Relation of Adult Height With Stroke Mortality in Japan
Background and Purpose— The age-adjusted stroke mortality rate in Japan was the highest in the world from 1950 to the 1970s, but it started to dramatically decrease after 1965. In addition to improved management of high blood pressure, the increase in average height might also contribute to this reduction. The present study investigates whether height is an independent risk for stroke mortality in Japan.
Methods— Among participants of the National Survey on Cardiovascular Diseases in 1980 who were randomly selected from the Japanese population, we followed up 3969 and 4955 Japanese men and women without prior cardiovascular disease for a maximum of 19 years and observed 158 and 132 stroke deaths.
Results— Height was inversely correlated with age and with crude stroke mortality. The relationship was attenuated in men when we adjusted for age or other possible confounders (multivariate adjusted relative hazards of a 5-cm increase of height for stroke mortality: 0.92, 95% CI: 0.79 to 1.08). For women, the inverse relationship (relative hazard: 0.77: 95% CI: 0.64 to 0.91) remained after multivariate adjustment. These relationships persisted when we stratified participants by age.
Conclusions— Height is inversely related to stroke mortality and the relationship is statistically significant among Japanese women.
Age-adjusted stroke mortality rate in Japan was the highest in the world from the 1950 to the 1970s1,2 and twice as high as that in the West at that time.1,2 However, age-adjusted stroke mortality in Japanese started decreasing dramatically after 1970. The rate was 175.8 per million in 1970 and 104.7 per million in 20011 among urban Japanese.3 A reduction in blood pressure (BP) and improved BP management are believed to be main factors in decreasing the incidence of stroke.4
The average height of Japanese people has simultaneously increased. The mean height at age 30 to 39 was lower in men and women born between 1936 and 1945 (163.8 cm for men and 152.7 cm for women) than between 1961 to 1970 (170.6 cm for men and 157.6 cm for women).5 Several previous studies,6–16 but not all,17 have reported that height is inversely related to stroke mortality. Therefore, increment of average height might be also associated with reduction of Japanese stroke mortality. However, no studies have been reported in Japanese and also in other Asians, excluding one article.16
To investigate whether height is an independent risk factor for stroke mortality, we performed prospective studies of a representative population of Japanese individuals.
The subjects of this cohort study were the participants in the National Cardiovascular Survey of 1980 that was conducted together with the National Nutrition Survey that is annually implemented using a similar method and a questionnaire. The standardized procedures used in this survey have been described elsewhere.18–20 All household members aged 30 years or older (up to 92 years) were surveyed in 300 census tracts that were randomly selected throughout Japan. The baseline survey included medical examinations, blood pressure measurements, blood tests, and a self-administered questionnaire about lifestyle. Trained staff at local health centers in the respective districts performed the examinations in community centers. A history of illnesses, including heart disease, stroke, and diabetes, as well as smoking and drinking habits were obtained from the questionnaire. Height and weight were measured when the subjects wore light clothing and no shoes. Subjects were asked to note whether they were current smokers, had quit smoking, or had never smoked, and smokers were asked to note the number of cigarettes smoked each day. Similarly, subjects were asked to answer whether they had never consumed alcohol, did so in the past, occasionally or regularly (daily) do so. Blood pressure was measured using a standard sphygmomanometer to obtain systolic and diastolic BP. Nonfasting blood samples were collected. The measurement precision and accuracy of the assay for serum total cholesterol (TC) was certified by the Lipid Standardization Program administered by the Centers for Disease Control and Prevention. Diabetes was defined as a nonfasting glucose value ≥200 mg/dL or a self-reported history of diabetes.21
A total of 10 546 individuals aged 30 years or older for whom complete baseline information regarding age, gender, and BP was complete in the 1980 data set was defined as the cohort (NIPPON DATA80).18–20 From these, we excluded two who did not have information about height; 755 with a history of stroke (N=117), coronary heart disease (N=163), or other heart disease (N=475); 16 who did not have complete information about confounding factors; and 849 participants who could not follow up because of incomplete residential access information after first survey. Consequently, we analyzed 8924 participants (3969 men and 4955 women). There was no significant difference in the mean age-adjusted height between the participants lost to follow up and those in the study.
As reported previously,18–20 we confirmed the participants who had died in each area by computer-matching data from the National Vital Statistics using area, gender, date of birth, and death as key codes. We then clarified causes of death using the National Vital Statistics. All death certificates issued by medical doctors in Japan are forwarded to a central database at the Ministry of Health and Welfare through public health centers in the area of residence.
