AJCN. First published ahead of print May 7, 2014 as doi: 10.3945/ajcn.113.079152.
Nut consumption in relation to cardiovascular risk and type 2 diabetes: a systematic review and meta-analysis of prospective studies1–5 Donghao Zhou, Haibing Yu, Fang He, Kathleen Heather Reilly, Jingling Zhang, Shuangshuang Li, Tao Zhang, Baozhen Wang, Yuanlin Ding, and Bo Xi ABSTRACT Background: Many prospective cohort studies have investigated the association between nut consumption and risk of coronary artery disease (CAD), stroke, hypertension, and type 2 diabetes (T2D). However, results have been inconsistent. Objective: We aimed to investigate the association between nut consumption and risk of CAD, stroke, hypertension, and T2D. Design: PubMed and EMBASE databases were searched up to October 2013. All prospective cohort studies of nut consumption and risk of CAD, stroke, hypertension, and T2D were included. Summary RRs with 95% CIs were estimated by using a fixed- or random-effects model. Results: A total of 23 prospective studies (9 studies for CAD, 4 studies for stroke, 4 studies for hypertension, and 6 studies for T2D) from 19 publications were included in the meta-analysis. There were 179,885 participants and 7236 CAD cases, 182,730 participants and 5669 stroke cases, 40,102 participants and 12,814 hypertension cases, and 342,213 participants and 14,400 T2D cases. The consumption of each 1-serving nuts/d was significantly associated with incident CAD (RR: 0.81; 95% CI: 0.72, 0.91; P , 0.001) and hypertension (RR: 0.66; 95% CI: 0.44, 1.00; P = 0.049). However, there was no association between the consumption of each 1 serving nuts/d and risk of stroke (RR: 0.90; 95% CI: 0.71, 1.14) or T2D (RR: 0.80; 95% CI: 0.57, 1.14). Conclusions: A higher consumption of nuts was associated with reduced risk of CAD and hypertension but not stroke or T2D. Large randomized controlled trials are warranted to confirm the observed associations. Am J Clin Nutr doi: 10.3945/ajcn.113.079152.
INTRODUCTION
Nuts contain a variety of nutrients, including unsaturated fatty acids, fiber, vitamins E and riboflavin, folate, minerals (potassium, copper, selenium, calcium, and magnesium), and other bioactive compounds such as phytosterols and polyphenols (1). Many prospective cohort studies have investigated the association between nut consumption and risk of coronary artery disease (CAD)6 (2–9). In 1992, Fraser et al (2) first reported an inverse relation between nut consumption and risk of CAD in a prospective study of 31,208 non-Hispanic whites. Since then, many cohort studies have attempted to replicate this association. However, results have been inconsistent (3–9). A previous metaanalysis of 4 cohort studies suggested an average risk reduction of CAD death by 37% for the highest compared with lowest amounts of nut consumption and a reduction of 8.3% in risk of
CAD death for each weekly serving of nuts (10). However, 2 of the included 4 cohort studies have updated the data with a longer duration of follow-up and more incident cases for each study (6, 7). Furthermore, several additional, large cohort studies have been published on nut consumption and CAD risk (5, 8, 9). Thus, a sufficient statistical power has been achieved to further clarify the association. In addition, a great number of large cohorts have investigated the association of nut consumption with stroke (11–13), hypertension (14–17), and type 2 diabetes (T2D) (18–22). However, these results have also been mixed. Although one recent meta-analysis assessed the association of nut consumption with risk of hypertension and T2D (23), the article was later retracted by the authors, and readers cannot access the full text of the related results again. Thus, it is necessary to further clarify the association. Therefore, in this study, we performed a systematic review and meta-analysis of prospective cohort studies to clarify the association 1
From the Department of Endocrinology, Linyi People’s Hospital, Linyi, China (DZ and JZ); the School of Public Health, Southern Medical University, Guangzhou, China (HY); the Institute of Medical Systems Biology, School of Public Health, Guangdong Medical College, Dongguan, China (HY and YD); the Department of Internal Medicine, Linyi People’s Hospital, Linyi, China (FH); Independent Consultant, New York (KHR); the Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, Jinan, China (SL, TZ, and BX); and the Department of Nutrition and Food Hygiene (BW), School of Public Health, Shandong University, Jinan, China. 2 DZ and HY are joint first authors. 3 Funders had no role in the study design, implementation, analysis, and interpretation of data. 4 Supported by the National Natural Science Foundation of China (81273166), Key Projects of Science and Technology Plan Medical and Scientific Research Units of Dongguan (2012105102010), the Science and Technology Innovation Team Project in Doctor Granting of Guangdong Medical College (STIF201121), the Science and Technology Innovation Project of Guangdong Province (2013KJCX0089), and the Key Disciplines of Guangdong Medical College Project (GDMCKSF201105). 5 Address correspondence to Y Ding, Institute of Medical Systems Biology, School of Public Health, Guangdong Medical College, Dongguan 523808, China. E-mail: [email protected]; or B Xi, Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, 44 Wenhuaxi Road, Jinan 250012, China. E-mail: [email protected]. 6 Abbreviations used: CAD, coronary artery disease; PREDIMED, Prevencio´n con Dieta Mediterra´nea; T2D, type 2 diabetes. Received October 30, 2013. Accepted for publication April 21, 2014. doi: 10.3945/ajcn.113.079152.
Am J Clin Nutr doi: 10.3945/ajcn.113.079152. Printed in USA. Ó 2014 American Society for Nutrition
Copyright (C) 2014 by the American Society for Nutrition
1 of 8
2 of 8
ZHOU ET AL
of nut consumption with risk of CAD, stroke, hypertension, and T2D. METHODS
Literature and search strategy Literature databases including PubMed (http://www.ncbi.nlm. nih.gov/pubmed) and EMBASE (http://www.elsevier.com/online-tools/embase) were searched. Search terms were (coronary disease OR myocardial infarction OR stroke OR hypertension OR type 2 diabetes) and (nut OR nuts). Reference lists of retrieved articles were manually searched. The literature search was limited to the English language. If more than one article was published that used the same cohort, only the study with the largest sample size was included. The literature search was updated on 10 October 2013. Inclusion criteria and data extraction Studies included in the meta-analysis met all of the following inclusion criteria: 1) evaluated the association between nut consumption and CAD, stroke, hypertension, or T2D; 2) used a prospective cohort design; and 3) provided the amount of nut consumption, distributions of cases and person years, and RRs or HRs with 95% CIs for $3 exposure categories. The following information was extracted from each study: 1) name of the first author, 2) year of publication, 3) study name, 4) origin of country, 5) sex of participants, 6) age of the study population at baseline, 7) follow-up period, 8) number of cases and study population, 9) endpoints, 10) the amount of nut consumption for each category, 11) RRs or HRs with 95% CIs for all categories of nut consumption, and 12) covariates used in adjustments. Two authors (BX and SL) independently assessed the articles for compliance with the inclusion and exclusion criteria and resolved disagreements through discussion. The quality of each study was assessed by using the NewcastleOttawa quality scale (24), which is a validated scale for nonrandomized studies in a meta-analysis. This scale was used to assign a maximum of 9 points for each study. Three broad perspectives were considered: the selection of the cohorts, comparability of cohorts; and ascertainment of the exposure and outcome of interest.
was assigned to the corresponding RR or HR with the 95% CI for each study. When nut consumption was reported by ranges of intakes, the midpoint of the range was used. When the highest category was open ended, we assumed the width of the category to be the same as the adjacent category. When the lowest category was open ended, we set the lower boundary to zero. If a linear dose-response result was already provided by the included study, it was used directly. A potential curvelinear doseresponse relation between nut consumption and risk of CAD, stroke, hypertension, or T2D was examined by using restricted cubic splines with 3 knots at percentiles 10%, 50%, and 90% of the distribution (30). A P value for curvelinearity or nonlinearity was calculated by testing the null hypothesis that the coefficient of the second spline was equal to zero. A meta-regression analysis was performed to identify the source of heterogeneity (31). A sensitivity analysis after the exclusion of one study at a time was performed to assess the stability of results. Publication bias was assessed by using Begg’s (32) and Egger’s (33) tests (P , 0.05 was considered significant). The statistical analysis was conducted with STATA version 11 software (StataCorp LP). RESULTS
Characteristics of included prospective studies After the literature search and selection, a total of 23 prospective studies from 19 publications were included in the metaanalysis of the association between nut consumption and risk of CAD, stroke, hypertension, and T2D (Figure 1). Note that 2 publications (2, 34) were excluded because neither study provided sufficient data for dose-response analyses. The duration of follow-up ranged from 3.8 to 26 y. For CAD, 3 studies were from Europe, and 6 studies were from the United States. For stroke, all 4 studies were from the United States. For hypertension, 3 studies were from the United States and 1 study was from Spain. For T2D, 5 studies were from the United States and 1 study was from China. Included studies have adjusted for most important covariates, but some studies did not adjust for energy intake, BMI, and age. Characteristics of included prospective studies are listed in Table 1 (also see Supplemental Table 1 under “Supplemental data” in the online issue).
