Nuts and Cardio-Metabolic Disease: A Review of Meta-Analyses

Objectives: Accumulating epidemiological and intervention evidence suggest that nut consumption is associated with reduced incidence of some cardiometabolic diseases. However, to date no review of meta-analyses of epidemiological and intervention studies has evaluated the effects of nut consumption on cardiometabolic disease. Design/Results: Electronic searches for meta-analyses of epidemiological and intervention studies were undertaken in PubMed®/MEDLINE®. Meta-analyses of prospective studies show that nut consumption appears to be associated with reduced all-cause mortality by 19–20% (n = 6), cardiovascular disease (CVD) incidence (19%; n = 3) and mortality (25%; n = 3), coronary heart disease (CHD) incidence (20–34%; n = 2) and mortality (27–30%; n = 2) and stroke incidence (10–11%; n = 7) and mortality (18%; n = 2). No association between nut consumption and the risk of type 2 diabetes mellitus (T2DM) was observed in meta-analyses of prospective studies, whereas a decrease in fasting blood glucose ranging from 0.08 to 0.15 mmol/L was observed in 3 meta-analyses of intervention studies. In the interventions, nut consumption also had favorable effects on total cholesterol (0.021 to 0.28 mmol/L reduction from 8 meta-analyses of interventions) and low-density lipoprotein cholesterol (0.017 to 0.26 mmol/L reduction from 8 meta-analyses of interventions) and endothelial function (0.79 to 1.03% increase in flow-mediated dilation from 4 meta-analyses of interventions). Nut consumption did not significantly affect body weight. Nut consumption had no effect on inflammatory markers in intervention studies. The effect on blood pressure was inconsistent. A higher nut consumption was associated with a lower incidence of hypertension in prospective studies, while nut consumption did not improve blood pressure in intervention studies. Conclusions: Nut consumption appeared to be associated with lower all-cause mortality and CVD and CHD mortality. There was no association between nut consumption and the incidence of T2DM although fasting blood glucose is decreased in intervention studies. In intervention studies nuts lower total cholesterol and low-density lipoprotein cholesterol (LDL-C).


Introduction
Cardiometabolic disease is a combination of disorders such as excessive visceral fat, hypertension, atherogenic dyslipidemia and glucose intolesrance which together result in type 2 diabetes mellitus (T2DM) and cardiovascular disease (CVD) [1].
Nuts have been reported to reduce the risk of T2DM and CVD [2][3][4][5]. High levels of unsaturated fatty acids and low levels of saturated fatty acids and bioactive compounds (polyphenols, carotenoids, phytosterols, fiber and minerals) may contribute to cardio-protection through improvement of glycemic control, lipid profiles, weight, blood pressure, endothelial function, oxidative status, antioxidant and anti-inflammatory action [6,7].
Diabetes, a key cardiometabolic disease is a global health concern with an expected increase in prevalence from 8.4% in 2017 to 9.9% in 2045 [8]. People with T2DM have an increased risk of CVD and up to 80% of individuals with T2DM die from CVD [9]. For this reason, primary prevention of both T2DM and CVD is vital. A healthy diet is important for the prevention and treatment of cardiometabolic disease such as T2DM and CVD [10,11].
This review aims to provide an update on the impact of nut consumption on preventable cardiometabolic diseases by reviewing meta-analyses of epidemiological studies and intervention trials. The evidence for nuts influencing diabetes risk and glycemic control as well as for nuts reducing strokes and weight is very mixed so this review provides a comprehensive review of these controversial areas.

