**2. Ethnic differences in lipid profiles in general populations**

The prevalence of dyslipidaemia varies depending on the population studied, geographic location, socioeconomic development and the definition used (Wood et al. 1972; Mann et al. 1988; Onat et al. 1992; Berrios et al. 1997; Ezenwaka et al. 2000; Foucan et al. 2000; Hanh et al. 2001; Zaman et al. 2001; Azizi et al. 2003; Florez et al. 2005; Li et al. 2005; Hertz et al. 2006; Pang et al. 2006; Pongchaiyakul et al. 2006; Tekes-Manova et al. 2006; Zhao et al. 2007; Erem et al. 2008; Steinhagen-Thiessen et al. 2008). Caucasians generally have higher mean TC concentrations than do populations of Asian or African origin (Fuentes et al. 2003; Tolonen et al. 2005). In general populations, the highest prevalence of hypercholesterolaemia (TC ≥ 6.5mmol/l) has been seen in Malta (up to 50% in women) and the lowest in China (2.7% in men) in the World Health Orgnization (WHO) Inter-Health Programme (Berrios et al. 1997). However, inhabitants of the developing world now have had access to more fats in their diets and more sedentary lives; therefore the disease is becoming an increasing problem there.

Ethnic differences in the risk of CVD and type 2 diabetes have consistently been identified, with the most studies comparing the risk between African-Americans and Whites. African-Americans usually display a more favorable lipid profile compared with Whites, despite having the highest overall mortality rates from CVD. In general, African-American men have similar or lower LDL-C and TG but higher HDL-C levels compared with White men. There is evidence that the difference in HDL-C between African-American and White men may be due to a relatively lower hepatic lipase activity in African-Americans (Vega GL 1998). The difference in TG may be related to increased activity of lipoprotein lipase in African-Americans (Sumner AE 2005). However, compared with Whites, Hispanics and Asians, African-Americans have less favorable levels of lipoprotein(a) (Lp[a]), which is structurally similar to LDL-C, with an additional disulfide linked glycoprotein termed ApoA. A number of studies have suggested that Lp(a) may be an important risk factor for CVD (Danesh J 2000; The Emerging Risk Factors C 2009).

Compared to non-Hispanic Whites, Hispanics, specifically Mexican-Americans, have demonstrated lower HDL-C and higher TG levels (Sundquist J 1999). Data from the Dallas Heart Study and a smaller cross-sectional analysis of healthy individuals confirm that levels of Lp(a) are likely similar or even lower in Hispanics compared with Whites (Tsimikas S 2009). Although Lp(a) levels have been associated with endothelial dysfunction in Hispanics, the relationship with coronary artery disease in this population is less clear.

Asian Indians exhibit a higher prevalence of diabetes mellitus than Chinese and Malays (Tan et al. 1999). They also have higher serum TG concentrations and lower HDL-C concentrations than Chinese (Gupta M 2006). In the HeartSCORE and IndiaSCORE studies (Mulukutla et al. 2008) where lipids were measured with the same assay procedures for Asian Indians as for Whites and Blacks, Asian Indians had lowest TC and HDL-C and highest TG among all the ethnic groups studied. In another multi-ethnic study of the 1992 Singapore National Health Survey (Tan et al. 1999), Asian Indians appeared to have lower HDL-C but higher TG levels compared with the Chinese group. Data in other racial/ethnic groups are somewhat limited. Mean total cholesterol and LDL-C levels are lower in American Indians compared with the US average, and levels of Lp(a) are reported to be lower than in Whites (Wang W 2002). East Asians tend to have lower LDL-C, HDL-C and TG as compared with non-Asians (Karthikeyan et al. 2009). East Asians have been reported to have low Lp(a) levels, whereas south Asians have higher mean Lp(a) levels (Geethanjali FS 2003; Berglund L 2004).