Statistical analysis
Association between nut consumption and risk of CAD
A fixed-effects (25) or random-effects (26) model, on the basis of whether there was heterogeneity, was used to calculate pooled RRs with 95% CIs for highest compared with lowest amounts of nuts consumption and the dose-response analysis. Heterogeneity was assessed by using the Q test and I2 statistic (27). The significance for the Q test was defined as P , 0.10. The I2 statistic represents the amount of total variation that could be attributed to heterogeneity. I2 values .25%, .50%, and .75% indicated little, moderate, and significant heterogeneity, respectively. For the dose-response analysis, the method reported by Greenland et al (28) and Orsini et al (29) was used to calculate study specific slopes (linear trends) on the basis of results across categories of nut consumption. We extracted data on the amount of nut consumption, distributions of cases and person years, and RRs or HRs with 95% CIs for $3 exposure categories. The median or mean amount of nut consumption in each category
Highest compared with lowest intakes Nine cohort studies from 7 publications, including 179,885 participants and 7236 CAD cases, were included. The highest intake of nuts was inversely associated with risk of CAD (RR: 0.83; 95% CI: 0.74, 0.93; see Supplemental Figure 1 under “Supplemental data” in the online issue) compared with the that of the lowest intake with modest evidence of heterogeneity (I2 = 59.9%, P = 0.010). After the sensitivity analysis, RRs (95% CIs) ranged from 0.86 (0.77, 0.96) to 0.81 (0.71, 0.91). There was no evidence of a publication bias (Begg’s test: P = 0.076; Egger’s test: P = 0.142). Dose-response analysis We did not find a curvelinear association between nut consumption and risk of CAD (P-nonlinearity . 0.05). The summary
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
3 of 8
each 1-serving/d consumption, Figure 3) with modest evidence of heterogeneity (I2 = 49.6%, P = 0.114). No publication bias was detected by using Begg’s (P = 0.308) or Egger’s (P = 0.273) test. We did not perform additional subgroup analyses by the type of stroke (hemorrhagic compared with ischemic) because there were only a few stroke studies. Association between nut consumption and risk of hypertension Highest compared with lowest intakes Four cohort studies from 4 publications, including 40,102 participants and 12,814 hypertension cases, were included. The highest intake of nuts was inversely associated with risk of hypertension (RR: 0.85; 95% CI: 0.79, 0.92; see Supplemental Figure 3 under “Supplemental data” in the online issue) compared with that for the lowest intake with no evidence of heterogeneity (I2 = 0.0%, P = 0.927). After the sensitivity analysis, RRs (95% CIs) ranged from 0.86 (0.79, 0.94) to 0.84 (0.76, 0.93). There was no evidence of a publication bias (Begg’s test: P = 0.734; Egger’s test: P = 0.537). Dose-response analysis
FIGURE 1. Flowchart of publications included in the systematic review. PubMed: http://www.ncbi.nlm.nih.gov/pubmed; EMBASE: http://www. elsevier.com/online-tools/embase. CAD, coronary artery disease; CVD, cardiovascular disease.
We showed no evidence of a curvelinear association between nut consumption and risk of hypertension (P-nonlinearity . 0.05). The summary RR for each 1-serving/d intake was 0.66 (95% CI: 0.44, 1.00; P = 0.049; Figure 4) with evidence of heterogeneity (I2 = 75.9%, P = 0.006). A sensitivity analysis was performed after the exclusion one study at a time. Summary RRs (95% CIs) ranged from 0.67 (0.38, 1.17) to 0.55 (0.38, 0.82). No publication bias was detected by using Begg’s (P = 0.734) or Egger’s (P = 0.195) test. Association between nut consumption and risk of T2D Highest compared with lowest intakes
RR for each 1-serving/d intake was 0.81 (95% CI: 0.72, 0.91; P , 0.001; Figure 2) with modest evidence of heterogeneity (I2 = 56.8%, P = 0.018). A sensitivity analysis was performed after the exclusion of one study at a time. Summary RRs (95% CIs) ranged from 0.84 (0.74, 0.94) to 0.78 (0.69, 0.89). No publication bias was detected by using Begg’s (P = 0.251) or Egger’s (P = 0.141) test.
Six cohort studies from 5 publications, including 342,213 participants and 14,400 T2D cases, were included. The highest intake of nuts was not associated with risk of T2D (RR: 0.92; 95% CI: 0.78, 1.09; see Supplemental Figure 4 under “Supplemental data” in the online issue) compared with that for the lowest intake with high evidence of heterogeneity (I2 = 78.7%, P , 0.001). There was no evidence of a publication bias (Begg’s test: P = 1.000, Egger’s test: P = 0.806).
Association between nut consumption and risk of stroke Highest compared with lowest intakes
Dose-response analysis
Four cohort studies from 3 publications, 182,730 participants, and 5669 stroke cases were included. There was no association between nut intake and risk of stroke (highest compared with lowest categories: RR, 0.87; 95% CI, 0.74, 1.03; see Supplemental Figure 2 under “Supplemental data” in the online issue), with no evidence of heterogeneity (I2 = 15.0%, P = 0.317). There was no evidence of a publication bias (Begg’s test: P = 0.734; Egger’s test: P = 0.550).
We showed no evidence of a curvelinear association between nut consumption and risk of T2D (P-nonlinearity . 0.05). There was no significant association between higher nut consumption and risk of T2D (for each 1-serving/d consumption: RR: 0.80 (95% CI: 0.57, 1.14); P = 0.217; Figure 5) with significant evidence of heterogeneity (I2 = 87.1%, P , 0.001). No publication bias was detected by using Begg’s (P = 0.707) or Egger’s (P = 0.697) test. Potential between-study heterogeneity
Dose-response analysis We showed no evidence of a curvelinear association between nut consumption and risk of stroke (P-nonlinearity . 0.05). There was no significant association between higher nut consumption and risk of stroke [RR: 0.90 (95% CI: 0.71, 1.14) for
Because there was evidence of between-study heterogeneity for CAD, stroke, and T2D, a meta-regression was performed to examine the source of heterogeneity. The introduced independent variables included the publication year, subtype endpoint, sex of participants, duration of follow-up, and quality score of studies.