Methods
We examined meta-analyses of the effects of nut consumption on cardiometabolic diseases in humans. The literature search was performed on the PubMed ® /MEDLINE ® (https://www.ncbi.nlm. nih.gov/pubmed/) database restricted to full articles in English up to 29 September 2018. The search terms included meta-analysis combined with nut(s) or tree nut(s) or almond(s) or Brazil nut(s) or cashew nut(s) or hazelnut(s) or macadamia(s) or peanut(s) or pistachio(s) or walnut(s). Reference lists of selected papers were also examined. The titles of articles were initially screened and then abstracts and full-texts of literature were reviewed for final study selection. Inclusion criteria were nuts as outlined above, meta-analysis, all-cause mortality, mortality, incidence, CVD, coronary heart disease (CHD), stroke, T2DM, hypertension, metabolic syndrome, obesity, adiposity, glycemic control, glucose, lipids, blood pressure, inflammatory markers, endothelial function, flow-mediated dilation. Meta-analyses of human studies were included and Non-English articles were excluded. Thirty-four meta-analyses that specifically addressed these inclusion criteria were selected and included in this review. A flow chart of the study screening and selection process is reported in Figure 1.

Prospective Cohort Studies
All cohort mortality studies are summarized in Table 1. Forest plots for Table 1 are presented in Supplementary Figures S1-S10.

All-Cause Mortality
Meta-analyses [5,[12][13][14][15][16] of prospective studies that have examined the effect of nut consumption on all-cause mortality have found a~19% reduction in all-cause mortality comparing the highest to the lowest nut intakes. Three meta-analyses [13,14,16] of prospective studies have reported a~21% reduction in all-cause mortality per one serving (28 g)/day increase. All meta-analyses show excellent consistency although the number of studies varied in each one.
Heterogeneity was higher in this analysis [13] but not in Chen et al. 2017 [12]. The higher heterogeneity is explained by between-study differences such as shorter-term follow-up and geographic location (with big differences between European and Asian studies) [13].

Cause-Specific Mortality Diabetes Mortality
Only one meta-analysis [14] examined the association between nut consumption and diabetes mortality showing a 32% reduction in diabetes deaths comparing the highest to the lowest nut intakes. It is difficult to know what the nature of a "diabetic death" was the International Classification of Diseases (ICD) code of 250 is very non-specific. The total number of deaths from diabetes were low in each paper. In addition, the study from Luu et al. 2015 was not significant in the original paper whereas it is shown as significant in the meta-analysis casting severe doubt on the conclusions of the meta-analysis.

CVD Mortality
Three meta-analyses [5,12,15] of prospective studies reported a 19-25% lower rate of CVD mortality comparing highest to lowest consumption.

Stroke Mortality
A meta-analysis including 12 prospective studies [12] showed an 18% reduction in stroke mortality while an earlier meta-analysis [5] which included only 3 studies showed a nonsignificant effect.
Chen et al. 2017 [12] showed a lower risk of stroke mortality after nut consumption with a RR of 0.82 (95% CI 0.73, 0.91; p-het = 0.82; I 2 = 0%) from 12 studies (9 publications) [17][18][19]21,25,27,28,33,34] involving 449,293 subjects with 4398 deaths in a comparison of highest with lowest total nut consumption. On the other hand, a statistically nonsignificant reduction (RR = 0.83; 95% CI 0.69, 1.00) in stroke mortality was observed in analyses of both highest-versus-lowest consumption and dose-response conducted by Mayhew et al. 2016 [5] which only included 4 studies. A summary of nut consumption and incidence of T2DM, CVD, hypertension, CHD and stroke is shown in Table 2. Most meta-analyses were analyzed with prospective studies while two meta-analyses by Afshine et al. 2014 [57] and Weng et al. 2016 [58] were analyzed with prospective and intervention studies in Table 2. Forest plots for Table 2 are presented in Supplementary Figures S11-S21.

Cardiovascular Disease
Three meta-analyses [5,14,16] of prospective studies that examined high and low consumption of nuts have shown a lower incidence of CVD. Aune et al. 2016 [14] showed a 19% reduction while Mayhew et al. 2016 [5] and Luo et al. 2014 [16] showed a 44% and a 30% reduction, respectively. However, Mayhew et al. 2016 [5] and Luo et al. 2014 [16] just analyzed only one and four studies, respectively.
Hypertension Meta-analyses of prospective studies [63,64,67] have found a~15% lower risk of hypertension comparing the highest to the lowest nut intakes.
Guo et al. 2015 [63] found that nut consumption, especially at >2 servings/week was associated with a lower risk of hypertension. Comparing the highest with the lowest categories of nut consumption, a 16% decrease in the incidence of hypertension was observed (RR = 0.84; 95% CI: 0.76, 0.93) in 3 studies from 3 publications [68,70,72].

