Nut consumption and risk of cardiovascular disease, total cancer, all-cause and cause-specific mortality: a systematic review and dose-response meta-analysis of prospective studies

Research article

Open Access

Open Peer Review

• Dagfinn AuneEmail author,
• NaNa Keum,
• Edward Giovannucci,
• Lars T. Fadnes,
• Paolo Boffetta,
• Darren C. Greenwood,
• Serena Tonstad,
• Lars J. Vatten,
• Elio Riboli and
• Teresa Norat

BMC Medicine201614:207

DOI: 10.1186/s12916-016-0730-3

©  The Author(s). 2016

Received: 28 January 2016

Accepted: 26 October 2016

Published: 5 December 2016

Open Peer Review reports

Abstract

Background

Although nut consumption has been associated with a reduced risk of cardiovascular disease and all-cause mortality, data on less common causes of death has not been systematically assessed. Previous reviews missed several studies and additional studies have since been published. We therefore conducted a systematic review and meta-analysis of nut consumption and risk of cardiovascular disease, total cancer, and all-cause and cause-specific mortality.

Methods

PubMed and Embase were searched for prospective studies of nut consumption and risk of cardiovascular disease, total cancer, and all-cause and cause-specific mortality in adult populations published up to July 19, 2016. Summary relative risks (RRs) and 95% confidence intervals (CIs) were calculated using random-effects models. The burden of mortality attributable to low nut consumption was calculated for selected regions.

Results

Twenty studies (29 publications) were included in the meta-analysis. The summary RRs per 28 grams/day increase in nut intake was for coronary heart disease, 0.71 (95% CI: 0.63–0.80, I² = 47%, n = 11), stroke, 0.93 (95% CI: 0.83–1.05, I² = 14%, n = 11), cardiovascular disease, 0.79 (95% CI: 0.70–0.88, I² = 60%, n = 12), total cancer, 0.85 (95% CI: 0.76–0.94, I² = 42%, n = 8), all-cause mortality, 0.78 (95% CI: 0.72–0.84, I² = 66%, n = 15), and for mortality from respiratory disease, 0.48 (95% CI: 0.26–0.89, I² = 61%, n = 3), diabetes, 0.61 (95% CI: 0.43–0.88, I² = 0%, n = 4), neurodegenerative disease, 0.65 (95% CI: 0.40–1.08, I² = 5.9%, n = 3), infectious disease, 0.25 (95% CI: 0.07–0.85, I² = 54%, n = 2), and kidney disease, 0.27 (95% CI: 0.04–1.91, I² = 61%, n = 2). The results were similar for tree nuts and peanuts. If the associations are causal, an estimated 4.4 million premature deaths in the America, Europe, Southeast Asia, and Western Pacific would be attributable to a nut intake below 20 grams per day in 2013.

Conclusions

Higher nut intake is associated with reduced risk of cardiovascular disease, total cancer and all-cause mortality, and mortality from respiratory disease, diabetes, and infections.

Keywords

Nuts Peanuts Cardiovascular disease Cancer All-cause mortality Cause-specific mortality Meta-analysis

Background

Cardiovascular disease and cancer remain the two most common causes of death, accounting for 25.5 million deaths worldwide in 2013 [1]. Epidemiological and intervention studies have shown that a high intake of nuts is associated with a reduced risk of coronary heart disease and possibly other health outcomes such as diabetes, overweight and obesity, gallstones, and colorectal cancer [2, 3, 4]. Tree nuts, botanically defined as dry fruit containing one seed (rarely two) within the ovary wall that becomes hard at maturity, include walnuts, almonds, hazelnuts, cashews, pistachios, and pecans [5]. While Brazil nuts and peanuts are botanically classified as seeds and legumes, respectively, all of tree nuts, Brazil nuts, and peanuts are collectively referred to as nuts due to their similar nutritional properties and culinary use. Nuts are good sources of dietary fiber, magnesium, polyunsaturated fats, vitamin E, and antioxidants, all of which may reduce risk of cardiovascular disease by reducing insulin resistance [6], cholesterol concentrations [4], lipid peroxidation [7], and oxidative stress [8]. Nuts also contain other bioactive compounds, such as ellagic acid, anacardic acid, genistein, resveratrol, and inositol phosphates, which may reduce cancer risk by inducing cell cycle arrest, apoptosis, inhibiting cell proliferation, migration, invasion, and angiogenesis [9]. However, epidemiological data on nuts and cancer risk are less extensive than for cardiovascular disease.