EPIC-Greece Study (CAD) Iowa Women’s Health Study Nurses’ Health Study (Diabetes) Nurses’ Health Study EPIC-Greece Study Iowa Women’s Health Study Physicians’ Health Study I Nurses’ Health Study Health Professionals Study CARDIA study Physicians’ Health Study I SUN project ARIC study Nurses’ Health Study Iowa Women’s Health Study Shanghai Women’s Health Study Physicians’ Health Study I Nurses’ Health Study Nurses’ Health Study II
Cohort of vegetarians and other health conscious people Physicians’ Health Study I
Study name
Greece United States United States United States Greece United States United States United States United States United States United States Spain United States United States United States China United States United States United States
United States
United Kingdom
Country
M F F F M F M F M M M M M F F F M F F
M
and F and F
and F
and F
and F
M and F
Sex
20–86 55–69 57 6 9 30–55 20–86 55–69 40–84 30–55 40–75 18–30 40–84 36 6 10 45–64 34–59 55–69 40–70 41–87 52–77 35–52
1302 31,778 6309 84,136 23,929 34,492 21,078 84,010 43,150 4304 15,966 9919 9913 83,818 35,988 64,227 20,224 58,063 79,893
21,454
10,977
y $16 40–84
No. of participants
Actual age at baseline
3.8 15 12 26 10 11 21.1 26 22 15 15 4.3 9 16 12 4.6 19.2 10 10
17
y 17
Duration of follow-up
Nonfatal myocardial infarction Coronary heart disease mortality Coronary heart disease mortality Coronary heart disease mortality Myocardial infarction Coronary heart disease Coronary heart disease Stroke mortality Stroke Stroke Stroke Hypertension3 Hypertension4 Hypertension4 Hypertension4 Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes
Ischemic heart disease mortality
Endpoints
1037 566 85 948 452 3162 636 215 1424 2633 1397 997 8423 541 2853 3206 1831 1605 1828 3166 2764
350
No. of cases
7 7 6 7 6 7 8 7 7 7 7 7 7 7 8 7 7 7 7 7 7
7
Study quality2
ARIC, Atherosclerosis Risk in Communities; CARDIA, Coronary Artery Risk Development in Young Adults; CAD, coronary heart disease; EPIC, European Prospective Investigation Into Cancer and Nutrition; ref, reference; SUN, Seguimiento Universidad de Navarra. 2 Assessed by using the Newcastle-Ottawa quality scale (24). 3 Systolic blood pressure/diastolic blood pressure $130/85 mm Hg or the use of antihypertensive medications. 4 Systolic blood pressure/diastolic blood pressure $140/90 mm Hg or the use of antihypertensive medications.
1
Steffen, 2006 (14) Djousse´, 2009 (15) Martı´nez -Lapiscina, 2010 (16) Weng, 2013 (17) Jiang, 2002 (18) Parker, 2003 (19) Villegas, 2008 (20) Kochar, 2010 (21) Pan, 2013 (22)
Trichopoulou, 2005 (5) Blomhoff, 2006 (6) Li, 2009 (7) Bernstein, 2010 (8) Dilis, 2012 (9) Yochum, 2000 (11) Djousse´, 2010 (12) Bernstein, 2012 (13)
Albert, 2002 (4)
Key, 1996 (3)
First author, publication year (ref)
TABLE 1 Characteristics of included prospective studies on the association of nut consumption with cardiovascular risk and type 2 diabetes1
4 of 8 ZHOU ET AL
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
5 of 8
FIGURE 2. Meta-analysis of nut consumption with risk of coronary heart disease (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. CHD, coronary heart disease; ID, identifier; MI, myocardial infarction.
However, none of these variables were identified as the source of heterogeneity for all 4 outcomes (all P . 0.05).
DISCUSSION
The current meta-analysis supported an inverse association between nut consumption and risk of CAD and hypertension, but we did not show any significant association between nut consumption and risk of stroke or T2D.
Comparison with other studies A previous meta-analysis that consisted of 4 cohort studies suggested that individuals who consume nuts $4 times/wk had a 37% reduction in risk of CAD in comparison with individuals who consumed nuts seldom or never with an average reduction of 8.3% for each weekly serving of nuts (10). In the current meta-analysis, we calculated summary RR with 95% CIs for each 1-serving/d increment of nut consumption. Our doseresponse meta-analysis suggested a 19% average risk reduction
FIGURE 3. Meta-analysis of nut consumption with risk of stroke (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
6 of 8
ZHOU ET AL
FIGURE 4. Meta-analysis of nut consumption with risk of hypertension (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
of CAD for each 1-serving/d consumption of nuts. In other words, our findings confirmed the beneficial effect of nut consumption on risk of CAD. In addition, the inverse association between nut consumption and hypertension was also observed. Note that the study by Djousse´ et al (15) accounted for the largest proportion in weights of all 3 studies (95.12%). However, after the exclusion of the study in the sensitivity analysis, the inverse association remained, which suggested the stability of the result. For stroke, although only 4 cohort studies were identified, the numbers of participants and incident stroke cases were
182,730 and 5669, respectively. Therefore, the null significant association of nut consumption with risk of stroke could not be attributed to the statistical power, and the possible reasons need additional investigation. It was also surprising to us that nut consumption was not shown to be significantly associated with T2D, although an inverse trend was observed. Because there was significant heterogeneity between studies, the pooled negative result should be interpreted with caution. More recently, 2 large cohort studies (35, 36) showed that an increased frequency of nut consumption was associated with
FIGURE 5. Meta-analysis of nut consumption with risk of type 2 diabetes (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
7 of 8
significantly reduced risk of all-cause mortality, although the significant association did not persist in some cause-specific mortality groups in one study (35). In the Prevencio´n con Dieta Mediterra´nea (PREDIMED) nutrition intervention trial (median follow-up: 4.8 y), a total of 7216 participants aged 55–80 y were randomly assigned to 1 of 3 interventions (Mediterranean diets supplemented with nuts or olive oil and a control diet). The multivariable-adjusted HR of main cardiovascular events was 0.72 (95% CI: 0.54, 0.96) for the group assigned to a Mediterranean diet with nuts compared with the control group (37). An additional analysis showed that the Mediterranean diet with nuts was only protective against the development of stroke but not myocardial infarction or deaths from CVDs (37). However, our findings in the current meta-analysis suggested that nut intake was not inversely associated with risk of stroke, but it was beneficial to CAD. This difference was not surprising to us because the PREDIMED trial was not powered for individual outcomes of the composite endpoint. The potential explanation was unknown. In the PREDIMED trial cohort, considered as a longitudinal study, individuals who consumed nuts $1 times/wk had significantly lower cardiovascular mortality than did individuals who never ate nuts (36). In addition, the Mediterranean diet enriched with nuts showed a beneficial effect on risk of T2D compared with for the control group in the PREDIMED nutrition intervention trial (RR: 0.48; 95% CI: 0.24, 0.96) (38). However, in our current meta-analysis, we did not show the protective effect of nuts on T2D. Additional large-scale randomized controlled trials are necessary to validate the findings of the PREDIMED nutrition intervention trial.
cluded studies did not provide data on types of nuts consumed and their preparation including salted, spiced, roasted, or raw nuts. Thus, we are unable to examine the influence of types of nuts or preparation methods on outcomes. Third, the majority of studies used the nut consumption at baseline as the dietary exposure. However, during the follow-up period, some individuals might have changed their diet habits. Thus, the misclassification of exposure might have biased the results. This possibility may help explain the null association with stroke or T2D but could not explain the significant results for CAD or hypertension because a misclassification is expected to bias the results toward the null. In a previous study, a diet rich in unsaturated fat that is the main component of nuts had a large beneficial effect on blood pressure and serum lipids (41). Fourth, there was evidence of betweenstudy heterogeneity for CAD, stroke, and T2D. Although a meta-regression model was applied to examine the source of heterogeneity for these diseases, none of the introduced variables could explain the heterogeneity, which suggested other unknown factors might have confounded these associations. Fifth, included studies adjusted for important covariates (see Table S1 under “Supplemental data” in the online issue) but not all studies adjusted for all covariates. For example, some but not all studies adjusted for energy intake, BMI, and age. Thus, the combined results should be interpreted with caution. In conclusion, our meta-analysis suggests a dose-response association between nut consumption and decreased risk of CAD and hypertension. The results do not support any significant association between nut consumption and risk of stroke or T2D. Large random controlled trials are warranted to confirm the observed association.
Potential mechanisms
The authors’ responsibilities were as follows—BX and YD: designed the research and had primary responsibility for the final content of the manuscript; SL, TZ, and JZ: conducted the research; FH, BX, BW, and KHR: analyzed data; DZ, HY, and BX: wrote the manuscript; and all authors: read and approved the final manuscript. None of the authors had a conflict of interest.