Stroke
All meta-analyses apart from two [14,75] which showed a 10-11% reduction in stroke incidence, have found no effect of nuts on stroke incidence. However, the meta-analysis of Aune et al. 2016 [14] is much more convincing as it included a greater number of studies compared with other meta-analyses [16,57,64,73].

Body Weight
A summary of nut consumption and anthropometric parameters in RCTs is shown in Table 3. Forest plots for Table 3 are presented in Supplementary Figures S22-S24.

Glycemic Control
Three meta-analyses [3,144,178] of intervention studies have found nuts lowered fasting glucose levels with a reduction ranging from 0.08 to 0.15 mmol/L compared with control diets. In three meta-analyses, control diets were a low-fat diet, a high-fat diet, a libitum diet without nuts, a refined carbohydrate snack, a diet without nuts and a usual diet. The type of nuts were almonds, walnuts, cashews, hazelnuts, cashew and pistachios. The study durations ranged from 4 weeks to 12 months.
Viguiliouk et al. 2014 [3] focused on subjects with T2DM in a meta-analysis of 12 RCTs (n = 450) of ≥3 weeks duration to investigate the effects of tree nut consumption on glycemic control (glycosylated hemoglobin (HbA1c), fasting glucose, fasting insulin and homeostasis model assessment of insulin resistance index (HOMA-IR)) compared with isocaloric diets without tree nuts. This meta-analysis showed that tree nut consumption at a median dose of 56 g/day over a median follow-up period of~8 weeks significantly decreased HbA1c (MD = −0.07%; p = 0.0003) and fasting glucose (MD = −0.15 mmol/L; p = 0.03) compared with control diets. However, fasting insulin and HOMA-IR were not altered. Most RCTs analyzed in this meta-analysis had short study periods and poor study quality [3].
In this review, the TC-lowering effect of nuts ranged from 0.021 to 0.28 mmol/L and LDL-C was lowered by 0.017 to 0.26 mmol/L comparing nuts to control diets. Interestingly, Banel et al. 2009 [112] found a bigger reduction in TC comparing walnuts with control diets. Part of the reason for the bigger reduction may be related to larger amount of walnuts consumed (30-108 g/day representing 5-25% of total energy).
A sub-analysis showed that this decreased effect of nuts was observed in the long-term studies (≥12 weeks) and in parallel RCTs. No significant effects on C-reactive protein (CRP), tumor necrosis factor alpha (TNF-α), vascular cell adhesion molecule (VCAM), IL-6, E-selectin were observed [184].
Overall it would appear that nuts have very little effects on inflammatory markers.

