Monounsaturated Fatty Acids and Risk of Cardiovascular Disease: Synopsis of the Evidence Available from Systematic Reviews and Meta-Analyses

No dietary recommendations for monounsaturated fatty acids (MUFA) are given by the National Institute of Medicine, the United States Department of Agriculture, European Food and Safety Authority and the American Diabetes Association. In contrast, the Academy of Nutrition and Dietetics, and the Canadian Dietetic Association both promote <25% MUFA of daily total energy consumption, while the American Heart Association sets a limit of 20% MUFA in their respective guidelines. The present review summarizes systematic reviews and meta-analyses of randomized controlled trials and cohort studies investigating the effects of MUFA on cardiovascular and diabetic risk factors, cardiovascular events and cardiovascular death. Electronic database Medline was searched for systematic reviews and meta-analyses using “monounsaturated fatty acids”, “monounsaturated fat”, and “dietary fat” as search terms with no restriction to calendar date or language. Reference lists and clinical guidelines were searched as well. Sixteen relevant papers were identified. Several studies indicated an increase of HDL-cholesterol and a corresponding decrease in triacylglycerols following a MUFA-rich diet. The effects on total and LDL-cholesterol appeared not consistent, but no detrimental effects on blood lipids were observed. Values for systolic and diastolic blood pressure were found to be reduced both during short- and long-term protocols using high amounts of MUFA as compared to low-MUFA diets. In type 2 diabetic subjects, MUFA exerted a hypoglycemic effect and reduced glycosylated hemoglobin in the long term. Data from meta-analyses exploring evidence from long-term prospective cohort studies provide ambiguous results with respect to the effects of MUFA on risk of coronary heart disease (CHD). One meta-analysis reported an increase in CHD events, however, most meta-analyses observed a lesser number of cases in participants subjected to a high-MUFA protocol. Although no detrimental side effects of MUFA-rich diets were reported in the literature, there still is no unanimous rationale for MUFA recommendations in a therapeutic regimen. Additional long-term intervention studies are required to characterized efficacy and effectiveness of recommending MUFA-rich diet among general and clinical populations.


Monounsaturated Fatty Acids
Monounsaturated fatty acids (MUFA) are chemically classified as fatty acids containing a single double bond (in contrast to polyunsaturated fatty acids (PUFA) containing two or more double bonds and saturated fatty acids (SFA) without double bonds).In the cis-configuration, the hydrogen atoms are on the same side as the double bond, whereas in trans-configuration the hydrogen atoms and the double bond are present on opposite sides.The cis-isomers are the predominant form of MUFA in food sources.The most common cis-configured MUFA in daily nutrition is oleic acid (18:1 n-9), followed by palmitoleic acid (16:1 n-7), and vaccenic acid (18:1 n-7).Moreover, oleic acid represents the topmost MUFA provided in the diet (~90% of all MUFAs) [1].The major trans-configured MUFA is elaidic acid (trans 18:1 n-9).Some MUFA-such as mystrioleic (14:1 n-5), gondoic (20:1 n-9), erucic (22:1 n-9) and nervonic (24:1 n-9) acid-are synthesized in minor concentrations endogenously using other MUFAs as precursors (see Table 1 for a summary of different types of MUFA).Various sources for MUFA in food are given in Table 2 (for comparison, PUFA and SFA contents are given as well).The most frequently consumed MUFA rich dietary oils are canola and olive oil.Furthermore, over the last decade commercial production of high oleic acid modified dietary oils with improved stability for the use in food processing has been markedly increased in order to replace dietary oils rich in SFA and trans fatty acids [2].It should be recognized that in some populations, MUFAs are provided in higher amounts in the form of erucic acid (C22:1 n-9), e.g., found in culinary oils derived from some Brassica spp.such as rapeseed and mustard seed [3].It is therefore not surprising that due to their widespread occurrence in oils nuts, seeds, fruits and meat, the predominant source of MUFA is largely depending on individual dietary habits.Like other fatty acids, MUFA are almost completely absorbed by the intestine and are oxidized for energy production, converted into other fatty acids, or incorporated into tissue lipids.