It has been well established that nuts are rich in MUFAs, PUFAs, fiber, minerals, vitamins, and many bioactive compounds. These healthy nutritional components contained in nuts may explain the inverse association with CAD and hypertension. The possible mechanisms include improving the blood lipid profile, decreasing insulin resistance, and modulating inflammation, oxidative stress, and endothelial function. For instance, in a cohort of participants at high risk of CAD, nut consumption was inversely associated with vascular cell adhesion molecule-1, intracellular adhesion molecule-1, C-reactive protein, and IL-6 serum concentrations, although this association was only significant for intracellular adhesion molecule-1 (39). In a clinical trial, a healthy diet supplemented with nuts decreased the insulin response by 22% from the initial value, and the response was significantly different from that of the control group (40). Additional experimental studies and clinical trials are necessary to clarify the mechanisms on protective effect of nut consumption on CAD and hypertension. Strengths and limitations The main strength of the current study was that the long duration of follow-up, high follow-up rate, and large number of participants in included studies provided sufficient power to detect the association. However, several limitations should be considered. First, because of the observational nature of the cohort study, the effect of unmeasured or residual confounding on the observed findings could not be ruled out. Second, most in-
REFERENCES 1. Ros E, Tapsell LC, Sabate J. Nuts and berries for heart health. Curr Atheroscler Rep 2010;12:397–406. 2. Fraser GE, Sabate´ J, Beeson WL, Strahan TM. A possible protective effect of nut consumption on risk of coronary heart disease. The Adventist Health Study. Arch Intern Med 1992;152:1416–24. 3. Key TJ, Thorogood M, Appleby PN, Burr ML. Dietary habits and mortality in 11,000 vegetarians and health conscious people: results of a 17 year follow up. BMJ 1996;313:775–9. 4. Albert CM, Gaziano JM, Willett WC, Manson JE. Nut consumption and decreased risk of sudden cardiac death in the Physicians’ Health Study. Arch Intern Med 2002;162:1382–7. 5. Trichopoulou A, Bamia C, Trichopoulos D. Mediterranean diet and survival among patients with coronary heart disease in Greece. Arch Intern Med 2005;165:929–35. 6. Blomhoff R, Carlsen MH, Andersen LF, Jacobs DR Jr. Health benefits of nuts: potential role of antioxidants. Br J Nutr 2006;96:S52–60. 7. Li TY, Brennan AM, Wedick NM, Mantzoros C, Rifai N, Hu FB. Regular consumption of nuts is associated with a lower risk of cardiovascular disease in women with type 2 diabetes. J Nutr 2009;139: 1333–8. 8. Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major dietary protein sources and risk of coronary heart disease in women. Circulation 2010;122:876–83. 9. Dilis V, Katsoulis M, Lagiou P, Trichopoulos D, Naska A, Trichopoulou A. Mediterranean diet and CHD: the Greek European Prospective Investigation into Cancer and Nutrition cohort. Br J Nutr 2012;108: 699–709.
8 of 8
ZHOU ET AL
10. Kelly JH Jr, Sabate´ J. Nuts and coronary heart disease: an epidemiological perspective. Br J Nutr 2006;96(suppl 2):S61–7. 11. Yochum LA, Folsom AR, Kushi LH. Intake of antioxidant vitamins and risk of death from stroke in postmenopausal women. Am J Clin Nutr 2000;72:476–83. 12. Djousse´ L, Gaziano JM, Kase CS, Kurth T. Nut consumption and risk of stroke in US male physicians. Clin Nutr 2010;29:605–9. 13. Bernstein AM, Pan A, Rexrode KM, Stampfer M, Hu FB, Mozaffarian D, Willett WC. Dietary protein sources and the risk of stroke in men and women. Stroke 2012;43:637–44. 14. Steffen LM, Kroenke CH, Yu X, Pereira MA, Slattery ML, Van Horn L, Gross MD, Jacobs DR Jr. Associations of plant food, dairy product, and meat intakes with 15-y incidence of elevated blood pressure in young black and white adults: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Clin Nutr 2005;82:1169–77. 15. Djousse´ L, Rudich T, Gaziano JM. Nut consumption and risk of hypertension in US male physicians. Clin Nutr 2009;28:10–4. 16. Martı´nez-Lapiscina EH, Pimenta AM, Beunza JJ, Bes-Rastrollo M, Martı´nez JA, Martı´nez-Gonza´lez MA. Nut consumption and incidence of hypertension: the SUN prospective cohort. Nutr Metab Cardiovasc Dis 2010;20:359–65. 17. Weng LC, Steffen LM, Szklo M, Nettleton J, Chambless L, Folsom AR. A diet pattern with more dairy and nuts, but less meat is related to lower risk of developing hypertension in middle-aged adults: the Atherosclerosis Risk in Communities (ARIC) study. Nutrients 2013;5:1719–33. 18. Jiang R, Manson JE, Stampfer MJ, Liu S, Willett WC, Hu FB. Nut and peanut butter consumption and risk of type 2 diabetes in women. JAMA 2002;288:2554–60. 19. Parker ED, Harnack LJ, Folsom AR. Nut consumption and risk of type 2 diabetes. JAMA 2003;290:38–9; author reply 39–40. 20. Villegas R, Gao YT, Yang G, Li HL, Elasy TA, Zheng W, Shu XO. Legume and soy food intake and the incidence of type 2 diabetes in the Shanghai Women’s Health Study. Am J Clin Nutr 2008;87:162–7. 21. Kochar J, Gaziano JM, Djousse´ L. Nut consumption and risk of type II diabetes in the Physicians’ Health Study. Eur J Clin Nutr 2010;64: 75–9. 22. Pan A, Sun Q, Manson JE, Willett WC, Hu FB. Walnut consumption is associated with lower risk of type 2 diabetes in women. J Nutr 2013; 143:512–8. 23. Guo K, Jiang Y, Zhou Z, Li Y. RETRACTED: Nut consumption with risk of hypertension and type 2 diabetes mellitus: a meta-analysis of prospective cohort studies. Eur J Prev Cardiol (Epub ahead of print 30 August 2013). 24. Wells GA, Shea B, O’connell D, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Available from: http://www.ohri.ca/ programs/clinical_epidemiology/oxford.asp (cited 14 October 2013). 25. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88.
26. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48. 27. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. 28. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301–9. 29. Orsini N, Bellocco R, Greenland S. Generalized least squares for trend estimation of summarized dose-response data. Stata J 2006;6:40–57. 30. Harrell FE Jr, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst 1988;80:1198–202. 31. Martin B. Meta-analysis: dealing with heterogeneity. York, United Kingdom: University of York, 2006:6. 32. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. 33. Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34. 34. Montonen J, Ja¨rvinen R, Helio¨vaara M, Reunanen A, Aromaa A, Knekt P. Food consumption and the incidence of type II diabetes mellitus. Eur J Clin Nutr 2005;59:441–8. 35. Baer HJ, Glynn RJ, Hu FB, Hankinson SE, Willett WC, Colditz GA, Stampfer M, Rosner B. Risk factors for mortality in the nurses’ health study: a competing risks analysis. Am J Epidemiol 2011;173:319–29. ´ , Ros E, Corella D, 36. Guasch-Ferre´ M, Bullo´ M, Martı´nez-Gonza´lez MA Estruch R, Fito´ M, Aro´s F, Wa¨rnberg J, Fiol M, et al.. Frequency of nut consumption and mortality risk in the PREDIMED nutrition intervention trial. BMC Med 2013;11:164. 37. Estruch R, Ros E, Salas-Salvado´ J, Covas MI, Corella D, Aro´s F, Go´mez-Gracia E, Ruiz-Gutie´rrez V, Fiol M, Lapetra J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 2013;368:1279–90. ´ , Ibarrola38. Salas-Salvado´ J, Bullo´ M, Babio N, Martı´nez-Gonza´lez MA Jurado N, Basora J, Estruch R, Covas MI, Corella D, Aro´s F, et al. Reduction in the incidence of type 2 diabetes with the Mediterranean diet: results of the PREDIMED-Reus nutrition intervention randomized trial. Diabetes Care 2011;34:14–9. 39. Salas-Salvado´ J, Garcia-Arellano A, Estruch R, Marquez-Sandoval F, Corella D, Fiol M, Go´mez-Gracia E, Vin˜oles E, Aro´s F, Herrera C, et al. Components of the Mediterranean-type food pattern and serum inflammatory markers among patients at high risk for cardiovascular disease. Eur J Clin Nutr 2008;62:651–9. 40. Casas-Agustench P, Bullo´ M, Salas-Salvado´ J. Nuts, inflammation and insulin resistance. Asia Pac J Clin Nutr 2010;19:124–30. 41. Appel LJ, Sacks FM, Carey VJ, Obarzanek E, Swain JF, Miller ER 3rd, Conlin PR, Erlinger TP, Rosner BA, Laranjo NM, et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. JAMA 2005;294:2455–64.
