Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Feb 1.
Published in final edited form as: Metabolism. 2009 Feb;58(2):212–219. doi: 10.1016/j.metabol.2008.09.016

Combined Hyperlipidemia in relation to Race/Ethnicity, Obesity, and Insulin Resistance in the Multi-Ethnic Study of Atherosclerosis (MESA)

Pathmaja Paramsothy 1, Robert Knopp 2, Alain G Bertoni 3, Michael Y Tsai 4, Tessa Rue 5, Susan R Heckbert 6
PMCID: PMC2677914  NIHMSID: NIHMS93067  PMID: 19154954

Abstract

Background

We have asked whether the prevalence of combined hyperlipidemia (CHL) differs by race/ethnicity, obesity, and insulin resistance in a contemporary, multi-ethnic, US cohort.

Methods & Results

We determined the prevalence and adjusted odds of CHL in a cohort of 5,923 men and women free of clinically-recognized cardiovascular disease and diabetes, according to race/ethnicity (White, Chinese, African-American, and Hispanic), obesity, and insulin resistance. Untreated lipid values were imputed for those on lipid lowering therapy. CHL was defined using age and gender-specific ≥75th percentile cut points for LDL-C and triglycerides obtained from a predominantly Caucasian North American population study. Compared to Whites, adjusted odds ratios (OR) for CHL were 0.48 in African Americans (95% confidence interval (CI): 0.30, 0.75), 1.33 in Hispanics (95% CI 0.93, 1.91), and 1.06 in Asians (95% CI 0.62, 1.82). Within the entire population, the adjusted odds of CHL was over 2-fold higher in overweight and obese participants compared with normal weight and more than 4-fold higher in quartiles 2 through 4 of insulin resistance, compared with quartile 1.

Conclusions

African-Americans had lower odds for CHL than Whites despite higher BMI and abdominal adiposity. Hispanics had a non-significantly higher trend and Asians had no significantly different odds than Whites. Modest increases in weight and insulin resistance were associated with significantly higher odds of CHL in a multi-ethnic US population. Further research is needed to determine the most efficacious diet, exercise and drug management to decrease the risk of CHL and CHD among racial/ethnic groups in the United States.

Introduction

Reference values for the definition of lipid abnormalities in the increasingly multi-ethnic US population are available from predominantly Caucasian (White) North American population studies. (1) Despite the availability of these data, there has been no study of the prevalence of lipid abnormalities by racial-ethnic groupings in the USA. Effects of race and ethnicity on cardiovascular disease (CVD) risk factors have been conducted in two other multi-ethnic nations, Great Britain and Canada. (24) Differences in US ethnic populations are suggested by D’Agostino et al. who demonstrated that the sex-specific Framingham risk prediction score estimated coronary heart disease (CHD) risk well for Caucasians and African-Americans but not for Japanese, Hispanics, or Native Americans in the US and Puerto Rico. (5)

The strongest risk factor for first myocardial infarction worldwide is dyslipidemia (6) primarily consisting of simple hypercholesterolemia or combined hyperlipidemia. Beginning with the studies of Goldstein et al. in 1973, familial combined hyperlipidemia (FCHL) was more strongly associated with prevalent and incident CHD than simple hypercholesterolemia. (710) The diagnosis of FCHL requires elevations of cholesterol, triglycerides, or both, as well as these elevations in a first degree relative. (710) Which lipid elevation predominates depends on cofactors such as body weight, abdominal adiposity, and diet. Overproduction of apo B-containing lipoproteins (VLDL, IDL, and LDL) and small dense LDL particles are also consistent features of FCHL. (11)

Family history of combined lipid elevations is often unavailable, in which case the condition is termed combined hyperlipidemia (CHL), and not FCHL. CHL is usually diagnosed based simply on LDL and triglyceride elevations irrespective of family history. For the purpose of this investigation we used a clinical trials definition of CHL: ≥75th percentile for both LDL cholesterol and triglycerides for age and gender based on a predominantly Caucasian population reference. (1,12,13) As an example, the 75th percentile for LDL in Caucasian men over 45 ranges from 163 to 170mg/dL and for triglyceride between 150 and 178 mg/dL. These cut points are consistent with the NCEP/ATP III guidelines which suggest drug treatment for LDL cholesterol ≥160 mg/dL (2 or more risk factors but <10% 10 year risk of CVD) and defines triglycerides ≥ 150 mg/dL as abnormally elevated. (14)

A strong relationship between FCHL and insulin resistance has been reported in predominantly people of European ancestry. (1520) We studied the relationship of race/ethnicity, obesity, and insulin resistance with CHL in a contemporary, multi-ethnic, US population of men and women free of CVD and diabetes (21) to answer the questions: (1) What is the prevalence and odds of CHL by ethnic group and 2) What are the independent associations of obesity and insulin resistance with CHL in this population?

