Total and specific fruit and vegetable consumption and risk of stroke: A prospective study

Atherosclerosis 227 (2013) 147e152

Contents lists available at SciVerse ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Total and specific fruit and vegetable consumption and risk of stroke: A prospective study Susanna C. Larsson a, *, Jarmo Virtamo b, Alicja Wolk a a b

Division of Nutritional Epidemiology, National Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE-17177 Stockholm, Sweden Department of Chronic Disease Prevention, National Institute for Health and Welfare, Helsinki, Finland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 August 2012 Received in revised form 15 November 2012 Accepted 4 December 2012 Available online 28 December 2012

Background: Fruit and vegetables is a heterogeneous food group with different content of dietary fiber, vitamins, minerals, carotenoids, and bioactive phytochemicals. Our objective was to examine the relation between specific consumption of fruit and vegetable subgroups and stroke risk in a cohort of Swedish women and men. Methods and results: We prospectively followed 74,961 participants (34,670 women and 40,291 men) who had completed a food frequency questionnaire in the autumn of 1997 and were free from stroke, coronary heart disease, and cancer at baseline. Diagnoses of stroke in the cohort during follow-up were ascertained from the Swedish Hospital Discharge Registry. A total of 4089 stroke cases, including 3159 cerebral infarctions, 435 intracerebral hemorrhages, 148 subarachnoid hemorrhages, and 347 unspecified strokes, were ascertained during 10.2 years of follow-up. The multivariable relative risk (RR) of total stroke for the highest vs. lowest category of total fruit and vegetable consumption was 0.87 (95% confidence interval [CI] 0.78e0.97; P for trend ¼ 0.01). The association was confined to individuals without hypertension (corresponding RR, 0.81; 95% CI, 0.71e0.93; P for trend ¼ 0.01). Among individual fruits and vegetable subgroups, inverse associations with total stroke were observed for apples/pears (RR, 0.89; 95% CI, 0.80e0.98; P for trend ¼ 0.02) and green leafy vegetables (RR, 0.92; 95% CI, 0.81e1.04; P for trend ¼ 0.03). Conclusion: This study shows an inverse association of fruit and vegetable consumption with stroke risk. Particularly consumption of apples and pears and green leafy vegetables was inversely associated with stroke. Ó 2013 Elsevier Ireland Ltd. All rights reserved.

Keywords: Fruits Prospective studies Vegetables Stroke

1. Introduction High consumption of fruit and vegetables has been associated with lower risk of stroke [1]. Fruit and vegetables is a heterogeneous food group with different content of dietary fiber, vitamins, minerals, carotenoids, and other bioactive phytochemicals. It remains unclear which fruit and vegetable subgroups that are most protective against stroke. Consumption of specific fruit and vegetable subgroups, such as apples and pears [2e4], citrus fruits [3,5,6], berries [7,8], cruciferous vegetables [5,6,9], leafy vegetables [5,9], and root vegetables [6,9] has been inconsistently associated with risk of stroke in previous studies. Moreover, studies on fruit and vegetable consumption in relation to risk of hemorrhagic stroke are limited [7,10].

* Corresponding author. Tel.: þ46 8 52486059; fax: þ46 8 304571. E-mail address: [email protected] (S.C. Larsson). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2012.12.022

To further examine the association between consumption of total fruits and vegetables, specific fruits, and vegetable subgroups and risk of total stroke and stroke types, we used data from two large prospective cohorts of Swedish women and men. 2. Methods 2.1. Study population We used data from two prospective population-based cohorts of Swedish women and men, namely the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM). In the autumn of 1997, 39 227 women (SMC) and 48 850 men (COSM) who lived in central Sweden (Uppsala, Västmanland, and Örebro counties) completed a 350-item questionnaire that sought information on diet, lifestyle factors, and other factors that could affect the risk of chronic diseases. For the present analyses, we excluded women and men with an erroneous or a missing National

