Dietary intake of saturated fatty acids and incident stroke and coronary heart disease in Japanese communities: the JPHC Study

Abstract

Aims

Although dietary saturated fatty acids (SFA) are considered atherogenic, associations between SFAs intake and stroke and coronary heart disease are still debated. We sought to test the hypothesis that SFA intake is associated inversely with risk of stroke and its subtypes and positively with coronary heart disease among Japanese, whose average SFA intake is lower than that of Westerners.

Methods and results

The Japan Public Health Center-based prospective Study involves two subcohorts: Cohort I, aged 45–64 in 1995 and followed-up through 2009, and Cohort II, aged 45–74 in 1998 and followed-up through 2007. A total of 38 084 men and 43 847 women were included in this report. Hazards ratios for incident total stroke, ischaemic stroke, intraparhenchymal haemorrhage, subarachnoid haemorrhage, myocardial infarction, and sudden cardiac death across quintiles of dietary SFAs were examined. We found inverse associations between SFA intake and total stroke [multivariable hazard ratio (95% confidence interval) for the highest vs. lowest quintiles = 0.77 (0.65–0.93), trend P = 0.002], intraparenchymal haemorrhage [0.61 (0.43–0.86), P for trend = 0.005], and ischaemic stroke [0.84 (0.67–1.06), trend P = 0.08], primarily for deep intraparenchymal haemorrhage [0.67 (0.45–0.99), P for trend = 0.04] and lacunar infarction [0.75 (0.53, 1.07), trend P = 0.02]. We also observed a positive association between SFAs intake and myocardial infarction [1.39 (0.93–2.08), trend P = 0.046] primarily among men. No associations were observed between SFAs intake and incidence of subarachnoid haemorrhage or sudden cardiac death.

Conclusions

In this Japanese population, SFAs intake was inversely associated with deep intraparenchymal haemorrhage and lacunar infarction and positively associated with myocardial infarction.

Epidemiology, Nutrition, Diet, Prospective study, Population

See page 1178 for the editorial comment on this article (doi:10.1093/eurheartj/eht057)

Introduction

Saturated fatty acid (SFA) intake has been considered to be atherogenic, and it is suggested that reducing SFA intake prevents atherosclerotic diseases.¹ Yet, recent meta-analyses have indicated that greater dietary intake of SFA per se may not be associated with increased risk of coronary disease,^2,3 though another meta-analysis showed that replacing SFA intake with polyunsaturated fatty acids (PUFA) intake was inversely associated with coronary events.⁴ Several cohort studies, but not all, have reported an inverse association between dietary SFA and risk of ischaemic stroke and/or intraparenchymal haemorrhage.³ Thus, the role of SFA intake in the development of both coronary diseases and stroke remains under debate.

Compared with Westerners, Asian people traditionally consume less SFA-containing foods, as shown in the Seven Countries Study⁵ and also in later studies.^6,7 The low SFA intake among Asians has been believed to be one reason why they have lower mortality from coronary disease than Westerners. However, no prospective studies have been conducted to examine the association between dietary SFA intake and incident stroke and coronary disease in Asians, except for intraparenchymal haemorrhage.⁶

The Japan Public Health Center-based prospective (JPHC) Study is one of the largest prospective studies in Japan, with a systematic incidence registry of cardiovascular disease. An advantage of this study included enough number of population to study subtypes of stroke (i.e. subarachnoid, deep or lobar intraparenchymal haemorrhage and lacunar, large-artery occlusive or embolic infarction), which is often hard to study in Western cohorts. The population is also unique in that they have very low SFA intake and high stroke and low coronary disease incidence. Our hypotheses were that low SFA intake is associated with elevated risk of deep intraparenchymal haemorrhage and lacunar infarction, and high SFA intake with elevated risk of coronary heart disease.