The prevalence of dyslipidaemia varies depending on the population studied, geographic location, socioeconomic development and the definition used (Wood et al. 1972; Mann et al. 1988; Onat et al. 1992; Berrios et al. 1997; Ezenwaka et al. 2000; Foucan et al. 2000; Hanh et al. 2001; Zaman et al. 2001; Azizi et al. 2003; Florez et al. 2005; Li et al. 2005; Hertz et al. 2006; Pang et al. 2006; Pongchaiyakul et al. 2006; Tekes-Manova et al. 2006; Zhao et al. 2007; Erem et al. 2008; Steinhagen-Thiessen et al. 2008). Caucasians generally have higher mean TC concentrations than do populations of Asian or African origin (Fuentes et al. 2003; Tolonen et al. 2005). In general populations, the highest prevalence of hypercholesterolaemia (TC ≥ 6.5mmol/l) has been seen in Malta (up to 50% in women) and the lowest in China (2.7% in men) in the World Health Orgnization (WHO) Inter-Health Programme (Berrios et al. 1997). However, inhabitants of the developing world now have had access to more fats in their diets and more sedentary lives; therefore the disease is becoming an increasing problem there. Ethnic differences in the risk of CVD and type 2 diabetes have consistently been identified, with the most studies comparing the risk between African-Americans and Whites. African-Americans usually display a more favorable lipid profile compared with Whites, despite having the highest overall mortality rates from CVD. In general, African-American men have similar or lower LDL-C and TG but higher HDL-C levels compared with White men. There is evidence that the difference in HDL-C between African-American and White men may be due to a relatively lower hepatic lipase activity in African-Americans (Vega GL 1998). The difference in TG may be related to increased activity of lipoprotein lipase in African-Americans (Sumner AE 2005). However, compared with Whites, Hispanics and Asians, African-Americans have less favorable levels of lipoprotein(a) (Lp[a]), which is structurally similar to LDL-C, with an additional disulfide linked glycoprotein termed ApoA. A number of studies have suggested that Lp(a) may be an important risk factor for

Compared to non-Hispanic Whites, Hispanics, specifically Mexican-Americans, have demonstrated lower HDL-C and higher TG levels (Sundquist J 1999). Data from the Dallas Heart Study and a smaller cross-sectional analysis of healthy individuals confirm that levels of Lp(a) are likely similar or even lower in Hispanics compared with Whites (Tsimikas S 2009). Although Lp(a) levels have been associated with endothelial dysfunction in Hispanics, the relationship with coronary artery disease in this population is less clear. Asian Indians exhibit a higher prevalence of diabetes mellitus than Chinese and Malays (Tan et al. 1999). They also have higher serum TG concentrations and lower HDL-C concentrations than Chinese (Gupta M 2006). In the HeartSCORE and IndiaSCORE studies (Mulukutla et al. 2008) where lipids were measured with the same assay procedures for Asian Indians as for Whites and Blacks, Asian Indians had lowest TC and HDL-C and highest TG among all the ethnic groups studied. In another multi-ethnic study of the 1992 Singapore National Health Survey (Tan et al. 1999), Asian Indians appeared to have lower HDL-C but higher TG levels compared with the Chinese group. Data in other racial/ethnic groups are somewhat limited. Mean total cholesterol and LDL-C levels are lower in American Indians compared with the US average, and levels of Lp(a) are reported to be lower than in Whites (Wang W 2002). East Asians tend to have lower LDL-C, HDL-C and TG as compared with non-Asians (Karthikeyan et al. 2009). East Asians have been reported to have low Lp(a) levels, whereas south Asians have higher mean Lp(a) levels (Geethanjali

**2. Ethnic differences in lipid profiles in general populations** 

CVD (Danesh J 2000; The Emerging Risk Factors C 2009).

FS 2003; Berglund L 2004).

Globalization of the western lifestyle contribute to worldwide increases of adiposity and type 2 diabetes not only in adults but also in children and adolescents (Kelishadi et al. 2006; Schwandt et al. 2010). In the BIG Study comparing the prevalence of the metabolic syndrome components in children and adolescents of European, Asian and South-American ethnicities, Iranian and Brazilian youths had considerably higher prevalence of dyslipidaemia than German youths. The most remarkable ethnic difference detected in this study is the high prevalence of low HDL-C levels in Iranian children and adolescents (38%) compared with German youths (7%) (Schwandt et al. 2010). Future longitudinal studies should seek the clinical importance of these ethnic differences.