General Nutrition Guidelines
Table 3 summarizes MUFA recommendations of national and international authorities and organizations.Limitation of total fat to 35% of TEC [22] International College of Nutrition of India 7 Diabetic patients [23] MUFA = monounsaturated fatty acids; SFA = saturated fatty acids; TEC = total energy content.
In 1999, the International Society for the Study of Fatty Acids and Lipids agreed upon a recommendation table on daily intake of fatty acids as a foundation for further discussions.Adequate intake levels for adults were specified with respect to α-linolenic acid, eicosapentaenoic acid, docosahexaenoic acid, as well as upper limits for linoleic acid, trans-fatty acids, and saturated, given as % of total energy content (TEC), respectively.Given a total fat range from 15% to 40% of TEC, these recommendations included the suggestion to provide the majority of fatty acids in the form of MUFAs.However, no precise value (i.e., % of TEC in the form of MUFA) was given by the panel [24].According to the Joint FAO/WHO Expert Consultation Committee, MUFA intakes should be determined by calculating the difference: MUFA (% of TEC) = total fat (% of TEC) − SFA (% of TEC) − PUFA (% of TEC) − TFA (% of TEC).Accordingly, MUFA intakes (% of TEC) will range with respect to the total fat and fatty acid composition of the diet [3].Based upon 13 peer-reviewed background papers dealing with fats and fatty acids in human nutrition, the Joint FAO/WHO Expert Consultation Committee concluded that replacement of carbohydrates by MUFA beneficially increases HDL-cholesterol, while the substitution of SFA with MUFA exerts favorable effects on LDL-cholesterol and the ratio of total cholesterol to HDL-cholesterol [3].In their position on dietary fatty acids of 2007, the American and Canadian Dietetic Association suggested a high maximum quota of MUFA, i.e., <25% of TEC [6].Less than 20% of TEC should be consumed in the form of MUFA as recommended by the American Heart Association (AHA) in 2006, which is interesting with respect to the fact that the corresponding value was only <15% in the year 2000-statement of the AHA [5,25].The National Cholesterol Education Program III suggested that <20% of TEC should be consumed in the form of MUFA [11].In their Dietary Guidelines for Americans, edition 2010 [16], the United States Department of Agriculture (USDA) gives no specific recommendations for MUFA [16].In addition, the National Institute of Medicine did not mention any specific recommendations for MUFA.In their statement, they concluded that -n-9 cis Monounsaturated fatty acids are synthesized by the body and have no known independent beneficial role in human health and are not required in the diet.‖Therefore, neither an Adequate Intake nor a Recommended Daily Allowance was set.Since there is insufficient evidence for an Upper Level as well, the Dietary Reference Intakes did not consider MUFA at all [12].In accordance with these proceedings and with a similar rationale, the European Food and Safety Authority (EFSA) skipped MUFA in their scientific opinions on dietary reference values for fat [8].On a national level, the recommendations given in European countries are far from being conclusive.The Italian Society of Human Nutrition did not list any specific references for MUFA [9].The Joint Committees of Germany, Austria, and Switzerland stated that MUFA consumption should be 10% of TEC, albeit with higher intakes being acceptable [10].The Nordic Nutrition Recommendations agreed on 10%-15% of TEC in the form of MUFA [13].The particulars of the Dutch Dietary Guidelines proposed a limit of 38% of TEC in the form of MUFA and PUFA for people with optimal body weight, whereas overweight and obese people should be more restrictive and limit their daily energy uptake in the form of MUFA/PUFA to 28% [7].The UK COMA Committee advocated that MUFA (in the form of oleic acid) should provide an average of 12% of TEC [15].The Nutritional Recommendations for the French Population promoted an intake of MUFA up to 20% of TEC for adults including pregnant and lactating women.It was emphasized that the neutrality of oleic acid represents a benefit and that its consumption was justified [14].