Methods

Study cohort

The JPHC Study is an ongoing cohort study comprising a community-based sample of 140 420 persons (68 722 men and 71 698 women) in Japan. Details of the JPHC Study protocol were given elsewhere.⁸ Briefly, the JPHC Study included two subcohorts based on Public Health Center areas; Cohort I (started in 1990, five Public Health Center areas, participants aged 40–59) and Cohort II (started in 1993, six PHC areas, aged 40–69); however, participants in two Public Health Center areas (Tokyo and Osaka) were excluded from the current study because the follow-up data were incomplete. Thus, 116 896 subjects, all residents living in the study areas at baseline, were eligible for follow-up. Participants were asked to complete self-administered questionnaires about their lifestyles and medical histories. Informed consent was obtained before participants completed the questionnaire, or sometimes from community leaders instead of individuals, as this had been in common practice for informed consent in Japan at that time. The JPHC study was approved by the institutional review boards of the National Cancer Center and the University of Tsukuba.

Baseline questionnaires

A self-administrated questionnaire was distributed to all residents aged 40–59 living in the study areas for Cohort I in 1990 and those aged 40–69 for Cohort II in 1993. The questionnaire included demographic characteristics, medical history, smoking, drinking, and dietary habits. A 5-year follow-up questionnaire, which includes food frequency questionnaire (FFQ), was distributed to all eligible study subjects in 1995 for Cohort I and in 1998 for Cohort II. Of these, 92 905 persons (79%) returned their follow-up questionnaire. We excluded persons who lost or refused follow-up (n = 226); who had histories of myocardial infarction, angina pectoris, stroke, or cancer based on the first or 5-year follow-up questionnaires or who were found in the stroke/coronary heart disease registry (n = 7964); or who did not satisfactorily complete their dietary questionnaire (n = 1114). We further excluded persons with total energy intake <1 or >99 percentile (n = 1670). Ultimately, 81 931 (38 084 men and 43 847 women) were included in this study.

The FFQ in the 5-year follow-up of Cohorts I and II was used to determine the dietary intake of SFA and other nutrients/foods. The FFQ included 138 food items and 4 choices for frequency of intake was offered for each item. The nutrients that each food item contained were mostly estimated by the Japan Food Table 5th version. Fatty acids were estimated by a JPHC study food table, which was largely based on the Japan Food Table 4th version,⁹ since the 5th version did not include fatty acids. The intake of SFA was then calculated by multiplying the frequency scores and estimated SFA for each food, and summing across all the 138 items. Validation of FFQ was tested comparing SFA determined by FFQ and that by dietary records: the Spearman's correlation coefficients between energy-adjusted SFA derived from FFQ and dietary records were 0.61 among men and 0.60 among women in Cohort I⁹ and 0.62 among men and 0.51 among women in Cohort II.¹⁰ The mean intake of SFA derived from FFQ was 11–31% higher than that from dietary records. Reproducibility of FFQ was also confirmed for Cohort II at ∼1-year interval and was fairly good (Spearman r = 0.61 among men and 0.53 among women).¹⁰ All dietary variables were adjusted for energy intake using the nutrient residual model.¹¹

Stroke and coronary heart disease registries

A total of 78 hospitals formed the register of events within the nine PHC areas. All were major hospitals with the capability of treating patients with acute coronary heart disease, stroke, or cancer events. Ninety-seven per cent of strokes and 92% of myocardial infarctions in the nine PHC areas were treated at these 78 registry hospitals. Physicians in the hospitals, PHCs, or investigators, blinded to the patients' lifestyle data, reviewed the medical records of cohort participants at each hospital, and extracted clinical information including brain images, electrocardiogram, and enzymes, onto cohort-specific registration forms.

Stroke was confirmed according to the criteria of the National Survey of Stroke,¹² which requires the presence of focal neurological deficits of sudden or rapid onset lasting at least 24 h or until death. Strokes were classified according to subtypes, i.e. intraparenchymal haemorrhage (deep or lobar), subarachnoid haemorrhage, or cerebral infarction (large-artery thrombotic, lacunar, or embolic).¹³ Almost all the registered hospitals were equipped with CT and/or MRI scans. Imaging was available for 98% of registered stroke events.

Myocardial infarction was confirmed in the medical records according to the criteria of the Monitoring Trends and Determinants of Cardiovascular Disease (MONICA) project,¹⁴ which requires typical chest pain and evidence from electrocardiogram and/or cardiac enzymes. For cases with typical prolonged chest pain (>20 min) but not confirmed by electrocardiograms or cardiac enzymes (11% of total myocardial infarctions), a possible myocardial infarction diagnosis was made and these were included in myocardial infarction cases. Sudden cardiac death was defined as a death of unknown origin that occurred within 1 h of the onset of the event.

