Downloaded from https://www.cambridge.org/core.

Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
British Journal of Nutrition, page 1 of 8 doi:10.1017/S0007114521000404
© The Author(s), 2021. Published by Cambridge University Press on behalf of The Nutrition Society

Consumption of flavonoid-rich fruits, flavonoids from fruits and stroke risk: a

prospective cohort study

Qi Gao1, Jia-Yi Dong1, Renzhe Cui1, Isao Muraki1, Kazumasa Yamagishi2, Norie Sawada3, Hiroyasu Iso1,2*
and Shoichiro Tsugane3 for the Japan Public Health Center-based Prospective Study Group
1
Public Health, Department of Social Medicine, Osaka University Graduate School of Medicine, Osaka, Japan
2
Department of Public Health Medicine, Faculty of Medicine, and Health Services Research and Development Center, University
of Tsukuba, Tsukuba, Japan
3
Epidemiology and Prevention Division, National Cancer Center, Tokyo, Japan
(Submitted 14 October 2020 – Final revision received 20 January 2021 – Accepted 21 January 2021)

Abstract
We sought to examine the prospective associations of specific fruit consumption, in particular flavonoid-rich fruit (FRF) consumption, with the
risk of stroke and subtypes of stroke in a Japanese population. A study followed a total of 39 843 men and 47 334 women aged 44–76 years, and
free of CVD, diabetes and cancer at baseline since 1995 and 1998 to the end of 2009 and 2012, respectively. Data on total and specific FRF
consumption for each participant were obtained using a self-administrated FFQ. The hazard ratios (HR) of stroke in relation to total and specific
FRF consumption were estimated through Cox proportional hazards regression models. During a median follow-up of 13·1 years, 4091 incident
stroke cases (2557 cerebral infarctions and 1516 haemorrhagic strokes) were documented. After adjustment for age, BMI, study area, lifestyles,
dietary factors and other risk factors, it was found that total FRF consumption was associated with a significantly lower risk of stroke in women
(HR = 0·70; 95 % CI 0·58, 0·84), while the association in men was not significant (HR = 0·93; 95 % CI 0·79, 1·09). As for specific FRF, consumptions
of citrus fruits, strawberries and grapes were found associated with a lower stroke risk in women. Higher consumptions of FRF, in particular citrus
fruits, strawberries and grapes, were associated with a lower risk of developing stroke in Japanese women.

Key words: Fruits: Flavonoid: Stroke: Incidence: Cohort study

Stroke is the second main cause of mortality(1) and the leading flavonoid-rich fruits (FRF) may contribute to the prevention of
cause of adult disability worldwide(2). The prevalence and bur- stroke.
den of stroke remain high globally(1), with the economic costs for Although many prospective cohort studies have been carried
post-stroke care keeping increasing due to the rapid rise of the out and shown that fruit consumptions may be protective against
population older than 65 years(3). Primary prevention strategies stroke(6), the relationships of specific types of fruits with stroke
are, therefore, needed to reduce the burden of stroke. risk were uncertain due to a limited number of available studies.
Fruits are rich in vitamins, minerals and polyphenols. Fruit For example, only one cohort study examined strawberries(8),
consumption has been widely considered essential to prevent and two cohort studies examined grapes(9,10), which are popular
diseases and to promote health. A number of observational stud- fruits rich in flavonoids. Furthermore, evidence from Asian pop-
ies have shown that higher fruit intakes are associated with a ulations linking specific fruits consumption and stroke risk has
lower risk of many chronic diseases, including hypertension(4), been limited and controversial. For example, a Japanese cohort
diabetes(5), CHD, stroke, cancer and all-cause mortality(6). study showed a strong inverse relationship between the fre-
Among various nutrients in fruits, flavonoids have received quency of citrus fruit intake and total stroke and cerebral infarc-
increasing interest for their potential effects on human health. tion risk(11). In contrast, a Chinese cohort study found no
Flavonoids are proposed to possess cardio-protective mecha- association between intakes of total fruit, citrus fruit and stroke
nisms, including having antioxidant and anti-inflammatory mortality(12).
effects, modulating lipid metabolism and platelet function and Considering the limited and conflicting reports on specific
improving hypertension(7). Such health effects indicate that fruits in Asian populations, we, therefore, aimed to examine

Abbreviation: FRF, flavonoid-rich fruit.

* Corresponding author: Hiroyasu Iso, email [email protected]

 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
2 Q. Gao et al.

103 802 men and women at baseline

-with a history of stroke, heart disease or cancer (n 7012)

-BMI <14 or > 40 kg/m2 (n 3854)

-implausible energy intake (n 5759)

39 843 men and 47 334 women for the analysis

Fig. 1 Flow chart for participant selection.

