivation, ethnicity,
comorbidities, smoking, alcohol intake,
sedentary time, physical activity and BMI
c3 BMI-unadjusted model
6
Appleby et al, 2016, OVS + EPIC-Oxford (only pooled data)
UK / Europe
Oxford vegetarian study (OVS) +
EPIC-Oxford study
20-89 > 1 million person- years
433 / 60 310 175 / 60 310 152 / 60 310
M/W M/W M/W
CVD mortality IHD mortality CBVD mortality
1.10 (0.95 - 1.27) 0.99 (0.79 - 1.23) 1.21 (0.94 - 1.56)
Age, smoking, alcohol, physical activity, married or cohabiting, nutritional
supplementation, method of recruitment, sex, parity, oral contraceptive use, HRT, prior diabetes, prior high blood pressure, receipt of long-term medical treatment
7
All RRs compare vegetarians to nonvegetarians. Main analysis use incidence and mortality data on both sexes combined but only incidence data if studies reported both.
a Data in table (RRs and 95% CIs) were calculated using the inverse method as original data compared non-vegetarians to vegetarians
b Data on IHD and stroke were pooled into CVD-RRs for these publications: Key et al. 1999 (AMS, AHS-1) and Tong et al 2019 (EPIC-Oxford)
c1 BMI-adjusted model, Tong et al 2019 (EPIC-Oxford), only used in subgroup analysis BMI-adjusted vs BMI-unadjusted
c2 BMI-unadjusted model, Chiu et al. 2020 (TCHS)
c3 BMI-unadjusted model, Petermann-Rocha et al. 2020 (UKB)
d confounders adjusted for in model with both sexes combined (both incidence- and mortality data have the same adjustments)
e confounders adjusted for in sex-stratified model
Data used in subgroup analysis on exclusion of early follow-up (vs no exclusion) are not listed (see supplementary material)
Table 2. Subgroup analysis vegetarian diet and CVD
Vegetarianism and cardiovascular disease (CVD)
n Relative risk (95% CI) I2 (%) Ph1 Ph2
All studies 8 0.85 (0.79 - 0.92) 68.0 0.003
Sex
Men 2 0.81 (0.66 - 1.00) 66.5 0.08 0.343A
0.313B
Women 2 0.94 (0.88 - 1.00) 0.0 0.51
Men + women 4 0.93 (0.77 - 1.12) 0.0 0.81
Follow-up
<10 years (1) 3 0.84 (0.73 - 0.96) 77.0 0.013 0.79
≥10 years (2) 5 0.86 (0.76 - 0.97) 69.3 0.011
Early follow-up years
Included 5 0.80 (0.71 - 0.90) 69.4 0.01 0.08
Excluded 3 0.92 (0.88 - 0.97) 0.0 0.48
Outcome
Incidence 2 0.91 (0.87 - 0.95) 0.0 0.84 0.70
Mortality 8 0.87 (0.77 - 0.98) 75.9 0.000
Excluding prevalent disease at baseline
Yes 3 0.90 (0.86 - 0.95) 0.0 0.85 0.30
No 5 0.82 (0.71 - 0.94) 74.5 0.003
Geographic location
Europe (1) 5 0.91 (0.87 - 0.95) 0.0 0.92 0.007
North America (2) 3 0.76 (0.66 - 0.87) 60.7 0.08
Asia (3) 0
Number of cases
Cases <250 (1) 1 0.83 (0.62 - 1.12) NC NC 0.81
Cases 250 – 499 (2) 1 0.93 (0.77 - 1.12) NC NC
Cases ≥ 500 (3) 6 0.84 (0.77 - 0.92) 76.6 0.001
Study quality
0-3 stars (1) 0 0.59
4-6 stars (2) 6 0.84 (0.75 - 0.94) 76.8 0.001
7-9 stars (3) 2 0.89 (0.83 - 0.96) 0.0 0.70
Adjustment for confounding factors
Age Yes 8 0.85 (0.79 - 0.92) 87.4 0.003 NC
No 0
Education Yes 3 0.89 (0.83 - 0.96) 0.0 0.84 0.66
No 5 0.84 (0.74 - 0.95) 81.3 0.000
Alcohol Yes 4 0.90 (0.86 - 0.94) 0.0 0.89 0.33
No 4 0.82 (0.69 - 0.96) 80.9 0.001
Smoking Yes 8 0.85 (0.79 - 0.92) 87.4 0.003 NC
No 0
BMI Yes 2 0.91 (0.86 - 0.96) 0.0 0.55 0.56
No 6 0.84 (0.76 - 0.93) 73.2 0.002
Physical activity Yes 4 0.90 (0.86 - 0.94) 0.0 0.89 0.33
No 4 0.82 (0.69 - 0.96) 80.9 0.001
n denotes the number of risk estimates (some publications reported RRs for more than one
study)
1 P for heterogeneity within each subgroup