For analysis, only first-ever stroke (regardless of subtype) or coronary heart disease (myocardial infarction or sudden cardiac death) events during the follow-up were included; recurrent events were excluded. If one person suffered both coronary heart disease and stroke, both events were included.

Statistical analyses

Follow-up started from the completion of the 5-year questionnaire and ended at the date of death, emigration, incident stroke or coronary heart disease event, or end of 2009 (for Cohort I) or end of 2007 (for Cohort II), whichever came first. The incidence rates of each outcome were calculated according to quintiles of energy-adjusted SFA intake. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated after adjustment for age, sex, and other covariates with Cox proportional hazard models. The covariates included cohort; baseline body mass index (quintile); smoking status (never, ex-, current smoker of 1 to <20, 20 to <30, and 30 cigarettes/day or more); alcohol intake (0, 1–150, 151–300, 301–450, and 451 g/week or more); sports at leisure time (rarely, 1–3/month, 1–2/week, 3–4/week, or more); walking and standing time (<1, 1–3, 3 h/day or more); perceived mental stress (low, medium, high); employment status (non-employed or employed); total energy intake (quintile); and quintiles of energy-adjusted dietary intakes of carbohydrate, cholesterol, vegetables, fruit, and calcium. We did not include MUFA and PUFA in the model, so the results should be interpreted as the association of SFA in exchange for the average combination of types of fat other than SFA (i.e. MUFA and PUFA). We also did not include serum lipids, hypertension, or diabetes, since these could be potential mediators for the relation between SFA intake and cardiovascular outcomes. The linear trend of HRs across quintiles was tested by an ordinal variable (−2, −1, 0, +1, +2) for successive quintiles. Multiplicative interactions with sex were tested using a cross-product term. The proportional hazards assumption was tested using time by SFA interaction terms and was not violated for each outcome.

We used SAS version 9.1.3 Service Pack 4 (SAS Institute, Inc., Cary, NC, USA) for the analyses. All probability values for statistical tests were two-tailed and values of P < 0.05 were regarded as statistically significant.

Results

The average age was 56.7 years and the median SFA intake was 16.3 g. Most risk/dietary factors at baseline were correlated with energy-adjusted SFA intake as shown in Table 1. During the follow-up of 81 932 persons for a mean of 11.1 years (median = 9.9 years), a total of 3192 incident strokes, including 894 intraparenchymal haemorrhages, 348 subarachnoid haemorrhages, and 1939 ischaemic strokes (871 lacunar, 403 large-artery occlusive, and 563 embolic), 610 myocardial infarctions, and 116 sudden cardiac deaths were registered. There were no statistical interactions between SFA and sex in relation to any cardiovascular outcomes. Therefore, we pooled men and women for analyses.

Table 1

Baseline cardiovascular risk factors and selected dietary variables in a cohort of 38 084 men and 43 847 women according to quintiles of saturated fatty acid (SFA) intake

