SEARCH

SEARCH BY CITATION

Keywords:

  • Body mass index;
  • coronary heart disease;
  • ischaemic stroke

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

Studying obesity in the Asia–Pacific region is difficult because of the diverse ethnic background and different stages of economic and nutrition transition. The burden of cardiovascular disease associated with overweight (defined as body mass index ≥25 kg m−2) was previously estimated for countries within the region. However, using the conventional cut-point of 25 kg m−2 ignores the continuous association between body mass index and cardiovascular disease from approximately 20 kg m−2. By estimating the proportion of cardiovascular disease that would be prevented if the theoretical mean body mass index in the population was shifted to 21 kg m−2, nationally representative data from 15 countries suggested the population attributable fractions for cardiovascular disease were approximately three times higher than the previous estimates. Coronary heart disease attributable to body mass index other than 21 kg m−2 ranged from 2% in India to 58% in American Samoa. Similarly, the population attributable fraction for ischaemic stroke ranged from 3% in India to 64% in American Samoa. If cardiovascular risk increases from 21 kg m−2 applies to all populations, most countries in the region will need to aim towards substantially reducing their current population mean body mass index in order to lower the burden of cardiovascular disease associated with excess weight.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

The Asia–Pacific region includes populations from diverse ethnic backgrounds that are undergoing different stages of economic and nutrition transitions. While the prevalence of overweight and obesity in Pacific Island countries is already high (1), the prevalence in Asian countries has increased substantially over the last decade (2,3). Coronary heart disease and ischaemic stroke mortalities attributable to overweight (defined as body mass index ≥25 kg m−2) were previously estimated to be 0.8–9.2% and 0.9–10.2%, respectively, for 14 countries in the Asia–Pacific region (4). However, cardiovascular disease risks associated with body mass index have been reported to increase from 20 kg m−2(5). Using the conventional cut-point of 25 kg m−2 in the estimation ignores the continuous nature of these associations. Hence, the reported population attributable fractions probably did not reflect the true contribution of excess body weight towards the burden of cardiovascular disease.

In this paper, our aim was to estimate the percentage of cardiovascular disease events that would be prevented in each population in the Asia–Pacific region (limited to countries with suitable data), should the population have the ideal (lowest risk) anthropometric structure. The World Health Organization has previously suggested that this ideal population would have a mean body mass index of 21 kg m−2, with a standard deviation of unity (6).

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

We searched Medline, WHO Global InfoBase (7) and government web sites for nationally representative studies in countries from the World Health Organization South-East Asia and Western Pacific regions (Asia–Pacific region). We included studies that were conducted on adults since 2000 and reported sex-specific mean body mass index and measure of variability (standard deviation, standard error or confidence intervals). For countries with more than one eligible study, we selected the study with the most recent data. In addition, age-specific data were collected, where available.

Countries that reported mean body mass index with standard error or confidence intervals were directly converted to standard deviation. Where necessary, standard deviations were imputed using simple linear regression models fitted to the data on mean and standard deviation published by the Asia Pacific Cohort Studies Collaboration (5). SAS 9.1 for Windows (SAS Institute, Inc., Cary, NC, USA) was used to perform simple linear regression.

We calculated population attributable fractions by comparing exposures to the theoretical minimum distribution: a normal distribution with mean 21.0 kg m−2 and standard deviation 1.0 kg m−2. Population attributable fractions were calculated by theoretically shifting the current distribution to the theoretical distribution (6,8). Hence, the population attributable fraction is the per cent estimated events of the selected disease that can be attributable to the amount by which current exposure exceeds the optimal minimum. The continuous relationship between body mass index and each of coronary heart disease and ischaemic stroke was characterized using a hazard ratio for a 1 kg m−2 increase in body mass index. This hazard ratio was obtained by converting the relative risk reduction of each outcome per 2 kg m−2 in body mass index published by the Asia Pacific Cohort Studies Collaboration (5). The Asia Pacific Cohort Studies Collaboration, with 44 studies from nine countries, has the largest cohort studies database on cardiovascular disease in the Asia–Pacific region. Hazard ratios used here were based on data from 33 studies in eight countries (Australia, China, Hong Kong, Japan, New Zealand, Singapore, South Korea and Taiwan), with over 2 million person years of follow-up and mean duration of follow-up of 6.9 years (5). Details of the Collaboration have been published elsewhere (http://www.apcsc.info/). Coronary heart disease included non-fatal myocardial infarction and fatal coronary heart disease events. Ischaemic stroke included both fatal and non-fatal events. Because of the weak continuous association between haemorrhagic stroke and body mass index, except at high body mass index levels (5), haemorrhagic stroke was not included in the analysis. Age-specific hazard ratios, from the same source (5), were used in the calculation of age-specific population attributable fractions for studies with mean body mass index by age group. To allow direct comparison between countries, an overall age-standardized population attributable fraction for each country with age-specific data was obtained by weighting each age-specific population attributable fraction by the age-specific deaths because of coronary heart disease and stroke, respectively, for the world(9).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

