Dietary Protein Intake and Stroke Risk in a General Japanese Population

Introduction

Stroke is one of the major causes of morbidity and disability in Japan and other developed countries.^1–3 Because diet is one of the modifiable factors affecting stroke, there is growing interest in discovering the influence of dietary factors on stroke risk.⁴ Several epidemiological studies have demonstrated that higher protein intake reduced the risk of stroke.^5–8 A recent meta-analysis of prospective cohort studies has also shown significant negative associations between protein intake and risks of stroke and intracerebral hemorrhage.⁹ Certainly, this association seems very plausible, but there have been inconsistent findings across studies.^10–13 It has been well established that food culture is quite different between Western countries and Asian countries, and the incidence of stroke is higher in Asian populations than in Western populations. Therefore, it would be clinically worthwhile to accumulate evidence of the influence of protein intake on the stroke risk for Asian populations, but there have been limited studies addressing this issue in Asian populations.^7,13 Further, there have not been enough studies investigating the association between the source of protein (eg, animal and vegetable protein) and the risk of stroke and its subtypes,⁵ which are important to clarify underlying mechanisms for the influence of protein intake on the development of stroke.

To shed the light on these issues, we performed a prospective cohort study followed up for 19 years in a general Japanese population to evaluate the association between dietary protein intake and the development of stroke and its subtypes. This study will provide additional evidence for preventive stroke medicines in Asian populations, with further consideration given to the source of protein.

Methods

Study Population

In 1988, a screening survey for this study was performed in the town of Hisayama, a suburb of the Fukuoka metropolitan area in the southern part of Japan.^14,15 A comprehensive assessment, including a dietary survey, was performed in 2587 residents aged 40 to 79 years (participation rate, 80.2%). After excluding 85 subjects with a past history of stroke or coronary heart disease, 50 who did not complete a dietary questionnaire at baseline, 48 who had implausible total energy intake (subjects in the highest or lowest 1% for total energy intake), and 4 who had died before the start of follow-up, the remaining 2400 subjects (1017 men and 1383 women) were enrolled in this study.

Follow-Up Survey

The subjects were followed up prospectively for 19 years from December 1988 through November 2007. During this time, health examinations were performed every 1 to 2 years. For subjects who did not have examinations or who had moved out of town, postal mail or phone was used to collect the health information. We also established a daily monitoring system for identifying stroke events among the study team and local physicians or members of the Health and Welfare Office of the town. Subjects with suspected stroke events were evaluated for their detailed clinical information by study team physicians. The clinical diagnosis of stroke was based on the patients’ history, physical and neurological examinations, and ancillary laboratory examinations. In addition, when a subject died, an autopsy was performed at the Departments of Pathology of Kyushu University. During the follow-up period, 653 subjects died, of whom 455 (69.7%) underwent autopsy examination.

Definition of Stroke Events

In principle, stroke was defined as a sudden onset of nonconvulsive and focal neurological deficit persisting for ≥24 hours. The diagnosis and classification of stroke were determined on the basis of clinical information, including brain computed tomography and magnetic resonance imaging, cerebral angiography, echocardiography, and carotid duplex imaging, and autopsy findings. During the follow-up period, we identified 254 cases of first-ever stroke: 172 cases of ischemic stroke, 58 cases of intracerebral hemorrhage, and 24 cases of subarachnoid hemorrhage.

Nutritional Survey

In the baseline screening examination, a dietary survey was conducted using a 70-item semiquantitative food frequency questionnaire concerning food intake.¹⁵ The validity of this questionnaire has been reported previously.¹⁶ The questionnaire was completed by each subject in advance and was checked by trained dieticians or nutritionists in the screening examination. The average food intake per day was calculated from the weekly frequency of various foods and the amount (quantity) of each food portion. Nutritional intake was calculated using the 4th revision of the Standard Tables of Food Composition in Japan.¹⁷ Each amount of nutrient intake was adjusted for total energy intake, using the residual method.¹⁸

Risk Factor Measurements

At baseline, each subject completed a self-administered questionnaire covering medical history, antihypertensive and antidiabetic treatments, smoking habits, alcohol intake, and physical activity. Smoking habits and alcohol intake were categorized as either current use or no current use. Physical activity during leisure time was defined using 4 categories: always sedentary, walking, exercise or sports 1 to 2 days a week, and exercise or sports ≥3 days a week; the last category was defined as the regular exercise group. Body height and weight were measured in light clothing without shoes, and body mass index (kg/m²) was calculated. Blood pressure was measured 3× using a standard mercury sphygmomanometer in the sitting position after rest for at least 5 minutes. The mean of 3 measurements was used for the analysis. Hypertension was defined as blood pressure ≥140/90 mm Hg or current use of antihypertensive drugs. Diabetes mellitus was defined by fasting plasma glucose ≥7.0 mmol/L, 2-hour postload glucose concentrations or postprandial glucose concentrations ≥11.1 mmol/L, or current use of insulin or oral medication for diabetes mellitus. Serum total cholesterol and high-density lipoprotein cholesterol concentrations were determined enzymatically. Freshly voided urine samples were collected at the screening, and proteinuria was defined as a value of ≥1+ using a reagent strip. ECG abnormalities were defined as high amplitude R waves (Minnesota Code, 3-1), ST depression (4-1, 2, 3), and atrial fibrillation (8-3).