the association of specific fruit consumption, in particular FRF oranges, other citruses, pears, strawberries, grapes, waterme-
consumption, with stroke risk in Japanese men and women. lons, melons, papayas, peaches, kiwifruits, pineapples, bananas
and Japanese persimmons. Their flavonoid contents are listed in
online Supplemental Table 1. Portion sizes of fruits ranged from
Methods 50 to 130 g. Apples, oranges, other citruses, pears, strawberries
Study design and population and grapes were defined as FRF according to their flavonoind
contents(16). The remaining fruits were grouped as other fruits.
The Japan Public Health Center-based prospective study (JPHC Nutrient intakes were computed based on the Standard Table
study) is an ongoing prospective study with the primary aim to of Food Composition in Japan(17). A validity study examining
provide evidence for the prevention of CVD, cancer and other the correlations between the FFQ and dietary records showed
non-communicable diseases. The detailed study design has that the Spearman correlation coefficients of citrus fruits and
been reported previously(13). The JPHC study was initiated in other fruits were 0·48 and 0·51 for men and 0·36 and 0·41 for
1990 (cohort I) and 1993 (cohort II). In the cohort I, we recruited women(18), respectively.
men and women aged 40–59 years from five public health
centres, while in the cohort II, men and women aged 40–69 years Covariate assessment
were enrolled from the other six public health centres. We
obtained socio-demographic information, self-rated physical Data on age, sex, occupations, personal history of diseases,
health and mental health, disease history, smoking status, drink- medication use, smoking status, drinking status and physical
ing status, physical activities and dietary habits through self- activities were obtained from a self-administrated questionnaire.
administered questionnaires at baseline and updated at BMI was calculated based on self-reported height and weight.
follow-up surveys 5 and 10 years after the baseline survey.
Dietary intake was assessed with a validated FFQ(14,15). The Identification of stroke incidence
FFQ was extended to 138 items at the 5-year follow-up survey. The outcome of interest was the first incidence of stroke. The
This extension made it possible to assess various food consump- administrative districts of the JPHC cohort registered a total of
tions, and we treated 103 802 men and women responding to the seventy-eight major hospitals with the capability of treating
5-year follow-up survey as the baseline population for this analy- patients from stroke. Medical records were reviewed to identify
sis (response rate 76·2 %). We excluded men and women report- stroke cases by physicians, hospital workers or investigators
ing a history of stroke, CVD or cancer at the baseline (n 7012), who were blinded to participant’s responses to the baseline
those having an extreme BMI (<14 or >40 kg/m2) (n 3854) and follow-up surveys. Stroke cases were diagnosed based on
and those reporting implausible energy intake (<800 or the criteria of the National Survey of Stroke(19), specifically, a sud-
>3600 kcal/d) (n 5759). Ultimately, 87 177 men and women den or rapid occurrence of a focal, non-convulsive neurological
were eligible for the present analysis. deficit lasting for at least 24 h or until death. Information obtained
from computed tomography scanning, MRI and autopsy was
Dietary assessment used to determine stroke subtypes (i.e. cerebral infarctions
and haemorrhagic stroke).
Dietary intakes were assessed using a semi-quantitative FFQ
with 138 items. Participants were asked their intake frequencies
Statistical analysis
of individual food items during the past 12 months. The FFQ
included nine categories ranging from never to 7 or more The main exposure of interest was FRF consumption in this
times/d for major food items. Fruits in the FFQ included apples, study. Because FRF intakes were considerably higher in women
 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
Flavonoid-rich fruits and stroke 3

than that in men, all the analyses were stratified by sex. We cal- vegetables, soya food, other fruits and total energy. Similar
culated age-adjusted means and proportions of the population results were found in women.
characteristics at the study baseline according to total FRF intake During a mean of 13·1 years of follow-up, a total of 4091 inci-
(quintiles). P for trend across groups use generalised linear mod- dent strokes were documented, including 2557 cerebral infarc-
els (linear regression). Person-years of follow-up were calcu- tions, 1516 haemorrhagic strokes and eighteen unknown
lated from the date at enrolment until the date of stroke subtypes. In the age-adjusted model, men in the highest group
incidence, death, moving out from the study area or end of fol- of FRF consumption, compared with those in the lowest group,
low-up (31 December 2009 for cohort I and 31 December 2012 experienced a lower risk of developing stroke (Table 2). The
for cohort II), whichever came first. Cox proportional hazards association was attenuated but remained significant after further
regression model was used to compute hazard ratios for stroke adjustment for sex, study area, BMI, medication use for hyper-
incidence according to FRF consumption. The group of lowest tension and hypercholesterolaemia, diabetes history, occupa-
consumption was treated as the reference group. All hazard tion, drinking status, smoking status and sports level
ratios were adjusted for age at study enrolment. In multivariable (model 2). Further adjustment for dietary factors including sea-
model 2, we further adjusted for the following variables: study food, red meat, processed meat, soya food, green tea, coffee, veg-
area, BMI (<18·5, 18·5–20·9, 21–22·9, 23–24·9, 25–29·9, etables and total energy led to a non-significant association
≥30 kg/m2 or missing), medication use for hypertension and (model 3). Table 3 shows the associations of FRF consumption
hypercholesterolaemia (yes or no), history of diabetes (yes or with stroke risk in women. Women in the highest group of FRF
no), alcohol use (never, 0–22·9, 23–45·9, 46–68·9, ≥69 g/d or consumption, compared with those in the lowest group, had a sig-
missing), smoking status (never, past, current <1 or ≥1 pack/d nificantly lower risk of stroke in all three models. Furthermore, the
or missing), sports level (never, 1–3 times/month, 1–2, 3–4, association in women tended to be dose-dependent. However,
≥5 times/week or missing) and occupation (unemployed, sex did not significantly modify the association as indicated by
white-collar worker, blue-collar worker or other). In multivari- the interaction analysis (Pfor interaction > 0·34 in all models).
able model 3, we further adjusted for dietary factors including Similarly, age, BMI and smoking status did not modify the associ-
coffee intake (never, 1–2, 3–6 cups/week, 1 or ≥2 cups/d), green ation (all Pfor interaction > 0·13 in all models).
tea intake (never, 1–2, 3–6 cups/week, 1, 2–3 or ≥4 cups/d) and We next examined the associations for specific FRF: apples/
quintile intakes of total energy, red meat, processed meat, sea- pears, citrus fruits, strawberries and grapes. None of these fruits
food, soya foods and vegetables. Interaction analyses testing was associated with stroke risk in men. In women, inverse asso-
whether sex could modify the association between FRF con- ciations were found for citrus fruits, strawberries and grapes but
sumption and stroke risk were performed by adding a product not for apples/pears (Table 3). In addition, other fruit consump-
term of FRF consumption and sex in Cox regression models. tion was not related to stroke risk in either men or women.
Similar interaction analyses were conducted for age, BMI and Additional analyses according to stroke subtypes yielded similar
smoking status. Trend tests were carried out by modelling FRF results as those for the total stroke (see online supplementary
consumption as a continuous variable and assigning the median material, Supplemental Table 2).
intake value of each group. The associations between specific Furthermore, we examined the association between flavo-
FRF, including apples/pears, citrus fruits, strawberries and noid intakes from fruits and risk of stroke among men and
grapes, and stroke risk were also performed. Consumptions of women (Table 4). The results were similar as our main analysis
strawberries and grapes were divided into quartiles because that higher flavonoid intakes from fruits were associated with a
of relatively lower consumptions. In addition, we examined lower risk of stroke among women but not among men.
the association between flavonoid intakes from fruits and risk Sensitivity analyses excluding cases diagnosed in the first 2
of stroke. Sensitivity analysis was performed by excluding cases years of the follow-up, excluding those with history of diabetes
diagnosed in the first 2 years of the follow-up, by excluding those and excluding those with medication use for hypertension and
with history of diabetes and by excluding adults those medica- hypercholesterolaemia yielded similar results. The fully adjusted
tion use for hypertension and hypercholesterolaemia. All analy- hazard ratios were from 0·86 (95 % CI 0·72, 1·04) to 0·92 (95 % CI
ses were performed using SAS software (version 9.4; SAS 0·79, 1·09) for men and from 0·72 (95 % CI 0·60, 0·87) to 0·75
Institute Inc.). All P values were two-sided, and P value < 0·05 (95 % CI 0·59, 0·94) for women.
was considered as statistically significant.