	Quintiles of SFA intakea
	0.8–11.7	11.8–14.8	14.9–17.7	17.8–21.5	21.6–96.7	P-value*
Men and women
Median SFA intakea, g/day	9.6	13.4	16.3	19.4	24.9
Number at risk	16 386	16 386	16 387	16 386	16 386
Age at baselineb, year	57.5 (7.7)	56.9 (7.7)	56.6 (7.8)	56.4 (7.8)	55.9 (7.9)	<0.001
Body mass index, kg/m2	23.5 (3.1)	23.6 (3.1)	23.6 (3.1)	23.5 (3.1)	23.6 (3.2)	0.005
Systolic blood pressurec, mmHg	133 (18)	131 (17)	130 (17)	130 (17)	129 (17)	<0.001
Diastolic blood pressurec, mmHg	79 (11)	78 (11)	78 (11)	78 (11)	77 (10)	<0.001
Treatment for hypertension, %	21.9	20.0	19.2	18.4	17.9	<0.001
Treatment for diabetes, %	2.2	2.6	3.0	3.3	3.8	<0.001
Serum cholesterolc, mg/dL	203 (36)	206 (36)	207 (35)	208 (35)	209 (34)	<0.001
Treatment for HC, %	5.7	5.6	5.5	4.5	4.3	<0.001
Current smoker, %	28.9	25.4	23.6	22.7	22.4	<0.001
Current drinker, %	51.2	45.8	41.8	39.7	36.2	<0.001
Sports ≥3 times/week, %	7.3	8.9	11.2	12.0	14.1	<0.001
Walking/standing ≥3 h/day, %	68.3	65.8	64.9	62.6	61.8	<0.001
High perceived mental stress, %	17.1	18.1	17.4	17.8	17.4	0.19
Non-employed, %	6.3	6.7	7.8	8.2	8.7	<0.001
Mean energy intake, Kcal/day	1956 (732)	2029 (688)	2038 (672)	2037 (653)	2057 (742)	<0.001
Dietary cholesterol, mg/day	198 (146)	275 (165)	313 (183)	345 (205)	388 (285)	<0.001
Protein, g/day	61 (29)	72 (30)	76 (30)	79 (31)	81 (35)	<0.001
Carbohydrate, g/day	289 (101)	286 (88)	275 (80)	260 (72)	232 (76)	<0.001
MUFA, g/day	13.6 (7.8)	19.5 (9.2)	22.5 (10.3)	25.4 (11.4)	30.8 (16.3)	<0.001
ω-3 PUFA, g/day	2.7 (1.7)	3.4 (2.0)	3.5 (1.9)	3.6 (1.9)	3.4 (1.8)	<0.001
ω-6 PUFA, g/day	7.4 (3.9)	9.2 (4.5)	9.9 (4.5)	10.5 (4.6)	11.2 (5.5)	<0.001
Salt, g/day	11.9 (6.8)	13.2 (6.5)	13.1 (6.3)	12.9 (5.9)	12.2 (5.8)	<0.001
Calcium, mg/day	389 (219)	497 (246)	553 (264)	602 (297)	750 (545)	<0.001
Vegetable intake, g/day	212 (192)	242 (188)	239 (172)	230 (155)	206 (148)	<0.001
Fruit intake, g/day	231 (254)	251 (226)	242 (207)	219 (184)	172 (160)	<0.001
Meat intake, g/day	28 (27)	49 (37)	63 (46)	78 (54)	103 (85)	<0.001
Dairy intake, g/day	59 (80)	122 (109)	171 (140)	221 (183)	379 (431)	<0.001

Age- and sex-adjusted means (unadjusted standard deviations), or age- and sex-adjusted percentages presented unless otherwise indicated.

HC, hypercholesterolaemia; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids.

^aEnergy-adjusted values by nutrient residual model.

^bSex-adjusted mean.

^cOnly available for subsample (8765 men and 16 047 women).

*P-values for overall difference among quintiles based on analysis of covariance.

As shown in Table 2, SFA intake was inversely associated with age-; sex-; and energy-adjusted incidence rates of total stroke, total and deep intraparenchymal haemorrhages, ischaemic stroke, lacunar infarction, and total cardiovascular disease. On the other hand, SFA intake was positively associated with myocardial infarction, but not associated with subarachnoid haemorrhage or sudden cardiac death. Further adjustment for potential cardiovascular risk factors and nutrients did not change these results substantially. There was no interaction between sex and SFA intake in relation to each outcome. The sex-specific baseline characteristics, the number of each event, multivariable HRs, and P for trend were presented in Supplementary material online, Tables S1–S3.

Table 2

Age-, sex-, and energy-adjusted and multivariate adjusted hazard ratios and 95% confidence intervals of incident cardiovascular outcomes according to quintiles of saturated fatty acids intake, JPHC Study, 1993–2008.