Nut consumption in relation to cardiovascular risk and type 2 diabetes: a systematic review and meta-analysis of prospective studies1–5 Donghao Zhou, Haibing Yu, Fang He, Kathleen Heather Reilly, Jingling Zhang, Shuangshuang Li, Tao Zhang, Baozhen Wang, Yuanlin Ding, and Bo Xi ABSTRACT Background: Many prospective cohort studies have investigated the association between nut consumption and risk of coronary artery disease (CAD), stroke, hypertension, and type 2 diabetes (T2D). However, results have been inconsistent. Objective: We aimed to investigate the association between nut consumption and risk of CAD, stroke, hypertension, and T2D. Design: PubMed and EMBASE databases were searched up to October 2013. All prospective cohort studies of nut consumption and risk of CAD, stroke, hypertension, and T2D were included. Summary RRs with 95% CIs were estimated by using a fixed- or random-effects model. Results: A total of 23 prospective studies (9 studies for CAD, 4 studies for stroke, 4 studies for hypertension, and 6 studies for T2D) from 19 publications were included in the meta-analysis. There were 179,885 participants and 7236 CAD cases, 182,730 participants and 5669 stroke cases, 40,102 participants and 12,814 hypertension cases, and 342,213 participants and 14,400 T2D cases. The consumption of each 1-serving nuts/d was significantly associated with incident CAD (RR: 0.81; 95% CI: 0.72, 0.91; P , 0.001) and hypertension (RR: 0.66; 95% CI: 0.44, 1.00; P = 0.049). However, there was no association between the consumption of each 1 serving nuts/d and risk of stroke (RR: 0.90; 95% CI: 0.71, 1.14) or T2D (RR: 0.80; 95% CI: 0.57, 1.14). Conclusions: A higher consumption of nuts was associated with reduced risk of CAD and hypertension but not stroke or T2D. Large randomized controlled trials are warranted to confirm the observed associations. Am J Clin Nutr doi: 10.3945/ajcn.113.079152.
INTRODUCTION
Nuts contain a variety of nutrients, including unsaturated fatty acids, fiber, vitamins E and riboflavin, folate, minerals (potassium, copper, selenium, calcium, and magnesium), and other bioactive compounds such as phytosterols and polyphenols (1). Many prospective cohort studies have investigated the association between nut consumption and risk of coronary artery disease (CAD)6 (2–9). In 1992, Fraser et al (2) first reported an inverse relation between nut consumption and risk of CAD in a prospective study of 31,208 non-Hispanic whites. Since then, many cohort studies have attempted to replicate this association. However, results have been inconsistent (3–9). A previous metaanalysis of 4 cohort studies suggested an average risk reduction of CAD death by 37% for the highest compared with lowest amounts of nut consumption and a reduction of 8.3% in risk of
CAD death for each weekly serving of nuts (10). However, 2 of the included 4 cohort studies have updated the data with a longer duration of follow-up and more incident cases for each study (6, 7). Furthermore, several additional, large cohort studies have been published on nut consumption and CAD risk (5, 8, 9). Thus, a sufficient statistical power has been achieved to further clarify the association. In addition, a great number of large cohorts have investigated the association of nut consumption with stroke (11–13), hypertension (14–17), and type 2 diabetes (T2D) (18–22). However, these results have also been mixed. Although one recent meta-analysis assessed the association of nut consumption with risk of hypertension and T2D (23), the article was later retracted by the authors, and readers cannot access the full text of the related results again. Thus, it is necessary to further clarify the association. Therefore, in this study, we performed a systematic review and meta-analysis of prospective cohort studies to clarify the association 1
From the Department of Endocrinology, Linyi People’s Hospital, Linyi, China (DZ and JZ); the School of Public Health, Southern Medical University, Guangzhou, China (HY); the Institute of Medical Systems Biology, School of Public Health, Guangdong Medical College, Dongguan, China (HY and YD); the Department of Internal Medicine, Linyi People’s Hospital, Linyi, China (FH); Independent Consultant, New York (KHR); the Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, Jinan, China (SL, TZ, and BX); and the Department of Nutrition and Food Hygiene (BW), School of Public Health, Shandong University, Jinan, China. 2 DZ and HY are joint first authors. 3 Funders had no role in the study design, implementation, analysis, and interpretation of data. 4 Supported by the National Natural Science Foundation of China (81273166), Key Projects of Science and Technology Plan Medical and Scientific Research Units of Dongguan (2012105102010), the Science and Technology Innovation Team Project in Doctor Granting of Guangdong Medical College (STIF201121), the Science and Technology Innovation Project of Guangdong Province (2013KJCX0089), and the Key Disciplines of Guangdong Medical College Project (GDMCKSF201105). 5 Address correspondence to Y Ding, Institute of Medical Systems Biology, School of Public Health, Guangdong Medical College, Dongguan 523808, China. E-mail: [email protected]; or B Xi, Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, 44 Wenhuaxi Road, Jinan 250012, China. E-mail: [email protected]. 6 Abbreviations used: CAD, coronary artery disease; PREDIMED, Prevencio´n con Dieta Mediterra´nea; T2D, type 2 diabetes. Received October 30, 2013. Accepted for publication April 21, 2014. doi: 10.3945/ajcn.113.079152.
Am J Clin Nutr doi: 10.3945/ajcn.113.079152. Printed in USA. Ó 2014 American Society for Nutrition
Copyright (C) 2014 by the American Society for Nutrition
1 of 8
2 of 8
ZHOU ET AL
of nut consumption with risk of CAD, stroke, hypertension, and T2D. METHODS
Literature and search strategy Literature databases including PubMed (http://www.ncbi.nlm. nih.gov/pubmed) and EMBASE (http://www.elsevier.com/online-tools/embase) were searched. Search terms were (coronary disease OR myocardial infarction OR stroke OR hypertension OR type 2 diabetes) and (nut OR nuts). Reference lists of retrieved articles were manually searched. The literature search was limited to the English language. If more than one article was published that used the same cohort, only the study with the largest sample size was included. The literature search was updated on 10 October 2013. Inclusion criteria and data extraction Studies included in the meta-analysis met all of the following inclusion criteria: 1) evaluated the association between nut consumption and CAD, stroke, hypertension, or T2D; 2) used a prospective cohort design; and 3) provided the amount of nut consumption, distributions of cases and person years, and RRs or HRs with 95% CIs for $3 exposure categories. The following information was extracted from each study: 1) name of the first author, 2) year of publication, 3) study name, 4) origin of country, 5) sex of participants, 6) age of the study population at baseline, 7) follow-up period, 8) number of cases and study population, 9) endpoints, 10) the amount of nut consumption for each category, 11) RRs or HRs with 95% CIs for all categories of nut consumption, and 12) covariates used in adjustments. Two authors (BX and SL) independently assessed the articles for compliance with the inclusion and exclusion criteria and resolved disagreements through discussion. The quality of each study was assessed by using the NewcastleOttawa quality scale (24), which is a validated scale for nonrandomized studies in a meta-analysis. This scale was used to assign a maximum of 9 points for each study. Three broad perspectives were considered: the selection of the cohorts, comparability of cohorts; and ascertainment of the exposure and outcome of interest.
was assigned to the corresponding RR or HR with the 95% CI for each study. When nut consumption was reported by ranges of intakes, the midpoint of the range was used. When the highest category was open ended, we assumed the width of the category to be the same as the adjacent category. When the lowest category was open ended, we set the lower boundary to zero. If a linear dose-response result was already provided by the included study, it was used directly. A potential curvelinear doseresponse relation between nut consumption and risk of CAD, stroke, hypertension, or T2D was examined by using restricted cubic splines with 3 knots at percentiles 10%, 50%, and 90% of the distribution (30). A P value for curvelinearity or nonlinearity was calculated by testing the null hypothesis that the coefficient of the second spline was equal to zero. A meta-regression analysis was performed to identify the source of heterogeneity (31). A sensitivity analysis after the exclusion of one study at a time was performed to assess the stability of results. Publication bias was assessed by using Begg’s (32) and Egger’s (33) tests (P , 0.05 was considered significant). The statistical analysis was conducted with STATA version 11 software (StataCorp LP). RESULTS
Characteristics of included prospective studies After the literature search and selection, a total of 23 prospective studies from 19 publications were included in the metaanalysis of the association between nut consumption and risk of CAD, stroke, hypertension, and T2D (Figure 1). Note that 2 publications (2, 34) were excluded because neither study provided sufficient data for dose-response analyses. The duration of follow-up ranged from 3.8 to 26 y. For CAD, 3 studies were from Europe, and 6 studies were from the United States. For stroke, all 4 studies were from the United States. For hypertension, 3 studies were from the United States and 1 study was from Spain. For T2D, 5 studies were from the United States and 1 study was from China. Included studies have adjusted for most important covariates, but some studies did not adjust for energy intake, BMI, and age. Characteristics of included prospective studies are listed in Table 1 (also see Supplemental Table 1 under “Supplemental data” in the online issue).