Discussion
Results
In this prospective cohort study of Japanese populations, we
Table 1 shows the baseline characteristics of the study popula- found that higher intakes of FRF were associated with a
tion according to FRF stratified by sex. Median FRF intakes were decreased risk of developing stroke in women but not in men.
9·1, 27·5, 57·4, 105·4 and 215·1 g/d for quintiles in men and were For specific FRF, consumptions of citrus fruits, strawberries
19·0, 61·2, 108·2, 173·1 and 329·7 g/d for quintiles in women. and grapes were found associated with lower stroke risk in
Compared with men with lower FRF consumption, those with women. For stroke subtypes, similar results were found as that
higher consumption were slightly older and more likely to be for total stroke.
never smokers and never drinkers. Also, higher consumption Consumptions of apples/pears were found to help reduce
of FRF was related to higher intakes of green tea, seafood, stroke risk in epidemiological studies(6). However, the absence
 4
Table 1. Sex-specific baseline characteristics according to flavonoid-rich fruit (FRF) intake*

Men Women
Q1 Q3 Q5 Q1 Q3 Q5
Mean SD Mean SD Mean SD Ptrend Mean SD Mean SD Mean SD Ptrend

Median intakes (g/d) 9·1 57·4 215·1 19·0 108·2 329·7

No. of participants 8159 7970 7968 9467 9466 9467
Age (years) 55·5 8 56·3 8 57·7 8 <0·001 56·6 8 56·7 8 57·0 8 <0·001
BMI (kg/m2) 23·7 3 23·5 3 23·6 3 0·003 23·7 3 23·4 3 23·5 3 <0·001
History of diabetes (%) 5·6 6·6 6·2 0·03 3·4 3·0 2·5 <0·001
Use of hypertension drug (%) 18·4 18·5 17·2 0·06 19·3 19·2 18·4 0·04

Q. Gao et al.
Use of hypercholester olaemia drug (%) 3·0 3·5 3·9 <0·001 4·8 7·1 7·9 <0·001
Never smoker (%) 29·3 34·2 39·7 <0·001 85·6 90·3 92·8 <0·001
Never drinker (%) 21·0 22·6 28·2 <0·001 77·5 75·9 80·0 <0·001
Alcohol (g/d) 35·5 40 27·9 32 21·5 27 <0·001 3·3 14 1·8 8 1·1 5 <0·001
Sports ≥ 3 times/week (%) 9·1 10·8 11·8 <0·001 8·8 10·6 12·1 <0·001
White collar (%) 40·4 44·4 43·3 <0·001 20·7 22·8 20·0 0·15
Coffee (cups/d) 0·8 1 0·8 1 0·8 1 0·42 0·9 1 0·8 1 0·8 1 0·37
Green tea (cups/d) 1·8 2 2·3 3 3·0 3 <0·001 1·8 3 2·7 3 3·5 3 <0·001
Seafood (g/d) 64·1 58 90·7 61 115·3 72 <0·001 60 52 91·8 58 114·2 73 <0·001
Red meat (g/d) 42·3 43 44·3 40 45·1 35 <0·001 42·4 46 40·2 37 40·7 35 <0·001
Processed meat (g/d) 6·3 10 6·6 9 6·7 9 0·16 6·8 10 6·5 10 6·7 9 <0·001
Vegetable (g/d) 124 111 185 125 283 172 <0·001 159 135 224 138 330 200 <0·001
Soya food (g/d) 47·9 67 59·2 61 72·7 62 <0·001 53·4 69 64·1 59 76·0 64 <0·001
Other fruits (g/d) 6·0 17 20·2 26 54·9 61 <0·001 10·7 23 32·6 32 74·9 81 <0·001
Flavonoid from fruits (mg/d) 4·6 3·0 29·3 6·6 129 71·0 <0·001 9·9 5·8 53·6 9·8 189 89·5 <0·001
Total energy (kcal/d) 1833 581 2077 548 2404 552 <0·001 1544 488 1836 500 2201 546 <0·001