	Men and women
	Quintiles of SFA intake (g/day)a
	Q1	Q2	Q3	Q4	Q5	Trend P
Median intake, g/day	9.6	13.4	16.3	19.4	24.9
Person-years	179 635	182 406	182 672	182 019	180 991
Total stroke (n)	817	695	594	540	546
Age, sex, energy-adjusted HR (95% CI)	1.0	0.91 (0.82–1.00)	0.83 (0.75–0.93)	0.80 (0.72–0.90)	0.86 (0.77–0.96)	<0.001
Multivariable HR (95% CI)	1.0	0.98 (0.88–1.10)	0.90 (0.79–1.03)	0.83 (0.71–0.97)	0.77 (0.65–0.93)	0.002
Intraparenchymal haemorrhage (n)	230	186	167	161	150
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.85 (0.70–1.03)	0.80 (0.66–0.98)	0.81 (0.66–1.00)	0.79 (0.64–0.98)	0.04
Multivariable HR (95% CI)	1.0	0.92 (0.74–1.14)	0.84 (0.65–1.07)	0.76 (0.57–1.01)	0.61 (0.43–0.86)	0.005
Deep intraparenchymal hemorrhage (n)	172	138	125	114	114
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.84 (0.67–1.05)	0.80 (0.64–1.01)	0.77 (0.60–0.98)	0.80 (0.62–1.02)	0.06
Multivariable HR (95% CI)	1.0	0.91 (0.71–1.17)	0.85 (0.64–1.13)	0.75 (0.54–1.05)	0.67 (0.45–0.99)	0.04
Lobar intraparenchymal haemorrhage (n)	52	42	39	42	31
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.86 (0.57–1.29)	0.84 (0.56–1.29)	0.95 (0.63–1.44)	0.74 (0.47–1.17)	0.33
Multivariable HR (95% CI)	1.0	0.93 (0.59–1.46)	0.85 (0.51–1.43)	0.81 (0.45–1.45)	0.49 (0.24–1.02)	0.07
Subarachnoid haemorrhage (n)	64	78	63	68	75
Age-, sex-, energy-adjusted HR (95% CI)	1.0	1.13 (0.81–1.57)	0.87 (0.61–1.24)	0.91 (0.64–1.29)	1.00 (0.70–1.41)	0.57
Multivariable HR (95% CI)	1.0	1.17 (0.81–1.68)	0.91 (0.59–1.39)	0.93 (0.58–1.50)	0.87 (0.50–1.52)	0.45
Ischaemic stroke (n)	520	427	364	309	319
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.90 (0.79–1.02)	0.85 (0.74–0.97)	0.78 (0.67–0.90)	0.86 (0.74–0.99)	0.006
Multivariable HR (95% CI)	1.0	0.98 (0.85–1.13)	0.94 (0.80–1.11)	0.85 (0.70–1.03)	0.84 (0.67–1.06)	0.08
Lacunar Infarction (n)	224	216	167	126	138
Age-, sex-, energy-adjusted HR (95% CI)	1.0	1.04 (0.86–1.26)	0.88 (0.72–1.08)	0.71 (0.57–0.89)	0.83 (0.67–1.03)	0.002
Multivariable HR (95% CI)	1.0	1.15 (0.93–1.41)	0.97 (0.76–1.25)	0.75 (0.55–1.01)	0.75 (0.53–1.07)	0.02
Large-artery occlusive infarction (n)	111	81	71	73	67
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.80 (0.60–1.07)	0.78 (0.58–1.06)	0.87 (0.65–1.18)	0.86 (0.63–1.18)	0.56
Multivariable HR (95% CI)	1.0	0.83 (0.60–1.14)	0.80 (0.55–1.16)	0.89 (0.58–1.35)	0.81 (0.49–1.34)	0.55
Embolic infarction (n)	155	116	102	90	100
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.83 (0.65–1.06)	0.82 (0.64–1.05)	0.79 (0.61–1.03)	0.95 (0.73–1.22)	0.59
Multivariable HR (95% CI)	1.0	0.92 (0.71–1.20)	0.96 (0.70–1.30)	0.94 (0.65–1.34)	1.04 (0.69–1.58)	0.84
Myocardial infarction (n)	142	104	125	115	124
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.85 (0.66–1.09)	1.17 (0.92–1.49)	1.20 (0.93–1.54)	1.41 (1.10–1.81)	<0.001
Multivariable HR (95% CI)	1.0	0.90 (0.68–1.18)	1.24 (0.92–1.67)	1.24 (0.88–1.75)	1.39 (0.93–2.08)	0.046
Sudden cardiac death (n)	43	24	13	19	17
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.64 (0.39–1.06)	0.40 (0.21–0.75)	0.65 (0.37–1.12)	0.64 (0.36–1.13)	0.18
Multivariable HR (95% CI)	1.0	0.53 (0.30–0.93)	0.29 (0.14–0.60)	0.42 (0.19–0.92)	0.39 (0.15–0.99)	0.06
Total cardiovascular disease (n)	996	812	724	664	671
Age-, sex-, energy-adjusted HR (95% CI)	1.0	0.88 (0.80–0.97)	0.86 (0.78–0.94)	0.84 (0.76–0.93)	0.90 (0.82–1.00)	0.03
Multivariable HR (95% CI)	1.0	0.94 (0.85–1.05)	0.91 (0.81–1.03)	0.86 (0.75–0.98)	0.82 (0.69–0.96)	0.01