## **3. Ethnic differences in lipid profiles in the state of hyperglycaemia**

#### **3.1 Lipid disorder and CVD risk in individuals with hyperglycaemia**

Lipids and lipoproteins abnormalities are major metabolic disorders, commonly including elevated levels of TC, LDL-C, Lp(a) and TG and reduced levels of HDL-C. In patients with type 2 diabetes, a CHD equivalent (Juutilainen et al. 2005), it is most commonly characterized by elevated TG and reduced HDL-C (Goldberg, I. J. 2001; Krauss 2004; Kendall 2005). There is increasing evidence that the diabetic dyslipidaemia pattern is common not only in patients with overt diabetes (Barrett-Connor et al. 1982) but also in individuals with different glucose categories, i.e., impaired glucose tolerance (IGT) or impaired fasting glucose (IFG) (Meigs et al. 2002; Novoa et al. 2005; Chen et al. 2006; Pankow et al. 2007). These abnormalities can be present alone or in combination with other metabolic disorders. It is well known that the risk of morbidity and mortality from CVD is increased by two- to four-fold in diabetic patients compared with the general population (Kannel 1985; Morrish et al. 1991; Almdal et al. 2004). A number of studies have determined the association of dyslipidaemia with cardiovascular risk in people with hyperglycaemia, and most of them were conducted in patients with diabetes. There is a large body of evidence linking dyslipidaemia and cardiovascular risk in patients with diabetes against quite few negative reports (Vlajinac et al. 1992; Roselli della Rovere et al. 2003) on this issue. Cross-sectional studies have found positive associations of atherosclerotic vascular disease with TC (Ronnemaa et al. 1989; Jurado et al. 2009), LDL-C (Reckless et al. 1978; Agarwal et al. 2009; Jurado et al. 2009), non-HDL-C (Jurado et al. 2009), TG (Santen et al. 1972; Ronnemaa et al. 1989; Gomes et al. 2009), apoB (Ronnemaa et al. 1989) and Lp(a) (Mohan et al. 1998; Murakami et al. 1998; Smaoui et al. 2004), but inverse associations with HDL-C (Reckless et al. 1978; Ronnemaa et al. 1989; Smaoui et al. 2004; Grant and Meigs 2007; Gomes et al. 2009; Jurado et al. 2009) and apoA-I (Seviour et al. 1988; Ronnemaa et al. 1989).

Prospective data have provided with further evidence. The UKPDS study (Turner et al. 1998) has demonstrated that high LDL-C and low HDL-C are potentially modifiable risk factors for coronary artery disease (CAD) in patients with type 2 diabetes. TG, however, was not independently associated with CAD risk in this study, possibly because of its close inverse relationship with HDL-C. Results from the MRFIT (Stamler et al. 1993), in which 356,499 nondiabetic and 5163 diabetic men without CHD at baseline were followed for 12 years, indicated that serum cholesterol is an independent predictor of CHD mortality in men with diabetes. Rosengren et al. (Rosengren et al. 1989) showed similar results in a prospective study of 6897 middle aged diabetic men. Patients with TC > 7.3 mmol/l had a significantly higher incidence of CHD during the 7-year follow up than those with TC ≤ 5.5 mmol/l (28.3% vs. 5.4%, p<0.05). Long term follow-up of the London cohort of the WHO

Ethnic Difference in Lipid Profiles 19

Similar to others (Harris and Eastman 2000; Hadaegh et al. 2008), we observed a worse lipid profile in individuals with undiagnosed diabetes than that of previously diagnosed patients in most of the ethnic groups, indicating individuals with undiagnosed diabetes are at increased CVD risk and need to be identified and treated early. On the other hand, glycaemic control is shown to be an important determinant of diabetic dyslipidaemia (Ismail et al. 2001). The better lipid profile in diagnosed diabetes as compared with undiagnosed diabetes might imply a benefit of lifestyle intervention or drug treatment targeting favorable metabolic profiles and hemoglobin A1c (HbA1c), a surrogate measure for average blood glucose. However, to what extent the levels of HbA1c have contributed to the differences is unknown due to the lack of information in the current study. In addition, the data on lipidlowering treatment is not available for most of the earlier studies conducted in the 1990s because the statins were not widely prescribed at that time. These deserve further