Methods

The purpose of the Multi-Ethnic Study of Atherosclerosis (MESA) study is to identify subclinical markers of cardiovascular disease and measure progression of these subclinical markers over time. (21) The complete MESA cohort consists of 6,814 men and women aged 45–84 from 6 different US communities (Baltimore County, MD; Chicago, IL; Forsyth County, NC; Los Angeles County, CA; New York, NY; and St. Paul, MN) representing 4 racial/ethnic categories including White, Chinese, African-American, and Hispanic. Appropriate human subjects IRB approval was obtained for each field center and the MESA coordinating center. Race/ethnicity was self reported and classified using standard National Institute of Health race/ethnicity classification. Subjects could belong only to one category. All members of the cohort were free of clinically recognized cardiovascular disease at baseline, 2000–2002. We excluded 13.2% (n=897) with diabetes and 0.1% (n=6) with missing, glucose, insulin, or waist circumference data.

At baseline, several cardiovascular risk factors were directly measured or assessed via questionnaire including height, weight, blood pressure, waist circumference, medical history including presence of diabetes, hypertension, dyslipidemia, family history of MI, current medication use including lipid lowering therapy, and assessment of personal habits such as tobacco and alcohol use. Diabetes was defined using the 2003 ADA criteria of a fasting glucose ≥ 126 mg/dL or taking medications for diabetes. (22) Hypertension was defined as SBP ≥140 mm Hg at baseline visit, or DBP ≥ 90 mm Hg, or by a history of physician-diagnosed hypertension. (23)

BMI was calculated as weight in kilograms divided by height in meters squared, and categorized according to the WHO criteria as normal weight (BMI <25 kg/m2), overweight (BMI 25–30 kg/m2), or obese (BMI ≥ 30 kg/m2). Waist circumference was measured using a standardized protocol. Waist circumference was categorized by 1) the metabolic syndrome classification in whites of enlarged waist ≥ 102 cm for men and ≥88 cm for women, and by 2) ethnic-specific cut points for men and women from Alberti et al. (14,24) Insulin resistance was defined as quartiles of HOMA-IR (glucose (mmol/L)*insulin (mu/mL))/22.5). (25) CHL was defined as ≥75th percentile for both LDL cholesterol and triglycerides for age and gender using the Lipid Clinics Prevalence Study data as a reference. (1)

All lipids were measured at a central location, Collaborative Studies Clinical Laboratory at Fairview-University Medical Center (Minneapolis, MN) using standardized methods and reagents. Triglycerides were measured in plasma using a glycerol blanked enzymatic method (Trig/GB, Roche Diagnostics Corporation, 115 Hague Road, Indianapolis, IN) on the Roche/Hitachi 911 Automatic Analyzer (Roche Diagnostics Corporation). Cholesterol was measured in plasma on the Hitachi 911 using a cholesterol esterase, cholesterol oxidase reaction (Chol R1, Roche Diagnostics Corporation). The same reaction was also used to measure HDL-C after precipitation of non-HDL-C with magnesium/dextran sulfate. The coefficient fo variance (CV) was 4.0% for triglyceride measurements, 1.6% for total cholesterol and 2.9% for HDL-C. LDL-C was calculated on specimens having a triglyceride of <400 mg/dL using the Friedewald formula. (26) Serum glucose was measured using the glucose oxidase method on the Vitros analyzer (Johnson & Johnson, Rochester, NY). Insulin was measured in serum by an immunoenzymatic sandwich assay using Access Ultrasensitive Insulin Reagent on the Access Immunoassay System (Beckman Instruments, Inc.). Proton NMR spectroscopy was used to determine HDL-C, LDL-C, and VLDL-C subclass size (Liposcience, Raleigh, NC).

For participants taking lipid lowering medications at baseline (n= 855), we imputed the underlying untreated levels of total, LDL, and HDL cholesterol and triglycerides, based upon their observed values under treatment and the observed changes in lipid levels associated with treatment among other MESA cohort members who started lipid-lowering therapy during cohort follow-up. A model relating untreated to treated lipid values (LDL-C, HDL-C, triglycerides, and total cholesterol) was created using a subset of participants who started taking lipid lowering medications between baseline and exam 2. This model included terms for the untreated lipid value, an indicator regarding medication type, and an interaction between these two terms to allow the relationship between treated and untreated cholesterol to vary by drug type. This model was then used to estimate untreated values for the participants who were on treatment at baseline. (27)

For 86 participants whose LDL-C was not able to be measured because triglycerides were >400 mg/dL, we substituted LDL-C levels measured using recalibrated NMR measurements of LDL-C. A calibration equation was developed for this purpose, using participants that had both NMR and regular LDL measurements and applied this to the NMR values in cases of missing LDL. The NMR measurements were highly correlated to the regular LDL measurement (correlation 0.86; r-squared for the calibration model 0.73). (27)

Statistical Analysis

We performed a cross-sectional analysis examining the association of race/ethnicity, obesity as defined by body mass index (BMI) categories, abdominal obesity as defined by waist circumference (cm) categories, and insulin resistance as defined by quartiles of HOMA-IR, with the prevalence of CHL and with the odds of CHL after adjustment for appropriate cofactors. Chi-square tests were used to compare categorical variables and analysis of variance (ANOVA) to compare means of continuous measurements across categories. Significance for comparisons of characteristics between pairs of race/ethnic groups was declared for p<0.01. Multivariate logistic regression was used to estimate the association of CHL with race/ethnicity (model 1), adiposity as defined by BMI (model 2), waist circumference using metabolic syndrome cutoffs (model 3), waist circumference using ethnic-specific cutoffs (model 4), and insulin resistance (model 5).