The underlying causes of death for Japan’s National Vital Statistics were to be coded according to the International Classification of Diseases, 9th Revision (ICD-9) by the close of 1994 and to the International Classification of Diseases, 10th Revision (ICD-10) from the beginning of 1995. Codes of 430 to 438 in ICD-9 and I60–I69 in ICD-10 were defined as death from total strokes, which included death from cerebral infarction (codes 433, 434, 437.7a, 7b in ICD-9, I61 and I69.1 in ICD-10) and from cerebral hemorrhage (codes of 431 to 432 in ICD-9, I63 and I69.3 in ICD-10). Permission to use the National Vital Statistics was obtained from the Management and Coordination Agency of the Government of Japan. Approval for this study was obtained from the Institutional Review Board of Shiga University of Medical Science (No. 12-18, 2000).
To examine the association between height and stroke mortality, participants were divided into quartiles. We compared the basic characteristics among height quartiles by mean as continuous variable and percentages as dichotomous variables. We also analyzed the crude and age-adjusted relationship between height and stroke risk factors. Pearson correlation coefficient was used for the crude analysis of continuous variables (systolic BP and TC), and the slope of the regression coefficient was used for age-adjusted analysis. As for the dichotomous variables (diabetes, current smoking, and daily drinking), OR estimated by logistic regression used both crude and age-adjusted analyses. The units of change in height were set at five centimeters in the logistic regression mentioned previously.
We estimated the relative hazards (RH) and the 95% CIs of height for stroke mortality using the Cox proportional hazard model. Because the average height of men and women is quite different, we separately analyzed men and women. We treated the lowest height quartile as a reference group. We used three models to estimate RH, namely crude, age-adjusted, and multivariate adjusted models. The multivariate adjusted model included the following possible confounding factors: age, body weight, systolic BP, use of antihypertensive medication, diabetes, TC, smoking category (never smoked; exsmoker; current smoker 1 to 20 cigarettes per day, 21 to 40, or 41+ cigarettes per day), and alcohol consumption category (never, past, occasional, and daily). We also analyzed relation of continuous height increase per 5-cm increase with stroke mortality. To investigate age-specific differences in the effect of height on stroke mortality, we established five age groups (30 to 49, 50 to 59, 60 to 69, 70 to 79, and 80 over) and performed age-specific analysis in each by Cox regression. A probability value of <0.05 was considered significant. SAS software (version 9.1) was used for analyses. To estimate the RH observed in previous estimates of RH per height reduction, we calculated the reciprocal of the RH to estimate the RH per height increase.
The mean age and height of the study participants were 50.0 (SD: 13.0) years and 162.3 (SD: 6.7) cm for men and 50.2 (SD: 13.1) years and 150.1 (SD: 6.1) cm for women.
Taller participants were younger than smaller individuals (Table 1). Correlations between height and age were close both in men (r=−0.44, P<0.01) and in women (r=−0.52, P<0.01). Mean systolic BP levels and the prevalence of diabetes were lower in taller participants. The mean TC level was higher in taller men. Conversely, the mean TC level was higher in smaller women. These characteristics were mostly attenuated after adjusting for age (Table 2).
After 19.0 years of follow up, 158 (97 ischemic, 37 hemorrhagic, and 24 unidentified) men and 132 (69 ischemic, 26 hemorrhagic, and 37 unidentified) women died of stroke. Table 3 shows that the crude stroke mortality rates were highest in the lowest height quartile and gradually decreased as the category increased both in men and in women.
Adjustment for age largely attenuated the inverse relationship and this relationship between height and stroke mortality in men was not statistically significant (RH of stroke mortality for 5-cm height increase, 0.90; 95% CI, 0.79 to 1.03). However, the inverse relationship between height and stroke mortality remained significant for women. This relationship was also unchanged after adjusting for other possible confounding factors (RH of stroke mortality for 5-cm height increase: 0.77; 95% CI, 0.64 to 0.91).
We also observed an inverse relationship between height and ischemic stroke in men and in women (multivariate adjusted RH [95% CI] for 5-cm height increase: men, 0.92 [0.75 to 1.13]; women, 0.66 [0.51 to 0.84]). Height was also inversely related to cerebral hemorrhage in men, but not in women among whom 26 had cerebral hemorrhage (multivariate adjusted RH [95% CI] for 5-cm height increase: men, 0.85 [0.62 to 1.16]; women, 1.05 [0.70 to 1.55]).
Because the correlations between age and height were highly significant for both men and women, we analyzed the relationship between height and stroke mortality according to age category (Table 4). The age-specific analyses of men produced inconsistent findings. However, height and stroke mortality in women were closely and inversely related (range of RH: 0.63 to 0.78) except for the age category 60 to 69 years (RH=1.16).
No relationship between height and heart disease, cancer, and total mortality was determined (RH and 95% CI of mortality for 5-cm height increase for men—heart disease: 1.03 [0.87 to 1.21]; cancer: 1.09 [0.98 to 1.22]; total mortality: 1.02 [0.95 to 1.08]) and that for women (heart disease: 1.07 [0.91 to 1.27]; cancer: 1.07 [0.93 to 1.23]; and total mortality: 1.02 [0.95 to 1.10]).
