Data from National Health and Nutrition Examination Survey (NHANES) II (1976 to 1980), NHANES III (1988 to 1994), and NHANES 1999 to 2006 were examined to assess trends in total cholesterol (TC), low-density lipoprotein (LDL) cholesterol, high-density lipoprotein cholesterol, triglycerides (TGs), lipid-lowering medication use, and obesity. Age-adjusted decreases in TC (210 to 200 mg/dl) and LDL cholesterol (134 to 119 mg/dl) were observed. Those with high TC showed a decrease of 9% from NHANES II to NHANES 1999 to 2006, whereas those with LDL cholesterol ≥160 mg/dl showed a decrease of 8%. A significant increase in mean high-density lipoprotein cholesterol was observed (50 to 53 mg/dl, p <0.001), most likely due to changes in methods. Those with TG levels ≥150 mg/dl showed a decrease from NHANES II to NHANES III from 30% to 27% but then an increase from NHANES III to NHANES 1999 to 2006 from 27% to 33%. Since NHANES III, mean TG levels have increased 12% from 130 to 146 mg/dl. In the 2 most recent surveys, self-reported “high cholesterol” increased from 17% to 27%, and self-reported lipid medication use by those with high cholesterol increased from 16% to 38%. Mean body mass index increased from 26 to 29 kg/m 2 , and prevalence of obesity doubled and was significantly associated with increased TG. In conclusion, recent favorable trends in TC and LDL cholesterol are likely due to increased awareness of high cholesterol and the greater use of lipid-lowering drugs. However, countertrends in obesity and TG levels, if continued, will likely have a negative impact on cardiovascular disease in the future.
The primary objective of this study was to assess trends in major lipid fractions for United States adults 20 to 74 years of age from 1976 through 2006. The secondary objective was to examine the prevalence of lipid levels based on optimal and nonoptimal classifications as defined by the Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). This study also examined trends and patterns of self-reported lipid medication use in respondents who reported a diagnosis of “high cholesterol” for the National Health and Nutrition Examination Survey (NHANES) III (1988 to 1994) and subsequent surveys (1999 to 2006). This study also assessed the relation between serum triglyceride (TG) levels and obesity.
Methods
This study used the cross-sectional, nationally representative NHANES data: NHANES II (1976 to 1980), NHANES III (1988 to 1994), and NHANES 1999 to 2006. The NHANES is a stratified probability sampling of noninstitutionalized United States civilians and derives data from personal interview, physical examination, and laboratory testing. For this study, respondents 20 to 74 years of age with ≥1 total cholesterol (TC) or lipid fraction result were used. Although the survey is intended to be a comprehensive cross section of the United States population, data are not provided by NHANES to define the representativeness of samples regarding educational level, race/ethnicity, or poverty status.
Response rates for NHANES II, NHANES III, and the 1999 to 2002 wave ranged from 64% to 84% and have been reported. NHANES 2003 to 2004 data are based on 6,916 adults ≥20 years old, of whom 5,041 (73%) were interviewed, and 4,742 were interviewed, underwent a physical examination, and were eligible for phlebotomy; the NHANES 2005 to 2006 data are based on 6,719 adults (≥20 years of age), of whom 4,979 (74%) were interviewed, and 4,773 were examined and eligible for phlebotomy.
Sampling procedures were similar among the various waves of NHANES, although some differences did exist. TC and lipid fraction levels were measured for adults ≥20 years of age in NHANES III and NHANES 1999 to 2006 but only for adults 20 to 74 years old in NHANES II. Methods for measuring, calculating, and analyzing all lipid fractions for NHANES II, NHANES III, and the surveys initiating in 1999 to 2002 have previously been reported. Methods for the NHANES 2003 to 2006 surveys were similar to the methods for previous surveys. TC, high-density lipoprotein (HDL) cholesterol, and TG measurements were standardized for each NHANES survey according to criteria from the Centers for Disease Control and Prevention Lipid Standardization Program. TC and HDL cholesterol levels are based on samples taken regardless of fasting state, whereas low-density lipoprotein (LDL) cholesterol and TG levels were based on a morning examination after fasting for 8 to 24 hours. Samples were analyzed by certified laboratories monitored by the Centers for Disease Control and Prevention.
Calculated LDL cholesterol was obtained using the Friedewald equation (TC − HDL cholesterol − TG/5) for those with TG levels ≤400 mg/dl and the Hattori equation (0.94 × TC − 0.94 × HDL cholesterol − 0.19 × TG) for those with TG concentrations >400 mg/dl. Direct LDL cholesterol was not available for NHANES II; LDL cholesterol was calculated for respondents with nonmissing TC, HDL cholesterol, and TG values (82% of sample). NHANES III included direct LDL cholesterol values for 6,118 respondents (87% of sample) and LDL cholesterol was calculated for 1 additional respondent. For NHANES 1999 to 2006, direct LDL cholesterol was reported for 7,132 respondents (87% of sample) and LDL cholesterol was calculated for 519 respondents (6% of sample). Respondents for whom LDL cholesterol was missing were excluded from the analysis of LDL cholesterol (n = 1,073 for NHANES II, n = 893 for NHANES III, n = 523 for NHANES 1999 to 2006).