Statistical Analysis

The SAS software package version 9.3 (SAS Institute, Cary, NC) was used to perform all statistical analyses. Subjects were divided into quartiles of protein intake. The trends in the mean values or the frequencies of risk factors across quartiles were tested by using the linear regression analysis or logistic regression analysis, respectively. The age- and sex-adjusted and multivariable-adjusted hazard ratios (HRs) with their 95% confidence intervals (CIs) of protein intake levels for the development of stroke were estimated using a Cox proportional hazards model. The assumption of the proportional hazards was checked graphically using the log cumulative hazard plots for outcomes according to the intake levels of protein. The heterogeneity in the association between subgroups was tested by adding a multiplicative interaction term in the relevant Cox model. Two-sided values of P<0.05 were considered statistically significant in all analyses.

Ethical Considerations

This study was conducted with the approval of the Kyushu University Institutional Review Board for Clinical Research. Written informed consent was obtained from the participants.

Results

Table 1 shows the age- and sex-adjusted potential risk factors of the study subjects according to the quartile of protein intake at baseline. Subjects with higher intake of protein were older and more likely to be female. The mean values of serum total cholesterol, high-density lipoprotein cholesterol, and body mass index and the frequency of regular exercise increased with higher protein intake levels, whereas subjects with higher intake of protein had a lower diastolic blood pressure and were less likely to have ECG abnormalities, smoking habits, and alcohol intake.

The age- and sex-adjusted incidence of total stroke decreased significantly with higher amount of total protein intake levels (P for trend = 0.03; Table 2). In regard to subtypes of stroke, the age- and sex-adjusted incidence of intracerebral hemorrhage decreased significantly with the elevating quartile (P for trend = 0.008), but no such association was observed for ischemic stroke or subarachnoid hemorrhage (both P for trend >0.23). Similar associations were observed after adjusting for potential confounding factors. As a consequence, the multivariable-adjusted HR for incident stroke decreased significantly with higher quartiles of total protein intake (P for trend = 0.04; Table 2). In the multivariable-adjusted analysis using protein intake as a continuous variable, every 10 g/d increment in total protein intake was associated with a 15% (95% confidence interval [CI], 3%–26%) lower risk of stroke. With regard to the subtypes of stroke, higher intake of total protein was significantly associated with lower risk of intracerebral hemorrhage after adjusting for the aforementioned confounding factors (P for trend = 0.01). However, there was no evidence of an association between total protein intake and risk of either ischemic stroke or subarachnoid hemorrhage (both P for trend >0.2).

We further examined the associations of vegetable and animal protein intakes with risks of stroke and its subtypes (Figure). The age- and sex-adjusted incidence of total stroke and ischemic stroke significantly decreased with the elevating quartile of vegetable protein intake (both P for trend <0.05), whereas there was no evidence of a linear association between vegetable protein intake and the risk of intracerebral hemorrhage and subarachnoid hemorrhage (both P for trend >0.23). On the other hand, the age- and sex-adjusted incidence of intracerebral hemorrhage decreased significantly with higher amount of animal protein intake (P for trend = 0.01), but such an association was not observed for total stroke, ischemic stroke, or subarachnoid hemorrhage (all P for trend >0.46). These associations were not altered substantially after adjusting for aforementioned confounders. The risks of total stroke and ischemic stroke significantly decreased in subjects with the highest quartile of vegetable protein intake compared with those with the lowest one (Table 3). In contrast, subjects with the highest quartile of animal protein intake had a significantly lower risk of intracerebral hemorrhage (Table 4). We also performed the sensitivity analysis after dividing the lowest quartile category into 2 categories by the median value to estimate the influence of very low vegetable (<27.3 g/d) and animal protein intakes (<13.0 g/d) on the risk of stroke and its subtypes. As a consequence, the findings were not altered substantially (data not shown).

• Download figure
• Download PowerPoint

Similar associations of vegetable protein intake with the risk of stroke and ischemic stroke were observed in either sex, although the associations between the animal protein intake and intracerebral hemorrhage were not statistically significant in either sex, probably because of the small sample size (Table I in the online-only Data Supplement). There is no evidence of significant heterogeneity of association between the sexes (all P for heterogeneity >0.2). The sensitivity analysis after excluding subjects with alcohol intake also showed a significant association of the vegetable protein intake with the risk of stroke and ischemic stroke (Table II in the online-only Data Supplement).