Statistical analysis
Association between nut consumption and risk of CAD
A fixed-effects (25) or random-effects (26) model, on the basis of whether there was heterogeneity, was used to calculate pooled RRs with 95% CIs for highest compared with lowest amounts of nuts consumption and the dose-response analysis. Heterogeneity was assessed by using the Q test and I2 statistic (27). The significance for the Q test was defined as P , 0.10. The I2 statistic represents the amount of total variation that could be attributed to heterogeneity. I2 values .25%, .50%, and .75% indicated little, moderate, and significant heterogeneity, respectively. For the dose-response analysis, the method reported by Greenland et al (28) and Orsini et al (29) was used to calculate study specific slopes (linear trends) on the basis of results across categories of nut consumption. We extracted data on the amount of nut consumption, distributions of cases and person years, and RRs or HRs with 95% CIs for $3 exposure categories. The median or mean amount of nut consumption in each category
Highest compared with lowest intakes Nine cohort studies from 7 publications, including 179,885 participants and 7236 CAD cases, were included. The highest intake of nuts was inversely associated with risk of CAD (RR: 0.83; 95% CI: 0.74, 0.93; see Supplemental Figure 1 under “Supplemental data” in the online issue) compared with the that of the lowest intake with modest evidence of heterogeneity (I2 = 59.9%, P = 0.010). After the sensitivity analysis, RRs (95% CIs) ranged from 0.86 (0.77, 0.96) to 0.81 (0.71, 0.91). There was no evidence of a publication bias (Begg’s test: P = 0.076; Egger’s test: P = 0.142). Dose-response analysis We did not find a curvelinear association between nut consumption and risk of CAD (P-nonlinearity . 0.05). The summary
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
3 of 8
each 1-serving/d consumption, Figure 3) with modest evidence of heterogeneity (I2 = 49.6%, P = 0.114). No publication bias was detected by using Begg’s (P = 0.308) or Egger’s (P = 0.273) test. We did not perform additional subgroup analyses by the type of stroke (hemorrhagic compared with ischemic) because there were only a few stroke studies. Association between nut consumption and risk of hypertension Highest compared with lowest intakes Four cohort studies from 4 publications, including 40,102 participants and 12,814 hypertension cases, were included. The highest intake of nuts was inversely associated with risk of hypertension (RR: 0.85; 95% CI: 0.79, 0.92; see Supplemental Figure 3 under “Supplemental data” in the online issue) compared with that for the lowest intake with no evidence of heterogeneity (I2 = 0.0%, P = 0.927). After the sensitivity analysis, RRs (95% CIs) ranged from 0.86 (0.79, 0.94) to 0.84 (0.76, 0.93). There was no evidence of a publication bias (Begg’s test: P = 0.734; Egger’s test: P = 0.537). Dose-response analysis
FIGURE 1. Flowchart of publications included in the systematic review. PubMed: http://www.ncbi.nlm.nih.gov/pubmed; EMBASE: http://www. elsevier.com/online-tools/embase. CAD, coronary artery disease; CVD, cardiovascular disease.
We showed no evidence of a curvelinear association between nut consumption and risk of hypertension (P-nonlinearity . 0.05). The summary RR for each 1-serving/d intake was 0.66 (95% CI: 0.44, 1.00; P = 0.049; Figure 4) with evidence of heterogeneity (I2 = 75.9%, P = 0.006). A sensitivity analysis was performed after the exclusion one study at a time. Summary RRs (95% CIs) ranged from 0.67 (0.38, 1.17) to 0.55 (0.38, 0.82). No publication bias was detected by using Begg’s (P = 0.734) or Egger’s (P = 0.195) test. Association between nut consumption and risk of T2D Highest compared with lowest intakes
RR for each 1-serving/d intake was 0.81 (95% CI: 0.72, 0.91; P , 0.001; Figure 2) with modest evidence of heterogeneity (I2 = 56.8%, P = 0.018). A sensitivity analysis was performed after the exclusion of one study at a time. Summary RRs (95% CIs) ranged from 0.84 (0.74, 0.94) to 0.78 (0.69, 0.89). No publication bias was detected by using Begg’s (P = 0.251) or Egger’s (P = 0.141) test.
Six cohort studies from 5 publications, including 342,213 participants and 14,400 T2D cases, were included. The highest intake of nuts was not associated with risk of T2D (RR: 0.92; 95% CI: 0.78, 1.09; see Supplemental Figure 4 under “Supplemental data” in the online issue) compared with that for the lowest intake with high evidence of heterogeneity (I2 = 78.7%, P , 0.001). There was no evidence of a publication bias (Begg’s test: P = 1.000, Egger’s test: P = 0.806).
Association between nut consumption and risk of stroke Highest compared with lowest intakes
Dose-response analysis
Four cohort studies from 3 publications, 182,730 participants, and 5669 stroke cases were included. There was no association between nut intake and risk of stroke (highest compared with lowest categories: RR, 0.87; 95% CI, 0.74, 1.03; see Supplemental Figure 2 under “Supplemental data” in the online issue), with no evidence of heterogeneity (I2 = 15.0%, P = 0.317). There was no evidence of a publication bias (Begg’s test: P = 0.734; Egger’s test: P = 0.550).
We showed no evidence of a curvelinear association between nut consumption and risk of T2D (P-nonlinearity . 0.05). There was no significant association between higher nut consumption and risk of T2D (for each 1-serving/d consumption: RR: 0.80 (95% CI: 0.57, 1.14); P = 0.217; Figure 5) with significant evidence of heterogeneity (I2 = 87.1%, P , 0.001). No publication bias was detected by using Begg’s (P = 0.707) or Egger’s (P = 0.697) test. Potential between-study heterogeneity
Dose-response analysis We showed no evidence of a curvelinear association between nut consumption and risk of stroke (P-nonlinearity . 0.05). There was no significant association between higher nut consumption and risk of stroke [RR: 0.90 (95% CI: 0.71, 1.14) for
Because there was evidence of between-study heterogeneity for CAD, stroke, and T2D, a meta-regression was performed to examine the source of heterogeneity. The introduced independent variables included the publication year, subtype endpoint, sex of participants, duration of follow-up, and quality score of studies.
EPIC-Greece Study (CAD) Iowa Women’s Health Study Nurses’ Health Study (Diabetes) Nurses’ Health Study EPIC-Greece Study Iowa Women’s Health Study Physicians’ Health Study I Nurses’ Health Study Health Professionals Study CARDIA study Physicians’ Health Study I SUN project ARIC study Nurses’ Health Study Iowa Women’s Health Study Shanghai Women’s Health Study Physicians’ Health Study I Nurses’ Health Study Nurses’ Health Study II
Cohort of vegetarians and other health conscious people Physicians’ Health Study I
Study name
Greece United States United States United States Greece United States United States United States United States United States United States Spain United States United States United States China United States United States United States
United States
United Kingdom
Country
M F F F M F M F M M M M M F F F M F F
M
and F and F
and F
and F
and F
M and F
Sex
20–86 55–69 57 6 9 30–55 20–86 55–69 40–84 30–55 40–75 18–30 40–84 36 6 10 45–64 34–59 55–69 40–70 41–87 52–77 35–52
1302 31,778 6309 84,136 23,929 34,492 21,078 84,010 43,150 4304 15,966 9919 9913 83,818 35,988 64,227 20,224 58,063 79,893
21,454
10,977
y $16 40–84
No. of participants
Actual age at baseline
3.8 15 12 26 10 11 21.1 26 22 15 15 4.3 9 16 12 4.6 19.2 10 10
17
y 17
Duration of follow-up
Nonfatal myocardial infarction Coronary heart disease mortality Coronary heart disease mortality Coronary heart disease mortality Myocardial infarction Coronary heart disease Coronary heart disease Stroke mortality Stroke Stroke Stroke Hypertension3 Hypertension4 Hypertension4 Hypertension4 Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes Type 2 diabetes
Ischemic heart disease mortality
Endpoints
1037 566 85 948 452 3162 636 215 1424 2633 1397 997 8423 541 2853 3206 1831 1605 1828 3166 2764
350
No. of cases
7 7 6 7 6 7 8 7 7 7 7 7 7 7 8 7 7 7 7 7 7
7
Study quality2
ARIC, Atherosclerosis Risk in Communities; CARDIA, Coronary Artery Risk Development in Young Adults; CAD, coronary heart disease; EPIC, European Prospective Investigation Into Cancer and Nutrition; ref, reference; SUN, Seguimiento Universidad de Navarra. 2 Assessed by using the Newcastle-Ottawa quality scale (24). 3 Systolic blood pressure/diastolic blood pressure $130/85 mm Hg or the use of antihypertensive medications. 4 Systolic blood pressure/diastolic blood pressure $140/90 mm Hg or the use of antihypertensive medications.