The present study uncovered a close inverse relationship between height and stroke mortality in women. We also observed an inverse but nonsignificant relationship between these two parameters in men.
The strengths of this study were as follows: (1) Japanese participants, (2) long follow up, and (3) separate analysis of age groups.
Several others have reported that height is inversely related to stroke mortality or incidence.6–17 Except for one follow-up study of an American nurse,17 height is consistently inversely related to stroke mortality and incidence. The range of RH per 5-cm increase was 0.84 to 0.93 for men and 0.74 to 0.90 for women.6,10,14,16 Similarly, the range of RH per 10-cm increase was 0.76 to 0.84 for men and 0.83 for women.11,12,15 These data were consistent with of ours, that is, the RH values per 5-cm increase were 0.92 (not significant) for men and 0.77 for women. The only study reported from Asia showed that the RH for stroke mortalities per 5-cm increase in height was 0.93 (95% CI: 0.88 to 0.98) in men,16 but there no data available for women.
To understand why height relates to stroke mortality, several explanations should be considered. First, because height is closely correlated with age, the possibility of an age effect cannot be excluded. However, we considered that the age or cohort effect did not fully explain the relationship between height and stroke mortality for the following reasons. Age adjustment did not fully attenuate the inverse relationship between height and stroke mortality in women and the association was inverse in most age categories of women. Second, height might relate to the classic stroke disease risk factors. However, our analyses did not identify a clear, statistically significant relationship between height and systolic BP, TC, diabetes, smoking, and drinking status after adjusting for age. Furthermore, adjusting these variables did not fully attenuate the relationship between height and stroke mortality in women. Third, because height is a function of both genetic and environmental factors, the relationship between height with stroke should be considered from two viewpoints. A recent study has shown an inverse relationship between height and mortality attributable to coronary heart disease, even within monozygotic discordant twins, and the authors concluded that the inverse relationship between height and coronary heart disease mortality can be explained by environmental factors.22 Of course, because their study focused on mortality attributable to coronary heart disease and not stroke, heredity, which relates to height itself, might affect stroke mortality. Further studies are required to clarify whether heredity itself affects the relationship between height and stroke. From the environmental viewpoint, an anthropologic measure such as height is one surrogate measure of childhood development. Throughout the fetal period and childhood, environmental factors such as nutrition, infection, and socioeconomic circumstance deeply influence development.23 Several, but not all, studies have shown that adverse socioeconomic conditions early in life are related to stroke mortality.24 Although height itself might be a risk for cardiovascular disease, we considered height as a surrogate measure of childhood development. Thus, although its contribution would be smaller than improved BP control or management, increasing height might partly contribute to the decrease in stroke mortality among Japanese over the past four decades, at least for women.
We observed no significant relationship between height and stroke mortality among men, which is inconsistent with previous findings. Although our results did not reach statistical significance, the RH was similar to those of a Korean study (RH=0.93 observed; 1263 stroke mortality).16 Thus, the inverse relationship between height and stroke mortality might not be very close in men and our study might not be sufficiently sensitive to detect it. A larger and longer follow-up study might be required to clarify this issue.
This study has several limitations. Because we did not have any information on socioeconomic status of the participants in this study, we could not adjust for socioeconomic status at baseline. However, all Japanese are covered by health insurance, which allows to access to all potential treatment. Therefore, treatment for participants of lower and higher socioeconomic status should not significantly differ and thus cannot explain the inverse relationship. We also did not have information on childhood socioeconomic status, which is more likely to affect adult height. Therefore, although we considered height as a surrogate measure of childhood development, we could not clarify whether the inverse relation between height and stroke mortality could be fully explained by childhood socioeconomic status. Second, because we did not have incidence data, height might not predict stroke incidence, but only be a marker of prognosis after a stroke event. However, the relationships between risk factors and stroke mortality and cardiovascular disease incidence were similar. Therefore, we did not consider this limitation critical.
Height is inversely related to stroke mortality and the relationship was statistically significant in women. Thus, the decreasing trend in stroke mortality among the Japanese population might be partly explained by an increase in height attributable to improved socioeconomic circumstances during fetal, childhood, and adolescent periods.
The authors thank all the staff of the public health centers.
Sources of Funding
This study was supported by a Grant-in-Aid from the Ministry of Health and Welfare under the auspices of the Japanese Association for Cerebro-cardiovascular Disease Control, a Research Grant for Cardiovascular Diseases (7A-2) from the Ministry of Health, Labor and Welfare, and a Health and Labor Sciences Research Grant, Japan (Comprehensive Research on Aging and Health: H11-chouju-046, H14-chouju-003, and H17,18-chouju-012).
