Dietary Calcium Intake and Risks of Stroke, Its Subtypes, and Coronary Heart Disease in Japanese
The JPHC Study Cohort I
Background and Purpose— Although it has been hypothesized that a high intake of dietary calcium may reduce the risk of cardiovascular disease (CVD), no prospective studies have been conducted to examine the specific association between calcium intake and incidence of stroke or coronary heart disease among Japanese with a low average calcium intake.
Methods— To investigate the association between calcium intake and risk of CVD, a total of 41 526 Japanese men and women age 40 to 59 years without a history of CVD or cancer and who had completed a food consumption frequency questionnaire were enrolled in this study. The subjects were followed up from 1990 to 1992 to 2003, and after 533 692 person-years of follow-up, 1321 incident cases of stroke (664 ischemic, 425 intraparenchymal hemorrhage, and 217 subarachnoid hemorrhage) and 322 of coronary heart disease were documented.
Results— Total calcium intake showed an inverse association with the risk of total stroke; the multivariable hazard ratio and 95% CIs for the highest versus the lowest quintile were 0.70 (95% CI, 0.56 to 0.88; P for trend=0.02). Dairy calcium intake was inversely associated with risks of total and ischemic stroke with respective multivariable hazard ratios (95% CIs) of 0.69 (0.56 to 0.85; P for trend=0.007) and 0.69 (0.52 to 0.93; P for trend=0.05). Dietary calcium intake was not significantly associated with risk of coronary heart disease.
Conclusions— Dietary calcium intake, especially calcium from dairy products, was found to be associated with a reduced incidence of stroke among middle-aged Japanese.
Findings concerning the association between calcium or dairy product intake and cardiovascular disease (CVD) have been inconsistent. Prospective studies of Japanese-American men1 and American women2 showed an inverse association between calcium intake or milk consumption and stroke incidence. However, a prospective study of American men did not show any significant association between milk or dairy product consumption and risk of stroke.3 As for the association of calcium or dairy product intake with coronary heart disease (CHD), none of the prospective studies have shown any significant association.
For Japanese, whose calcium intake is lower than that in Western populations,4,5 an inverse association has been found between calcium intake in the form of dairy products and mortality from total stroke, hemorrhagic stroke, and ischemic stroke among men and women age 40 to 79 years.6 However, it remains uncertain whether calcium intake is also inversely associated with the incidence of stroke.
Inverse associations between calcium intake and the risk of stroke found in previous studies may be primarily due to a hypotensive effect of calcium.7–10 Because high blood pressure is also a risk factor for CHD, we examined the association between dietary calcium intake and risk of CHD as well.
The mean calcium intake in Japan is far lower than that in Western countries. According to national nutrition surveys, the mean dairy calcium intake by Japanese was 550 mg for men and 528 mg for women in 2004,4 whereas the NHANES of 1999 to 2000 showed that the corresponding figures for Americans were 966 and 765 mg, respectively.5 This difference may be due to differences in dairy product consumption. Because stroke is the most common CVD in Japan,11 the aforementioned findings indicate that it is of vital importance for public health to examine the association between calcium intake and the risk of stroke among Japanese.
Our a priori hypothesis was that calcium intake is inversely associated with the risk of stroke and its subtypes among Japanese whose average calcium intake is known to be low. We investigated this hypothesis in a large, prospective study of middle-aged Japanese men and women.
Subjects and Methods
The Japan Public Health Center (JPHC)–based study (cohort I) was a population-based study that included 27 063 men and 27 435 women who were age 40 to 59 years as of January 1, 1990 (ie, born between 1930 and 1949). All participants were registered in 14 administrative districts supervised by 4 PHC areas.12
The population comprised 6022 men and 6269 women from Ninohe City and Karumai Town in the Ninohe PHC area of Iwate Prefecture in northern Japan; 7559 men and 8223 women from Yokote City and Omonogawa Town in the Yokote PHC area of Akita Prefecture in northern Japan; 6173 men and 6046 women from 8 districts of Minami-Saku County in the Saku PHC area of Nagano Prefecture in central Japan; and 7309 men and 6897 women from Gushikawa City and Onna Village in the Ishikawa PHC area of Okinawa Prefecture in southern Japan.12 This study was approved by the human ethics review committees of the National Cancer Center and the University of Tsukuba.