* Values are age-adjusted. Continuous variables are expressed as mean and SD and categorical variables are expressed as percentages.

Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
Flavonoid-rich fruits and stroke 5

Table 2. Intakes of flavonoid-rich fruits (FRF) and other fruits in relation to stroke risk in men
(Hazard ratios and 95 % confidence intervals)

Q2 Q3 Q4 Q5
Q1 HR 95 % CI HR 95 % CI HR 95 % CI HR 95 % CI Ptrend

FRF
Median intakes (g/d) 9·1 27·5 58·4 105·4 215·1
Person-year 102 982 99 170 101 406 102 238 102 832
No. of cases 513 429 472 466 476
Model 1 1·00 0·84 0·74, 0·96 0·89 0·78, 1·01 0·83 0·73, 0·94 0·80 0·71, 0·91 0·001
Model 2 1·00 0·88 0·78, 1·01 0·93 0·82, 1·06 0·88 0·77, 1·00 0·85 0·75, 0·97 0·03
Model 3 1·00 0·94 0·82, 1·07 1·00 0·87, 1·15 0·96 0·83, 1·11 0·93 0·79, 1·09 0·80
Apples/pears
Median intakes (g/d) 1·9 5·3 13·0 27·8 63·0
Person-year 100 604 100 740 102 223 101 984 103 077
No. of cases 540 420 482 460 454
Model 1 1·00 0·86 0·76, 0·96 0·86 0·77, 0·97 0·87 0·77, 0·98 0·82 0·71, 0·94 0·002
Model 2 1·00 0·89 0·79, 1·00 0·91 0·81, 1·03 0·93 0·82, 1·06 0·88 0·77, 1·01 0·08
Model 3 1·00 0·94 0·84, 1·07 0·99 0·86, 1·12 1·01 0·88, 1·16 0·96 0·82, 1·13 0·92
Citrus fruits
Median intakes (g/d) 3·6 11·6 26·6 58·6 118·9
Person-year 96 775 110 066 97 230 104 958 99 599
No. of cases 497 494 402 502 461
Model 1 1·00 0·92 0·82, 1·03 0·88 0·78, 0·99 0·83 0·73, 0·94 0·81 0·71, 0·93 <0·001
Model 2 1·00 0·96 0·85, 1·07 0·91 0·80, 1·02 0·86 0·76, 0·98 0·84 0·73, 0·96 <0·001
Model 3 1·00 1·00 0·89, 1·12 0·95 0·83, 1·08 0·91 0·8, 1·05 0·88 0·76, 1·03 0·12
Strawberries
Median intakes (g/d) 0 2·4 7·7 18·0
Person-year 111 334 177 359 121 997 97 938
No. of cases 673 673 673 673
Model 1 1·00 0·80 0·72, 0·89 0·77 0·68, 0·86 0·72 0·63, 0·81 <0·001
Model 2 1·00 0·86 0·77, 0·95 0·84 0·74, 0·94 0·82 0·72, 0·94 0·002
Model 3 1·00 0·91 0·81, 1·02 0·90 0·79, 1·02 0·88 0·76, 1·02 0·18
Grapes
Median intakes (g/d) 0 2·0 6·4 15·0
Person-year 122 591 170 870 109 173 105 996
No. of cases 683 743 461 469
Model 1 1·00 0·82 0·74, 0·91 0·80 0·71, 0·90 0·83 0·73, 0·93 <0·001
Model 2 1·00 0·88 0·79, 0·98 0·87 0·77, 0·98 0·90 0·79, 1·01 0·07
Model 3 1·00 0·93 0·83, 1·04 0·94 0·83, 1·07 0·97 0·85, 1·12 0·94
Other fruits
Median intakes (g/d) 3·1 10·4 19·0 34·5 71·2
Person-year 101 356 101 522 101 458 102 562 101 733
No. of cases 547 447 469 410 483
Model 1 1·00 0·82 0·73, 0·93 0·85 0·76, 0·97 0·70 0·61, 0·79 0·78 0·69, 0·88 <0·001
Model 2 1·00 0·87 0·76, 0·98 0·92 0·82, 1·05 0·77 0·67, 0·87 0·86 0·76, 0·98 0·29
Model 3 1·00 0·89 0·78, 1·01 0·96 0·84, 1·09 0·80 0·70, 0·92 0·90 0·78, 1·05 0·38
Total fruits
Median intakes (g/d) 23·9 78·1 132·2 205·1 364·8
Person-year 101 083 101 506 101 602 102 866 102 280
No. of cases 508 490 444 438 477
Model 1 1·00 0·93 0·82, 1·05 0·84 0·74, 0·95 0·79 0·70, 0·90 0·82 0·72, 0·93 <0·001
Model 2 1·00 0·97 0·86, 1·10 0·91 0·80, 1·03 0·87 0·76, 0·99 0·90 0·79, 1·03 0·03
Model 3 1·00 1·00 0·88, 1·14 0·94 0·82, 1·08 0·91 0·79, 1·05 0·96 0·82, 1·12 0·31

* Model 1: adjusted for age.