148

S.C. Larsson et al. / Atherosclerosis 227 (2013) 147e152

Registration Number and those with implausible values for total energy intake (i.e., 3 SDs from the loge-transformed mean energy intake). We further excluded those with a previous stroke, coronary heart disease, or cancer at baseline. After these exclusions, 74 961 participants (34 670 women and 40 291 men), 45e83 years of age, remained for the analyses. The Regional Ethical Review Board at Karolinska Institutet in Stockholm, Sweden, approved this study. 2.2. Baseline data collection Information about education, weight, height, smoking, physical activity, aspirin use, history of hypertension and diabetes, family history of myocardial infarction before 60 years, alcohol consumption (type of alcoholic beverage and frequency of consumption), and diet was obtained through a self-administered questionnaire. We classified participants into three groups based on their highest reported education/school (elementary school, secondary school/ vocational school, and university). Participants were classified as having diabetes if they self-reported diabetes on the questionnaire or had a diagnosis of diabetes recorded in Swedish National Inpatient Register or the Swedish National Diabetes Register. History of hypertension was based on self-report only. We calculated packyears of smoking history by multiplying the number of packs of cigarettes smoked per day by the number of years of smoking. Body mass index was calculated as the weight in kilograms divided by the square of height in meters. Participants reported their level of activity at work, home/housework, walking/bicycling, and exercise as well as inactivity (watching TV/reading) and hours per day of sitting/lying down and sleeping. The time per day reported by the subject to have engaged in each activity was multiplied by the activity’s typical energy expenditure requirement expressed in metabolic equivalents. The metabolic equivalent-hours for all individual activities (including inactivity) reported by the subject were added together to create a metabolic equivalent-hours per day (24-h) score [11]. Spearman correlation coefficient between total activity score estimated from the questionnaire and two 7-day activity records, performed 6 months apart, was 0.56 [11]. 2.3. Dietary assessment Fruit and vegetable consumption was assessed with a foodfrequency questionnaire (FFQ) on which participants indicated their average consumption of 96 foods and beverages over the previous year. Participants could choose from 8 predefined frequency categories, ranging from never to 3 or more times per day. The FFQ included the following fruit and vegetable items: apples and pears; bananas; orange and other citrus fruits; other fruits; berries; spinach; lettuce and green salad; cabbages (cabbage and red cabbage); cauliflower; broccoli and Brussels sprouts; carrots; beetroots; tomatoes and tomato juice; sweet pepper; onion and leek; garlic; green peas; and mixed vegetables. For the present study, the exposures of interest were consumption of total fruit and vegetables, total fruit, total vegetables, specific fruits and vegetables, and vegetable subgroups. We combined specific vegetables into the following subgroups: green leafy vegetables (including spinach, lettuce, and green salad), cruciferous vegetables (white cabbage and red cabbage, cauliflower, broccoli, and Brussels sprouts), and root vegetables (carrots and beetroots). The remaining vegetable items on the FFQ were examined separately, as were the fruit items. We defined the reported times eaten per day as the number of servings. For example, consumption of apples one time per day was defined as one serving of apples per day. In a validation study in a subsample of 129 women from the SMC using a similar FFQ (including 60 foods), the Pearson correlation coefficients between the FFQ and four 1-week diet records (completed 3e4 months apart) ranged

from 0.4 to 0.5 for fruit items and from 0.4 to 0.6 for vegetable items (A. Wolk, unpublished data). 2.4. Case ascertainment Cases of first stroke that occurred in the cohort were ascertained by linkage to the Swedish Hospital Discharge Registry. The strokes were classified as cerebral infarction (International Classification of Diseases 10th Revision code I63), hemorrhagic stroke (I61 and I60), and unspecified stroke (I64). Information on dates of death for deceased participants was obtained from the Swedish Death Register. 2.5. Statistical analysis Participants contributed follow-up time from January 1, 1998 until the date of diagnosis of stroke, death, or end of follow-up (December 2008), whichever occurred first. We categorized participants into five frequency categories of consumption of total fruit and vegetables (1st to 99th percentile: 0.5e13.0 servings/d), total fruit (0.1e5.5 servings/d), and total vegetables (0.1e9.3 servings/d). Because of the narrow range of consumption of specific fruits and vegetables and the uneven distribution of individuals across frequency categories, the number of individuals (and person-years) in each category varied. Cox proportional hazards regression models were used to estimate relative risks (RR) with 95% confidence intervals (CI) of total stroke, cerebral infarction, intracerebral hemorrhage, and subarachnoid hemorrhage. Separate analyses of women and men showed similar associations. Therefore, we report results for women and men combined, adjusting for sex as a stratum variable in the Cox model. We also adjusted all models for age (in months) as a stratum variable. In the multivariable models, we additionally adjusted for known risk factors for stroke, including smoking status and pack-years of smoking (never; past < 20, 20e39, or 40 pack-years; or current < 20, 20e39, or 40 pack-years), education (less than high school, high school, or university), body mass index (<20, 20e24.9, 25e29.9, or 30 kg/m2), physical activity (metabolic equivalent-hours/day, quintiles), aspirin use (never, 1e6 tablets/week, 7 tablets/week), history of hypertension (yes or no), history of diabetes (yes or nor), family history of myocardial infarction before 60 years of age (yes or no). All multivariable models were also adjusted for total energy intake (kcal/day, continuous variable) (to account for over- or underreporting in the FFQ) as well as dietary factors that are associated with risk of stroke in these cohorts and previous studies, including consumption of alcohol, coffee, fresh red meat, processed meat, and total fish (all in quintiles). Further adjustment for intakes of low-fat dairy foods, poultry, whole grains, nuts, legumes, chocolate, sugar, dietary fat, fiber, and sodium did not change the results materially. Therefore, those variables were not included in the multivariable model. Total fruit and total vegetable consumption was mutually adjusted by including both variables in the same multivariable model. Likewise, in the multivariable analyses of specific fruits and vegetable subgroups, the individual fruit items and vegetable subgroups were included in the same model and adjusted for each other. Tests for trends were conducted by assigning the median value for each category and modeling this variable as a continuous variable. We conducted analyses stratified by history of hypertension to assess potential interaction with this variable. Test for interaction was performed by using the likelihood ratio test. We also performed a sensitivity analysis confined to individuals without diabetes at baseline because diabetics may have changed their fruit and vegetable consumption after their diabetes diagnosis and they are at higher risk for stroke. All analyses were conducted with SAS Version 9.2 (SAS Institute, Cary, NC). All statistical tests were