Baseline and Follow-Up Surveys
A self-administrated questionnaire was distributed to all registered, noninstitutionalized community residents in 1990 and 1995. The questionnaire included items regarding demographic characteristics, medical history, smoking and drinking habits, and diet. A total of 20 665 men (76%) and 22 484 women (82%) returned their questionnaires between January 1990 and May 1992, primarily between February 1990 and October 1990.12 We excluded men and women with histories of myocardial infarction (MI), angina pectoris, stroke, and cancer at baseline. Finally, 19 947 men and 21 579 women were enrolled in the study. In 1995, a 5-year follow-up survey was conducted, to which 15 753 men and 18 366 women replied.
The baseline food consumption frequency questionnaire dealt with 44 foods, and the follow-up questionnaire, with 147 foods. In both the baseline and follow-up questionnaires, participants were asked how often on average they had consumed each of the items during the previous month. A unit or portion size for each item was specified in the follow-up questionnaire but not in the baseline questionnaire. Possible responses for each food item were “rarely,” “1 to 2 days/wk,” “3 to 4 days/wk,” and “almost every day” in the baseline questionnaire and “<1 day/mo,” “1 to 3 days/mo,” “1 to 2 days/wk,” “3 to 4 days/wk,” “5 to 6 days/wk,” “once per day,” “2 to 3 times per day,” “4 to 6 times per day,” and “≥7 times per day” in the follow-up questionnaire. The average daily intake of nutrients was calculated by multiplying the frequency of consumption of each item by its nutrient content per serving and totaling the nutrient intake for all items, and a residual model was used for adjusting total energy intake.
The reproducibility and validity of the questionnaires have been reported elsewhere.13,14 As for validity, the Spearman rank correlation coefficients of calcium intake for the baseline questionnaire and four 1-week dietary records obtained after an interval of 3 to 4 months were 0.56 for men and 0.37 for women,13 and those for the 1-year follow-up questionnaire and the four 1-week dietary records were 0.51 for men and 0.45 for women.14 The Spearman rank correlation coefficients of total calcium intake for the baseline and 5-year follow-up questionnaires were 0.44 for men and 0.43 for women.
To calculate dairy calcium intake, we specified 2 kinds of dairy products, ie, cheese and dairy products except cheese, for the baseline questionnaire, and 4 kinds, ie, whole milk, lowfat milk, cheese, and yogurt, for the 5-year follow-up questionnaire. The Spearman rank correlation coefficients of dairy calcium intake for the baseline and 5-year follow-up questionnaires were 0.41 for both men and women. Main sources of nondairy calcium in Japanese were vegetables, soybean products, and fish.15
Confirmation of CVD
Our study involved a total of 30 major hospitals in the 4 PHC areas (10 for the Ninohe PHC area, 4 for the Yokote PHC area, 3 for the Saku PHC, and 13 for the Ishikawa PHC area). Medical records at each hospital were reviewed by registered hospital physicians or PHC physicians who were blinded to the lifestyle data.
Incidences of stroke were confirmed according to the criteria of the National Survey of Stroke, which requires a constellation of neurologic deficits of sudden rapid onset lasting at least 24 hours (or until death). For each subtype of stroke, ie, subarachnoid hemorrhage, intraparenchymal hemorrhage, and ischemic stroke, a definite diagnosis was established on the basis of examination of data from computed tomography scans, magnetic resonance images, or autopsy.
To complete the surveillance for fatal stroke, we conducted a systematic search for death certifications. For all fatal strokes (International Classification of Diseases 9th Revision 430 to 438; International Classification of Diseases 10th Revision I60 to I69) listed on the death certificate but that had not been registered, medical workers, PHC physicians, or research physician/epidemiologists reviewed medical records at the registered hospitals. When no medical records were available, we regarded these fatal strokes as probable stroke. As for nonfatal strokes, when subjects reported a history of nonfatal stroke on the 10-year follow-up questionnaire and had not been registered as stroke cases, we asked by letter or telephone about the onset of stroke and for permission to review the medical records.
CHD consisted of MI and sudden cardiac death. MI was confirmed according to the criteria of the Monitoring Trends and Determinants of Cardiovascular Disease project.16 These criteria specify evidence of MI from electrocardiograms, cardiac enzyme values, and/or autopsy. When such examinations were not performed and there was typical chest pain, a diagnosis of probable MI was made. To complete the surveillance for MI, we conducted a systematic search as we did for stroke.