† Model 2: adjusted further for study area, BMI, occupation, medication use for hypertension and hypercholesterolaemia, diabetes history, drinking status, smoking status and sports.
‡ Model 3: adjusted further for dietary intakes of seafood, red meat, processed meat, soya food, green tea, coffee, vegetables and total energy.

of such an association in the present study was not surprising. An inverse association between citrus fruits and stroke risk in
Peeling is common among Japanese when they consume apples Japanese women was consistent with previous studies con-
or pears, while the peels of apples have been shown to contain ducted in Western populations. In a Finnish cohort of 10 054
much greater amount of total phenolic compounds, the ingre- men and women, the incidence of cerebrovascular disease
dient responsible for the antioxidant activity, than the fleshes was lower in the group taking the dietary sources rich in flavo-
do(20). With antioxidant-rich peels frequently discarded as a noids, orange (0·79; 95 % CI 0·64, 0·98; P = 0·02) and grapefruit
waste before eating, it may explain the null association between (0·63; 95 % CI 0·36, 1·09; P = 0·07)(21). A prospective study
the consumption of apples/pears and stroke risk in the Japanese among 69 622 women during 14 years of follow-up from the
population. Nurses’ Health Study also indicated that women’s intake of citrus
 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
6 Q. Gao et al.

Table 3. Intakes of flavonoid-rich fruits (FRF) and other fruits in relation to stroke risk in women
(Hazard ratios and 95 % confidence intervals)

Q2 Q3 Q4 Q5
Q1 HR 95 % CI HR 95 % CI HR 95 % CI HR 95 % CI Ptrend

FRF
Median intakes (g/d) 19·0 61·2 108·2 173·1 329·7
Person-year 125 150 126 427 127 759 128 395 128 543
No. of cases 423 352 338 314 423
Model 1 1·00 0·83 0·72, 0·96 0·78 0·68, 0·90 0·73 0·63, 0·84 0·70 0·61, 0·81 <0·001
Model 2 1·00 0·84 0·73, 0·97 0·80 0·69, 0·92 0·74 0·63, 0·86 0·71 0·61, 0·83 <0·001
Model 3 1·00 0·83 0·72, 0·97 0·79 0·67, 0·92 0·73 0·61, 0·86 0·70 0·58, 0·84 <0·001
Apples/pears
Median intakes (g/d) 3·5 11·6 25·9 47·9 83·8
Person-year 126 192 115 846 134 964 129 503 129 770
No. of cases 412 340 312 354 317
Model 1 1·00 0·97 0·83, 1·14 0·80 0·68, 0·93 0·85 0·73, 0·99 0·85 0·73, 0·99 0·03
Model 2 1·00 1·01 0·86, 1·19 0·83 0·70, 0·97 0·90 0·77, 1·05 0·88 0·75, 1·04 0·11
Model 3 1·00 1·02 0·87, 1·20 0·85 0·71, 1·01 0·93 0·78, 1·11 0·93 0·77, 1·11 0·43
Citrus fruits
Median intakes (g/d) 7·2 27·0 58·6 97·1 215·2
Person-year 123 138 127 966 131 897 122 963 130 309
No. of cases 388 391 322 320 314
Model 1 1·00 0·99 0·84, 1·16 0·82 0·70, 0·96 0·77 0·66, 0·90 0·74 0·63, 0·86 <0·001
Model 2 1·00 1·01 0·86, 1·18 0·84 0·72, 0·99 0·78 0·67, 0·92 0·76 0·65, 0·89 <0·001
Model 3 1·00 1·02 0·86, 1·20 0·86 0·72, 1·02 0·80 0·67, 0·95 0·78 0·66, 0·94 <0·001
Strawberries
Median intakes (g/d) 0 2·4 7·7 27·0
Person-year 139 034 172 638 187 777 136 826
No. of cases 524 485 424 302
Model 1 1·00 0·91 0·80, 1·03 0·74 0·65, 0·84 0·68 0·59, 0·79 <0·001
Model 2 1·00 0·93 0·82, 1·05 0·78 0·68, 0·89 0·74 0·63, 0·85 <0·001
Model 3 1·00 0·93 0·82, 1·06 0·79 0·69, 0·91 0·74 0·63, 0·87 <0·001
Grapes
Median intakes (g/d) 0 2·0 6·4 22·5
Person-year 167 580 165 075 159 475 144 144
No. of cases 570 467 373 325
Model 1 1·00 0·94 0·83, 1·07 0·78 0·69, 0·89 0·74 0·65, 0·85 <0·001
Model 2 1·00 0·97 0·85, 1·09 0·82 0·72, 0·94 0·77 0·67, 0·89 <0·001
Model 3 1·00 0·97 0·85, 1·10 0·83 0·72, 0·96 0·77 0·66, 0·90 <0·001
Other fruits
Median intakes (g/d) 6·0 16·7 30·8 52·1 101·1
Person-year 126 231 127 520 126 986 127 708 127 831
No. of cases 425 328 338 305 339
Model 1 1·00 0·81 0·70, 0·94 0·83 0·72, 0·96 0·73 0·63, 0·84 0·79 0·69, 0·91 0·07
Model 2 1·00 0·86 0·74, 0·99 0·88 0·76, 1·02 0·78 0·67, 0·91 0·85 0·74, 0·99 0·37
Model 3 1·00 0·87 0·75, 1·01 0·93 0·79, 1·08 0·83 0·71, 0·98 0·93 0·79, 1·11 0·28
Total fruits
Median intakes (g/d) 53·5 129·1 198·3 296·6 499·9
Person-year 125 837 126 910 127 594 127 896 128 561
No. of cases 399 376 316 311 329
Model 1* 1·00 0·95 0·82, 1·09 0·80 0·69, 0·92 0·78 0·67, 0·90 0·82 0·71, 0·95 <0·001
Model 2† 1·00 0·98 0·85, 1·14 0·84 0·72, 0·97 0·81 0·70, 0·95 0·85 0·73, 0·99 0·003
Model 3‡ 1·00 0·99 0·86, 1·15 0·86 0·73, 1·00 0·85 0·72, 1·00 0·91 0·76, 1·09 0·11