In contrast to the lower HDL-C and higher TG profiles, Asian Indians had considerably lower TC and LDL-C concentrations than others. As shown in Table 2, 71% non-diabetic and 57.6% diabetic Asian Indians had low LDL-C (< 3.0 mmol/l), while the corresponding figures were 19.2% and 24.6% (p < 0.01) for C&N Europeans and 46.6% and 38.8% (p < 0.01) for Qingdao Chinese. However, even within the low LDL-C category, there was still a higher proportion of Asian Indians having low HDL-C compared with others (Table 2). The results were confirmed in the same analysis conducted separately for men and women. There is a large body of evidence showing that diabetes is associated with a high prevalence of dyslipidaemia (Kannel 1985; Cowie et al. 1994; 1997; Jacobs et al. 2005; Bruckert et al. 2007; Abdel-Aal et al. 2008; Ahmed et al. 2008; Okafor et al. 2008; Surana et al. 2008; Agarwal et al. 2009; Jurado et al. 2009; Papazafiropoulou et al. 2009; Roberto Robles et al. 2009; Temelkova-Kurktschiev et al. 2009; Zhang et al. 2009; Seyum et al. 2010). In the Framingham Heart Study (Kannel 1985), the prevalence of low HDL-C (21% vs. 12% in men and 25% vs. 10% in women, respectively) and high TG levels (19% vs. 9% in men and 17% vs. 8% in women, respectively) in people with diabetes was almost twice as high as the prevalence in non-diabetic individuals. By contrast, TC and LDL-C levels did not differ from those of nondiabetic counterparts. A similar pattern of lipid profiles was observed in the UK Prospective Diabetes Study (UKPDS) (1997). In this study, the plasma TG levels were substantially increased whereas HDL-C levels were markedly reduced in both men and women with diabetes compared with the non-diabetic controls. Higher prevalence has been reported in other studies. Data from a primary care-based 7692 patients with type 2 diabetes in the United States showed nearly half of the patients had low HDL-C (Grant and Meigs 2007). The figure was even worse in an urban Indian cohort of 5088 type 2 diabetes patients, with more than half having low HDL-C (52.3%) or high TG (57.9%) (Surana et al. 2008). In addition to the traditional lipid measurement, increased levels of apoB were also seen in patients with diabetes compared with non-diabetic individuals (Bangou-Bredent et al. 1999). It has been shown that the prevalence of lipid and/or glucose abnormality differs between ethnic groups. It is clear that certain ethnic groups have differences in lipid profiles in general. Elevated TG and reduced HDL-C, as the components of the metabolic syndrome and atherogenic dyslipidaemia, was seen more common in Asian Indians than in the Whites (Anand et al. 2000; Razak et al. 2005; Chandalia et al. 2008; Mulukutla et al. 2008), Chinese (Tan et al. 1999; Anand et al. 2000; Razak et al. 2005; The DECODA Study Group 2007; Karthikeyan et al. 2009), Japanese (The DECODA Study Group 2007; Karthikeyan et al. 2009) or Africans (Mulukutla et al. 2008). In a nationally representative sample of seven

investigation in future studies.

Multinational Study of Vascular Disease in Diabetics, consisting of 254 type 2 diabetic patients, has showed that TC was associated with incidence of MI (Morrish et al. 1991) and overall cardiovascular mortality (Morrish et al. 1990). The role of TC in predicting CHD was also confirmed in women patients with diabetes (Schulze et al. 2004).

### **3.2 Ethnic difference in lipid profiles across glucose categories**

Although the ethnic variation in lipid patterns has been wided studied in general populations, the ethnic differences in lipid profiles given the same glucose levels have not been well investigated. This issue has been recently studied in the DECODE (Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe) and DECODA (Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Asia) study, which consisted of 64 cohorts of mainly population-based from 24 countries and regions around the world, with about 84 000 Europeans and 84 207 Asians of Chinese, Japanese, Indians, Mongolians and Filipinos.

In the collaborative analysis of seven ethnic groups of European and Asian populations (studies included see Appendix 1), considerable ethnic differences in lipid profiles were observed within each glucose category. Asian Indians exhibited an adverse lipid pattern consisting of low HDL-C and high TG across all glucose categories as compared with other ethnic groups. Reduced HDL-C is prevalent even in Asian Indians with desirable LDL-C levels regardless of the diabetic status. In addition, in most of the ethnic groups, individuals detected with undiagnosed diabetes had a worse lipid profile than did diagnosed cases. Age-, cohort- and BMI adjusted mean TC, LDL-C and TG increased while the mean HDL-C decreased with more pronounced glucose intolerance in most of the ethnic groups in individuals without a prior history of diabetes (Fig. 1 a-h). Subjects with undiagnosed diabetes, however, had a worse lipid profile than those with known disease. Within individuals with normoglycaemia, mean lipid and lipoprotein concentrations differed among the ethnic groups. The Europeans had highest TC (Fig. 1 a-b) and LDL-C (Fig. 1 c-d), while Qingdao Chinese had highest HDL-C levels among all ethnic groups (Fig. 1 e-f). In contrast, Asian Indians had the lowest TC (Fig. 1 a-b), LDL-C (Fig. 1 c-d) and HDL-C (Fig. 1 e-f) but the highest TG (Fig. 1 g-h) among the ethnic groups (p <0.05 for all comparisons). These ethnic differences were consistently found in all glucose categories.