Changes in residence status were identified through the residential registry in each area. Subjects who moved from their original residential areas (2% of the total participants) were treated as censored at that time.
Statistical analyses were based on incidence rates of stroke and of CHD during the 13-year follow-up from 1990 to the end of 2003. For each individual, person-years of follow-up were calculated from January 1, 1990 until 1 of 4 end points had been achieved: incidence of stroke and CHD, death, emigration, or December 31, 2003, whichever came first. The hazard ratios (HRs) of stroke and CHD incidence were calculated according to sex-specific quintiles of dietary calcium intake, after adjusting for age, sex, and other covariates by using conventional and time-dependent Cox proportional-hazards models. A test for linear trends across the quintiles of calcium intake was conducted by using the Wald test. We tested for the calcium intake×sex interaction by using an interaction term generated by multiplying the median of each quintile of calcium intake at baseline with sex and found that there were no significant interactions. Thus, we present the combined data for men and women.
We updated the data of dietary intake and covariates except for age, sex, history of hypertension, and PHC. Dietary intakes obtained from baseline questionnaire responses were therefore related to incidences from 1990 through 1995, and those from the 5-year follow-up questionnaire were related to incidences from 1995 through 2003. For those who had not responded to the 5-year follow-up questionnaire, we used the information from the baseline questionnaire. Covariates for statistical adjustment were baseline values for age (5-year groups), sex, body mass index (sex-specific quintiles), history of hypertension (none, nonmedicated, medicated), history of diabetes (none, nonmedicated, medicated), medication for hypercholesterolemia (yes or no), menopause (yes or no), smoking status (never, ex-smoker, and current smokers smoking 1 to 19, 20 to 29, or ≥30 cigarettes/d), ethanol intake (nondrinker, occasional drinker, and current drinkers consuming ≤150, 151 to 300, 301 to 450, or ≥451 g ethanol per week), sodium intake (sex-specific quintiles), potassium intake (sex-specific quintiles), n-3 fatty acid intake (sex-specific quintiles), and PHC (4 areas). Of these covariates, history of hypertension has been proposed as a mediator for the relation between calcium intake and risk of stroke and CHD7–10; therefore, we conducted the analyses with and without adjustment for this potential mediator. We used SAS version 8.02 software (SAS Institute Inc, Cary, NC) for all analyses. Probability values <0.05 were regarded as statistically significant.
During an average 12.9-year follow-up period of 41 526 middle-aged persons (19 947 men and 21 579 women), 1321 incident cases of total stroke were documented. They comprised 664 cases of ischemic stroke, 425 of intraparenchymal hemorrhage, 217 of subarachnoid hemorrhage, and 15 of undetermined type. We also documented 322 incident cases of CHD.
Table 1 shows age-adjusted cardiovascular risk factors based on sex-specific quintiles of total, dairy, and nondairy calcium intake obtained from the baseline questionnaire. Total, dairy, and nondairy calcium intakes were positively associated with age, history of diabetes, and n-3 fatty acid intake and inversely associated with body mass index and current smoking. Dairy calcium intake was also inversely associated with history of hypertension. Total, dairy, and nondairy calcium intakes were positively associated with sodium and potassium intakes.
Table 2 shows HRs for total stroke, subtypes of stroke, and CHD in relation to total calcium intake. There were inverse associations of total calcium intake with age- and sex-adjusted risks of total stroke, intraparenchymal hemorrhage, and ischemic stroke. After adjustment for cardiovascular risk factors and other cation (sodium and potassium) intakes, total calcium intake was inversely associated with risk of total stroke; the multivariable HR based on a comparison of the highest and the lowest quintile of total calcium intake was 0.70 (95% CI, 0.56 to 0.88), P for trend=0.02. The multivariable HR of intraparenchymal hemorrhage for the highest versus the lowest quintile of total calcium intake was 0.70 (95% CI, 0.46 to 1.06), P=0.09, but its linear trend did not reach statistical significance. The multivariable HR of subarachnoid hemorrhage for the highest versus the lowest quintile of total calcium intake did not reach statistical significance, whereas its linear trend was of borderline statistical significance, P for trend=0.09. The multivariable HR of ischemic stroke for the highest versus the lowest quintile of total calcium intake was 0.72 (95% CI, 0.52 to 1.00), P=0.05, but its linear trend did not reach statistical significance. Total calcium intake was not significantly associated with risk of CHD. These associations were similar after further adjustment for history of hypertension.