S.C. Larsson et al. / Atherosclerosis 227 (2013) 147e152

2-sided. P values < 0.05 were considered statistically significant. This study had about 80% power (at a ¼ 0.05) to detect an RR of total stroke of 0.87 for the highest vs. lowest category of total fruit and vegetable consumption. The corresponding RRs were 0.85 for cerebral infarction, 0.6 for intracerebral hemorrhage, and 0.4 subarachnoid hemorrhage. 3. Results Over a mean follow-up of 10.2 years, we ascertained a total of 4089 cases of stroke (1680 in women and 2409 in men), including 3159 cerebral infarctions, 435 intracerebral hemorrhages, 148 subarachnoid hemorrhages, and 347 unspecified strokes. Baseline characteristics of the study population according to total fruit and vegetable consumption are shown in Table 1. Compared with participants with a low fruit and vegetable consumption, those with a high consumption were more likely to have a postsecondary education but were less likely to be overweight, physically inactive, to have a history of hypertension, and to be current smokers. On average, those with a high fruit and vegetable consumption also consumed more fresh red meat and processed meat. Mean total fruit and vegetable consumption was higher in women (5.0  2.8 servings/d) than in men (3.8  2.3 servings/d). Spearman correlation between total fruit and total vegetables was 0.44. Baseline characteristics across categories of total fruit and total vegetables were similar to that observed for total fruit and vegetables. Total fruit and vegetable consumption was significantly inversely associated with risk of total stroke, cerebral infarction, and hemorrhagic stroke (Table 2). For example, women and men in the highest category of total fruit and vegetable consumption had a 13% lower risk of total stroke compared with those in the lowest category. Total fruit consumption was significantly inversely associated with total stroke and hemorrhagic stroke, whereas

149

total vegetable consumption was nonsignificantly inversely associated with total stroke and cerebral infarction (Table 2). In a dosee response analysis, the risk of total stroke decreased with increasing fruit and vegetable consumption to a consumption of about 5 servings per day where the association seemed to reach a plateau (Fig. 1). The inverse association between fruit and vegetable consumption and risk of stroke was similar when women and men with a history of diabetes were excluded. In analyses restricted to individuals without diabetes at baseline, the multivariable RRs of total stroke for the highest vs. the lowest category of consumption were 0.85 (95% CI, 0.76e0.96; P for trend ¼ 0.004) for total fruit and vegetables, 0.85 (95% CI, 0.76e0.96; P for trend ¼ 0.007) for total fruit, and 0.88 (95% CI, 0.78e1.00; P for trend ¼ 0.09) for total vegetables. The association between total fruit and vegetable consumption and risk of stroke was modified by hypertension (P for interaction ¼ 0.04). The multivariable RRs of total stroke for the highest vs. lowest category of total fruit and vegetable consumption were 0.81 (95% CI, 0.71e0.93; P for trend ¼ 0.01) among those without hypertension at baseline and 1.03 (95% CI, 0.85e1.25; P for trend ¼ 0.82) among those with a history of hypertension. All results were identical or similar when we removed history of hypertension (a potential intermediate) from the multivariable model. For example, the RR of total stroke for the highest compared with the lowest category of total fruit and vegetable consumption was 0.87 (95% CI, 0.78e0.97) after removing hypertension from the model. The association between specific fruits and vegetable subgroups and risk of stroke is presented in Table 3. Consumption of apples/ pears, banana, citrus fruits, leafy vegetables, and onions and leeks was significantly inversely associated with risk of total stroke in the age- and sex-adjusted analysis. After adjustment for other risk

Table 1 Age-standardized baseline characteristics of 74 961 Swedish women and men by total fruit and vegetable consumption. Characteristicsa

Total fruit and vegetables (servings/d) <2.3 (n ¼ 14 998)

2.3e3.3 (n ¼ 15,002)

3.4e4.4 (n ¼ 14,846)

4.5e6.0 (n ¼ 15,156)

>6.0 (n ¼ 14,959)