Fifteen countries with 330 374 participants from the Asia–Pacific region were included (Table 1). The median study commencement year was 2005. Seven studies included participants aged 15–17 years as adults. Body mass indexes were calculated from measured height and weight data. Standard deviations for five studies were imputed. Regardless of sex, mean body mass index was highest in American Samoa and lowest in India.

Table 1.  Summary of the most recent national representative studies with mean body mass index (BMI) for countries in the Asia–Pacific region
CountryStudy nameStudy yearStudy sizeAge (year)Mean (SD) BMI (kg m−2)
MenWomen
  • *

    Standard deviation imputed from simple linear regression.

  • Standard deviation was computed from the raw data of the Australian Diabetes, Obesity and Lifestyle Study.

  • NCD, Noncommunicable disease; SD, standard error.

American Samoa (10)American Samoa NCD Risk Factors STEPS Report20041 99525–6433.7 (7.8)36.2 (5.0)
Australia (11)Australian Diabetes, Obesity and Lifestyle Study1999–200011 247≥2526.9 (4.1)26.4 (5.6)
China (12)China National Diabetes and Metabolic Disorders Study2007–200846 239≥2024.0 (6.9)23.4 (8.5)
Fiji (13)Fiji NCD Risk Factors STEPS Report20025 54415–6424.2 (3.6)*26.7 (4.3)*
India (7)National Family Health Survey2005–2006177 52315–4920.2 (2.4)*20.5 (2.5)*
Indonesia (14)Indonesia NCD Risk Factors STEPS Report200113 133≥1520.4 (4.0)21.3 (4.3)
Japan (7)National Nutrition Survey20008 305≥2023.2 (3.2)22.5 (3.6)
Kiribati (15)Kiribati NCD Risk Factors STEPS Report2004–20061 35125–6429.4 (7.5)31.5 (7.0)
Malaysia (16)Malaysia NCD Risk Factors STEPS Report2005–20062 57225–6424.7 (7.4)25.6 (7.0)
Mongolia (17)Mongolia NCD Risk Factors STEPS Report20063 40415–6423.3 (3.3)*24.5 (3.7)*
Nauru (18)Nauru NCD Risk Factors STEPS Report20042 25415–6431.7 (5.8)*32.5 (6.0)*
New Zealand (19)2006/07 New Zealand Health Survey2006–200712 488≥1527.3 (7.4)27.1 (8.7)
South Korea (20)3rd Korea National Health & Nutrition Examination Survey20055 429≥2024.0 (3.1)23.5 (3.4)
Thailand (21)Thai Third National Health Examination Survey200438 323≥1822.7 (3.9)23.9 (4.6)
Tokelau (22)Tokelau NCD Risk Factors STEPS Report200556715–6431.5 (5.7)*33.2 (6.2)*

While sex-specific hazard ratios per 1 kg m−2 was available for coronary heart disease (1.07 in men and 1.05 in women), an overall hazard ratio was used for ischaemic stroke (1.07). Because the hazard ratios were identical in men, the population attributable fractions for both cardiovascular outcomes ranged from 2% in India to 62% in American Samoa (Figs 1 and 2). In women, the population attributable fractions ranged from 2% in India to 53% in American Samoa for coronary heart disease (Fig. 1) and 3% in India to 66% in American Samoa for ischaemic stroke (Fig. 2). In general, Asian countries had lower cardiovascular events attributable to body mass index and Pacific Island countries had higher cardiovascular events attributable to body mass index. Australia and New Zealand, with predominantly Caucasians, were in between.

image

Figure 1. Population attributable fraction for coronary heart disease related to body mass index of 21 kg m−2 as theoretical minimum threshold by sex and country in the Asia–Pacific region.