Important co-variates for the relationship between 1) race/ethnicity and CHL included age, gender, and BMI; 2) for the relationship of adiposity with CHL included age, gender, and race/ethnicity; and 3) for the relationship of insulin resistance with CHL included age, gender, race/ethnicity, and BMI.

Three sensitivity analyses were performed. In one, the analysis was performed excluding those on lipid lowering therapy and those with missing values for lipid values using only observed lipid values. In another sensitivity analysis, we re-defined CHL using NCEP/ATP III cut points (LDL-C ≥ 160 mg/dL and triglycerides ≥150 mg/dL). Exogenous estrogen has known effects on LDL-C and triglycerides and could possibly be associated with adiposity. The final sensitivity analysis was performed by excluding women on estrogen (n=901) to evaluate whether estrogen use confounded the associations examined. Statistical analyses were performed using Stata Statistical Software 9.0 (Stata Corp., College Station, Texas).

Results

After applying the exclusion criteria, there remained 5,923 participants. Table 1 describes the baseline characteristics of the entire population and as classified by race/ethnicity. The mean age of the study population was 62 years (range 44–84 years), 53% were women, and the mean BMI was 28.0 kg/m2 (range 15.4–54.5 kg/m2). The racial/ethnic breakdown of the 5,923participants was consistent with the entire MESA cohort: 41% White, 12% Asian, 26% African-American, and 21% Hispanic. There were no significant differences observed in the number of women vs. men among race/ethnicity categories. Whites were more likely to take lipid lowering medications than the other race/ethnicity groups.

Table 1.

Metabolic, anthropometric, and lipid parameters in all subjects and as classified by race/ethnicity

All (n=5923) White (n=2452) Asian (Chinese) (n=695) African- American (n=1546) Hispanic (n=1230)
% Women 53 53 52 56 53
Age, yrs.* 61.8 (10.3) 62.4 (10.3) 61.8 (10.4) 61.7 (10.2) 60.6 (10.4)
Hypertension, %* 42 37 37 56 37
Ever smoked, %* 49 55 24 54 45
Current Smokers, %* 13 12 5.0 18 14
Education, %*
< High School education 16 5 23^ 11 42
Completed High School 18 16 17 18 21
> High School education 66 79 60 71 37
Family History of Myocardial Infarction, %* 43 51 20 42 40
Lipid Medication Use, %* 14 17 11 14 11
Systolic Blood Pressure, mmHg* 127 (21) 124 (20) 125 (22) 132 (22) 127 (22)
BMI, kg/m2* 28.0 (5.4) 27.5 (5.0) 23.8 (3.3) 29.9 (5.9) 29.0 (4.8)
Waist Circumference, cm* 97.2 (14.1) 97.3 (14.2) 86.6 (9.9) 99.9 (14.5) 99.4 (12.6)
Glucose mg/dL * 89.6 (10.5) 87.9 (10.1) 91.6 (10.0) 90.3 (10.8) 90.9 (10.9)
Insulin mu/U* 5.2 (3.4–8.0) 4.5 (3.1–7.2) 4.8 (3.5–7.3) 5.5 (3.6–8.6) 6.1 (4.1–9.6)
HOMA-IR* 1.1 (0.7–1.8) 1.0 (0.6–1.6) 1.1 (0.8–1.7) 1.2 (0.8–2.0) 1.4 (0.9–2.2)
[Lipid], Mean (SD), mg/dL (fasting)§
Total Cholesterol* 200.6 (34.4) 203.2 (33.0) 198.1 (31.5) 195.8(36.3) 203.1 (35.7)
LDL-C* 123.4 (30.6) 124.4 (29.4) 120.5 (28.4) 121.9 (33.0) 125.1 (30.9)
HDL-C* 51.3 (14.7) 52.3 (15.5) 50.0 (12.6) 53.1 (15.3) 48.1 (12.8)
Triglycerides* 112 (79–159) 112 (77–161) 119 (86–167) 88 (65–119) 131 (94–184)
Non-HDL-C* 149.2 (35.1) 150.9 (34.4) 148.0 (32.1) 142.6 (36.2) 155.0 (35.8)
LDL size, nm* 20.8 (0.77) 20.9 (0.75) 20.7 (0.81) 21.0 (0.77) 20.7 (0.76)
HDL size, nm* 9.17 (0.41) 9.16 (0.42) 9.17 (0.37) 9.24 (0.43) 9.09 (0.38)