Age (y) Sex (% men) Postsecondary education (%) Current smokers (%) Pack-years among current smokers (20 pack-years, %) Body mass index (overweight %)b Physical activity (inactive %) History of hypertension (%) History of diabetes (%) Family history of MI (%) Aspirin use (7 tablets/wk, %) Dietary variables Energy intake (kcal/d) Alcohol (g/d) Coffee (cups/d) Fresh red meat (g/d) Processed meat (g/d) Fish (servings/d) Total fruit (servings/d) Apple/pears (servings/d) Banana (servings/d) Orange (servings/d) Berries (servings/d) Total vegetables (servings/d) Root vegetables (servings/d) Leafy vegetables (servings/d) Cruciferous vegetables (servings/d) Onion and leek (servings/d)

61.6  9.9 71.0 10.1 34.1 25.1

60.2  9.5 61.8 15.1 25.3 22.4

59.9  9.3 53.4 18.4 23.0 21.1

59.9  9.1 46.6 21.9 20.2 20.3

59.9  9.0 36.9 25.4 18.9 18.3

54.3 52.3 22.2 7.0 15.1 6.5

51.3 46.0 20.9 5.8 15.1 6.6

49.4 42.2 20.0 5.3 15.3 6.8

47.8 38.6 20.6 5.6 15.6 6.6

46.8 33.5 20.5 6.1 15.8 7.4

2124  838 7.5  9.6 3.4  2.3 46.9  36 33.3  28 0.2  0.3 0.5  0.3 0.2  0.2 0.1  0.2 0.1  0.1 0.1  0.1 1.0  0.5 0.2  0.2 0.1  0.2 0.1  0.1 0.1  0.1

2229  841 7.9  9.2 3.2  2.1 51.0  34 35.6  25 0.3  0.2 1.0  0.5 0.5  0.3 0.3  0.3 0.2  0.2 0.1  0.1 1.9  0.5 0.3  0.3 0.3  0.2 0.2  0.2 0.2  0.2

2238  832 7.8  8.8 3.1  2.0 51.6  37 35.8  26 0.3  0.2 1.3  0.6 0.5  0.4 0.3  0.3 0.3  0.3 0.1  0.2 2.6  0.6 0.4  0.3 0.4  0.3 0.3  0.2 0.3  0.2

2275  828 7.7  8.6 3.1  1.9 51.7  37 35.2  27 0.3  0.3 1.8  0.8 0.6  0.5 0.4  0.4 0.3  0.3 0.2  0.2 3.4  0.8 0.5  0.4 0.5  0.4 0.4  0.3 0.4  0.3

2417  900 7.2  8.7 3.0  1.9 54.0  52 37.0  38 0.4  0.4 2.9  1.4 1.0  0.7 0.6  0.5 0.5  0.5 0.3  0.3 5.4  2.3 0.8  0.6 0.8  0.6 0.8  0.7 0.6  0.4

a b

Values are means  SD if not otherwise indicated. Overweight was defined as body mass index  25 kg/m2.

150

S.C. Larsson et al. / Atherosclerosis 227 (2013) 147e152

Table 2 Relative risks of stroke by quintiles of total fruit and vegetable consumption among 74,961 Swedish women and men, 1998e2008. Servings/d (median)

Total fruit and vegetables Q1 1.6 Q2 2.8 Q3 3.9 Q4 5.2 Q5 7.6 P for trend Total fruit Q1 0.4 Q2 0.9 Q3 1.2 Q4 1.8 Q5 3.1 P for trend Total vegetables Q1 0.9 Q2 1.8 Q3 2.5 Q4 3.4 Q5 5.1 P for trend

Total stroke

Cerebral infarction

Intracerebral hemorrhage

Subarachnoid hemorrhage

Cases

Person-years

RR (95% CI)a

RR (95% CI)b

Cases

RR (95% CI)b

Cases

RR (95% CI)b

Cases

RR (95% CI)b

1076 845 758 722 688

148 153 152 157 155

186 239 820 023 385

1.00 0.88 (0.81e0.97) 0.85 (0.77e0.93) 0.81 (0.73e0.89) 0.80 (0.72e0.88) <0.0001

1.00 0.93 0.91 0.87 0.87 0.01

(0.85e1.03) (0.82e1.00) (0.79e0.97) (0.78e0.97)

838 642 577 565 537

1.00 0.91 0.88 0.86 0.87 0.04

(0.82e1.01) (0.79e0.99) (0.79e1.00) (0.77e0.99)

119 92 89 89 46

1.00 0.94 1.03 1.02 0.57 0.01

(0.71e1.24) (0.77e1.38) (0.76e1.38) (0.39e0.84)

28 37 29 15 39

1.00 1.13 0.89 0.43 1.10 0.75

(0.68e1.89) (0.52e1.55) (0.22e0.84) (0.63e1.93)

1078 712 754 836 709

178 127 143 163 153

934 926 359 354 079

1.00 0.90 (0.82e0.99) 0.89 (0.81e0.98) 0.84 (0.76e0.92) 0.78 (0.70e0.86) <0.0001

1.00 0.94 0.95 0.92 0.87 0.02

(0.85e1.04) (0.86e1.05) (0.83e1.01) (0.78e0.97)