Specific Guidelines for the Prevention and Treatment of Diabetes
On closer examination, the MUFA recommendations of the American Diabetes Association (ADA) evolved in a -downhill‖ fashion.In 2002, a consumption of 10%-20% of TEC in the form of MUFA was proposed [18].Two years later, a carbohydrate plus MUFA intake of 60%-70% of TEC was regarded as an evidence-based nutrition principle for the prevention and treatment of diabetes [26].In 2008, the ADA's position statement did not offer a specific value for MUFA as a preventive or therapeutic tool any longer [19].Correspondingly, the American Association of Clinical Endocrinologists excluded MUFA in their medical guidelines for the management of diabetes [17].However, the Canadian Diabetes Association suggested frequently replacing SFA with MUFA for a successful nutritional management of diabetes mellitus [21].Likewise, the Joslin Diabetes Center recommended the consumption of MUFA, again without any specific reference values [27].The Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes stated that MUFA should provide 10%-20% of TEC with total fat to be limited to 35% of TEC [22].The British Diabetes Association, probably referring to the 2004 nutrition principles of the ADA, recommended a daily amount of 60%-70% of TEC in the form of carbohydrates and MUFA, with MUFA values specified separately to aim at 10%-15% of TEC [20].In Japan, no specific quota of MUFA is given in as a nutritional reference, while other Asian nations like India allow for up to 7% of TEC in the form of MUFA [23,28].In South Africa, the corresponding authorities recommended <13% MUFA for diabetic subjects [29].

Risk Factors for Diabetes and Cardiovascular Disease
The National Cholesterol Education Program guidelines have outlined risk factors that increase CHD risk over a 10 year period.Elevated LDL-cholesterol (>100 mg/dL) remains the strongest primary factor in predicting CHD and therefore is a primary target of therapy [11].However, as circulating triacyglycerols (TG) and HDL-cholesterol concentrations are critical risk factors in metabolic syndrome, the TC:HDL-cholesterol ratio has been expressed as a more valuable marker in determining CHD risk [30].Summing-up, elevated levels of TC, LDL-cholesterol and TG as well low levels of HDL-cholesterol are evidence-based risk factors of CVD [31][32][33].High levels of blood pressure are also associated with an increased mortality risk [34].In addition, the Emerging Risk Factor Collaboration indicated FG levels >100 mg/dL as a predictor of mortality [35].The Framingham Heart Study showed that impaired fasting glucose was associated with an aggravated risk of CHD in women [36].A meta-analysis of cohort studies including 44,158 individuals without diabetes found a significant association between glycosylated hemoglobin (HbA1c) and cardiovascular events as well as death [37].In another meta-analysis of observational studies, it was concluded that chronic hyperglycemia is associated with an increased risk of CVD in patients with type 2 diabetes mellitus (T2D) [38].Among women, high-sensitive C-reactive protein (CRP) was the strongest predictor of CVD, accompanied by TC, LDL-cholesterol, TC:HDL-cholesterol, and Apo B 100 [39].A recent meta-analysis indicated that Apo B is a more accurate marker of cardiovascular risk as compared to non-HDL-cholesterol (=TC-HDL-cholesterol), while the latter is still more accurate in comparison to LDL-cholesterol [40].Strong associations between low serum HDL-cholesterol/high serum LDL-cholesterol and the onset of abdominal aortic aneurysms prove the continuous validity of both markers as predictive risk factors [41].A collaborative analysis of individual data from 36 prospective studies involving more than 126,000 individuals, has demonstrated that circulating Lp(a) concentrations are correlated with an increased incidence of CHD and stroke independent from several conventional risk factors (including TC) [42].

Data Sources and Search Strategy
Electronic database MEDLINE (between 1966 and November 2012) was searched for systematic review and meta-analysis using following search terms -monounsaturated fatty acids‖, -monounsaturated fat‖ and -dietary fat‖ with no restriction to calendar data and language.Reference lists and relevant clinical guidelines were also searched.