Mean (SD) or percents presented unless

median and interquartile range presented

*

p<0.001 for race/ethnicity categories by chi-square tests for categorical variables and by ANOVA for continuous variables

p<0.01 for analyses comparing Chinese, Hispanic, or African-American to White

§

Total cholesterol, LDL-C, HDL-C, Non-HDL-C, and triglycerides are imputed values for those on lipid lowering therapy; observed values were used for all others

Compared to whites, African-Americans had higher mean BMI, waist circumference, and HOMA-IR. Yet, their mean triglycerides and non-HDL-C were lower and LDL size was larger than in Whites. Compared to Whites, Asians had lower mean BMI and waist circumference, LDL-C, and HDL-C. However, markers of insulin resistance such as HOMA-IR and triglycerides were not different between Asians and Whites. Compared to Whites, Hispanics had higher mean BMI, waist circumference, non-HDL-C, HOMA-IR, and triglycerides. LDL-C levels were not different but LDL size was significantly smaller for Hispanics compared to Whites.

Table 2 describes the prevalence of CHL and the characteristics of the subjects with and without CHL, including anthropometric, metabolic, and lipid parameters. CHL prevalence in the entire cohort was 2.9%. CHL subjects were more likely to be women, younger, heavier, have larger waist circumference and be more insulin resistant compared to the non CHL subjects. CHL subjects had higher mean total cholesterol, LDL-C, and triglycerides by definition compared to the non CHL subjects. CHL subjects also had lower mean HDL-C, higher non-HDL-C, smaller LDL size, and smaller HDL size compared to the non CHL subjects.

Table 2.

Characteristics of subjects by the absence or presence of CHL

No CHL (n=5749) CHL (n=174)
Women, % 53 63
Age, yrs 61.9 (10.3) 59.7 (10.2)
BMI, kg/m2* 28.0 (5.3) 29.4 (5.3)
Waist Circumference, cm 97.1 (14.1) 99.9(12.7)
HOMA-IR 1.1 (0.7–1.8) 1.4 (1.0–2.0)
Total Cholesterol, mg/dL* 198.5 (32.2) 270.7 (32.5)
LDL-C, mg/dL* 121.7 (29.1) 182.3 (21.8)
HDL-C, mg/dL * 51.5 (14.8) 45.5 (9.1)
Triglycerides, mg/dL* 109 (78–155) 189 (167–227)
Non-HDL-C, mg/dL* 147.0 (32.6) 225.2 (31.7)
LDL size, nm* 20.9 (0.77) 20.3 (0.61)
HDL size, nm* 9.2 (0.41) 8.9 (0.35)

Mean (SD) and percentages presented unless

median and interquartile range presented

*

p<0.001 for Not CHL vs. CHL

p<0.05 for Not CHL vs. CHL using chi-square tests for categorical variables and student t-tests for continuous variables

Table 3 demonstrates the prevalence and adjusted odds of CHL by race/ethnicity. Hispanics had the highest prevalence of CHL at 4.6%. After adjusting for age and gender, the likelihood of CHL compared to Whites was 47% lower for African-Americans, 52% higher for Hispanics, and not significantly different for Asians. After further adjustment for BMI, the likelihood of CHL compared to Whites remained lowest for African-Americans (52% lower) and not significantly different for Hispanics or Asians.

Table 3.

Prevalence and Odds of CHL by Race/Ethnicity

CHL (n=174) No CHL (n=5749) Prevalence of CHL,% * OR (95% CI) OR (95% CI)
White 74 2,379 3.0 1.0 1.0
Asian (Chinese) 18 677 2.6 0.84 (0.50, 1.42) 1.06 (0.62, 1.82)
African American 26 1520 1.7 0.53 (0.34, 0.84) 0.48 (0.30, 0.75)
Hispanic 56 1174 4.6 1.48 (1.04, 2.11) 1.33 (0.93, 1.91)
*

p<0.001 by chi-square test

adjusted for age and gender

adjusted for age, gender, and BMI categories

Table 4 describes the prevalence and adjusted odds of CHL by BMI categories, waist circumference categories, and quartiles of insulin resistance. The likelihood of CHL was significantly greater with elevated BMI and enlarged waist circumference, and demonstrated a threshold effect with increasing insulin resistance after adjustment for age and gender. Results for the association with BMI and waist circumference were similar after further adjustment for race/ethnicity, and results for insulin resistance were similar after further adjustment for race/ethnicity and BMI category. CHL was significantly associated with both the overweight and obese categories compared to the normal weight group. Similarly, CHL was significantly associated with insulin resistance in quartiles 2, 3, and 4 of HOMAIR compared to quartile 1.

Table 4.