The multivariate-adjusted odds ratio (95% CI) of having low HDL-C was significantly higher for Asian Indians, Mauritian Indians, Hong Kong Chinese and Southern Europeans but lower for Qingdao Chinese compared with Central & North (C&N) Europeans, across all glucose categories from normal to diabetes (Table 1). Asian Indians and Mauritian Indians tended to have higher but Southern Europeans lower odds ratios for having high-TG compared with the reference group. Unlike that for HDL-C or TG, the odds ratio for having high LDL-C was consistently lower in all Asian ethnic groups compared with the reference, across most of the glucose categories.

In the HeartSCORE and IndiaSCORE studies (Mulukutla et al. 2008) where lipids were measured with the same assay procedures for Asian Indians as for whites and blacks, Asian Indians had lowest TC and HDL-C and highest TG among all the ethnic groups studied. In another multi-ethnic study of the 1992 Singapore National Health Survey (Tan et al. 1999), Asian Indians appeared to have lower HDL-C but higher TG levels compared with Chinese. The findings of these previous studies are consistent with ours although glucose status was not controlled in the previous studies.

Multinational Study of Vascular Disease in Diabetics, consisting of 254 type 2 diabetic patients, has showed that TC was associated with incidence of MI (Morrish et al. 1991) and overall cardiovascular mortality (Morrish et al. 1990). The role of TC in predicting CHD was

Although the ethnic variation in lipid patterns has been wided studied in general populations, the ethnic differences in lipid profiles given the same glucose levels have not been well investigated. This issue has been recently studied in the DECODE (Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Europe) and DECODA (Diabetes Epidemiology: Collaborative analysis Of Diagnostic criteria in Asia) study, which consisted of 64 cohorts of mainly population-based from 24 countries and regions around the world, with about 84 000 Europeans and 84 207 Asians of Chinese, Japanese, Indians,

In the collaborative analysis of seven ethnic groups of European and Asian populations (studies included see Appendix 1), considerable ethnic differences in lipid profiles were observed within each glucose category. Asian Indians exhibited an adverse lipid pattern consisting of low HDL-C and high TG across all glucose categories as compared with other ethnic groups. Reduced HDL-C is prevalent even in Asian Indians with desirable LDL-C levels regardless of the diabetic status. In addition, in most of the ethnic groups, individuals detected with undiagnosed diabetes had a worse lipid profile than did diagnosed cases. Age-, cohort- and BMI adjusted mean TC, LDL-C and TG increased while the mean HDL-C decreased with more pronounced glucose intolerance in most of the ethnic groups in individuals without a prior history of diabetes (Fig. 1 a-h). Subjects with undiagnosed diabetes, however, had a worse lipid profile than those with known disease. Within individuals with normoglycaemia, mean lipid and lipoprotein concentrations differed among the ethnic groups. The Europeans had highest TC (Fig. 1 a-b) and LDL-C (Fig. 1 c-d), while Qingdao Chinese had highest HDL-C levels among all ethnic groups (Fig. 1 e-f). In contrast, Asian Indians had the lowest TC (Fig. 1 a-b), LDL-C (Fig. 1 c-d) and HDL-C (Fig. 1 e-f) but the highest TG (Fig. 1 g-h) among the ethnic groups (p <0.05 for all comparisons).

The multivariate-adjusted odds ratio (95% CI) of having low HDL-C was significantly higher for Asian Indians, Mauritian Indians, Hong Kong Chinese and Southern Europeans but lower for Qingdao Chinese compared with Central & North (C&N) Europeans, across all glucose categories from normal to diabetes (Table 1). Asian Indians and Mauritian Indians tended to have higher but Southern Europeans lower odds ratios for having high-TG compared with the reference group. Unlike that for HDL-C or TG, the odds ratio for having high LDL-C was consistently lower in all Asian ethnic groups compared with the

In the HeartSCORE and IndiaSCORE studies (Mulukutla et al. 2008) where lipids were measured with the same assay procedures for Asian Indians as for whites and blacks, Asian Indians had lowest TC and HDL-C and highest TG among all the ethnic groups studied. In another multi-ethnic study of the 1992 Singapore National Health Survey (Tan et al. 1999), Asian Indians appeared to have lower HDL-C but higher TG levels compared with Chinese. The findings of these previous studies are consistent with ours although glucose status was

also confirmed in women patients with diabetes (Schulze et al. 2004).