Study Quality Assessment
Review quality was rated using a modified version of the Overview of Quality Assessment Questionnaire (OQAQ) including a bias tool [43] (Supplemental material, Table S1) as described recently [44].Results of OQAQ assessments are summarized in Table 4.It should be noted that the analyses considered were in some cases based on overlapping sets of trials.1993 [45] Overall 1995 [47] The purpose of this investigation was to 1995 [48] Conducted 1997 [49] The aim of this meta-analysis of metabolic ward studies is to provide reliable quantitative estimates of the relevance of dietary intake of fatty acids and dietary cholesterol to blood concentrations of total cholesterol and cholesterol fraction Design: dietary intervention studies conducted under controlled conditions that ensured compliance not analyzed / 9 Garg 1998 [50] Examining the effects of high carbohydrate low fat diets vs. high MUFA diets on metabolic indexes in T2D subjects Design: randomized, crossover trials using isoenergetic, weight maintaining diets not analyzed / 9 Mensink et al.
2003 [30] Combining 2007 [46] Comparing The present review included meta-analyses of intervention trials (randomized, non-randomized and crossover trials) and cohort studies.A common problem associated with cross-over trials is that of carry-over (a type of period-by-intervention interaction), but it seems only justifiable to exclude cross-over trials from a systematic review if the design is inappropriate within the clinical context [59].Duration of studies varied remarkably between the different meta-analyses as well as between the different within each meta-analysis.This represents a major problem especially when comparing intervention trials.Sensitivity analyses comparing short-vs.long-term studies might be used as an alternative approach to resolve this issue.Another issue associated with meta-analyses is heterogeneity of various aspects and characteristics of the study protocols, especially in nutritional intervention trials.Therefore, it is not surprising that the literature chosen for the present review varies regarding type(s) of diets used (MUFA vs. carbohydrates/PUFA/SFA/trans fatty acids), definitions of MUFA diets, and study population (healthy, overweight, or obese subjects, patients with T2D, abnormal glucose metabolism, or CVD).In addition, in most of the included meta-analyses differential compliance (drop outs) was not investigated.Another potential source of bias is measurement issues (especially of self-reported data, e.g., 24 h recalls, food records).Only few systematic reviews screened for the presence of publication bias by assessing the symmetry of the funnel plots in which mean differences were plotted against their corresponding standard errors.

Healthy Subjects
See Table 5 summarizes the study characteristics of the meta-analyses included in this review.For a better understanding of the categorization of meta-analyses and other scientific studies, the Levels of evidence by the Scottish Intercollegiate Guidelines Network are given in Table 6 [60].In their meta-analysis, Clarke et al. (1997) [49] investigated the effects of MUFA as well as SFA and PUFA on cardiovascular risk factors in non-diabetic subjects.In addition, liquid formula diets were included, although they were analyzed separately.Dietary protocols were mostly iso-energetic but differed with respect to study design: they included randomized crossover, randomized or matched parallel, non-randomized Latin square and non-randomized sequential attempts.The authors concluded that substitution of carbohydrates by MUFA (5% of TEC) had no significant effect on TC and LDL-cholesterol, but managed to increase HDL-cholesterol.With respect to PUFA-rich diets, TC and LDL-cholesterol were both decreased and HDL-cholesterol was augmented in solid food experiments [49].Clarke and Mozaffarian (2009) [55] observed that replacing hydrogenated fats with MUFA (1% of TEC) resulted in advantageous changes of several CVD risk factors like TC, LDL-cholesterol, HDL-cholesterol, TG, apoproteins A-1, B as well as B/A1, and lipoprotein (a) in 12 crossover and 1 parallel designed trials.Yu and co-workers (1995) [48] explored the results of 18 studies (again including crossover and parallel designed set-ups) enrolling a total of 804 healthy and normocholesterolemic participants.Following meta-regression, they observed that MUFA increased HDL-cholesterol and decreased TC and LDL-cholesterol.The corresponding effects of PUFA were more pronounced with respect to TC and LDL-cholesterol, but not to HDL-cholesterol [48].In 1992, a meta-analysis of short-term RCTs investigated the effects of dietary fatty acids as an iso-caloric substituent for carbohydrates on CVD risk factors.HDL-cholesterol levels were significantly augmented following the MUFA-rich diet, while levels of TG and the ratio of TC to HDL-cholesterol were significantly reduced, respectively [61].In 2003, the authors published an updated meta-analysis including 1672 instead of 682 participants and were able to confirm their previous results.In addition, they observed a significant improvement in LDL-cholesterol, Apo A-1, and Apo B following high-MUFA regimens [30].In a recent meta-analysis investigating the long-term (≥6 months) effects of high-(>12% MUFA) vs. low-(≤12% MUFA) MUFA diets on cardiovascular risk factors, we could show that high-MUFA diets significantly reduced systolic and diastolic blood pressure in overweight/obese subjects [57] thus confirming data previously reported by Shah et al. in 2007 [46].When MUFA-rich diets were compared with PUFA-rich onsets, no effects on HDL-cholesterol and LDL-cholesterol, but a borderline increase (p = 0.05) in TG could be observed [47].Hegsted et al. [45] analyzed metabolic studies and field trials and could not observe any impact of MUFA on TC, LDL-cholesterol, and HDL-cholesterol in their meta-regression.