Table 3 shows HRs of total stroke, subtypes of stroke, and CHD according to quintiles of dairy calcium intake, after adjustment for cardiovascular risk factors and other cation intakes. Dairy calcium intake was inversely associated with risks of total and ischemic strokes. The multivariable HRs based on a comparison of the highest and lowest quintiles of dairy calcium intake were 0.69 (95% CI, 0.56 to 0.85), P for trend=0.007, for total stroke and 0.69 (95% CI, 0.52 to 0.93), P for trend=0.05, for ischemic stroke. The multivariable HR of intraparenchymal hemorrhage for the highest versus the lowest quintile of dairy calcium intake was 0.64 (95% CI, 0.43 to 0.96), but its linear trend did not reach statistical significance. These associations were similar after further adjustment for history of hypertension.
Table 4 shows HRs of total stroke, subtypes of stroke, and CHD according to quintiles of nondairy calcium intake. Nondairy calcium was not significantly associated with risks for any end point, although an inverse association with risk of CHD was of borderline statistical significance. The multivariable HR for the highest versus the lowest quintile of dairy calcium intake was 0.63 (95% CI, 0.39 to 1.03), P=0.07 and P for trend=0.11. These associations were similar after further adjustment for history of hypertension.
We found a significant inverse association between calcium intake and risk of stroke, but not with CHD, in the current large, prospective study of Japanese middle-aged men and women. Total calcium intake was inversely associated with risk of total stroke, whereas dairy calcium intake was inversely associated with risks of total stroke and ischemic stroke. There were inconsistent inverse associations of total and dairy calcium intakes with risk of hemorrhagic stroke. Nondairy calcium intake was not associated with risk of any end point except for a suggestive inverse association with CHD.
Several prospective studies have examined the association between calcium intake and incidence of or mortality from stroke or CHD.1,2,6,17 As for total calcium, a study of American women showed a 28% lower risk of ischemic stroke for the highest quintile of calcium intake (median intake 1145 mg/d), compared with the lowest quintile (median intake 395 mg/d),2 whereas another study of Dutch civil servants showed no association between total calcium intake and mortality from total stroke.17 As for dairy calcium intake, 3 studies showed an inverse association between dairy calcium intake and incidence of or mortality from ischemic stroke.1,2,4 A study of Japanese-Americans showed a 33% lower risk of thromboembolic stroke for the highest quartile of dairy calcium intake (range of intake 296 to 2373 mg/d), compared with the lowest quartile (range of intake 0 to 1 mg/d).1 A study of American women showed a 30% lower risk of ischemic stroke for the highest versus the lowest quintile of dairy calcium intake (median intake 844 mg/d vs 108 mg/d).2 A recent Japanese study showed a 47% lower mortality from total stroke, a 54% lower mortality from hemorrhagic stroke, and a 47% lower mortality from ischemic stroke for the highest versus the lowest quintile of dairy calcium intake (median intake 150 mg/d for men and 173 mg/d for women vs 0 mg/d for both men and women).6 These results suggest that calcium intake, especially calcium intake from dairy products, has a protective effect against ischemic stroke.
Although the precise mechanisms for the inverse associations between calcium intake and risk of stroke remain unclear, physiologic and epidemiologic studies have suggested some possible mechanisms.7–10,18–20 One of these is the hypotensive effect of calcium as reported by observational studies7,8 and 2 large meta-analyses of randomized clinical trials.9,10 In the Nurses’ Health Study, calcium intake, both dairy and nondairy, showed an inverse association with risk of development of hypertension.7 A previous cross-sectional study of Japanese showed that total and dairy calcium intakes were inversely associated with systolic blood pressure levels, whereas no association was found for nondairy calcium intake.8 In the present study, dairy calcium intake was inversely associated with a history of hypertension. Calcium in milk and dairy products can be absorbed more efficiently than that in other foods, probably because casein, abundant in milk and dairy products, enhances calcium absorption.18 In addition to its hypotensive effect, calcium reduces platelet aggregation19 and lowers plasma total cholesterol levels by forming insoluble complexes with fatty acids and reducing the absorption of fatty acids.20 The hypotensive effect of dietary calcium helps to explain a protective effect against total stroke, especially hemorrhagic stroke. Potential effects of dietary calcium on reduced platelet aggregation and total cholesterol levels may result in a further reduction of ischemic stroke risk. These beneficial effects may also contribute to a reduction in risk of CHD, but they may be counterbalanced by an adverse effect of saturated fat, rich in whole milk and dairy products. In the present study, saturated fat intake tended to be positively associated with a risk of CHD: the multivariable risk for the highest versus the lowest quintile of saturated fat intake was 1.51 (95% CI, 0.98 to 2.35), P=0.06 and P for trend=0.12.