Prevalence and Odds of CHL by BMI, Waist Circumference (WC), Ethnic Specific WC

Prevalence of CHL,%* OR (95% CI) OR (95% CI) , §
BMI Categories
Normal Weight (n=1,808) 1.8 1.0 1.0
Overweight (n=2,355) 3.4 2.00 (1.32, 3.04) 2.05 (1.34, 3.16)
Obese (n=1,760) 3.6 1.97 (1.28, 3.03) 2.18 (1.38, 3.44)
Waist Circumference (WC)||
Normal WC (n=2,812) 2.0 1.0 1.0
Enlarged WC (n=3,111) 3.8 1.83 (1.30, 2.55) 1.87 (1.32, 2.65)
Ethnic Specific WC
Normal WC (n=2,238) 1.7 1.0 1.0
Enlarged WC (n=3,685) 3.7 2.05 (1.42, 2.96) 1.90 (1.29, 2.79)
HOMA-IR Categories
Quartile 1 (n=1,479) 0.7 1.0 1.0
Quartile 2 (n=1,480) 3.5 4.85 (2.51, 9.33) 4.50 (2.31, 8.76)
Quartile 3 (n=1,484) 3.9 5.51 (2.88, 10.6) 4.90 (2.49, 9.64)
Quartile 4 (n=1,480) 3.6 4.99 (2.60, 9.60) 4.40 (2.18, 8.90)
*

p<0.01 by chi-square tests (prevalence of CHL)

Adjusted for age and gender

BMI & WC analyses adjusted for age, gender, race/ethnicity

§

Insulin resistance analysis adjusted for age, gender, race/ethnicity, and BMI categories

||

NCEP/ATP III Age and gender cutoff for metabolic syndrome definition for waist circumference (14)

Age, gender, and race/ethnic cutoff for metabolic syndrome definition for waist circumference (24)

In the sensitivity analysis excluding those on lipid lowering therapies and those with missing values for lipid parameters, the overall prevalence of CHL was 3.4%, very similar to the primary analysis (2.9%). The odd ratios (ORs) for the relationship of race/ethnicity with CHL after adjusting for age, gender, and BMI categories were very similar to the primary analysis: Asians (OR: 1.26 95% CI 0.75 to 2.10), African-Americans (OR: 0.36 95% CI 0.22 to 0.58), and Hispanics (OR 1.31 95% CI 0.92 to 1.85). The ORs for the relationship of BMI categories, waist circumference, and insulin resistance with CHL were also very similar to the primary analysis.

In the sensitivity analysis using the NCEP/ATP III cut points to define CHL, the overall prevalence of CHL was 3.9%. The odd ratios (ORs) for the relationship of race/ethnicity with CHL after adjusting for age, gender, and BMI categories were very similar to the primary analysis for Asians (OR: 1.16 95% CI 0.73 to 1.85), African-Americans (OR: 0.52 95% CI 0.35 to 0.77), and Hispanics (OR 1.38 95% CI 1.00 to 1.89). The ORs for the relationship of BMI categories, waist circumference, and insulin resistance with CHL were very similar to those of the primary analysis. The threshold effect for odds of CHL based on quartile of HOMA-IR also remained. Finally, there were no important differences in the reported associations when women on estrogen were excluded from the analyses.

Discussion

This is the first study, to our knowledge, to examine the association of race/ethnicity with the odds of CHL in the US. In this cohort, African-Americans were on average heavier, had larger waist circumference, and were more insulin resistant compared to Whites. However, the odds of CHL were lower for African-Americans compared to whites before and after adjustment for age, gender, and BMI categories. Lower triglyceride levels associated with increased levels of lipoprotein lipase (LPL) activity have been previously observed in African-Americans compared to Whites. (28) Increased levels of LPL activity, however, do not explain the lower LDL levels seen in African-Americans compared to Whites. In our cohort, there was a small but statistically significant difference in LDL-C between African-Americans (121.9 mg/dL) and Whites (124.4 mg/dL) (p<0.01). Increased body weight (BMI and waist circumference) predicts insulin resistance well in African-Americans but triglycerides do not. (29) These findings suggest the importance of other shared genetic and/or environmental factors among African Americans, such as a lower rate of lipoprotein entry into the circulation or a greater rate of LDL removal that protects them from the development of CHL even in a setting of obesity and insulin resistance.

In contrast, the odds of CHL were higher for Hispanics compared to Whites after adjusting for age and gender. The association was no longer significant after also adjusting for BMI, indicating that a large proportion of the increased odds of CHL in Hispanics is explained by obesity. However, other shared genetic and/or environmental factors beyond obesity in Hispanics may also be present, as the point estimate for the OR remained >1 after the full adjustment. In Asians, the odds of CHL were not statistically different compared to Whites, possibly because of a lack of power due to low numbers of Asians with CHL.