Patients with Abnormal Glucose Metabolism/Diabetes Mellitus
In a recent meta-analysis of short-term RCTs (crossover and parallel study designs) with a duration between 10 days and 6 weeks enrolling 306 subjects with type 2 diabetes mellitus, a significant decrease in TG values following a MUFA-rich dietary regimen could be observed when compared with a low-fat/high carbohydrate diet [53].This is in congruence with data presented by Garg (1998) [50] reporting reduced fasting TG in patients with type 2 diabetes mellitus subjected to a weight maintenance diet following replacement of carbohydrates by MUFA [50].Moreover, improvements in FG and pre-prandial plasma glucose were shown, while no significant changes in fasting plasma insulin concentrations, fructosamine and HbA1c were observed.The high-MUFA protocols were accompanied with significantly lower values for TC and VLDL-cholesterol as well as increases in HDL-cholesterol, but were not correlated to changes in LDL-cholesterol.Comparison of high-(>12% MUFA) vs. low-(≤12% MUFA) MUFA diets on glycemic control in subjects with abnormal glucose metabolism revealed improvements in HbA1c and fasting glucose in diabetic subjects, but no differences in blood lipids were found [58,62].
With respect to short-term studies (2-12 weeks duration), comparison of low vs. moderate dietary fat content was performed in a meta-analysis by Cao et al. (2009) [51].Participants with and without diabetes and a body mass index ranging from 21.1 to 30.2 kg/m 2 were enrolled.The mean MUFA content in a correspondingly modified diet was 23.6% of TEC and 11.4% in the low-fat versions.In the healthy collective, HDL-cholesterol was significantly increased and TG levels were significantly decreased in the moderate fat groups as compared to low-fat diets.TC and LDL-cholesterol were reduced in a similar fashion following both dietary protocols (moderate and low fat).Patients with diabetes adopting the diet with a higher MUFA content established a significant increase in HDL-cholesterol as well, accompanied by a significant reduction in TG and a non-significant reduction in TC as compared to the low fat diets.TG response was even more pronounced in participants with diabetes as compared to healthy subjects [51].

Patients with CVD
In a prospective trial investigating the effects of a Mediterranean diet, the Lyon Diet Heart Study reported a benefit of increased MUFA intake in survivors first time myocardial infarction [63].Three recent meta-analyses of cohort studies investigated the effects of dietary fats on CHD events and cardiovascular death.Skeaff and Miller [56] did not observe any effects of MUFA-rich diets on relative risks of CHD events and death.Moreover, no differences between of high-and low-fat intake were registered [56].Jakobsen [52] performed a meta-analysis of cohort studies including 344,696 subjects.They postulated a positive correlation between MUFA-rich diets and risk of coronary events, but not between MUFA-rich diets and risk of coronary deaths.The authors explain that in the western diet, the MUFA supply is predominantly of animal origin resulting in a confounder that should be taken into consideration when comparing dietary fats.The usual source of MUFA/oleic acid is of vegetable origin.These results are in strong discrepancy with another recent meta-analysis of cohort studies, were Mente et al. [54] reported a correlation between MUFA uptake and a significant decrease in the relative risk for CHD.None of these three meta-analyses reported information regarding stroke or arrhythmic diseases, but included data for -hard‖ CHD endpoints like angina pectoris, sudden death, fatal and non-fatal myocardial infarction.In June 2012, the Cochrane Collaboration updated their systematic review and meta-analysis on the effects of low vs. modified fat diets on cardiovascular disease.The findings are suggestive of a small but potentially important reduction in cardiovascular risk on modification of dietary fat (but not reduction of total fat) in longer trials.However, no association between total fat content and risk of cardiovascular death and events were reported [64].