822 528 589 658 562

1.00 0.91 0.98 0.96 0.91 0.24

(0.81e1.02) (0.88e1.10) (0.86e1.08) (0.80e1.03)

119 89 77 96 54

1.00 1.11 0.90 1.00 0.67 0.02

(0.83e1.47) (0.66e1.23) (0.74e1.34) (0.47e0.96)

43 25 25 20 35

1.00 0.77 0.63 0.43 0.73 0.30

(0.47e1.29) (0.37e1.05) (0.24e0.76) (0.43e1.25)

1106 871 720 749 643

148 152 150 160 154

358 754 808 369 364

1.00 0.90 (0.82e0.99) 0.83 (0.76e0.92) 0.87 (0.79e0.95) 0.80 (0.72e0.88) <0.0001

1.00 0.97 0.91 0.98 0.90 0.12

(0.88e1.06) (0.82e1.01) (0.88e1.08) (0.80e1.01)

869 672 548 569 501

1.00 0.95 0.88 0.94 0.88 0.07

(0.85e1.06) (0.78e0.98) (0.83e1.06) (0.77e1.00)

121 95 81 79 59

1.00 0.95 0.96 0.98 0.88 0.62

(0.72e1.27) (0.71e1.31) (0.71e1.36) (0.62e1.27)

24 37 26 28 33

1.00 1.55 1.15 1.18 1.45 0.67

(0.91e2.67) (0.64e2.08) (0.65e2.16) (0.78e2.70)

a

Adjusted for age and sex. Adjusted for age, sex, smoking status and pack-years of smoking, education, body mass index, total physical activity, aspirin use, history of hypertension, diabetes, family history of myocardial infarction, and intakes of total energy, alcohol, coffee, fresh red meat, processed meat, and fish. Total fruit and total vegetable consumption was mutually adjusted by including both variables in the same multivariable model. b

factors for stroke and mutually for other fruits and vegetable subgroups, only the inverse relation between apple and pear consumption and risk of total stroke remained significant, whereas results for the other fruit and vegetable subgroups became nonsignificant. Leafy vegetables consumption was inversely associated with total stroke and cerebral infarction in the trend analysis but not when comparing highest vs. lowest category. There was a positive association between berry consumption and risk of total stroke and cerebral infarction. When we stratified the analysis by history of hypertension, the multivariable RRs of total stroke for the highest vs. lowest category of berry consumption were 1.07 (95% CI, 0.92e1.23) among those without hypertension and 1.22 (95% CI, 0.99e1.49) among those with hypertension. We observed no

Fig. 1. Adjusted relative risks of stroke as a function of total fruit and vegetable consumption. Data were fitted by a restricted cubic spline Cox proportional hazards model. The median consumption in the lowest category of fruit and vegetable consumption (1.6 servings/d) was used as the reference for all relative risks. The relative risks are indicated by the solid line, and the 95% confidence intervals are indicated by the dashed lines. Adjusted for the same variables as in the multivariable model in Table 2.

association between specific vegetables and risk of total stroke (data not shown), but onion and leek consumption was associated with a nonsignificant lower risk (Table 3). 4. Discussion This prospective study of Swedish women and men confirms an inverse association between total fruit and vegetable consumption and stroke risk. The association appeared to be confined to participants without a history of hypertension. Consumption of total fruits but not total vegetables was significantly inversely associated with stroke. Among fruit and vegetable subgroups, only consumption of apples and pears and green leafy vegetables was significantly inversely associated with risk of total stroke after adjustment for potential confounders. Other fruits and vegetable subgroups were not significantly inversely associated with risk of stroke. The reduction in stroke risk associated with increasing fruit and vegetable consumption was most pronounced up to a consumption of around 5 servings per day. Above that level of consumption, the risk of stroke did not decrease materially with increasing fruit and vegetable consumption. Total fruit and vegetable consumption has been inversely associated with risk of total stroke in most previous prospective studies. In a meta-analysis of 9 cohort studies, with a total of 4917 stroke cases, the summary RR of total stroke was 0.74 (95% CI, 0.69e0.79) for more than 5 servings of fruit and vegetables per day compared with less than 3 servings per day [1]. In analysis stratified by stroke types, the corresponding RRs were 0.72 (95% CI, 0.66e0.79) for ischemic stroke (7 studies) and 0.73 (95% CI, 0.61e0.87) for total hemorrhagic stroke (3 studies) [1]. In the present study, which was based on a similar number of cases as the meta-analysis, total fruit and vegetable consumption was inversely associated with risk of cerebral infarction and intracerebral hemorrhage but not subarachnoid hemorrhage. In another meta-analysis of 7 cohort studies, including 2955 stroke cases, the risk of stroke decreased by 11% (95% CI, 7%e15%) and 3% (95%