We have also established in this multi-ethnic cohort free of CVD that greater BMI was associated with a higher odds of CHL, after adjusting for age, gender, and race/ethnicity. The odds of CHL was approximately twice as high with abdominal adiposity, whether defined by traditional metabolic syndrome cut offs or by ethnic-specific cutoffs for waist circumference, and remained so after adjusting for age, gender, and race/ethnicity. The odds of CHL was substantially higher with worsening insulin resistance but demonstrated a threshold effect, as the odds for quartiles 2 through 4 were similarly elevated compared to quartile 1. The mean BMI in the entire MESA population was rather high at 28.0 kg/m2 and it is likely that CHL is expressed in an appropriate genetic background except in the context of very good insulin sensitivity.

The relationship of obesity and insulin resistance with FCHL has been previously described. (15, 1720) Our findings support a similar relationship with CHL. Although increasing obesity worsens insulin resistance, obesity is not necessarily always seen with FCHL. (19) However, insulin resistance as defined by various methods including HOMA-IR, is a common concomitant finding with FCHL. (15,30) Our findings strongly support the same relationship of insulin resistance to CHL. Mechanistically, insulin resistance increases free fatty acid flux, causing hepatic overproduction of triglycerides and apolipoprotein B, key features of FCHL.

We demonstrated that the largest increment in CHL odds occurred in the transition between normal weight to overweight and between quartile 1 and quartile 2 of HOMA-IR. Current suggested obesity cutoffs using BMI of 30 appear to be less useful when considering increased cardio-metabolic risk in a multi-ethnic population and especially populations of Asian origin. (31,32) This observation is supported by the Study of Health Assessment and Risk in Ethnic Groups (SHARE) and Risk Evaluation in Aboriginal Peoples (SHARE-AP) groups’ findings that a BMI much less than 30 is associated with increased cardio-metabolic risk in various ethnic populations in Canada. (4)

The present analysis is cross-sectional and thus a causal relationship cannot be directly inferred, especially since the temporal sequence between adiposity, insulin resistance, and CHL is unknown. However, the biological plausibility and magnitude of the association (for BMI categories, waist circumference categories, and HOMA-IR quartiles) are consistent with a causal association between obesity and CHL as well as insulin resistance and CHL independent of obesity in this multi-ethnic population. Furthermore, our findings are strengthened by very similar findings using only observed lipid values and excluding those on lipid lowering therapies and with missing values for lipid values.

We have also demonstrated that CHL, as defined by ≥75th percentile for both LDL cholesterol and triglycerides for age and gender based on a predominantly Caucasian population reference (1), is very consistent with current NCEP/ATP III cutoffs for consideration of treatment for LDL (≥ 160 mg/dL) and triglycerides (≥ 150 mg/dL) in terms of prevalence and associations with race/ethnicity, obesity, and insulin resistance. The benefit of identifying people based on age and gender standardized percentiles is that we can recognize dyslipidemia that is otherwise not obvious in younger people who traditionally are at lower Framingham/NCEP 10 year CHD risk. The percentile definitions allow young people to have enhanced recognition and screening for other cardiac risk factors, family history, and appropriate advice regarding diet, lifestyle, and when indicated pharmacologic intervention. Although, the association between obesity and insulin resistance and FCHL has been clearly documented, we have demonstrated that CHL as defined by 75% cut points is strongly associated with only modest degrees of increasing weight and insulin resistance (second quartile and above) without the need of family history for lipid abnormalities. Thus evaluating for CHL in overweight and/or insulin resistant people may help practitioners better identify this dyslipidemia and propose appropriate treatment.

We acknowledge additional limitations. HOMA-IR is a valid marker of insulin resistance in epidemiological studies but is subject to misclassification. (33) The mean HOMA-IR in this study was lower than in previous population studies. (34) The difference may be due to differences in populations but also to differences in the method of measuring insulin. Nonetheless, the division of HOMA-IR into quartiles remained meaningful as demonstrated by the robust and significant association between insulin resistance and CHL beginning in the second quartile. Misclassification of abdominal obesity is also possible, using the traditional metabolic syndrome cutoffs of waist circumference in Caucasians and probably African Americans. We analyzed the relationship of abdominal obesity using proposed ethnic-specific cutoffs in order to address this possible problem. The direction and magnitude of association of abdominal obesity and CHL was similar using traditional metabolic syndrome and ethnic-specific cutoffs for waist circumference.

Classification of race/ethnicity was self-reported and subject to misclassification. Many people may belong to multiple ethnic categories but could only be classified as one category in MESA. We may not presume race/ethnicity to represent common genetic traits and can only say that race/ethnicity as broadly classified could represent shared genetic and/or environmental influences which include social behaviors such as diet and activity.