Multivariable model includes age; sex; energy intake; cohort; cigarette smoking status; alcohol intake; body mass index; sports at leisure time; walking and standing time; perceived mental stress; energy-adjusted dietary intakes of carbohydrate, protein, cholesterol, vegetables, fruit, and calcium.

^aSFA intake was energy-adjusted by nutrient residual model.

We consider hypertension and diabetes to be on the causal pathway between SFA and cardiovascular outcomes. When we included treatment for hypertension and diabetes in the model, the results were similar for total stroke [multivariable HR for highest vs. lowest SFA quintiles = 0.82 (0.69–0.98), P for trend = 0.01], intraparenchymal haemorrhage [HR = 0.63 (0.45–0.89), P for trend = 0.009], lacunar infarction [HR = 0.81 (0.57–1.14), P for trend = 0.047], and myocardial infarction [HR = 1.49 (0.99–2.22), P for trend = 0.02], but attenuated for ischaemic stroke [HR = 0.91 (0.72–1.14), P for trend = 0.25].

Discussion

We observed an inverse relationship between dietary intake of SFA and stroke, consistent with some previous studies,^6,15–18 but not all.^19–22 The major strengths of this study were the large sample size and the availability of stroke subtypes, with 98% of cases diagnosed by CT/MRI imaging. We revealed that the inverse association of SFA intake with stroke incidence was mainly due to its association with the incidence of deep intraparenchymal haemorrhage and lacunar infarction. We also found that the association between SFA intake and myocardial infarction was positive, the first epidemiological observation in Asia, but the overall association between SFA intake and total cardiovascular disease was inverse and was driven by its inverse association with stroke.

According to the national statistics, Japanese consumed SFA far less than Americans (mean intakes were 16–17 g among Japanese in 1988–93 vs. 33 and 22 g for American men and women, respectively, in 1989–91).^23,24 In contrast, they have much higher incidence of intraparenchymal haemorrhage and lacunar infarction.²⁵ This ecological association suggested that their high incidence of both intraparenchymal haemorrhage and lacunar infarction could be attributed to low SFA intake since haemorrhage and ischaemia in the perforator area of the brain is reported to share a common pathological condition (i.e. cerebral small vessel pathology called arteriolosclerosis),²⁶ and the presence of arteriolosclerosis has been associated with very low blood total cholesterol levels²⁷ attributable in part to low SFA consumption. The present finding would be consistent with previous observations that very low blood total or LDL cholesterol levels,^28–31 as well as low SFA intake,^6,16 were associated with increased risk of intraparenchymal haemorrhage.