S.C. Larsson et al. / Atherosclerosis 227 (2013) 147e152

151

Table 3 Relative risks of stroke by categories of specific fruits and vegetables among 74 961 Swedish women and men, 1998e2008. Servings/d (median)

Apples/pears Q1 0.1 Q2 0.2 Q3 0.5 Q4 1.0 P for trend Banana Q1 0.1 Q2 0.2 Q3 0.5 Q4 1.0 P for trend Citrus fruits Q1 0 Q2 0.1 Q3 0.2 Q4 0.8 P for trend Berries Q1 0 Q2 0.1 Q3 0.2 Q4 0.5 P for trend Root vegetables Q1 0.1 Q2 0.3 Q3 0.6 Q4 1.1 P for trend Leafy vegetables Q1 0.1 Q2 0.2 Q3 0.6 Q4 1.1 P for trend Cruciferous vegetables Q1 0.1 Q2 0.2 Q3 0.4 Q4 0.8 P for trend Onion and leek Q1 0 Q2 0.2 Q3 0.5 Q4 0.9 P for trend

Total stroke

Cerebral infarction

Cases

Person-years

RR (95% CI)a

RR (95% CI)b

1125 965 1164 835

181 190 236 158

837 551 192 073

1.00 0.91 (0.84e0.99) 0.82 (0.76e0.90) 0.81 (0.74e0.89) <0.0001

1.00 0.96 (0.88e1.06) 0.88 (0.80e0.97) 0.89 (0.80e0.98) 0.02

1668 1205 583 633

255 238 129 142

967 431 719 536

1.00 0.93 (0.87e1.01) 0.90 (0.81e0.99) 0.84 (0.77e0.92) 0.0003

1051 1122 994 922

134 249 205 176

923 787 805 139

974 1558 978 579

137 363 183 82

452 1981 1065 591

Intracerebral hemorrhage

Subarachnoid hemorrhage

RR (95% CI)b

Cases

RR (95% CI)b

Cases

RR (95% CI)b

850 741 913 655

1.00 0.98 (0.88e1.09) 0.91 (0.82e1.01) 0.92 (0.82e1.03) 0.11

125 103 125 82

1.00 0.94 (0.71e1.26) 0.93 (0.70e1.23) 0.91 (0.66e1.24) 0.60

44 39 34 31

1.00 0.76 (0.47e1.22) 0.56 (0.34e0.92) 0.63 (0.37e1.08) 0.14

1.00 0.98 (0.90e1.06) 0.99 (0.90e1.10) 0.94 (0.85e1.03) 0.27

1275 932 462 490

1.00 1.00 (0.91e1.09) 1.04 (0.93e1.17) 0.96 (0.85e1.07) 0.57

193 124 64 54

1.00 0.85 (0.66e1.08) 0.93 (0.69e1.26) 0.74 (0.53e1.02) 0.14

50 51 16 31

1.00 1.16 (0.76e1.78) 0.70 (0.39e1.28) 1.18 (0.72e1.94) 0.93

1.00 0.89 (0.81e0.97) 0.90 (0.82e0.98) 0.85 (0.78e0.93) 0.009

1.00 0.94 (0.85e1.03) 0.98 (0.89e1.08) 0.95 (0.86e1.05) 0.49

807 858 771 723

1.00 0.96 (0.86e1.07) 1.01 (0.90e1.13) 0.97 (0.87e1.09) 0.69

116 128 106 85

1.00 0.82 (0.62e1.08) 0.86 (0.63e1.16) 0.81 (0.59e1.10) 0.37

20 58 39 31

1.00 1.66 (0.95e2.90) 1.49 (0.81e2.71) 1.23 (0.66e2.30) 0.46

509 586 293 266

1.00 0.91 (0.84e0.99) 0.90 (0.82e0.98) 0.94 (0.84e1.04) 0.38

1.00 1.01 (0.93e1.11) 1.05 (0.95e1.16) 1.13 (1.00e1.26) 0.05

746 1187 773 453

1.00 1.01 (0.91e1.12) 1.07 (0.96e1.20) 1.14 (0.99e1.30) 0.03

104 167 110 54

1.00 1.01 (0.77e1.33) 1.16 (0.86e1.57) 1.03 (0.72e1.48) 0.86

21 74 27 26

1.00 1.34 (0.79e2.28) 0.97 (0.52e1.80) 2.17 (1.14e4.11) 0.04

86 395 192 91

354 364 990 946

1.00 0.98 (0.89e1.09) 0.94 (0.84e1.05) 0.91 (0.80e1.03) 0.07

1.00 1.07 (0.95e1.19) 1.05 (0.93e1.19) 1.04 (0.90e1.19) 0.96

340 1518 847 455

1.00 1.09 (0.96e1.24) 1.11 (0.96e1.28) 1.04 (0.89e1.22) 0.96

58 214 103 60

1.00 0.87 (0.63e1.19) 0.81 (0.56e1.16) 0.96 (0.64e1.44) 0.99

14 82 29 23

1.00 1.23 (0.66e2.28) 0.92 (0.45e1.87) 1.37 (0.64e2.94) 0.68

1095 1596 901 497

139 293 224 108

739 608 729 578

1.00 0.92 (0.85e1.00) 0.80 (0.73e0.88) 0.81 (0.73e0.90) <0.0001

1.00 1.00 (0.92e1.09) 0.89 (0.80e0.98) 0.92 (0.81e1.04) 0.03

852 1248 672 387

1.00 1.02 (0.92e1.12) 0.86 (0.77e0.97) 0.94 (0.81e1.08) 0.04

117 168 110 40

1.00 0.94 (0.72e1.23) 1.02 (0.75e1.38) 0.77 (0.51e1.16) 0.41

23 60 38 27

1.00 1.27 (0.75e2.17) 1.19 (0.66e2.15) 1.39 (0.72e2.69) 0.45