* Model 1: adjusted for age.

† Model 2: adjusted further for study area, BMI, occupation, medication use for hypertension and hypercholesterolaemia, diabetes history, drinking status, smoking status and sports.
‡ Model 3: adjusted further for dietary intakes of seafood, red meat, processed meat, soya food, green tea, coffee, vegetables and total energy.

fruits/juices tended to be associated with a reduced risk for We are aware of one observational cohort study which exam-
ischaemic stroke (RR = 0·90 and 95 % CI 0·77, 1·05)(22). A cohort ined the association between strawberry intake and risk of
study during a 24-year follow-up based on data from the Finnish stroke. In the Women’s Health Study of 38 176 women with a
Mobile Clinic Health Examination Survey on 3932 men and mean age of 54·4 years during a mean follow-up of 10·1 years,
women aged 40–74 years indicated that the intake of citrus fruits consuming strawberry ≥2 servings/week compared with having
helped to reduce the risks of both ischaemic stroke and intra- no strawberry consumption did not appear to affect risk of stroke
cerebral haemorrhage(23). Another prospective cohort study of (hazard ratio = 1·40; 95 % CI 0·86, 2·28) [10], which was inconsis-
54 506 men and women in the Danish showed the reduced risk tent with our finding. The authors noted that the low level of
of ischaemic stroke was most obvious in the highest quintiles of strawberry consumption in their study may be responsible for
intake of citrus fruit(24). the null association. In addition, the number of incident stroke
 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
Flavonoid-rich fruits and stroke 7

Table 4. Intakes of flavonoids from fruits in relation to stroke risk in men and women
(Hazard ratios and 95 % confidence intervals)

Q2 Q3 Q4 Q5
Q1 HR 95 % CI HR 95 % CI HR 95 % CI HR 95 % CI Ptrend

Men
Median intakes (mg/d) 4·6 13·8 29·0 51·8 107·6
Person-year 100 925 101 635 102 068 101 947 102 759
No. of cases 514 428 467 468 480
Model 1* 1·00 0·83 0·73, 0·95 0·87 0·77, 0·99 0·84 0·74, 0·95 0·80 0·71, 0·91 0·007
Model 2† 1·00 0·87 0·76, 0·99 0·92 0·81, 1·04 0·89 0·78, 1·01 0·86 0·76, 0·98 0·16
Model 3‡ 1·00 0·91 0·80, 1·04 0·97 0·85, 1·11 0·96 0·83, 1·11 0·95 0·82, 1·11 0·71
Women
Median intakes (mg/d) 9·5 29·8 52·9 86·3 162·7
Person-year 125 637 126 670 127 824 127 906 128 758
No. of cases 423 359 319 338 292
Model 1 1·00 0·86 0·75, 0·99 0·76 0·65, 0·88 0·80 0·69, 0·92 0·68 0·58, 0·79 <0·001
Model 2 1·00 0·87 0·75, 1·00 0·77 0·66, 0·90 0·80 0·69, 0·93 0·68 0·58, 0·80 <0·001
Model 3 1·00 0·87 0·75, 1·01 0·78 0·66, 0·92 0·83 0·70, 0·98 0·72 0·60, 0·86 0·005

* Model 1: adjusted for age.

† Model 2: adjusted further for study area, BMI, occupation, medication use for hypertension and hypercholesterolaemia, diabetes history, drinking status, smoking status and sports.
‡ Model 3: adjusted further for dietary intakes of seafood, red meat, processed meat, soya food, green tea, coffee, vegetables and total energy.