Acknowledgments

This research was supported by contracts N01-HC-95159 through N01-HC-95166 from the National Heart, Lung, and Blood Institute. The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. The authors gratefully acknowledge Robyn McClelland, PhD at the University of Washington for providing the imputed lipid values for the MESA participants. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org. Dr. Paramsothy was supported by a Pfizer Preventive Cardiology Award and is funded by grant Number 1KL2RR025015-01 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH) and NIH Roadmap for Medical Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. Information on NCRR is available at http://www.ncrr.nih.gov/.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • 1.The Lipid Research Clinics Population Studies Data Book, Volume I: The Prevalence Study. US Dept of Health and Human Services. 1980 NIH publication 80–1527. [Google Scholar]
  • 2.Brindle P, May M, Gill P, et al. Primary prevention of cardiovascular disease: a web-based risk score for seven British black and minority ethnic groups. Heart. 2006 Nov;92(11):1595–1602. doi: 10.1136/hrt.2006.092346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Anand SS, Yusuf S, Vuksan V, et al. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE) Lancet. 2000 Jul 22;356(9226):279–284. doi: 10.1016/s0140-6736(00)02502-2. [DOI] [PubMed] [Google Scholar]
  • 4.Razak F, Anand SS, Shannon H, et al. Defining obesity cut points in a multiethnic population. Circulation. 2007 Apr 24;115(16):2111–2118. doi: 10.1161/CIRCULATIONAHA.106.635011. [DOI] [PubMed] [Google Scholar]
  • 5.D’Agostino RB, Sr, Grundy S, Sullivan LM, Wilson P. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. Jama. 2001 Jul 11;286(2):180–187. doi: 10.1001/jama.286.2.180. [DOI] [PubMed] [Google Scholar]
  • 6.Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004 Sep 11–17;364(9438):937–952. doi: 10.1016/S0140-6736(04)17018-9. [DOI] [PubMed] [Google Scholar]
  • 7.Austin MA, McKnight B, Edwards KL, et al. Cardiovascular disease mortality in familial forms of hypertriglyceridemia: A 20-year prospective study. Circulation. 2000 Jun 20;101(24):2777–2782. doi: 10.1161/01.cir.101.24.2777. [DOI] [PubMed] [Google Scholar]
  • 8.Brunzell JD, Schrott HG, Motulsky AG, Bierman EL. Myocardial infarction in the familial forms of hypertriglyceridemia. Metabolism. 1976 Mar;25(3):313–320. doi: 10.1016/0026-0495(76)90089-5. [DOI] [PubMed] [Google Scholar]
  • 9.Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG. Hyperlipidemia in coronary heart disease. II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest. 1973 Jul;52(7):1544–1568. doi: 10.1172/JCI107332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hopkins PN, Heiss G, Ellison RC, et al. Coronary artery disease risk in familial combined hyperlipidemia and familial hypertriglyceridemia: a case-control comparison from the National Heart, Lung, and Blood Institute Family Heart Study. Circulation. 2003 Aug 5;108(5):519–523. doi: 10.1161/01.CIR.0000081777.17879.85. [DOI] [PubMed] [Google Scholar]
  • 11.Veerkamp MJ, de Graaf J, Hendriks JC, Demacker PN, Stalenhoef AF. Nomogram to diagnose familial combined hyperlipidemia on the basis of results of a 5-year follow-up study. Circulation. 2004 Jun 22;109(24):2980–2985. doi: 10.1161/01.CIR.0000130646.93255.86. [DOI] [PubMed] [Google Scholar]
  • 12.Knopp RH, Walden CE, Retzlaff BM, et al. Long-term cholesterol-lowering effects of 4 fat-restricted diets in hypercholesterolemic and combined hyperlipidemic men. The Dietary Alternatives Study. Jama. 1997 Nov 12;278(18):1509–1515. [PubMed] [Google Scholar]
  • 13.Walden CE, Retzlaff BM, Buck BL, Wallick S, McCann BS, Knopp RH. Differential effect of National Cholesterol Education Program (NCEP) Step II diet on HDL cholesterol, its subfractions, and apoprotein A-I levels in hypercholesterolemic women and men after 1 year: the beFIT Study. Arterioscler Thromb Vasc Biol. 2000 Jun;20(6):1580–1587. doi: 10.1161/01.atv.20.6.1580. [DOI] [PubMed] [Google Scholar]
  • 14.Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III) Jama. 2001 May 16;285(19):2486–2497.19. doi: 10.1001/jama.285.19.2486. [DOI] [PubMed] [Google Scholar]
  • 15.Veerkamp MJ, de Graaf J, Stalenhoef AF. Role of insulin resistance in familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol. 2005 May;25(5):1026–1031. doi: 10.1161/01.ATV.0000160612.18065.29. [DOI] [PubMed] [Google Scholar]