980 1300 1041 768

138 246 213 167

503 245 935 970

1.00 0.97 (0.89e1.06) 0.92 (0.84e1.00) 0.96 (0.87e1.06) 0.45

1.00 1.02 (0.93e1.12) 1.00 (0.91e1.11) 1.10 (0.97e1.23) 0.13

771 1001 802 585

1.00 1.00 (0.90e1.11) 0.96 (0.86e1.08) 1.04 (0.91e1.18) 0.57

95 149 118 73

1.00 1.31 (0.98e1.75) 1.40 (1.03e1.92) 1.35 (0.94e1.93) 0.25

28 52 30 38

1.00 0.96 (0.58e1.59) 0.62 (0.35e1.10) 1.10 (0.61e1.98) 0.61

1486 1425 797 381

200 284 183 97

882 636 485 651

1.00 0.91 (0.84e0.98) 0.92 (0.84e1.00) 0.83 (0.74e0.93) 0.003

1.00 0.95 (0.88e1.03) 0.98 (0.89e1.08) 0.89 (0.79e1.01) 0.15

1150 1093 624 292

1.00 0.95 (0.87e1.04) 1.02 (0.91e1.13) 0.89 (0.77e1.03) 0.32

164 165 73 33

1.00 0.96 (0.76e1.21) 0.77 (0.57e1.04) 0.73 (0.49e1.10) 0.06

31 64 31 22

1.00 1.53 (0.96e2.42) 1.21 (0.70e2.10) 1.43 (0.77e2.68) 0.68

Cases

a

Adjusted for age and sex. Adjusted for age, sex, smoking status and pack-years of smoking, education, body mass index, total physical activity, aspirin use, history of hypertension, diabetes, family history of myocardial infarction, and intakes of total energy, alcohol, coffee, fresh red meat, processed meat, and fish. Apples/pears, banana, citrus fruits, and berries were mutually adjusted and adjusted for total vegetable consumption. The vegetable subgroups were mutually adjusted and were further adjusted for total fruit consumption. b

CI, -2%-8%) for each additional portion per day of fruits and vegetables, respectively [12]. Our finding for apples and pears is consistent with results from previous prospective studies showing an inverse association between consumption of apples and pears and risk of stroke [2e4]. Moreover, in a population-based cohort study in the Netherlands, consumption of white fruits (apples and pears were the most commonly consumed white fruit and vegetables contributing with 55% of the consumption) was the only fruit and vegetable subgroup that was significantly inversely associated with risk of stroke [13]. We observed an unexpected increased risk of stroke associated with berry consumption. The positive association was restricted to those with hypertension. In this study population, strawberry was the most commonly consumed berry contributing with about 60% of total berry consumption. In a cohort of US women, strawberry

consumption was positively associated with stroke after adjustment for lifestyle factors but was attenuated after further adjustment for dietary factors [8]. In a cohort of Finnish men, total berry consumption was inversely associated with risk of cerebral infarction (RR, 0.81; 95% CI, 0.66e1.00, for >49 g/day vs. <12 g/day) [7]. We observed a nonsignificant inverse association between onion consumption and stroke in the present study. In addition, green leafy vegetable consumption was statistically significantly inversely associated with stroke risk, a finding that is consistent with previous studies [5,9]. Fruit and vegetables contain many nutrients and bioactive phytochemicals that may lower the risk of stroke. Increased fruit and vegetable consumption has been shown to significantly reduce blood pressure [14]. As raised blood pressure is a strong risk factor for stroke, the reduction in blood pressure by fruit and vegetable