**3.2 Ethnic difference in lipid profiles across glucose categories** 

These ethnic differences were consistently found in all glucose categories.

reference, across most of the glucose categories.

not controlled in the previous studies.

Mongolians and Filipinos.

Similar to others (Harris and Eastman 2000; Hadaegh et al. 2008), we observed a worse lipid profile in individuals with undiagnosed diabetes than that of previously diagnosed patients in most of the ethnic groups, indicating individuals with undiagnosed diabetes are at increased CVD risk and need to be identified and treated early. On the other hand, glycaemic control is shown to be an important determinant of diabetic dyslipidaemia (Ismail et al. 2001). The better lipid profile in diagnosed diabetes as compared with undiagnosed diabetes might imply a benefit of lifestyle intervention or drug treatment targeting favorable metabolic profiles and hemoglobin A1c (HbA1c), a surrogate measure for average blood glucose. However, to what extent the levels of HbA1c have contributed to the differences is unknown due to the lack of information in the current study. In addition, the data on lipidlowering treatment is not available for most of the earlier studies conducted in the 1990s because the statins were not widely prescribed at that time. These deserve further investigation in future studies.

In contrast to the lower HDL-C and higher TG profiles, Asian Indians had considerably lower TC and LDL-C concentrations than others. As shown in Table 2, 71% non-diabetic and 57.6% diabetic Asian Indians had low LDL-C (< 3.0 mmol/l), while the corresponding figures were 19.2% and 24.6% (p < 0.01) for C&N Europeans and 46.6% and 38.8% (p < 0.01) for Qingdao Chinese. However, even within the low LDL-C category, there was still a higher proportion of Asian Indians having low HDL-C compared with others (Table 2). The results were confirmed in the same analysis conducted separately for men and women.

There is a large body of evidence showing that diabetes is associated with a high prevalence of dyslipidaemia (Kannel 1985; Cowie et al. 1994; 1997; Jacobs et al. 2005; Bruckert et al. 2007; Abdel-Aal et al. 2008; Ahmed et al. 2008; Okafor et al. 2008; Surana et al. 2008; Agarwal et al. 2009; Jurado et al. 2009; Papazafiropoulou et al. 2009; Roberto Robles et al. 2009; Temelkova-Kurktschiev et al. 2009; Zhang et al. 2009; Seyum et al. 2010). In the Framingham Heart Study (Kannel 1985), the prevalence of low HDL-C (21% vs. 12% in men and 25% vs. 10% in women, respectively) and high TG levels (19% vs. 9% in men and 17% vs. 8% in women, respectively) in people with diabetes was almost twice as high as the prevalence in non-diabetic individuals. By contrast, TC and LDL-C levels did not differ from those of nondiabetic counterparts. A similar pattern of lipid profiles was observed in the UK Prospective Diabetes Study (UKPDS) (1997). In this study, the plasma TG levels were substantially increased whereas HDL-C levels were markedly reduced in both men and women with diabetes compared with the non-diabetic controls. Higher prevalence has been reported in other studies. Data from a primary care-based 7692 patients with type 2 diabetes in the United States showed nearly half of the patients had low HDL-C (Grant and Meigs 2007). The figure was even worse in an urban Indian cohort of 5088 type 2 diabetes patients, with more than half having low HDL-C (52.3%) or high TG (57.9%) (Surana et al. 2008). In addition to the traditional lipid measurement, increased levels of apoB were also seen in patients with diabetes compared with non-diabetic individuals (Bangou-Bredent et al. 1999). It has been shown that the prevalence of lipid and/or glucose abnormality differs between ethnic groups. It is clear that certain ethnic groups have differences in lipid profiles in general. Elevated TG and reduced HDL-C, as the components of the metabolic syndrome and atherogenic dyslipidaemia, was seen more common in Asian Indians than in the Whites (Anand et al. 2000; Razak et al. 2005; Chandalia et al. 2008; Mulukutla et al. 2008), Chinese (Tan et al. 1999; Anand et al. 2000; Razak et al. 2005; The DECODA Study Group 2007; Karthikeyan et al. 2009), Japanese (The DECODA Study Group 2007; Karthikeyan et al. 2009) or Africans (Mulukutla et al. 2008). In a nationally representative sample of seven