Download figure to PowerPoint

image

Figure 2. Population attributable fraction for ischaemic stroke related to body mass index of 21 kg m−2 as theoretical minimum threshold by sex and country in the Asia–Pacific region.

Download figure to PowerPoint

Ten studies reported age-specific mean body mass index. Compared with the crude population attributable fractions, the age-standardized population attributable fractions were higher. The age-standardized population attributable fractions for men ranged from 5% in Indonesia to 84% in American Samoa for coronary heart disease and from 3% in Indonesia to 81% in American Samoa for ischaemic stroke. For women, the age-standardized population attributable fractions ranged from 9% in Indonesia to 91% in American Samoa for coronary heart disease and from 10% in Indonesia and Japan to 89% in America Samoa for ischaemic stroke.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

We estimated the burden of coronary heart disease and ischaemic stroke that could be prevented if the population distribution of body mass index was shifted towards a mean of 21 kg m−2. Our results suggest between 2% and 64% of these cardiovascular disease outcomes in the Asia–Pacific region is because of population mean body mass index higher than 21 kg m−2. These estimates are substantially higher than those previously estimated for overweight (body mass index ≥25 kg m−2) with body mass index <25 kg m−2 as the referent group (4). However, only estimates for Australia and Japan can be directly compared, as different study data were used for other countries that also appeared in both reports. Nevertheless, these estimates should be compared bearing in mind that the previous study estimated the burden of fatal cardiovascular disease only while the current study estimated the burden of both fatal and non-fatal cardiovascular outcomes. Compared with the population attributable fractions associated with overweight and obesity, the current estimates were approximately three times greater for Australia and Japan, respectively.

Coronary heart disease and cerebrovascular disease mortality attributable to body mass index >21 kg m−2 were estimated by the World Health Organization for the year 2000. The population attributable fraction for coronary heart disease mortality ranged from 5% in the high mortality South-East Asian region to 11–12% in the Western Pacific region, and 15% in the low mortality South-East Asian region (23). For cerebrovascular disease mortality, the population attributable fractions ranged from 3% in the high mortality South-East Asian region to 7% in the low mortality South-East Asian region. Compared with our estimates, the lower attributable fractions are likely because of the lower mean body mass index for most, if not all, countries over a decade ago. The World Health Organization's estimates also did not include non-fatal events.

Because the age for study inclusion varied between countries, it is technically incorrect to compare the population attributable fractions between countries. However, we were able to compare those countries with mean body mass index by age group. A similar trend was observed from population attributable fractions that were standardized to the world age-specific deaths because of coronary heart disease and stroke, respectively. However, it should be pointed out that the association between body mass index and cardiovascular disease is stronger in the young (age <60 years) and weak in the old (age ≥70 years). For example, hazard ratios for coronary heart disease were 1.13 in those age <60 years, 1.05 in those age 60–69 years and 1.03 in those age ≥70 years (5). Therefore, in theory, a country with a young population will have greater burden of coronary heart disease associated with body mass index than a country with an ageing population with the same mean body mass index. However, the absolute risk of developing coronary heart disease is greater in the old; hence, the burden of coronary heart disease is higher in an ageing population even though the mean body mass index is the same as a young population.