152

S.C. Larsson et al. / Atherosclerosis 227 (2013) 147e152

consumption may, at least in part, explain the observed inverse association between fruit and vegetable consumption and risk of stroke. Fruit and vegetables are also rich sources of antioxidants, such as vitamin C, carotenoids, and polyphenols (e.g., flavonols). A meta-analysis of 6 prospective studies showed that a high intake of flavonols, present in for example apples and onions (as well as tea), was associated with a significant 20% lower risk of stroke [15]. Besides their antioxidant effects, polyphenols have antiinflammatory, vasodilating, and antithrombotic properties [16] that may explain the beneficial effect of fruit and vegetable consumption on stroke. Major strengths of this study include its prospective and population-based design, large sample size, large number of stroke cases, and the virtually complete follow-up of participants through linkage to population-based Swedish registries. This is the largest prospective study to date on fruit and vegetable consumption in relation to risk of stroke. This study also has limitations. First, because of the observational design, we cannot exclude the possibility of residual confounding due to unmeasured or imprecise measurement of other risk factors for stroke. Second, consumption of fruits and vegetables was assessed using a self-administered questionnaire and was measured only once, at baseline. This will inevitably lead to some measurement error in the assessment of fruit and vegetable consumption. Because fruit and vegetable consumption was assessed before diagnosis of stroke, any measurement error would be nondifferential and most likely lead to an underestimation of the true relation between fruit and vegetable consumption and stroke. Finally, we did not have information on some fruits for example grapes (a source of quercetin) or types of berries. In summary, findings from this prospective study indicate that consumption of apples and pears and green leafy vegetables is associated with a lower risk of stroke. Furthermore, this study confirms an inverse relation between total fruit and vegetable consumption and risk of total stroke, cerebral infarction, and intracerebral hemorrhage. Results from this study lend further support to the recommendation to increase fruit and vegetable consumption. Conflict of interest None.

Acknowledgments This study was supported by research grants from the Swedish Council for Working Life and Social Research (FAS), the Swedish Research Council/Committee for Infrastructure, and by a Research Fellow grant from Karolinska Institutet (to Dr. Larsson). References [1] He FJ, Nowson CA, MacGregor GA. Fruit and vegetable consumption and stroke: meta-analysis of cohort studies. Lancet 2006;367(9507): 320e6. [2] Keli SO, Hertog MG, Feskens EJ, Kromhout D. Dietary flavonoids, antioxidant vitamins, and incidence of stroke: the Zutphen study. Arch Intern Med 1996; 156(6):637e42. [3] Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. Am J Clin Nutr 2007;85(3):895e909. [4] Knekt P, Kumpulainen J, Jarvinen R, et al. Flavonoid intake and risk of chronic diseases. Am J Clin Nutr 2002;76(3):560e8. [5] Joshipura KJ, Ascherio A, Manson JE, et al. Fruit and vegetable intake in relation to risk of ischemic stroke. JAMA 1999;282(13):1233e9. [6] Mizrahi A, Knekt P, Montonen J, Laaksonen MA, Heliovaara M, Jarvinen R. Plant foods and the risk of cerebrovascular diseases: a potential protection of fruit consumption. Br J Nutr 2009;102(7):1075e83. [7] Hirvonen T, Virtamo J, Korhonen P, Albanes D, Pietinen P. Intake of flavonoids, carotenoids, vitamins C and E, and risk of stroke in male smokers. Stroke 2000;31(10):2301e6. [8] Sesso HD, Gaziano JM, Jenkins DJ, Buring JE. Strawberry intake, lipids, C-reactive protein, and the risk of cardiovascular disease in women. J Am Coll Nutr 2007;26(4):303e10. [9] Johnsen SP, Overvad K, Stripp C, Tjonneland A, Husted SE, Sorensen HT. Intake of fruit and vegetables and the risk of ischemic stroke in a cohort of Danish men and women. Am J Clin Nutr 2003;78(1):57e64. [10] Gillman MW, Cupples LA, Gagnon D, et al. Protective effect of fruits and vegetables on development of stroke in men. Jama 1995;273(14):1113e7. [11] Norman A, Bellocco R, Bergstrom A, Wolk A. Validity and reproducibility of self-reported total physical activityedifferences by relative weight. Int J Obes Relat Metab Disord 2001;25(5):682e8. [12] Dauchet L, Amouyel P, Dallongeville J. Fruit and vegetable consumption and risk of stroke: a meta-analysis of cohort studies. Neurology 2005;65(8): 1193e7. [13] Oude Griep LM, Verschuren WM, Kromhout D, Ocke MC, Geleijnse JM. Colors of fruit and vegetables and 10-year incidence of stroke. Stroke 2011;42(11): 3190e5. [14] Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH collaborative research group. N Engl J Med 1997;336(16):1117e24. [15] Hollman PC, Geelen A, Kromhout D. Dietary flavonol intake may lower stroke risk in men and women. J Nutr 2010;140(3):600e4. [16] Landete JM. Updated knowledge about polyphenols: functions, bioavailability, metabolism, and health. Crit Rev Food Sci Nutr 2012;52(10):936e48.