1
Steffen, 2006 (14) Djousse´, 2009 (15) Martı´nez -Lapiscina, 2010 (16) Weng, 2013 (17) Jiang, 2002 (18) Parker, 2003 (19) Villegas, 2008 (20) Kochar, 2010 (21) Pan, 2013 (22)
Trichopoulou, 2005 (5) Blomhoff, 2006 (6) Li, 2009 (7) Bernstein, 2010 (8) Dilis, 2012 (9) Yochum, 2000 (11) Djousse´, 2010 (12) Bernstein, 2012 (13)
Albert, 2002 (4)
Key, 1996 (3)
First author, publication year (ref)
TABLE 1 Characteristics of included prospective studies on the association of nut consumption with cardiovascular risk and type 2 diabetes1
4 of 8 ZHOU ET AL
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
5 of 8
FIGURE 2. Meta-analysis of nut consumption with risk of coronary heart disease (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. CHD, coronary heart disease; ID, identifier; MI, myocardial infarction.
However, none of these variables were identified as the source of heterogeneity for all 4 outcomes (all P . 0.05).
DISCUSSION
The current meta-analysis supported an inverse association between nut consumption and risk of CAD and hypertension, but we did not show any significant association between nut consumption and risk of stroke or T2D.
Comparison with other studies A previous meta-analysis that consisted of 4 cohort studies suggested that individuals who consume nuts $4 times/wk had a 37% reduction in risk of CAD in comparison with individuals who consumed nuts seldom or never with an average reduction of 8.3% for each weekly serving of nuts (10). In the current meta-analysis, we calculated summary RR with 95% CIs for each 1-serving/d increment of nut consumption. Our doseresponse meta-analysis suggested a 19% average risk reduction
FIGURE 3. Meta-analysis of nut consumption with risk of stroke (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
6 of 8
ZHOU ET AL
FIGURE 4. Meta-analysis of nut consumption with risk of hypertension (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
of CAD for each 1-serving/d consumption of nuts. In other words, our findings confirmed the beneficial effect of nut consumption on risk of CAD. In addition, the inverse association between nut consumption and hypertension was also observed. Note that the study by Djousse´ et al (15) accounted for the largest proportion in weights of all 3 studies (95.12%). However, after the exclusion of the study in the sensitivity analysis, the inverse association remained, which suggested the stability of the result. For stroke, although only 4 cohort studies were identified, the numbers of participants and incident stroke cases were
182,730 and 5669, respectively. Therefore, the null significant association of nut consumption with risk of stroke could not be attributed to the statistical power, and the possible reasons need additional investigation. It was also surprising to us that nut consumption was not shown to be significantly associated with T2D, although an inverse trend was observed. Because there was significant heterogeneity between studies, the pooled negative result should be interpreted with caution. More recently, 2 large cohort studies (35, 36) showed that an increased frequency of nut consumption was associated with
FIGURE 5. Meta-analysis of nut consumption with risk of type 2 diabetes (for 1-serving/d increment). Size of symbol is proportional to inverse of variance of RR; horizontal line represents 95% CI. ID, identifier.
NUTS AND CARDIOVASCULAR RISK, TYPE 2 DIABETES
7 of 8
significantly reduced risk of all-cause mortality, although the significant association did not persist in some cause-specific mortality groups in one study (35). In the Prevencio´n con Dieta Mediterra´nea (PREDIMED) nutrition intervention trial (median follow-up: 4.8 y), a total of 7216 participants aged 55–80 y were randomly assigned to 1 of 3 interventions (Mediterranean diets supplemented with nuts or olive oil and a control diet). The multivariable-adjusted HR of main cardiovascular events was 0.72 (95% CI: 0.54, 0.96) for the group assigned to a Mediterranean diet with nuts compared with the control group (37). An additional analysis showed that the Mediterranean diet with nuts was only protective against the development of stroke but not myocardial infarction or deaths from CVDs (37). However, our findings in the current meta-analysis suggested that nut intake was not inversely associated with risk of stroke, but it was beneficial to CAD. This difference was not surprising to us because the PREDIMED trial was not powered for individual outcomes of the composite endpoint. The potential explanation was unknown. In the PREDIMED trial cohort, considered as a longitudinal study, individuals who consumed nuts $1 times/wk had significantly lower cardiovascular mortality than did individuals who never ate nuts (36). In addition, the Mediterranean diet enriched with nuts showed a beneficial effect on risk of T2D compared with for the control group in the PREDIMED nutrition intervention trial (RR: 0.48; 95% CI: 0.24, 0.96) (38). However, in our current meta-analysis, we did not show the protective effect of nuts on T2D. Additional large-scale randomized controlled trials are necessary to validate the findings of the PREDIMED nutrition intervention trial.
cluded studies did not provide data on types of nuts consumed and their preparation including salted, spiced, roasted, or raw nuts. Thus, we are unable to examine the influence of types of nuts or preparation methods on outcomes. Third, the majority of studies used the nut consumption at baseline as the dietary exposure. However, during the follow-up period, some individuals might have changed their diet habits. Thus, the misclassification of exposure might have biased the results. This possibility may help explain the null association with stroke or T2D but could not explain the significant results for CAD or hypertension because a misclassification is expected to bias the results toward the null. In a previous study, a diet rich in unsaturated fat that is the main component of nuts had a large beneficial effect on blood pressure and serum lipids (41). Fourth, there was evidence of betweenstudy heterogeneity for CAD, stroke, and T2D. Although a meta-regression model was applied to examine the source of heterogeneity for these diseases, none of the introduced variables could explain the heterogeneity, which suggested other unknown factors might have confounded these associations. Fifth, included studies adjusted for important covariates (see Table S1 under “Supplemental data” in the online issue) but not all studies adjusted for all covariates. For example, some but not all studies adjusted for energy intake, BMI, and age. Thus, the combined results should be interpreted with caution. In conclusion, our meta-analysis suggests a dose-response association between nut consumption and decreased risk of CAD and hypertension. The results do not support any significant association between nut consumption and risk of stroke or T2D. Large random controlled trials are warranted to confirm the observed association.
Potential mechanisms
The authors’ responsibilities were as follows—BX and YD: designed the research and had primary responsibility for the final content of the manuscript; SL, TZ, and JZ: conducted the research; FH, BX, BW, and KHR: analyzed data; DZ, HY, and BX: wrote the manuscript; and all authors: read and approved the final manuscript. None of the authors had a conflict of interest.