2 P for heterogeneity between subgroups with meta-regression analysis
3A P for heterogeneity between men, women and men/women combined with meta-regression analysis
3B P for heterogeneity between men and women (excluding studies with both sexes combined) with meta-regression analysis
4 Restricted to studies reporting both BMI-adjusted and BMI-unadjusted results
BMI body mass index NC not calculable because no studies were present in one of the subgroups
Table 3. Subgroup analysis IHD and total stroke (incl cerebrovascular disease)
Vegetarianism and ischemic heart disease (IHD) Vegetarianism and total stroke / cerebrovascular disease
n Relative risk (95%
CI)
I2 (%) Ph1 Ph2 n Relative risk (95% CI) I2 (%) Ph1 Ph2
All studies 8 0.79 (0.71 – 0.80) 67.3 0.003 9 0.88 (0.73 - 1.07) 70.8 0.001
Sex
Men 5 0.74 (0.63 - 0.86) 62.9 0.03 0.243A
0.193B
2 0.91 (0.72 - 1.16) 0.0 0.41 0.543A
0.473B
Women 5 0.86 (0.74 – 0.99) 66.9 0.02 2 1.09 (0.76 - 1.57) 73.8 0.05
Men, women 3 0.84 (0.73 - 0.97) 0.0 0.78 7 0.83 (0.66 - 1.04) 61.8 0.02
Follow-up
<10 years 3 0.84 (0.70 - 1.00) 70.5 0.03 0.38 5 0.71 (0.54 - 0.94) 51.9 0.08 0.03
≥10 years 5 0.76 (0.67 - 0.86) 55.3 0.06 4 1.07 (0.94 - 1.21) 19.2 0.29
Early follow-up years
Included 4 0.71 (0.62 - 0.80) 27.5 0.25 0.04 5 0.73 (0.53 - 1.00) 62.2 0.03 0.11
Excluded 4 0.89 (0.80 -0.99) 45.9 0.14 4 1.02 (0.86 - 1.21) 47.7 0.13
Outcome
Incidence 2 0.87 (0.71 - 1.06) 84.9 0.01 0.48 4 0.78 (0.53 - 1.15) 83.7 0.000 0.37
Mortality 8 0.79 (0.71 - 0.89) 52.8 0.04 7 0.97 (0.81 - 1.16) 49.9 0.06
Stroke subtype
Ischemic stroke 3 0.56 (0.22 - 1.42) 81.9 0.004 0.67
Hemorrhagic stroke 2 0.77 (0.19 - 3.09) 84.8 0.01
Excluding prevalent disease at baseline
Yes 3 0.85 (0.73 - 0.99) 70.8 0.03 0.25 4 0.78 (0.53 - 1.15) 83.7 0.000 0.50
No 5 0.75 (0.65 - 0.86) 52.5 0.08 5 0.92 (0.75 - 1.13) 51.1 0.09
Geographic location
Europe 5 0.85 (0.77 - 0.95) 44.3 0.13 0.09 5 1.05 (0.89 - 1.23) 45.3 0.12 0.01
North America 3 0.71 (0.61- 0.83) 51.7 0.13 2 0.79 (0.55 - 1.12) 70.0 0.07
Asia 0 2 0.52 (0.35 - 0.76) 0.0 0.96
Number of cases
Cases <250 1 0.70 (0.41 - 1.19) NC NC 0.97 5 0.76 (0.52 - 1.09) 65.4 0.02 0.21
Cases 250 – 499 2 0.83 (0.70 - 0.99) 0.0 0.74 2 0.96 (0.82 - 1.14) 0.0 0.71
Cases ≥ 500 5 0.79 (0.68 - 0.91) 80.8 0.000 2 1.02 (0.72 - 1.44) 82.9 0.02
Study quality
0-3 stars 0 1.00 0 0.53
4-6 stars 6 0.79 (0.67 - 0.93) 76.2 0.001 6 0.90 (0.77 - 1.06) 41.5 0.13
7-9 stars 2 0.79 (0.71 - 0.87) 0.0 0.77 3 0.72 (0.36 - 1.41) 87.3 0.000
Reported cerebrovascular disease estimates (and not stroke)
Yes 5 0.92 (0.75 - 1.13) 51.1 0.09 0.50
No 4 0.78 (0.53 - 1.15) 83.7 0.000
Adjustment for confounding factors
Age Yes 8 0.79 (0.71 - 0.89) 67.3 0.003 NC 9 0.88 (0.73 - 1.07) 87.8 0.001 NC
No 0 0
Education Yes 3 0.78 (0.71 - 0.86) 0.0 0.88 0.86 3 0.72 (0.36 - 1.41) 87.3 0.000 0.53
No 5 0.80 (0.67 - 0.94) 80.6 0.000 6 0.90 (0.77 - 1.06) 41.5 0.13
Alcohol Yes 4 0.84 (0.73 - 0.97 59.3 0.06 0.32 4 0.78 (0.53 - 1.15) 83.7 0.000 0.50
No 4 0.75 (0.64 - 0.88) 64.2 0.04 5 0.92 (0.75 - 1.13) 51.1 0.09
Smoking Yes 8 0.79 (0.71 - 0.89) 86.6 0.003 NC 9 0.88 (0.73 - 1.07) 87.8 0.001 NC
No 0 0
BMI Yes 3 0.88 (0.77 - 1.00) 52.1 0.12 0.20 3 0.90 (0.62 - 1.32) 80.2 0.01 0.83
No 5 0.76 (0.67 - 0.86) 55.1 0.06 6 0.86 (0.68 - 1.08) 63.3 0.02
BMI4 Yes 2 0.89 (0.77 - 1.03) 70.9 0.06 0.60 2 1.02 (0.71 - 1.46) 83.6 0.01 0.96
No 2 0.84 (0.73 - 0.96) 67.7 0.08 2 1.00 (0.69 - 1.46) 84.5 0.01