cases was limited in their study (n 339). The differences in the The strengths of our study included a large sample size and a
levels of strawberry consumption and statistical power may have prospective population-based cohort design, which reduced the
contributed to the inconsistent results between that study risk of selection and recall biases. However, our study also had
and ours. limitations. First, the FFQ used in this cohort did not include other
In various animal experiments, polyphenols from grape pow- FRF such as blueberries and cherries. However, because the con-
der, taken as a diet supplement, have been found to deter macro- sumptions of blueberries and cherries were relatively low in
phage cholesterol accumulation, inhibit enhanced development Japan, a lack of data on them was not likely to change our find-
of atherosclerosis(25) and protect the brain against ischaemic ings materially. Second, residual confounding is always an alter-
damage(26). However, previous studies in humans found no native explanation for findings from observational studies. In our
association between grape consumption and risk of stroke(10,27). study, higher FRF intakes were associated with a healthier life-
The inverse association of grape consumption with stroke risk in style, including less smoking, less drinking and higher intakes
Japanese women warranted further research. of seafood, vegetables, soya food, green tea and other fruits.
Flavonoids have been suggested to have potential effects on Although we controlled many risk factors, unmeasured factors
stroke prevention. The association between flavonoids intake may have an influence on our findings. Moreover, we only cal-
and stroke risk has been evaluated in several epidemiological culated the content of flavonoids in fruits and did not include the
studies. The Zutphen Study, a cohort study examining 552 content of flavonoids in vegetables and other foods (such as
men aged 50–69 years for 15 years since 1970, showed that onions, chocolate or cocoa, tea and soya)(32). A lack of validity
the habitual intake of flavonoids (the highest v. the lowest quar- of flavonoid estimates may also affect our findings. Third, dietary
tile of flavonoid intake (≥28·6 mg/d < 18·3 mg/d)) may protect habits may change during the follow-up period. As only baseline
against stroke (RR = 0·27; 95 % CI 0·11, 0·70)(21). The Finnish dietary data were used for analysis, the influence of dietary
cohort on a sample of 10 054 participants showed that the flavo- changes on the associations could not be examined. Finally,
noid groups (flavonols, flavones and flavanones) were inversely our findings may be limited in generalisation to other Western
associated with the incidence of cerebrovascular disease leading populations in which dietary habits and lifestyles are different
to hospitalisation or death(28). However, no association between from those of the Japanese population.
total flavonoid intake and stroke mortality (Pfor trend = 0·83) was In conclusion, this prospective cohort study provided evi-
found in a prospective study on 34 492 postmenopausal women dence that higher consumptions of FRF, in particular citrus fruits,
in Iowa over 10 years of follow-up(29). In addition, the cohort strawberries and grapes, were associated with a lower risk of
study on 26 593 male smokers aged 50–69 years during a 6·1- developing stroke in Japanese women.
year follow-up performed by the Alpha-Tocopherol, Beta-
Carotene Cancer Prevention Study in Finland showed that
dietary flavonols and flavones were not associated with risk
Acknowledgements
for stroke in male smokers(30). Recently, it was found that dietary
flavonoid intake may be inversely associated with the risk of Members of the Japan Public Health Center-based prospective
stroke in one meta-analysis of eleven prospective studies. In study (JPHC study) group are listed in https://epi.ncc.go.jp/
addition, it is indicated that a 100 mg/d increment in flavonoid en/jphc/781/8233.html.
intake led to a 9 % lower risk of stroke (RR 0·91; 95 % CI 0·77, J-Y. D. designed the study, analyzed the data, and edited the
1·08) in a dose–response analysis(31). manuscript. Q.G. analyzed the data and drafted the manuscript.
 Downloaded from https://www.cambridge.org/core. Carleton University Library, on 04 Jun 2021 at 18:24:39, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0007114521000404
8 Q. Gao et al.