There is a growing body of evidence suggesting a role of nut consumption in reducing risk of coronary heart disease [2, 10, 11, 12, 13, 14, 15, 16, 17] and mortality [10, 12, 13, 14, 15, 16, 18, 19, 20, 21, 22]. However, whether a high intake of nuts is associated with risk of stroke [13, 14, 15, 23, 24, 25, 26, 27] or overall cancer risk [13, 14, 15, 21, 24, 28] is not clear, as most studies reported no significant association [13, 14, 15, 21, 23, 25, 26, 27] and only a few reported significant inverse associations [14, 16, 21]. Nevertheless, the possibility that a weak association may have been missed because of low statistical power cannot be excluded. Although a few previous reviews reported a reduced risk of coronary heart disease and mortality [29, 30, 31] with higher nut intake, associations with stroke have been unclear, with one meta-analysis finding no statistically significant association [32], but another meta-analysis reporting a significant inverse association [33]. However, in the latter, the Nurses’ Health Study and the Health Professionals Follow-up Study had been included twice, thus, questions remain with regards to whether there is an association between nut intake and stroke. In addition, several large cohort studies including 47,061 deaths and > 748,000 additional participants were either not included [20, 28] or have been published [14, 15, 22, 34, 35, 36, 37] since these reviews, and more detailed and updated analyses have since been published from the Physicians’ Health Study [15] and the Netherlands Cohort Study [16]. Associations between nut consumption and less common causes of death have not been systematically assessed. Therefore, we conducted a systematic review and meta-analysis of prospective studies of nut consumption and the risk of coronary heart disease, stroke, cardiovascular disease, total cancer, and all-cause mortality as well as less common causes of death to provide a more up-to-date and comprehensive assessment of the available evidence. We aimed to clarify the strength and shape of the dose–response relationship between nut consumption and these outcomes, identify potential differences by type of nuts consumed (total nuts, tree nuts, peanuts), as well as potential sources of heterogeneity between studies by geographic location. To examine the health impact of low nut consumption we also estimated the number of deaths in North and South America, Europe, Southeast Asia, and the Western Pacific attributable to low nut consumption based on regional studies [14, 38, 39, 40, 41, 42, 43, 44] and data on mortality from the Global Burden of Disease Study [1].

Methods

Search strategy and inclusion criteria

The PubMed and EMBASE databases were searched from their inception (1966 and 1947, respectively) to July 19, 2016. The search terms used for the PubMed search are provided in Additional file 1: Table S1 and a similar search was conducted in EMBASE. Published prospective studies (cohort studies, case-cohort studies, nested case–control studies within cohort studies, and randomized trials) of nut intake (any type of edible nut consumption, including all dosages) among mainly adult populations and incidence or mortality from coronary heart disease, stroke, cardiovascular disease, total cancer, and all-cause (primary outcomes) and cause-specific mortality (secondary outcomes) from any cause of death investigated by at least two studies were included if they reported adjusted relative risk (RR) estimates and 95% confidence intervals (CIs). For the dose–response analyses, a quantitative measure of the intake for at least three categories of nut intake or a risk estimate on a continuous scale had to be available. Retrospective case–control studies were excluded because of the greater potential for recall and selection bias, while cross-sectional studies were excluded because of the difficulty of drawing conclusions with regard to the cause and effect. Reviews, meta-analyses, duplicate publications, studies with an unspecific exposure (e.g., nut intake was combined with fruits or legumes), studies on other outcomes, studies that did not report adjusted risk estimates, ecological studies, letters, and studies with unusable data, as well as abstracts, grey literature, and unpublished studies, were not included. When duplicate publications were published from the same studies we chose the publication with the largest number of cases or deaths for inclusion. We searched the references of the retrieved reports for any additional studies. The first author of one study [20] was contacted to obtain information with regard to the amount of nut intake for each category of intake and this information was provided. DA conducted the literature search and the screening of the studies and TN screened in duplicate the 89 potentially relevant studies identified from the initial screening (Fig. 1). Any discrepancies were resolved by discussion. Study quality was assessed by two authors (DA, DCG) using the Newcastle–Ottawa Scale, which awards a score of 0–9 based on the selection, comparability, and outcome assessment [45]. We considered studies with a score of 0–3, 4–6, and 7–9 to represent low, medium, and high quality studies, respectively. We followed the PRISMA criteria for reporting of meta-analyses of observational studies [46]. A list of the excluded studies is provided in Additional file 1: Table S2. Although there was no protocol for the current review, we followed standard methods and analytic approaches similar to our previous meta-analysis [47].