  • 16.van der Kallen CJ, Voors-Pette C, Bouwman FG, et al. Evidence of insulin resistant lipid metabolism in adipose tissue in familial combined hyperlipidemia, but not type 2 diabetes mellitus. Atherosclerosis. 2002 Oct;164(2):337–346. doi: 10.1016/s0021-9150(02)00109-0. [DOI] [PubMed] [Google Scholar]
  • 17.Skoumas J, Papadimitriou L, Pitsavos C, et al. Metabolic syndrome prevalence and characteristics in Greek adults with familial combined hyperlipidemia. Metabolism. 2007 Jan;56(1):135–141. doi: 10.1016/j.metabol.2006.09.007. [DOI] [PubMed] [Google Scholar]
  • 18.Ascaso JF, Real JT, Merchante A, Rodrigo A, Carmena R. Lipoprotein phenotype and insulin resistance in familial combined hyperlipidemia. Metabolism. 2000 Dec;49(12):1627–1631. doi: 10.1053/meta.2000.18568. [DOI] [PubMed] [Google Scholar]
  • 19.Bredie SJ, Tack CJ, Smits P, Stalenhoef AF. Nonobese patients with familial combined hyperlipidemia are insulin resistant compared with their nonaffected relatives. Arterioscler Thromb Vasc Biol. 1997 Jul;17(7):1465–1471. [PubMed] [Google Scholar]
  • 20.Aitman TJ, Godsland IF, Farren B, Crook D, Wong HJ, Scott J. Defects of insulin action on fatty acid and carbohydrate metabolism in familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol. 1997 Apr;17(4):748–754. doi: 10.1161/01.atv.17.4.748. [DOI] [PubMed] [Google Scholar]
  • 21.Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol. 2002 Nov 1;156(9):871–881. doi: 10.1093/aje/kwf113. [DOI] [PubMed] [Google Scholar]
  • 22.Genuth S, Alberti KG, Bennett P, et al. Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care. 2003 Nov;26(11):3160–3167. doi: 10.2337/diacare.26.11.3160. [DOI] [PubMed] [Google Scholar]
  • 23.The Sixth Report of the Joint National Committee on Prevention, DetectionEvaluation, and Treatment of High Blood Pressure. Arch Intern Med. 1997;157:2413–2446. doi: 10.1001/archinte.157.21.2413. [DOI] [PubMed] [Google Scholar]
  • 24.Alberti KG, Zimmet P, Shaw J. The metabolic syndrome--a new worldwide definition. Lancet. 2005 Sep 24–30;366(9491):1059–1062. doi: 10.1016/S0140-6736(05)67402-8. [DOI] [PubMed] [Google Scholar]
  • 25.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985 Jul;28(7):412–419. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
  • 26.Demacker PN, Hijmans AG, Brenninkmeijer BJ, Jansen AP, van ‘t Laar A. Five methods for determining low-density lipoprotein cholesterol compared. Clin Chem. 1984 Nov;30(11):1797–1800. [PubMed] [Google Scholar]
  • 27.McClelland RL, Kronmal RA, Haessler J, Blumenthal RS, Goff DC., Jr Estimation of risk factor associations when the response is influenced by medication use: an imputation approach. Statistics in Medicine. Published online (epub ahead of press) in Wiley Interscience. doi: 10.1002/sim.3341. www.interscience.wiley.com. [DOI] [PMC free article] [PubMed]
  • 28.Despres JP, Couillard C, Gagnon J, et al. Race, visceral adipose tissue, plasma lipids, and lipoprotein lipase activity in men and women: the Health, Risk Factors, Exercise Training, and Genetics (HERITAGE) family study. Arterioscler Thromb Vasc Biol. 2000 Aug;20(8):1932–1938. doi: 10.1161/01.atv.20.8.1932. [DOI] [PubMed] [Google Scholar]
  • 29.Sumner AE, Finley KB, Genovese DJ, Criqui MH, Boston RC. Fasting triglyceride and the triglyceride-HDL cholesterol ratio are not markers of insulin resistance in African Americans. Arch Intern Med. 2005 Jun 27;165(12):1395–1400. doi: 10.1001/archinte.165.12.1395. [DOI] [PubMed] [Google Scholar]
  • 30.Cabezas MC, et al. Impaired fatty acid metabolism in familial combined hyperlipidemia. J Clin Invest. 1993;92:160–168. doi: 10.1172/JCI116544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004 Jan 10;363(9403):157–163. doi: 10.1016/S0140-6736(03)15268-3. [DOI] [PubMed] [Google Scholar]
  • 32.Lear SA, Toma M, Birmingham CL, Frohlich JJ. Modification of the relationship between simple anthropometric indices and risk factors by ethnic background. Metabolism. 2003 Oct;52(10):1295–1301. doi: 10.1016/s0026-0495(03)00196-3. [DOI] [PubMed] [Google Scholar]
  • 33.Bonora E, Targher G, Alberiche M, et al. Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity. Diabetes Care. 2000 Jan;23(1):57–63. doi: 10.2337/diacare.23.1.57. [DOI] [PubMed] [Google Scholar]
  • 34.Rutter MK, Meigs JB, Sullivan LM, D’Agostino RB, Sr, Wilson PW. Insulin resistance, the metabolic syndrome, and incident cardiovascular events in the Framingham Offspring Study. Diabetes. 2005 Nov;54(11):3252–3257. doi: 10.2337/diabetes.54.11.3252. [DOI] [PubMed] [Google Scholar]

RESOURCES