The limitations of the present study warrant discussion. First, we were unable to evaluate the effect of calcium supplementation, because the questionnaire did not ask about the use of calcium supplements. However, the use of calcium supplements is not popular among Japanese populations: according to a national nutritional survey, only 3% of Japanese subjects used a calcium supplement.4 Second, we used only 2 food items for dairy products in the baseline questionnaire. On the other hand, we evaluated the contribution to calcium intake of each food entered in the dietary records, which confirmed that major sources of dietary calcium for Japanese were mostly limited to milk, yogurt, and processed cheese.21 We therefore believe that we were able attain an adequate evaluation of dairy calcium intake. Third, the proportion of history of hypertension in the present study was 16%, which seems to be lower than that reported for a national representative sample, 25%.22 That proportion in our study was almost the same as the prevalence of hypertension, 17%, defined as systolic blood pressure ≥160 mm|Hg, diastolic blood pressure ≥95 mm|Hg, or taking antihypertensive medication among ≈15 000 subjects who reported their blood pressures at the baseline survey.
The methodology of this study is superior to that of previous studies. We evaluated a large prospective cohort enrolled from the Japanese general populations. A prospective study has little recall bias, and results from the general population are more relevant than results from occupational, hospital-based, or volunteer samples. We examined the incidence of stroke and CHD rather than mortality. Incidence is a more direct measure of CVD risk than mortality because mortality is influenced by treatment.
Previous cohort studies showed that a high sodium intake and a low potassium intake were associated with increased risk of death from stroke,23,24 partly due to increased blood pressure levels.25,26 The present study added the evidence that calcium intake may reduce the risk of stroke, independent of sodium and potassium intakes.
In conclusion, our large, prospective study of middle-aged Japanese men and women showed that dietary intakes of total and dairy calcium were inversely associated with the incidence of stroke. Clinical trials are necessary to confirm a beneficial effect of dietary calcium intake on stroke prevention.
S. Tsugane (principal investigator), T. Hanaoka, M. Inoue and T. Sobue, Epidemiology and Prevention Division, National Cancer Center, Tokyo; J. Ogata, S. Baba, T. Mannami, Y. Kokubo, and A. Okayama, National Center for Circulatory Diseases, Suita; K. Miyakawa, F. Saito, A. Koizumi, Y. Sano, and I. Hashimoto, Iwate Prefectural Ninohe Public Health Center, Ninohe; Y. Miyajima, N. Suzuki, S. Nagasawa, and Y. Furusugi, Akita Prefectural Yokote Public Health Center, Yokote; H. Sanada, Y. Hatayama, F. Kobayashi, H. Uchino, Y. Shirai, T. Kondo, R. Sasaki, and Y. Watanabe, Nagano Prefectural Saku Public Health Center, Saku; Y. Kishimoto, E. Takara, M. Kinjo, T. Fukuyama, and M. Irei, Okinawa Prefectural Ishikawa Public Health Center, Ishikawa; S. Matsushima and S. Natsukawa, Saku General Hospital, Usuda; S. Watanabe and M. Akabane, Tokyo University of Agriculture, Tokyo; M. Konishi and K. Okada, Ehime University, Matsuyama; S. Tominaga, Aichi Cancer Center Research Institute, Nagoya; M. Iida and W. Ajiki, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka; S. Sato, Osaka Medical Center for Health Science and Promotion, Osaka; the late M. Yamaguchi, Y. Matsumura, and S. Sasaki, National Institute of Health and Nutrition, Tokyo; Y. Tsubono, Tohoku University, Sendai; H. Iso and Y. Honda, University of Tsukuba, Tsukuba; H. Sugimura, Hamamatsu University, Hamamatsu; M. Kabuto, National Institute for Environmental Studies, Tsukuba; N. Yasuda, Kochi Medical School; S. Kono, Kyushu University; K. Suzuki, Research Institute for Brain and Blood Vessels Akita, Akita; Y. Takashima, Kyorin University; and E. Maruyama, Kobe University.