All authors conducted the technique review and edited the 13. Tsugane S & Sawada N (2014) The JPHC study: design and
manuscript. The JPHC study was supported by National some findings on the typical Japanese diet. Japan J clin
Cancer Center Research and Development Fund (23-A-31[toku] Oncol 44, 777–782.
14. Ishihara J, Sobue T, Yamamoto S, et al. (2003) Validity and repro-
and 26-A-2) (since 2011) and a Grant-in-Aid for Cancer Research
ducibility of a self-administered food frequency questionnaire in
from the Ministry of Health, Labour and Welfare of Japan (from the JPHC Study Cohort II: study design, participant profile and
1989 to 2010). This work was supported by the JSPS KAKENHI results in comparison with Cohort I. J Epidemiol 13, S134–147.
(A18H06391, T19K214700) to Dr. Dong. 15. Sasaki S, Matsumura Y, Ishihara J, et al. (2003) Validity of a self-
The authors declare that there are no conflicts of interest. administered food frequency questionnaire used in the 5-year
follow-up survey of the JPHC Study Cohort I to assess dietary
fiber intake: comparison with dietary records. J Epidemiol
13, S106–114.
Supplementary material 16. McCullough ML, Peterson JJ, Patel R, et al. (2012) Flavonoid
intake and cardiovascular disease mortality in a prospective
For supplementary material referred to in this article, please visit cohort of US adults. Am J Clin Nutr 95, 454–464.
https://doi.org/10.1017/S0007114521000404 17. The Council for Science and Technology, Ministry of Education,
Culture, Sports, Science, and Technology, the Government of
Japan (2005) Standard Tables of Food Composition in Japan:
Fifth Revised and Enlarged Edition. Tokyo, Japan: Ministry of
References
Finance.
1. Johnson CO, Nguyen M, Roth GA, et al. (2019) Global, regional, 18. Nanri A, Shimazu T, Ishihara J, et al. (2012) Reproducibility and
and national burden of stroke, 1990–2016: a systematic analysis validity of dietary patterns assessed by a food frequency question-
for the Global Burden of Disease Study 2016. Lancet Neurol 18, naire used in the 5-year follow-up survey of the Japan Public
439–458. Health Center-Based Prospective Study. J Epidemiol 22, 205–215.
2. Rajsic S, Gothe H, Borba HH, et al. (2019) Economic burden of 19. Robins M & Baum HM (1981) The National Survey of Stroke.
stroke: a systematic review on post-stroke care. Eur J Health Incidence. Stroke 12, I45–57.
Econ: HEPAC: Health Econ Prev Care 20, 107–134. 20. Escarpa A & Gonzalez MC (1998) High-performance liquid
3. Ovbiagele B, Goldstein LB, Higashida RT, et al. (2013) chromatography with diode-array detection for the determina-
Forecasting the future of stroke in the United States: a policy tion of phenolic compounds in peel and pulp from different
statement from the American Heart Association and apple varieties. J Chromatogr A 823, 331–337.
American Stroke Association. Stroke 44, 2361–2375. 21. Keli SO, Hertog MG, Feskens EJ, et al. (1996) Dietary flavo-
4. Schwingshackl L, Schwedhelm C, Hoffmann G, et al. (2017) noids, antioxidant vitamins, and incidence of stroke: the
Food groups and risk of hypertension: a systematic review Zutphen study. Arch Intern Med 156, 637–642.
and dose-response meta-analysis of prospective studies. Adv 22. Cassidy A, Rimm EB, O’Reilly EJ, et al. (2012) Dietary flavonoids
Nutr 8, 793–803. and risk of stroke in women. Stroke 43, 946–951.
5. Schwingshackl L, Hoffmann G, Lampousi AM, et al. (2017) 23. Mizrahi A, Knekt P, Montonen J, et al. (2009) Plant foods and
Food groups and risk of type 2 diabetes mellitus: a systematic the risk of cerebrovascular diseases: a potential protection of
review and meta-analysis of prospective studies. Eur J fruit consumption. Br J Nutr 102, 1075–1083.
Epidemiol 32, 363–375. 24. Johnsen SP, Overvad K, Stripp C, et al. (2003) Intake of fruit and
6. Aune D, Giovannucci E, Boffetta P, et al. (2017) Fruit and veg- vegetables and the risk of ischemic stroke in a cohort of Danish
etable intake and the risk of cardiovascular disease, total men and women. Am J Clin Nutr 78, 57–64.
cancer and all-cause mortality-a systematic review and dose- 25. Fuhrman B, Volkova N, Coleman R, et al. (2005) Grape powder
response meta-analysis of prospective studies. Int J polyphenols attenuate atherosclerosis development in apolipo-
Epidemiol 46, 1029–1056. protein E deficient (E0) mice and reduce macrophage athero-
7. Weseler AR & Bast A (2012) Pleiotropic-acting nutrients require genicity. J Nutr 135, 722–728.
integrative investigational approaches: the example of flavo- 26. Wang Q, Simonyi A, Li W, et al. (2005) Dietary grape supple-
noids. J Agr Food Chem 60, 8941–8946. ment ameliorates cerebral ischemia-induced neuronal death
8. Sesso HD, Gaziano JM, Jenkins DJ, et al. (2007) Strawberry in gerbils. Mol Nutr Food Res 49, 443–451.
intake, lipids, C-reactive protein, and the risk of cardiovascular 27. Tognon G, Lissner L, Saebye D, et al. (2014) The Mediterranean
disease in women. J Am Coll Nutr 26, 303–310. diet in relation to mortality and CVD: a Danish cohort study. Br J
9. Hjartaker A, Knudsen MD, Tretli S, et al. (2015) Consumption of Nutr 111, 151–159.
berries, fruits and vegetables and mortality among 10 000 28. Knekt P, Kumpulainen J, Jarvinen R, et al. (2002) Flavonoid
Norwegian men followed for four decades. Eur J Nutr 54, intake and risk of chronic diseases. Am J Clin Nutr 76, 560–568.
599–608. 29. Yochum L, Kushi LH, Meyer K, et al. (1999) Dietary flavonoid
10. Lai HT, Threapleton DE, Day AJ, et al. (2015) Fruit intake and intake and risk of cardiovascular disease in postmenopausal
cardiovascular disease mortality in the UK Women’s Cohort women. Am J Epidemiol 149, 943–949.
Study. Eur J Epidemiol 30, 1035–1048. 30. Hirvonen T, Virtamo J, Korhonen P, et al. (2000) Intake of fla-
11. Yamada T, Hayasaka S, Shibata Y, et al. (2011) Frequency of vonoids, carotenoids, vitamins C and E, and risk of stroke in
citrus fruit intake is associated with the incidence of cardio- male smokers. Stroke 31, 2301–2306.
vascular disease: the Jichi Medical School cohort study. J 31. Tang Z, Li M, Zhang X, et al. (2016) Dietary flavonoid intake and
Epidemiol 21, 169–175. the risk of stroke: a dose-response meta-analysis of prospective
12. Wang JB, Fan JH, Dawsey SM, et al. (2016) Dietary components cohort studies. BMJ Open 6, e008680.
and risk of total, cancer and cardiovascular disease mortality in 32. Hooper L, Kroon PA, Rimm EB, et al. (2008) Flavonoids, flavo-
the Linxian Nutrition Intervention Trials cohort in China. Sci Rep noid-rich foods, and cardiovascular risk: a meta-analysis of ran-
6, 22619. domized controlled trials. Am J Clin Nutr 88, 38–50.