Moreover, the ethnic distribution of a country can affect the population attributable fraction estimate. Population attributable fractions can be calculated directly from the national mean body mass index for countries with homogeneous populations such as China and Japan. For countries with diverse ethnic backgrounds such as Fiji and New Zealand, population attributable fractions should be calculated using mean body mass index for each ethnic group. For example, the mean body mass index was 25.3 kg m−2 for Fijian men and 22.5 kg m−2 for Indo-Fijian men (12). Therefore, the population attributable fractions for coronary heart disease were 27.0% in Fijian men and only 12.0% in Indo-Fijian men assuming that the risk in both ethnic groups has the same magnitude. We were unable to calculate population attributable fractions that are standardized by ethnic groups for countries with populations from diverse ethnic backgrounds given that not all of such countries reported mean body mass index by ethnic group.

Because of the higher mean body mass index in Pacific Island populations, the population attributable fractions for cardiovascular disease were higher in Pacific Island countries compared with Asian countries. However, the age-standardized death rates for coronary heart disease and cerebrovascular disease in Pacific Island countries such as Kiribati and Nauru were comparable or lower than in India and Indonesia, the two countries with the lowest mean body mass index (24). There are two possible explanations for these observed differences.

First, the magnitude of cardiovascular risk associated with each 1 kg m−2 increase in body mass index may be different between Asians and Pacific Islanders. We applied the same hazard ratios to all populations because the Asia Pacific Cohort Studies Collaboration reported no significant difference between Asian and Australasian (Australian and New Zealand) studies (5). However, the Asian studies included in the analysis were confined to East Asia. Cardiovascular risk may be significantly different for South Asians (e.g. Indians) and Pacific Islanders, or even between East Asians and South Asians.

Second, the continuous association between cardiovascular disease and body mass index may not begin at 21 kg m−2 for all populations. A systematic review on the coronary heart disease and body mass index relationship reported the body mass index with lowest coronary heart disease risk ranged from <20 kg m−2 to ≤24 kg m−2 and one study even reported <27 kg m−2(25). The Prospective Studies Collaboration, with 92% of participants from Europe, Israel, the USA and Australia, reported no evidence of a positive association between body mass index and stroke (total and by subtype) for body mass index <25 kg m−2(26). The universally adopted definition of overweight, which was developed from Caucasian studies, has certainly been challenged over the last few years. Based on different body fat percentage in Asians and Pacific Islanders compared with Caucasians of the same body mass index, suggestions have been made to change the definition of overweight to ≥23 kg m−2 for Asians and ≥26 kg m−2 for Pacific Islanders (27). Results from dual-energy X-ray absorptiometry suggested that Indian men with body mass index of 24 kg m−2, Indian women with body mass index of 26 kg m−2, European men and women with body mass index of 30 kg m−2, Pacific Island men with body mass index of 34 kg m−2 and Pacific Island women with 35 kg m−2, all had the same body fat percentage (28). Furthermore, the Obesity in Asia Collaboration reported that the prevalence of diabetes, hypertension and dyslipidaemia were consistently higher in Asians than Caucasians at any given level of body size (29,30). If the theoretical mean body mass index was shifted to 19 kg m−2 for India and Indonesia, the population attributable fractions for these two countries will be increased by 93% to 352%. Conversely, if the theoretical mean body mass index was shifted to 23 kg m−2 for Pacific Island countries, the population attributable fractions for these countries will be decreased by 9% to 36%.

The validity of these two explanations will require investigation. To the best of our knowledge, there has been no study, with longitudinal data, on the relationship between cardiovascular disease and body mass index in Indian or Pacific Island populations.

Because of the weak continuous association between haemorrhagic stroke and body mass index, we did not estimate the burden of haemorrhagic stroke associated with body mass index in the Asia–Pacific region. Contrary to Caucasian populations, however, haemorrhagic stroke has been more common than ischaemic stroke in East Asian populations (31). With lifestyle changes in favour of high-fat food and physical inactivity occurring in these populations, the incidence of ischaemic stroke is on the rise (32). Nevertheless, the burden of haemorrhagic stroke will remain substantial because high blood pressure is more strongly associated with haemorrhagic than ischaemic stroke and that it is also associated with body mass index.

