Which should we eat, fruit or vegetables? The association with abdominal aortic aneurysm

Long-term consumption of a Mediterranean diet (including high consumption of fruit and vegetables) increases circulating levels of endothelial progenitor cells more than a low-fat diet in patients with type-2 diabetes.1 Furthermore, adherence to the Mediterranean diet and adoption of an active lifestyle are inversely associated with incidence of cardiovascular disease.2 Consumption of fruit and/or vegetables has been known to be inversely associated with risk of hypertension,3,4 type 2 diabetes mellitus,5,–7 and metabolic syndrome.8 Fruit/vegetable consumption may be also inversely associated with risk of cardiovascular disease (not only incidence9,10 but also mortality11), especially coronary heart disease12 and stroke.13 The association of fruit/vegetable consumption with risk of abdominal aortic aneurysm (AAA), however, has been less investigated. In the present article, to assess whether consumption of either fruit or vegetables, both fruit and vegetables, or neither fruit nor vegetables is associated with incidence of AAA, we performed the first meta-analysis of currently available all prospective cohort studies. To identify all prospective cohort studies researching the association of fruit/vegetable consumption with AAA incidence, we searched PubMed, Web of Science and Google Scholar up to July 2019. Search terms included fruit(s) or vegetable(s); aorta or aortic; and aneurysm(s). From each study, we extracted multivariable-adjusted hazard ratios (with their 95% confidence intervals (CIs)) of AAA incidence for higher versus the lowest fruit or vegetable consumption, then combined the specific estimates by means of inverse variance-weighted averages of logarithmic hazard ratios in the random-effects model. All analyses were conducted using Review Manager version 5.3. We identified three large population-based prospective cohort studies14,–16 (Table 1) enrolling 120,055 participants. Incidence rates of AAA during follow-up of 1314 to 23 years13 was 0.52%16 to 3.83%.15 All studies provided multivariable-adjusted hazard ratios for fruit and vegetables separately. We excluded a study17 reporting hazard ratios for fruit and vegetables together. In all studies, smoking and socioeconomic status (education and physical activity) were entered into the multivariable Cox proportional hazards regression as potential covariates in addition to fruit/vegetables consumption (Table 1), which means that the inverse association of fruit, not vegetable, consumption with AAA incidence (demonstrated in the present meta-analysis) is ‘independent’ of smoking and socioeconomic status. A sex-stratified analysis was performed in only one study,16 which suggests that men (hazard ratio for the fourth versus lowest quartile consumption, 0.79: 95% CI, 0.64 to 0.99), not women (corresponding hazard ratio, 0.71; 95% CI, 0.45 to 1.11), benefit from fruit consumption. Unfortunately, no study reported patient ethnicity. Fruit and vegetable consumption was assessed by a 168-item food frequency questionnaire (FFQ) and a 1-h interview,14 a 66-item FFQ,15 and a 96-item FFQ.16 In a study by Bergwall et al.,14 fruits were divided into citrus and non-citrus fruits; vegetables were, in addition to total vegetables, divided into root vegetables (including carrots), leaf vegetables (including kale), carrots, and cabbage; and juices (including all types of vegetable and citrus juices) were excluded. In other studies,15,16 however, fruit/vegetable consumption was not especially defined. Bergwall et al.14 provided hazard ratios for citrus fruit, non-citrus fruit, leaf vegetables, and root vegetables separately, and Stackelberg et al.16 reported those for men and women separately. Bergwall et al.14 and Haring et al.15 divided participants into quintiles according to consumption amount14/frequency,15 and Stackelberg et al.16 categorized those into quartiles in accordance with consumption amount. We combined hazard ratios for the second lowest (second quartile and quintile, Group 2) versus lowest (first quartile and quintile, Group 1) consumption and those for the second highest (third quartile and quintile, Group 3) versus lowest (Group 1) consumption separately. In the primary meta-analysis for the highest versus lowest consumption, we pooled hazard ratios for the fourth quartile (Group 4) versus lowest (Group 1) consumption together with those for the fourth quintile (Group 4-1) versus lowest (Group 1) consumption. As the sensitivity analysis, we combined hazard ratios for the fourth quartile (Group 4) versus lowest (Group 1) consumption together with those for the fifth quintile (Group 4-2) versus lowest (Group 1) consumption. Study design and results. AAA: abdominal aortic aneurysm; AP: angina pectoris; ARIC: Atherosclerosis Risk in Communities; BMI: body mass index; CI: confidence interval; CVD: cardiovascular disease; DM: diabetes mellitus; HR: hazard ratio; ICD 9/10: International Classification of Disease version 9/10; IQR: interquartile range; MDCS: Malmö Diet and Cancer Study; MI: myocardial infarction. Study design and results. AAA: abdominal aortic aneurysm; AP: angina pectoris; ARIC: Atherosclerosis Risk in Communities; BMI: body mass index; CI: confidence interval; CVD: cardiovascular disease; DM: diabetes mellitus; HR: hazard ratio; ICD 9/10: International Classification of Disease version 9/10; IQR: interquartile range; MDCS: Malmö Diet and Cancer Study; MI: myocardial infarction. Forest plot of hazard ratios of abdominal aortic aneurysm incidence for higher versus the lowest fruit or vegetable consumption. CI: confidence interval; IV: inverse variance. Forest plot of hazard ratios of abdominal aortic aneurysm incidence for higher versus the lowest fruit or vegetable consumption. CI: confidence interval; IV: inverse variance. The present meta-analysis suggests the inverse association of fruit, not vegetable, consumption with AAA incidence. The inflammatory process, which brings about protease-mediated degradation of the extracellular matrix and apoptosis of smooth muscle cells (predominant matrix synthesizing cells) in the vascular wall and accordingly advances to impair the aortic wall, has been well-known to play an essential role in the pathogenesis of AAA.18 Oxidative stress, which is constantly increased in the process, profoundly involves the pathophysiology of inflammation, and reactive oxygen species also play a momentous role in regulating matrix metalloproteinases and inducing smooth muscle cell apoptosis.18 Both fruit and vegetables are rich in antioxidants, thus may potentially reduce AAA risk. The association of fruit, not vegetables, with reduced AAA incidence demonstrated by the present meta-analysis may be potentially explained by different kinds of antioxidants contained in fruit and vegetables, which could influence oxidative stress more or less in the aortic wall.16 Fruit is the major source (apples 32.0% followed by chocolate 17.9% and grapes 17.8%) of procyanidins (polyphenol compounds from the flavonoid group) in the diet, whereas vegetables are not an important source.19 Procyanidins from grape pomace restrain nicotinamide adenine dinucleotide phosphate (NADPH) oxidase in the human endothelium irrespective of their extent of polymerization and proportion of galloylation.20 NADPH oxidase, playing an important role as a source of oxidative stress in AAA pathogenesis, is suggested to be a potential molecular target for treatment of AAA.21,,–24 Furthermore, grapes and berries are some sources of polyphenols, which may protectively impact vascular function and cardiovascular disease risk via their cholesterol-modulating properties.25 In conclusion, consumption of fruit, not vegetables, may be inversely associated with AAA incidence. ‘I suppose I ought to eat or drink something or other; but the great question is, what?’ said Alice.