The authors thank all staff members in each of the areas studied and in the central office for their painstaking efforts in conducting the baseline survey and follow-up.
Sources of Funding
This study was supported by grants-in-aid for cancer research and by the Third Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labor and Welfare of Japan.
- Received December 10, 2007.
- Revision received February 3, 2008.
- Accepted February 19, 2008.
Abbott RD, Curb JD, Rodriguez BL, Sharp DS, Burchfiel CM, Yano K. Effect of dietary calcium and milk consumption on risk of thromboembolic stroke in older middle-aged men: the Honolulu Heart Program. Stroke. 1996; 27: 813–818.
Iso H, Stampfer MJ, Manson JE, Rexrode K, Hennekens CH, Colditz GA, Speizer FE, Willett WC. Prospective study of calcium, potassium, and magnesium intake and risk of stroke in women. Stroke. 1999; 30: 1772–1779.
Ascherio A, Rimm EB, Hernan MA, Giovannucci EL, Kawachi I, Stampfer MJ, Willett WC. Intake of potassium, magnesium, calcium, and fiber and risk of stroke among US men. Circulation. 1998; 98: 1198–1204.
Ministry of Health, Labour and Welfare. Kokumin Eiyo no Genjyo (The National Health and Nutrition Survey in Japan, 2004). Tokyo: Daiichi Shuppan; 2006.[In Japanese].
Umesawa M, Iso H, Date C, Yamamoto A, Toyoshima H, Watanabe Y, Kikuchi S, Koizumi A, Kondo T, Inaba Y, Tanabe N, Tamakoshi A. Dietary intake of calcium in relation to mortality from cardiovascular disease: the JACC Study. Stroke. 2006; 37: 20–26.
Witteman JC, Willett WC, Stampfer MJ, Colditz GA, Sacks FM, Speizer FE, Rosner B, Hennekens CH. A prospective study of nutritional factors and hypertension among US women. Circulation. 1989; 80: 1320–1327.
Iso H, Terao A, Kitamura A, Sato S, Naito Y, Kiyama M, Tanigaki M, Iida M, Konishi M, Shimamoto T, Komachi Y. Calcium intake and blood pressure in seven Japanese populations. Am J Epidemiol. 1991; 133: 776–783.
Health and Welfare Statistics Association. Annual statistical report of national health conditions 2006. J Health Welfare Stat. 2006: 430–433.[in Japanese].
Yoko F. Trends in dietary intake of calcium and NaCl in nine geographical and occupational populations. J Med Soc Toho. 1990; 37: 283–305.[In Japanese].
Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas AM, Pajak A. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project: registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in four continents. Circulation. 1994; 90: 583–612.
Van der Vijver LP, van der Waal MA, Weterings KG, Dekker JM, Schouten EG, Kok FJ. Calcium intake and 28-year cardiovascular and coronary heart disease mortality in Dutch civil servants. Int J Epidemiol. 1992; 21: 36–39.
Jolma P, Koobi P, Kalliovalkama J, Kahonen M, Fan M, Saha H, Helin H, Lehtimaki T, Porsti I. Increased calcium intake reduces plasma cholesterol and improves vasorelaxation in experimental renal failure. Am J Physiol Heart Circ Physiol. 2003; 285: 1882–1889.
Ministry of Welfare. The National Study for Circulatory Disorders, 1990. Tokyo: Ministry of Welfare; 1993.[in Japanese].
Nagata C, Takatsuka N, Shimizu N, Shimizu H. Sodium intake and risk of death from stroke in Japanese men and women. Stroke. 2004; 35: 1543–1547.
Siani A, Strazzullo P, Russo L, Guglielmi S, Iacoviello L, Ferrara LA, Mancini M. Controlled trial of long term oral potassium supplements in patients with mild hypertension. BMJ. 1987; 294: 1453–1456.