Discussion
Nuts are comprised of unique macro and micronutrients (e.g., monounsaturated fatty acids, polyunsaturated fatty acids, fiber, arginine and magnesium) and bioactive compounds (e.g., phytosterols and polyphenols). The synergistic beneficial effects of nuts on cardiometabolic disease result from these nutrient-and non-nutrient compositions of nuts. The potential mechanisms underlying these effects can be explained through the improvement of glucose, lipids, weight, endothelial function and gut microbiota modification [7,233,234]. The most important and most convincingly demonstrated is the reduction in LDL cholesterol which appears to be similar with all nuts and to be clinically significant.
This review highlights findings from meta-analyses of prospective studies that indicate that nut consumption appears to be associated with reduced all-cause mortality, CVD mortality and incidence, coronary heart disease incidence and mortality stroke incidence and mortality. No association between nut consumption and the incidence of T2DM was found in meta-analyses of prospective studies, even though the fasting glucose levels were reduced by a small degree in intervention studies. Moreover, the association between nut consumption and the incidence of hypertension was found in meta-analyses of prospective studies, whereas blood pressure was not favorably altered in meta-analyses of intervention studies.
In this review we have highlighted differences between meta-analyses and explained why they may differ and come to a conclusion about the likely effects of nuts. No other review has focused entirely on meta-analyses both of all nuts and individual nuts. In general, the more recent meta-analyses are larger and hopefully contain all the previous primary studies in earlier meta-analyses but sometimes this is not the case. Where the meta-analyses are reasonably large the results should be similar even if a few studies are added or omitted. All of the meta-analyses on total mortality and cause-specific mortality assessed the study quality using the Newcastle-Ottawa scale or a very similar scale and found no effect of study quality on the outcomes of the meta-analyses. There appears to be little differences between different nut types.
The major areas of new work required are intervention studies to examine the effects of nuts on glucose homeostasis and the incidence of T2DM in at risk populations as well as clarifying the effects on weight and blood pressure as here there are disparities between cohort studies and interventions.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/10/12/1935/ s1. Figure S1. Total nuts and all-cause mortality, high vs. low analysis, Figure S2. Total nuts and all-cause mortality, dose-response analysis, Figure S3. Peanuts and all-cause mortality, high vs. low analysis, Figure S4. Tree nuts and all-cause mortality, high vs. low analysis, Figure S5. Total nuts and coronary heart disease mortality, high vs. low analysis, Figure S6. Total nuts and coronary heart disease mortality, dose-response analysis, Figure S7. Total nuts and cardiovascular disease mortality, high vs. low analysis, Figure S8. Total nuts and cardiovascular disease mortality, dose-response analysis, Figure S9. Total nuts and stroke mortality, high vs. low analysis, Figure  S10. Total nuts and stroke mortality, dose-response analysis, Figure S11. Total nuts and type 2 diabetes, high vs. low analysis, Figure S12. Total nuts and type 2 diabetes, dose-response analysis, Figure S13. Total nuts and total cardiovascular disease, high vs. low analysis, Figure S14. Total nuts and total cardiovascular disease, dose-response analysis, Figure S15. Total nuts and hypertension, high vs. low analysis, Figure S16. Total nuts and hypertension, dose-response analysis, Figure S17. Total nuts and total coronary heart disease, high vs. low analysis, Figure S18. Total nuts and total coronary heart disease, dose-response analysis, Figure S19. Total nuts and non-fatal/fatal coronary heart disease, dose-response analysis, Figure S20. Total nuts and total stroke, high vs. low analysis, Figure S21. Total nuts and total stroke, dose-response analysis, Figure S22. Nuts and body weight, Figure S23. Nuts and waist circumference, Figure S24. Nuts and body mass index, Figure S25. Nuts and fasting blood glucose, Figure S26. Nuts and total cholesterol, Figure S27. Nuts and low-density lipoprotein cholesterol, Figure S28. Nuts and high-density lipoprotein cholesterol, Figure S29. Nuts and triglycerides, Figure S30. Nuts and apolipoprotein A-1, Figure S31. Nuts and apolipoprotein B, Figure S32. Nuts and apolipoprotein B 100, Figure S33. Nuts and systolic blood pressure, Figure S34. Nuts and diastolic blood pressure, Figure S35. Nuts and adiponectin, Figure S36. Nuts and C-reactive protein (CRP), Figure S37. Nuts and interleukin-6 (IL-6), Figure S38. Nuts and tumor necrosis factor-alpha (TNF-α), Figure S39. Nuts and intracellular adhesion molecule-1 (ICAM-1), Figure S40. Nuts and vascular cell adhesion molecule-1 (VCAM-1), Figure S41. Nuts and E-selectin, and Figure  S42. Nuts and flow-mediated dilation (FMD).
Author Contributions: All authors (Y.K., J.K. and P.M.C.) conceived of the manuscript structure and contributed to the writing and editing.
Funding: This research received no external funding.