We found a positive association between SFA intake and myocardial infarction incidence, primarily for men, but not for women, although there was no sex interaction. This is compatible with the result of a previous Japanese study showing a similar sex difference in the association between LDL cholesterol and mortality from myocardial infarction.³² Major reasons for no association among women in the present study would be a small number of incident myocardial infarction, but could also relate to the attenuated LDL cholesterol levels by sex hormones and/or the low incident rate of myocardial infarction, which might obscure the adverse coronary effect of high SFA intake. Much lower smoking rate in women may also be related.³³ The null associations in women were in line with the results of other studies such as the Nurses' Health Study.³⁴

The distribution of SFA intake in the present study population was significantly lower than that for USA or European populations, and the cardiovascular disease profile is also different. Therefore, we plotted SFA intake and crude incidence of or mortality from haemorrhagic stroke, ischaemic stroke, and coronary heart disease from the present and other published literature,^{6,17,18,21,35,36} where information on absolute amount of daily SFA intake was available (Figure 1). Taken together with the other studies, there seems to be a threshold around 20 g/day of SFA intake for inverse relation of SFA intake with stroke, especially with haemorrhagic stroke. There could be an optimal level of SFA intake, which should be confirmed by meta-analyses.

Figure 1

Comparison of the association between saturated fatty acid intake and crude incidence/mortality rates (per 100 000 person-years) of haemorrhagic stroke (A), ischaemic stroke (B), and coronary heart disease/myocardial infarction (C). NHS: Nurses' Health Study,¹⁶ ages 34–59, USA; HPFS: Health Professional Follow-up Study,^21,35 ages 40–75, USA; CIRCS: Circulatory Risk in Communities Study,⁶ ages 40–69, Japan; JPHC: Japan Public Health-based Cohort Study (this study), ages 45–74, Japan; JACC: Japan Collaborative Cohort Study,¹⁸ ages 40–79, Japan; LSS: Life Span Study,¹⁷ ages 35–89, Japan; ATBC: Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study,³⁶ ages 50–69, Finland. The JACC and LSS were mortality studies, which made large difference in crude event rates with JPHC Study. The other studies as well as JPHC Study were incidence studies. The ATBC was an intervention study for male smokers. Note that characteristics of study population (e.g. age range) varied by study.

Haemorrhage in the perforator area and lacunar infarction in Western populations are not as common as those in Asians.³⁷ The present results may be extrapolated to Asian populations that share dietary habits and cardiovascular disease profiles with Japan, but not necessarily to populations with Western lifestyles. Although there is no clinical trial evidence, we speculate that people consuming below optimal level, if any, of SFA might benefit by modestly increasing their SFA intake. However, even in Japanese, this could increase the risk of myocardial infarction especially in men. Therefore, recommendation to increase SFA intake cannot made in current Japan, since both SFA intake and coronary heart disease incidence rate are increasing among urban Japanese men.^38,39

We did not include hypertension and diabetes, since we consider them mediators of the associations, as SFA was reported to affect insulin resistance and blood pressure.^40–42 In the present study, the adjustment for treatment of hypertension and diabetes weaken the association between SFA intake and risk of ischaemic stroke, suggesting that the association was mediated largely by hypertension and diabetes.

As far as we know, this is the first study to examine the association between SFA intake and stroke subtypes. In addition, this is the first study to examine the association between SFA intake and risk of myocardial infarction among Asians. Other strengths of this study include the prospective cohort design, systematic registration of stroke and coronary heart disease incidence, high availability of CT/MRI imaging for stroke diagnosis, reasonable number of events for statistical power, and multiple adjustments for relevant confounders. Limitations of this study, however, warrant discussion. First, the use of a FFQ inevitably leads to SFA misclassification, although the FFQ was validated by dietary records. Random misclassification of SFA would typically attenuate observed associations. Secondly, persons with low SFA intake in the present study were older (women), leaner (men), more hypertensive, less diabetic (men), less active, and more likely to be current smokers and drinkers (men) and to have fruit (women). Although we adjusted for a number of covariates, there would still be a potential residual confounding by other dietary and/or socioeconomic factors. Thirdly, the present study was not sufficient to prove the causality. The effects of increasing or decreasing SFA intake on the incidence of different types of cardiovascular disease should be evaluated ideally by randomized controlled trial.

In conclusion, we found an inverse association between dietary intake of SFA and risk of stroke, primarily for deep intraparenchymal haemorrhage and lacunar infarction, and a positive association between SFA intake and risk of myocardial infarction among Japanese.

Supplementary material

Supplementary material is available at European Heart Journal online.

Funding

This study was supported by Grants-in-Aid for Cancer Research and for the Third Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare of Japan, and Grants-in-Aid for Scientific Research on Priority Areas (17015049) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Conflict of interest: none declared.