A large burden of cardiovascular disease is associated with increased body mass index in the Asia–Pacific region. Assuming that cardiovascular risk increases from a body mass index of 21 kg m−2 across all populations, we estimated cardiovascular disease burden by shifting body weight to an ideal distribution with mean body mass index of 21 kg m−2. The results indicate that most countries in the region could reduce their cardiovascular disease burden by reducing the population mean body mass index towards 21 kg m−2. Meeting this target will be a challenge for the region over the coming decades. Rapid nutrition transition, rising childhood obesity rates and prevalent abdominal obesity even in Asians with low body mass index are examples of the obesity problem in this region. In addition, considering the large body sizes of some ethnic groups, a reduction to this target would be unrealistic. Nevertheless, any reduction towards this target would decrease the population burden of cardiovascular disease attributable to excess body weight.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References

C.M.Y.L. is supported by the National Health and Medical Research Council Training Fellowship. M.W. is supported by National Health and Medical Research Council programme grant 571281.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conflict of Interest Statement
  8. Acknowledgements
  9. References
  • 1
    Secretariat of the Pacific Community. Obesity in the Pacific. Too Big to Ignore. Secretariat of the Pacific Community: Noumea, 2002.
  • 2
    Aekplakorn W, Mo-suwan L. Prevalence of obesity in Thailand. Obes Rev 2009; 10: 589592.
  • 3
    Park HS, Park CY, Oh S, Yoo HJ. Prevalence of obesity and metabolic syndrome in Korean adults. Obes Rev 2008; 9: 104107.
  • 4
    Asia Pacific Cohort Studies Collaboration. The burden of overweight and obesity in the Asia-Pacific region. Obes Rev 2007; 8: 191196.
  • 5
    Asia Pacific Cohort Studies Collaboration. Body mass index and cardiovascular disease in the Asia-Pacific Region: an overview of 33 cohorts involving 310 000 participants. Int J Epidemiol 2004; 33: 751758.
  • 6
    Ezzati M, Lopez AD, Rodgers A, Murray CJL (eds). Comparative Quantification of Health Risks. World Health Organization: Geneva, 2004.
  • 7
    World Health Organization. WHO Global InfoBase. [WWW document]. URL https://apps.who.int/infobase/report.aspx (accessed 25 May 2010).
  • 8
    Woodward M, Martiniuk A, Lee CMY, Lam TH, Vanderhoorn S, Ueshima H, Fang X, Kim HC, Rodgers A, Patel A, Jamrozik K, Huxley R, For the Asia Pacific Cohort Studies Collaboration. Elevated total cholesterol: its prevalence and population attributable fraction for mortality from coronary heart disease and ischaemic stroke in the Asia-Pacific region. Eur J Cardiovasc Prev Rehabil 2008; 15: 397401.
  • 9
    World Health Organization. The Global Burden of Disease 2004 Update. World Health Organization: Geneva, 2008.
  • 10
    World Health Organization Western Pacific Region. American Samoa NCD Risk Factors STEPS Report. World Health Organization: Suva, 2007.
  • 11
    Cameron AJ, Welborn TA, Zimmet PZ, Dunstan DW, Owen N, Salmon J, Dalton M, Jolley D, Shaw JE. Overweight and obesity in Australia: the 1999–2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Med J Aust 2003; 178: 427432.
  • 12
    Yang W, Lu J, Weng J, Jia W, Ji L, Xiao J, Shan Z, Liu J, Tian H, Ji Q, Zhu D, Ge J, Lin L, Chen L, Guo X, Zhao Z, Li Q, Zhou Z, Shan G, He J, for the China National Diabetes and Metabolic Disorders Study Group. Prevalence of diabetes among men and women in China. N Engl J Med 2010; 362: 10901101.
  • 13
    Ministry of Health, World Health Organization, the Fiji School of Medicine, the Menzies Center for Population Health Research of the University of Tasmania. Fiji non-communicable diseases (NCD) STEPS survey 2002. [WWW document]. URL http://www.who.int/chp/steps/FijiSTEPSReport.pdf (accessed 25 May 2010).