Conclusions
In comparison, a considerably larger number of meta-analyses explored the effects of PUFAs on maintenance or reduction of body weight as well as biomarkers of impaired glucose metabolism or CVD/CHD than there are systematic reviews and meta-analyses dealing with the corresponding impact of MUFAs.Consequently, the international recommendations for PUFA are more consistent than those for MUFA, averaging a value of 10% of TEC for healthy persons for the most part.If MUFA recommendations are given at all, they vary between 12% and 25% of TEC, equaling a remarkable range of ~30-70 g/day for a 2.500 kcal-diet.Prestigious authorities and organizations such as the National Institute of Medicine, the EFSA, the USDA and the ADA do not provide specific recommendation for MUFAs either for healthy people or for patients in need of diabetic or cardiovascular management.
In the present review, only meta-analyses were included, which indicates a high level of evidence, i.e., from 2+ to 1+++ according to the Scottish Intercollegiate Guidelines Network indicate levels of evidence (Table 6).Apart from the fact that several meta-analyses and meta-regressions observed benefits of MUFA on cardiovascular risk factors, it should be noted that most meta-analyses did not report significant negative effects of a MUFA-rich diet on CVD risk factors.With respect to the favorable influences of MUFA found in studies recruiting healthy volunteers or patients with diabetes and CHD respectively, some reservations still remain.Due to various inhomogeneities, the results of different studies are far from being conclusive.Thus, MUFA were compared to carbohydrate-rich diets, low fat diets or regimens focusing on PUFA or SFA.Moreover, the term MUFA-rich diet lacks a concrete definition leading to inconsistent amounts of MUFA used in the corresponding protocols.Some of the discrepancies in the findings of different studies can be explained by their uneven and maybe incompatible durations.Long-term biomarkers of glucose metabolism such as HbA1c will be most likely not or just slightly improved following short-term interventions of 2-6 weeks Nevertheless, in view of the importance of dietary interventions for the prevention and therapy of cardiovascular disease, monounsaturated fatty acid may represent a valuable tool in the modification of dietary regimens.There is strong evidence that by replacing SFA and carbohydrates with MUFA, various cardiovascular risk factors will be significantly improved.The results of the different meta-analyses addressed in this review point to a beneficial effect of MUFA-rich diets on systolic and diastolic blood pressure as well as parameters of glycemic control.On the other hand, the impact of MUFA on blood lipids is still discussed controversially.While TG levels were decreased and HDL-cholesterol levels were increased following short-term interventions with higher amounts of MUFA, these findings could not be confirmed in long-term study protocols.Thus, there is no unanimous rationale for MUFAs in a therapeutic regimen.However, since no detrimental effects of MUFA-rich diets were reported in the literature to date, there is no evidence speaking against the consideration of MUFAs in dietary guidelines.Further studies dealing with long-term effects of MUFA on biomarkers of obesity, diabetes, and cardiovascular diseases as well as clinical endpoints are needed to clarify the potential benefits of MUFA-rich diets in primary and secondary prevention.
evaluation of the rather extensive literature on the effects of dietary fatty acid composition and cholesterol on serum lipid concentration Design: metabolic studies (appear to have been done under rather careful control in which food was prepared and fed to the subjects); field trials (diet was modified by instructions or by a combination of instructions and provision of some foods) Combining results to derive equations that relate changes in the dietary fatty acid intake to changes in serum HDL-C, LDL-C, TC and TG Design: parallel design, crossover or Latin-square; -before and after‖ designs that lacked a control group were excluded.Diets enriched with very-long-chain (n-3) PUFA were also excluded address the controversy regarding a differential effect of MUFA vs. PUFA on serum lipids Design: randomized trials comparing a high-mono and high-poly fat diet; similar in all respects (isoenergetic, total fat content, SFA) except for levels of monounsaturated and polyunsaturated fat intake; minimum 10 subjects on each diet results to derive equations that relate changes in the dietary fatty acid intake to changes in serum HDL-C, LDL-C, TC and TG, Apo-B and Apo A-I, TC:HDL-C Design: parallel design, crossover or Latin-square; -before and after‖ designs that lacked a control group were excluded.Diets enriched with very-long-chain (n-3) PUFA were also excluded

Table 2 .
Fatty acid content of different oils, nuts, fruits, seeds and animal products.

Table 3 .
National and international MUFA recommendations for healthy adults and patients with diabetes.

Table 4 .
Qualitative aspects of the included systematic reviews and meta-analyses.
Clarke et al.

Table 5 .
Study characteristics of meta-analyses.

Table 6 .
Levels of evidence by the Scottish Intercollegiate Guidelines Network.1++High quality meta-analyses, systematic reviews of RCTs, or RCTs with a very low risk of bias 1+ Well conducted meta-analyses, systematic reviews, or RCTs with a low risk of bias 1− Meta-analyses, systematic reviews, or RCTs with a high risk of bias 2++ High quality systematic reviews of case control or cohort studies