Acknowledgements

The authors thank all staff members in each study area and in the central and cardiovascular offices for their cooperation and technical assistance. We also thank Dr Aaron R. Folsom, University of Minnesota, for valuable comments on this manuscript.

Appendix

JPHC Study Group Members

Members of the JPHC Study Group: S.T. (principal investigator), M.I., T. Sobue, and T. Hanaoka, Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo; J. Ogata, S. Baba, T. Mannami, A. Okayama, and Y. K. National Cerebral and Cardiovascular Center, Suita; K. Miyakawa, F. Saito, A. Koizumi, Y. Sano, I. Hashimoto, T. Ikuta, and Y. Tanaba, Iwate Prefectural Ninohe Public Health Center, Ninohe; Y. Miyajima, N. Suzuki, S. Nagasawa, Y. Furusugi, and N. Nagai, Akita Prefectural Yokote Public Health Center, Yokote; H. Sanada, Y. Hatayama, F. Kobayashi, H. Uchino, Y. Shirai, T. Kondo, R. Sasaki, Y. Watanabe, Y. Miyagawa, Y. Kobayashi, and M. Machida, Nagano Prefectural Saku Public Health Center, Saku; Y. Kishimoto, E. Takara, T. Fukuyama, M. Kinjo, M. Irei, and H. Sakiyama, Okinawa Prefectural Chubu Public Health Center, Okinawa; K. Imoto, H. Yazawa, T. Seo, A. Seiko, F. Ito, F. Shoji, and R. Saito, Katsushika Public Health Center, Tokyo; A. Murata, K. Minato, K. Motegi, and T. Fujieda, Ibaraki Prefectural Mito Public Health Center, Mito; T. Abe, M. Katagiri, M. Suzuki, and K. Matsui, Niigata Prefectural Kashiwazaki and Nagaoka Public Health Center, Kashiwazaki and Nagaoka; M. Doi, A. Terao, Y. Ishikawa, and T. Tagami, Kochi Prefectural Chuo-higashi Public Health Center, Tosayamada; H. Doi, M. Urata, N. Okamoto, F. Ide, and H. Sueta, Nagasaki Prefectural Kamigoto Public Health Center, Arikawa; H. Sakiyama, N. Onga, H. Takaesu, and M. Uehara, Okinawa Prefectural Miyako Public Health Center, Hirara; F. Horii, I. Asano, H. Yamaguchi, K. Aoki, S. Maruyama, M. Ichii, and M. Takano, Osaka Prefectural Suita Public Health Center, Suita; S. Matsushima and S. Natsukawa, Saku General Hospital, Saku; M. Akabane, Tokyo University of Agriculture, Tokyo; the late M. Konishi, K. Okada, and I.S., Ehime University, Tōon; H. I., Osaka University, Suita; Y. Honda, K. Y. S. Sakurai, and N. Tsuchiya, University of Tsukuba, Tsukuba; H. Sugimura, Hamamatsu University, Hamamatsu; Y. Tsubono, Tohoku University, Sendai; the late M. Kabuto, National Institute for Environmental Studies, Tsukuba; S. Tominaga, Aichi Cancer Center Research Institute, Nagoya; M. Iida W. Ajiki, and A. Ioka, Osaka Medical Center for Cancer and Cardiovascular Disease, Osaka; S. Sato, Chiba Prefectural Institute of Public Health, Chiba; N. Yasuda, Kochi University, Nankoku; K. Nakamura, Niigata University, Niigata; S. Kono, Kyushu University, Fukuoka; K. Suzuki, Research Institute for Brain and Blood Vessels Akita, Akita; the late Y. Takashima and M. Yoshida, Kyorin University, Mitaka; E. Maruyama, Kobe University, Kobe; the late M. Yamaguchi, Y. Matsumura, S. Sasaki, and S. Watanabe, National Institute of Health and Nutrition, Tokyo; T. Kadowaki, University of Tokyo, Tokyo; M. Noda and T. Mizoue, International Medical Center of Japan, Tokyo; Y. Kawaguchi, Tokyo Medical and Dental University, Tokyo; H. Shimizu, Sakihae Institute, Gifu.