Ethnic Difference in Lipid Profiles 21

(Kuller 2004) and diabetes (Ramachandran et al. 2008; Snehalatha and Ramachandran 2009) in this ethnic group. In addition, it may also reflect the genetic variation, for example, at the apoE locus (Tan et al. 2003) and an excess of other risk factors such as homocysteine, Lp(a)

As suggested by previous research, dietary factors may play a role in both lipid and insulin profiles, although these patterns may be mediated by body fat content (Ku CY 1998). Total fat (and saturated fat) intake has been shown to adversely affect total cholesterol concentrations in children, adolescents, and young adults (Post GB 1997). The difference in HDL-C concentrations between Qingdao and Hong Kong Chinese subgroups observed in the DECODA study cannot be simply explained by the difference in assay methods. It may largely attribute to the differences in dietary structure and preference, geographic and environmental factors. Shellfish and beer, for example, are commonly consumed all the year round in Qingdao. Nevertheless, whether other factors exist and contribute to the high

Mexican Americans have been previously reported to have greater adiposity, higher TG levels and lower HDL-C levels than Anglos. The relationship between behavioral variables (caloric balance, cigarette and alcohol consumption, exercise, post-menopausal estrogen or oral contraceptive use) and lipid pattern has been investigated in the San Antonio Heart Study (1979–1982) (n=2,102) to explain the ethnic difference in lipids and lipoproteins. Adjustment for caloric balance (as reflected by body mass index) narrowed the ethnic difference in TG and HDL-C levels for both sexes, while adjustment for smoking widened the ethnic difference. For females, the ethnic difference was also decreased by adjustment for alcohol and estrogen use. However, adjustment for these behavioral variables did not completely eliminate the ethnic difference in lipids and lipoproteins in either sex. Increased central adiposity, more characteristic of Mexican Americans than Anglos, was positively associated with triglycerides and negatively associated with HDL-C levels, especially in females. Fat patterning made a more important contribution to the prediction of TG and HDL-C levels than did the other behavioral variables (except for caloric balance) and, in general, eliminated ethnic differences in lipids and lipoproteins (Steven H 1986). Epidemiologists should consider the use of a centrality index to distinguish different types

or dietary fat (France et al. 2003).

HDL-C in Qingdao needs to be further investigated.

of adiposity since it is easy and inexpensive to measure.

**5. Implications for management and prevention of dyslipidaemia** 

Epidemiological investigations of human populations have revealed a robust relationship between lipids and CVD risk. Furthermore, the benefit of lipid-modifying strategy on cardiovascular events has been demonstrated from a large number of randomized clinical trials (Thavendiranathan et al. 2006; Mills et al. 2008), especially from those using 3 hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) reductase inhibitors (i.e., statins) (Goldberg, R. B. et al. 1998; Collins et al. 2003; Colhoun et al. 2004; Pyorala et al. 2004; Sever et al. 2005; Knopp et al. 2006; Shepherd et al. 2006). Intensive control of dyslipidaemia has been greatly emphasized in the prevention and management of CVD. Current guidelines from the National Cholesterol Education Program Adult Treatment Panel III (ATP III) (Adult Treatment Panel III 2002), the European Society of Cardiology (Graham et al. 2007) and the American Diabetes Association (American Diabetes Association 2009) consistently

**4.2 Environmental factors** 

ethnic groups in the UK (Zaninotto et al. 2007), the prevalence of low HDL-C was highest in south Asian groups such as Bangladeshi, Indian and Pakistani, followed by Chinese, Irish and those from the general population living in private households; In contrast, the lowest prevalence was seen in Black Caribbean. Similar finding was reported in another study where the comparison was made between non-South-Asians and South Asians (France et al. 2003). In addition, African Americans have been reported to have less adverse lipid profiles than Whites or Hispanics despite the presence of diabetes (Werk et al. 1993; Cowie et al. 1994; Sharma and Pavlik 2001). The causes of ethnic difference in levels of CVD risk factor are complex and may include genetic, environmental and cultural factors (Zaninotto et al. 2007). However, little is known about such ethnic differences in lipid profiles at comparable glucose tolerance status.