In subgroup analyses of age (<65 years or ≥65 years), sex, and status of hypertension, and diabetes mellitus (presence or absence), we observed no heterogeneities in the association of vegetable protein intake with the risk of ischemic stroke between subgroups of age, sex, and hypertension status (all P for heterogeneity >0.16), but the magnitude of the association was significantly stronger in subjects with diabetes mellitus than in those without diabetes mellitus (nondiabetes mellitus: HR, 0.87; 95% CI, 0.52–1.45 for Q2, HR, 0.98; 95% CI, 0.59–1.60 for Q3, and HR, 0.73; 95% CI, 0.43–1.22 for Q4; diabetes mellitus: HR, 0.49; 95% CI, 0.21–1.19 for Q2, HR, 0.48; 95% CI, 0.20–1.15 for Q3, and HR, 0.32; 95% CI, 0.12–0.80 for Q4; P for interaction = 0.02). There were no heterogeneities in the associations of animal protein intake and the risk of intracerebral hemorrhage between all subgroups (all P for heterogeneity >0.23).

Finally, we also evaluated the correlation coefficients between types of protein intake and food groups, as shown in Table 5. This table indicates which food mainly contributed to each type of protein intake, with the Pearson correlation coefficients of ≥0.20 or ≤−0.20 as shown in Table 5. Total protein intake was negatively correlated with intakes of rice and alcohol beverages and positively correlated with intakes of soybeans and soybean products, vegetable, algae, fish, eggs, and milk and dairy products. Vegetable protein intake was positively correlated with intake of soybeans and soybean products, vegetable, and algae and inversely correlated with intake of alcohol beverages, whereas animal protein intake showed positive correlations with intakes of fish, meat, eggs, and milk and dairy products. Intakes of both types of protein were negatively correlated with the intake of rice.

Discussion

This study showed that higher protein intake was associated with a lower risk of total stroke in a general Japanese population. In addition, subjects with higher vegetable protein intake had a lower risk of total stroke and ischemic stroke, whereas those with higher animal protein intake had a lower risk of intracerebral hemorrhage. To the best of our knowledge, this is the first study to clearly show a protective influence of vegetable protein intake on the risks of total stroke and ischemic stroke. Importantly, vegetable protein intake was positively correlated with intakes of soybeans and soybean products, vegetable, and algae, whereas higher intake of animal protein was most strongly correlated with higher intakes of fish, followed in order by milk and dairy products, meat, and eggs, in this study, suggesting that the consumption of these food groups may have some favorable influence on stroke incidence. While considering the fact that the amount of red meat intake is lower, and the consumption of white meat and fish is higher in Japan than in Western countries,¹⁹ these food groups are similar to the Dietary Approaches to Stop Hypertension diet, which has been recommended to be a healthy dietary pattern on the basis of the scientific evidence of reducing risk of hypertension and stroke.²⁰ These findings should enhance our current knowledge of the association between protein intake and stroke risk.

Several prospective studies have assessed the association between dietary protein intake and the risk of stroke.^5–12 The Swedish Mammography Cohort⁵ and the Women’s Health Initiative⁶ demonstrated an inverse association between total protein intake and stroke risk. In the Adult Health Study, higher intake of total protein was significantly associated with a lower risk of mortality from stroke in a Japanese population.⁷ These results are consistent with ours. Although some studies, such as the Health Professionals Follow-up Study of US men,¹⁰ the ARIC Study (Atherosclerosis Risk in Communities) of US middle-aged adults,¹¹ and a cohort study from southern CA,¹² failed to reveal a significant association between protein intake and stroke, a recent meta-analysis of 7 prospective studies found that the highest dietary protein intake level was associated with a 20% lower risk of stroke.⁹ Therefore, the optimal amount of dietary protein intake may be an important protective factor against the development of stroke.

This study showed that higher intake of vegetable protein clearly reduced the risk of total stroke and ischemic stroke. However, in other epidemiological studies, such as the Swedish Mammography Cohort,⁵ the Women’s Health Initiative,⁶ and the aforementioned meta-analysis,⁹ the vegetable protein intake was not associated with the risk of stroke. This discrepancy in the findings may have been caused by the difference in the amount of vegetable protein intake among the populations (ie, subjects in this study consumed more vegetable protein and less animal protein than those in the other studies). Vegetable protein contains more nonessential amino acids than animal protein.²¹ Arginine, which is one of the nonessential amino acids, has been reported to have a blood pressure–lowering effect.²² A recent meta-analysis of cohort studies and randomized controlled trials has demonstrated that higher intake of dietary protein, especially vegetable protein, significantly decreased blood pressure levels.²³ Arginine also stimulates insulin and glucagon secretion in diabetic subjects.²⁴ Intriguingly, this study found that the magnitude of the negative association between vegetable intake and the risk of ischemic stroke was stronger in subjects with diabetes mellitus than those without. Therefore, arginine intake from vegetable protein may have a protective influence against ischemic stroke through the improvement in blood pressure and glucose tolerance. In our study, soybean and soybean products were a major food source of vegetable protein. Soybean protein has been thought to have a beneficial effect on cardiovascular disease through the lowering of low-density lipoprotein cholesterol.²⁵ In addition, soybeans are abundant dietary sources of isoflavone and magnesium, likely reducing the risk of ischemic stroke.²⁶