It has been well established that nuts are rich in MUFAs, PUFAs, fiber, minerals, vitamins, and many bioactive compounds. These healthy nutritional components contained in nuts may explain the inverse association with CAD and hypertension. The possible mechanisms include improving the blood lipid profile, decreasing insulin resistance, and modulating inflammation, oxidative stress, and endothelial function. For instance, in a cohort of participants at high risk of CAD, nut consumption was inversely associated with vascular cell adhesion molecule-1, intracellular adhesion molecule-1, C-reactive protein, and IL-6 serum concentrations, although this association was only significant for intracellular adhesion molecule-1 (39). In a clinical trial, a healthy diet supplemented with nuts decreased the insulin response by 22% from the initial value, and the response was significantly different from that of the control group (40). Additional experimental studies and clinical trials are necessary to clarify the mechanisms on protective effect of nut consumption on CAD and hypertension. Strengths and limitations The main strength of the current study was that the long duration of follow-up, high follow-up rate, and large number of participants in included studies provided sufficient power to detect the association. However, several limitations should be considered. First, because of the observational nature of the cohort study, the effect of unmeasured or residual confounding on the observed findings could not be ruled out. Second, most in-
REFERENCES 1. Ros E, Tapsell LC, Sabate J. Nuts and berries for heart health. Curr Atheroscler Rep 2010;12:397–406. 2. Fraser GE, Sabate´ J, Beeson WL, Strahan TM. A possible protective effect of nut consumption on risk of coronary heart disease. The Adventist Health Study. Arch Intern Med 1992;152:1416–24. 3. Key TJ, Thorogood M, Appleby PN, Burr ML. Dietary habits and mortality in 11,000 vegetarians and health conscious people: results of a 17 year follow up. BMJ 1996;313:775–9. 4. Albert CM, Gaziano JM, Willett WC, Manson JE. Nut consumption and decreased risk of sudden cardiac death in the Physicians’ Health Study. Arch Intern Med 2002;162:1382–7. 5. Trichopoulou A, Bamia C, Trichopoulos D. Mediterranean diet and survival among patients with coronary heart disease in Greece. Arch Intern Med 2005;165:929–35. 6. Blomhoff R, Carlsen MH, Andersen LF, Jacobs DR Jr. Health benefits of nuts: potential role of antioxidants. Br J Nutr 2006;96:S52–60. 7. Li TY, Brennan AM, Wedick NM, Mantzoros C, Rifai N, Hu FB. Regular consumption of nuts is associated with a lower risk of cardiovascular disease in women with type 2 diabetes. J Nutr 2009;139: 1333–8. 8. Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major dietary protein sources and risk of coronary heart disease in women. Circulation 2010;122:876–83. 9. Dilis V, Katsoulis M, Lagiou P, Trichopoulos D, Naska A, Trichopoulou A. Mediterranean diet and CHD: the Greek European Prospective Investigation into Cancer and Nutrition cohort. Br J Nutr 2012;108: 699–709.
8 of 8
ZHOU ET AL
10. Kelly JH Jr, Sabate´ J. Nuts and coronary heart disease: an epidemiological perspective. Br J Nutr 2006;96(suppl 2):S61–7. 11. Yochum LA, Folsom AR, Kushi LH. Intake of antioxidant vitamins and risk of death from stroke in postmenopausal women. Am J Clin Nutr 2000;72:476–83. 12. Djousse´ L, Gaziano JM, Kase CS, Kurth T. Nut consumption and risk of stroke in US male physicians. Clin Nutr 2010;29:605–9. 13. Bernstein AM, Pan A, Rexrode KM, Stampfer M, Hu FB, Mozaffarian D, Willett WC. Dietary protein sources and the risk of stroke in men and women. Stroke 2012;43:637–44. 14. Steffen LM, Kroenke CH, Yu X, Pereira MA, Slattery ML, Van Horn L, Gross MD, Jacobs DR Jr. Associations of plant food, dairy product, and meat intakes with 15-y incidence of elevated blood pressure in young black and white adults: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Clin Nutr 2005;82:1169–77. 15. Djousse´ L, Rudich T, Gaziano JM. Nut consumption and risk of hypertension in US male physicians. Clin Nutr 2009;28:10–4. 16. Martı´nez-Lapiscina EH, Pimenta AM, Beunza JJ, Bes-Rastrollo M, Martı´nez JA, Martı´nez-Gonza´lez MA. Nut consumption and incidence of hypertension: the SUN prospective cohort. Nutr Metab Cardiovasc Dis 2010;20:359–65. 17. Weng LC, Steffen LM, Szklo M, Nettleton J, Chambless L, Folsom AR. A diet pattern with more dairy and nuts, but less meat is related to lower risk of developing hypertension in middle-aged adults: the Atherosclerosis Risk in Communities (ARIC) study. Nutrients 2013;5:1719–33. 18. Jiang R, Manson JE, Stampfer MJ, Liu S, Willett WC, Hu FB. Nut and peanut butter consumption and risk of type 2 diabetes in women. JAMA 2002;288:2554–60. 19. Parker ED, Harnack LJ, Folsom AR. Nut consumption and risk of type 2 diabetes. JAMA 2003;290:38–9; author reply 39–40. 20. Villegas R, Gao YT, Yang G, Li HL, Elasy TA, Zheng W, Shu XO. Legume and soy food intake and the incidence of type 2 diabetes in the Shanghai Women’s Health Study. Am J Clin Nutr 2008;87:162–7. 21. Kochar J, Gaziano JM, Djousse´ L. Nut consumption and risk of type II diabetes in the Physicians’ Health Study. Eur J Clin Nutr 2010;64: 75–9. 22. Pan A, Sun Q, Manson JE, Willett WC, Hu FB. Walnut consumption is associated with lower risk of type 2 diabetes in women. J Nutr 2013; 143:512–8. 23. Guo K, Jiang Y, Zhou Z, Li Y. RETRACTED: Nut consumption with risk of hypertension and type 2 diabetes mellitus: a meta-analysis of prospective cohort studies. Eur J Prev Cardiol (Epub ahead of print 30 August 2013). 24. Wells GA, Shea B, O’connell D, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Available from: http://www.ohri.ca/ programs/clinical_epidemiology/oxford.asp (cited 14 October 2013). 25. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88.
26. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48. 27. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. 28. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301–9. 29. Orsini N, Bellocco R, Greenland S. Generalized least squares for trend estimation of summarized dose-response data. Stata J 2006;6:40–57. 30. Harrell FE Jr, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst 1988;80:1198–202. 31. Martin B. Meta-analysis: dealing with heterogeneity. York, United Kingdom: University of York, 2006:6. 32. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. 33. Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34. 34. Montonen J, Ja¨rvinen R, Helio¨vaara M, Reunanen A, Aromaa A, Knekt P. Food consumption and the incidence of type II diabetes mellitus. Eur J Clin Nutr 2005;59:441–8. 35. Baer HJ, Glynn RJ, Hu FB, Hankinson SE, Willett WC, Colditz GA, Stampfer M, Rosner B. Risk factors for mortality in the nurses’ health study: a competing risks analysis. Am J Epidemiol 2011;173:319–29. ´ , Ros E, Corella D, 36. Guasch-Ferre´ M, Bullo´ M, Martı´nez-Gonza´lez MA Estruch R, Fito´ M, Aro´s F, Wa¨rnberg J, Fiol M, et al.. Frequency of nut consumption and mortality risk in the PREDIMED nutrition intervention trial. BMC Med 2013;11:164. 37. Estruch R, Ros E, Salas-Salvado´ J, Covas MI, Corella D, Aro´s F, Go´mez-Gracia E, Ruiz-Gutie´rrez V, Fiol M, Lapetra J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 2013;368:1279–90. ´ , Ibarrola38. Salas-Salvado´ J, Bullo´ M, Babio N, Martı´nez-Gonza´lez MA Jurado N, Basora J, Estruch R, Covas MI, Corella D, Aro´s F, et al. Reduction in the incidence of type 2 diabetes with the Mediterranean diet: results of the PREDIMED-Reus nutrition intervention randomized trial. Diabetes Care 2011;34:14–9. 39. Salas-Salvado´ J, Garcia-Arellano A, Estruch R, Marquez-Sandoval F, Corella D, Fiol M, Go´mez-Gracia E, Vin˜oles E, Aro´s F, Herrera C, et al. Components of the Mediterranean-type food pattern and serum inflammatory markers among patients at high risk for cardiovascular disease. Eur J Clin Nutr 2008;62:651–9. 40. Casas-Agustench P, Bullo´ M, Salas-Salvado´ J. Nuts, inflammation and insulin resistance. Asia Pac J Clin Nutr 2010;19:124–30. 41. Appel LJ, Sacks FM, Carey VJ, Obarzanek E, Swain JF, Miller ER 3rd, Conlin PR, Erlinger TP, Rosner BA, Laranjo NM, et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. JAMA 2005;294:2455–64.