Members of the NIPPON DATA Research Group are listed in reference 18.
- Received August 30, 2006.
- Accepted September 10, 2006.
Vital statistics (Chapter 2). In: Journal of Health and Welfare Statistics. 2003; 50: 40–64. Tokyo: Health and Welfare Statistics Association, Tokyo [in Japanese].
Kubo M, Kiyohara Y, Kato I, Tanizaki Y, Arima H, Tanaka K, Nakamura H, Okubo K, Iida M. Trends in the incidence, mortality, and survival rate of cardiovascular disease in a Japanese community: the Hisayama study. Stroke. 2003; 34: 2349–2354.
The National Health and Nutrition Survey in Japan. Tokyo: Daiichi Syuppan Publisher, 2003[in Japanese].
Njolstad I, Arnesen E, Lund-Larsen PG. Body height, cardiovascular risk factors, and risk of stroke in middle-aged men and women. A 14-year follow-up of the Finnmark Study. Circulation. 1996; 94: 2877–2882.
Wannamethee SG, Shaper AG, Whincup PH, Walker M. Adult height, stroke, and coronary heart disease. Am J Epidemiol. 1998; 148: 1069–1076.
Parker DR, Lapane KL, Lasater TM, Carleton RA. Short stature and cardiovascular disease among men and women from two southeastern New England communities. Int J Epidemiol. 1998; 27: 970–975.
Hart CL, Hole DJ, Smith GD. Risk factors and 20-year stroke mortality in men and women in the Renfrew/Paisley study in Scotland. Stroke. 1999; 30: 1999–2007.
Jousilahti P, Tuomilehto J, Vartiainen E, Eriksson J, Puska P. Relation of adult height to cause-specific and total mortality: a prospective follow-up study of 31 199 middle-aged men and women in Finland. Am J Epidemiol. 2000; 151: 1112–1120.
Davey Smith G, Hart C, Upton M, Hole D, Gillis C, Watt G, Hawthorne V. Height and risk of death among men and women: aetiological implications of associations with cardiorespiratory disease and cancer mortality. J Epidemiol Community Health. 2000; 54: 97–103.
McCarron P, Greenwood R, Ebrahim S, Elwood P, Smith GD. Adult height is inversely associated with ischaemic stroke. The Caerphilly and Speedwell collaborative studies. J Epidemiol Community Health. 2000; 54: 239–240.
McCarron P, Hart CL, Hole D, Smith GD. The relation between adult height and haemorrhagic and ischaemic stroke in the Renfrew/Paisley study. J Epidemiol Community Health. 2001; 55: 404–405.
Goldbourt U, Tanne D. Body height is associated with decreased long-term stroke but not coronary heart disease mortality? Stroke. 2002; 33: 743–748.
McCarron P, Okasha M, McEwen J, Smith GD. Height in young adulthood and risk of death from cardiorespiratory disease: a prospective study of male former students of Glasgow University, Scotland. Am J Epidemiol. 2002; 155: 683–687.
Song YM, Smith GD, Sung J. Adult height and cause-specific mortality: a large prospective study of South Korean men. Am J Epidemiol. 2003; 158: 479–485.
Rich-Edwards JW, Manson JE, Stampfer MJ, Colditz GA, Willett WC, Rosner B, Speizer FE, Hennekens CH. Height and the risk of cardiovascular disease in women. Am J Epidemiol. 1995; 142: 909–917.
Ueshima H, Choudhury SR, Okayama A, Hayakawa T, Kita Y, Kadowaki T, Okamura T, Minowa M, Iimura O. Cigarette smoking as a risk factor for stroke death in Japan: NIPPON DATA80. Stroke. 2004; 35: 1836–1841.
Okamura T, Tanaka H, Miyamatsu N, Hayakawa T, Kadowaki T, Kita Y, Nakamura Y, Okayama A, Ueshima H, for the NIPPON DATA80 Research Group. The relationship between serum total cholesterol and all-cause or cause-specific mortality in a 17.3-year study of a Japanese cohort. Atherosclerosis. 2006; Mar 8 [Epub ahead of print].
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2005; 28: S37–42.
Silventoinen K, Zdravkovic S, Skytthe A, McCarron P, Herskind AM, Koskenvuo M, de Faire U, Pedersen N, Christensen K, Kaprio J; GenomEUtwin Project. Association between height and coronary heart disease mortality: a prospective study of 35 000 twin pairs. Am J Epidemiol. 2006; 163: 615–621.
Miura K, Nakagawa H, Greenland P. Invited commentary: height–cardiovascular disease relation: where to go from here? Am J Epidemiol. 2002; 155: 688–689.