  • 14
    Soemantri S, Pradono J, Hapsari D. SURKESNAS 2001. National Household Health Survey Morbidity Study. NCD risk factors in Indonesia. [WWW document]. URL http://www.who.int/chp/steps/STEPS_Report_Indonesia_National_2001.pdf (accessed 25 May 2010).
  • 15
    World Health Organization Western Pacific Region. Kiribati NCD Risk Factors STEPS Report. World Health Organization: Suva, 2009.
  • 16
    Disease Control Division (NCD). Malaysia NCD Surveillance 2006: NCD Risk Factors in Malaysia. Ministry of Health: Putrajaya, Malaysia, 2006.
  • 17
    World Health Organization Western Pacific Region. Mongolian STEPS Survey on the Prevalence of Noncommunicable Disease Risk Factors 2006. World Health Organization: Ulaanbaatar, 2007.
  • 18
    World Health Organization Western Pacific Region. Nauru NCD Risk Factors STEPS Report. World Health Organization: Suva, 2007.
  • 19
    Ministry of Health. A Portrait of Health. Key Results of the 2006/07 New Zealand Health Survey. Ministry of Health: Wellington, 2008.
  • 20
    Park SH, Choi SJ, Lee KS, Park HY. Waist circumference and waist-to-height ratio as predictors of cardiovascular disease risk in Korean adults. Circ J 2009; 73: 16431650.
  • 21
    Aekplakorn W, Hogan MC, Chongsuvivatwong V, Tatsanavivat P, Chariyalertsak S, Boonthum A, Tiptaradol S, Lim SS. Trends in obesity and associations with education and urban or rural residence in Thailand. Obesity 2007; 15: 31133121.
  • 22
    World Health Organization Western Pacific Region. Tokelau NCD Risk Factors STEPS Report. World Health Organization: Suva, 2007.
  • 23
    World Health Organization. Comparative quantification of health risks for the year 2000. [WWW document]. URL http://www.who.int/healthinfo/global_burden_disease/risk_factors_2000/en/index.html (accessed 25 May 2010).
  • 24
    World Health Organization. Disease and injury country estimates. Death and DALY estimates for 2004 by cause for WHO Member States. [WWW document]. URL http://www.who.int/healthinfo/global_burden_disease/gbddeathdalycountryestimates2004.xls (accessed 25 May 2010).
  • 25
    Whitlock G, Lewington S, Ni Mhurchu C. Coronary heart disease and body mass index: a systematic review of the evidence from larger prospective cohort studies. Semin Vasc Med 2002; 2: 369381.
  • 26
    Prospective Studies Collaboration. Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet 2009; 373: 10831096.
  • 27
    World Health Organization Western Pacific Region. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Health Communications Australia: Sydney, 2000.
  • 28
    Rush EC, Freitas I, Plank LD. Body size, body composition and fat distribution: comparative analysis of European, Maori, Pacific Island and Asian Indian adults. Br J Nutr 2009; 102: 632641.
  • 29
    Huxley R, James WPT, Barzi F, Patel JV, Lear SA, Suriyawongpaisal P, Janus E, Caterson I, Zimmet P, Prabhakaran D, Reddy S, Woodward M, Obesity in Asia Collaboration. Ethnic comparisons of the cross-sectional relationships between measures of body size with diabetes and hypertension. Obes Rev 2008; 9 (Suppl.1): 5361.
  • 30
    Barzi F, Woodard M, Czernichow S, Lee CMY, Kang JH, Janus E, Lear S, Patel A, Caterson I, Patel J, Lam TH, Suriyawongpaisal P, Huxley R. The discrimination of dyslipidaemia using anthropometric measures in ethnically diverse populations of the Asia-Pacific Region: the obesity in Asia Collaboration. Obes Rev 2010; 11: 127136.
  • 31
    Liu M, Wu B, Wang WZ, Lee LM, Zhang SH, Kong LZ. Stroke in China: epidemiology, prevention, and management strategies. Lancet Neurol 2007; 6: 456464.
  • 32
    Hankey GJ, Wong KSL, Chankrachang S, Chen C, Crimmins D, Frayne J, Kim JS, Li Y, Liou CW, Merican JS, Misbach J, Navarro J, Shinohara Y, Wang Y, Yoon BW. Management of cholesterol to reduce the burden of stroke in Asia: consensus statement. Int J Stroke 2010; 5: 209216.