In our subjects, the risk of intracerebral hemorrhage was significantly decreased with increasing dietary intake of animal protein. The aforementioned meta-analysis also revealed that higher intake of animal protein was significantly associated with a 43% lower risk of intracerebral hemorrhage.⁹ Similarly, the Nurses’ Health Study showed that higher animal protein intake significantly reduced the risk of intracerebral hemorrhage.⁸ An animal study also demonstrated that stroke-prone spontaneously hypertensive rats fed higher dietary protein intakes had dose dependently lower incidence of hemorrhagic stroke.²⁷ Although the exact mechanisms underlying the significant association between higher protein intake, especially animal protein intake, and decreased risk of intracerebral hemorrhage are unclear, the protective influence of animal protein could be explained by the fish intake,⁴ which was most strongly correlated with animal protein intake in our subjects. In line with our study, the Hiroshima/Nagasaki Life Span Study in Japan reported that higher intake of fish products significantly reduced the risk of total stroke and intracerebral hemorrhage.²⁸ A systematic review also demonstrated that higher fish consumption was significantly associated with lower incidence of stroke.²⁹ In an animal study, the proportion of elastin, which is a major component of the arterial wall, was higher in the stroke-prone spontaneously hypertensive rats fed a fish diet than those fed a stock chow diet.³⁰ These findings suggest that consuming a fish diet may maintain the elasticity of the vessels, which is likely to inhibit the formation and rupture of microaneurysms.^28,30 In this study, the amount of intake of milk and dairy products had the second strongest correlation with the amount of animal protein intake. An animal experiment with stroke-prone spontaneously hypertensive rats showed that a diet rich in milk protein significantly delayed the onset of hemorrhagic stroke compared with a diet rich in soybean or egg white protein.³¹ Another study in stroke-prone spontaneously hypertensive rats reported that a milk-rich protein diet was associated with significantly lower blood pressure levels than a regular stock diet, with the milk-rich protein diet possibly contributing to the decreasing incidence of hemorrhagic stroke.³² Further investigations will be required to elucidate the mechanism underlying the association between animal protein intake and the development of intracerebral hemorrhage.

The advantages of this study include its longitudinal population-based study design, long duration of follow-up, perfect follow-up of study subjects, and accurate diagnoses of stroke and its subtypes. However, some potential limitations of this study should be noted. First, information on the intake of dietary nutrients derived from a semiquantitative food frequency questionnaire may not be fully valid. Second, the dietary assessment was performed only once at baseline. These limitations were likely to lead to some extent of misclassification of food intake. Such misclassification would weaken the association found in this study, biasing the results toward the null hypothesis. Third, we may not have been able to completely exclude the influence of the residual confounders (eg, socioeconomic status, health awareness, intakes of isoflavone and potassium, etc) on the association between the protein intake and the stroke risk. Finally, this study had a relatively small sample size and was conducted at only 1 site in Japan. Therefore, the generalizability of the findings is limited. Especially, the results from the analysis for subarachnoid hemorrhage, in which the event number was very small, should be carefully interpreted, although the sensitivity analysis using the Cox proportional hazards model with Firth penalized likelihood³³ did not altered the findings substantially (data not shown).

Summary/Conclusions

In conclusion, the findings of this study suggest that individuals with a higher intake of protein had a lower risk of stroke in the general Japanese population. In addition, when vegetable and animal protein intakes were analyzed separately, the intake of vegetable protein was inversely associated with the risk of total stroke and ischemic stroke, whereas the animal protein intake was inversely related to the risk of intracerebral hemorrhage. Subjects with higher vegetable protein intake ate soybeans and soybean products, vegetables, and algae, and subjects with higher intake of animal protein were more likely to eat fish, meat, eggs, and milk and dairy products, than those with lower intake of either type of protein. Therefore, an increased intake of protein, with an attempt to balance the animal and vegetable protein intakes, can be recommended as an effective strategy for preventing stroke.

Acknowledgments

We thank the staff of the Division of Health and Welfare of Hisayama for their cooperation in this study. We also would like to gratefully and sincerely thank Koji Yonemoto in Advanced Medical Research Center, Faculty of Medicine, University of the Ryukyus, who gave us statistical advice.

Footnotes