An instrument change occurred in NHANES 2005 to 2006 for TC, but the method and laboratory site were the same as the 2003 to 2004 survey. From 2003 to 2006, HDL cholesterol was analyzed using a direct HDL cholesterol immunoassay method compared to the precipitation method used for participants in the 1999 to 2002 surveys. A correction procedure was employed to control for these differences in methods and instrumentation and this 1999 to 2002 method bias is further described by Carroll et al. From 2005 to 2006, there was a change in equipment to measure HDL cholesterol, but the laboratory method was similar and the laboratory site was the same as in NHANES 2003 to 2004.
Outcomes were examined for adults 20 to 74 years of age and trends were examined across all NHANES waves, unless otherwise specified. Demographics (gender, age, race/ethnicity) and body mass index were evaluated. Race/ethnicity was self-reported based on census categories from the National Vital Statistics System, and NHANES III and forward included non-Hispanic white, non-Hispanic black, Mexican-American, or other (NHANES II classified subjects as white, black, and others using primarily interviewer observation). Body mass index was stratified as normal/underweight (<25 kg/m 2 ), overweight (25 to 29 kg/m 2 ) and obese (≥30 kg/m 2 ). Mean serum TC and individual lipid fractions (LDL cholesterol, HDL cholesterol, TGs) were evaluated by age group and gender for all 3 waves.
TC was classified as high (≥240 mg/dl), borderline high (200 to 239 mg/dl), and desirable (<200 mg/dl). LDL cholesterol was categorized as high or very high (≥160 mg/dl), borderline high (130 to 160 mg/dl), near or above optimal (100 to 130 mg/dl), and optimal (<100 mg/dl). Low HDL cholesterol (<40 mg/dl) and high HDL cholesterol (≥60 mg/dl) were also defined based on the criteria. TGs were divided into normal (<150 mg/dl), borderline high (150 to 199 mg/dl), and high or very high (≥200 mg/dl). To assess the association between increased TG and body mass index, increased TG was defined as ≥150 mg/dl.
Lipid medication use was assessed in respondents with a self-reported high cholesterol diagnosis for the NHANES III and NHANES 1999 to 2006 surveys. For NHANES III, high cholesterol and lipid medication use were established using interview questions. Similar questions were used for the NHANES 1999 to 2006 surveys. The proportion of respondents with confirmed possession of a lipid medication, using the prescription datasets for the NHANES III and NHANES 1999 to 2006 surveys, was also calculated. Prescriptions for NHANES III and NHANES 1999 to 2002 were assigned a National Drug Code up to 6 digits from the Master Drug Database for each product. Prescriptions for NHANES 2003 to 2006 were assigned a unique generic drug code from the Multum Lexicon Drug Database. Generic drug codes for any of the following drugs were included as lipid medications: atorvastatin, cerivastatin, cholestyramine, clofibrate, colesevelam, colestipol, ezetimibe, fenofibrate, fluvastatin, gemfibrozil, lovastatin, niacin, ω-3 acid ethyl esters, pravastatin, probucol, rosuvastatin, simvastatin, Vytorin (simvastatin–ezetimibe combination) (Merck/ Schering Plough Pharmaceuticals, North Wales, Pennsylvania), or cholesterol-lowering drug not otherwise stated.
Frequencies and means ± SDs are presented to describe the study population. Sample weights, which account for the differential probabilities of selection and nonrespondent, are incorporated into all estimation analyses. Based on the “least-common denominator” approach, the final oral glucose tolerance test weights were used in this study. In addition, age adjustment was employed for estimation of means using the direct estimate method, with United States census population projected estimates for the year 2000 (age groups 20 to 39 years, 40 to 59 years, and 60 to 74 years), to decrease different age structures between surveys. Age-adjusted means ± SEMs are presented for TC, LDL cholesterol, HDL cholesterol, and TG levels. Changes in age-adjusted mean lipid levels were assessed between waves using analysis of variance. In addition, estimates are presented for TC and individual lipid fractions stratified by age groups and gender, as are prevalence estimates for individual lipid fractions by National Cholesterol Education Program/Adult Treatment Panel III categories. Multiple logistic regressions were performed to assess the association between increased TG values and obesity, controlling for age, gender, race/ethnicity, NHANES wave, and self-reported smoking status.
All statistical hypotheses were tested at the alpha level of 0.05. To generalize the statistics to the United States population, SAS survey procedures were used (SAS 9.1, SAS Institute, Cary, North Carolina).
Results
Demographics and sample sizes are presented in Table 1 . Notable differences between surveys with respect to age and race/ethnicity are due to oversampling of certain population subgroups. Mean body mass index increased from 26 to 29 kg/m 2 , a 13% increase, from NHANES II to the 1999 to 2006 survey, and the percentage of respondents classified as obese (body mass index ≥30 kg/m 2 ) increased from 15% to 34% during the same period.
| Variable | NHANES II (1976–1980) | NHANES III (1988–1994) | NHANES 1999–2006 |
|---|---|---|---|
| (n = 5,792) | (n = 7,012) | (n = 8,174) | |
| Men | 49% | 46% | 48% |
| Age (years), mean ± SD | 50.1 ± 16.0 | 43.9 ± 16.3 | 45.4 ± 16.1 |
| Age range (percentage of respondents) | |||
| 20–29 | 15% | 24% | 21% |
| 30–39 | 16% | 23% | 19% |
| 40–49 | 14% | 17% | 19% |
| 50–59 | 16% | 12% | 14% |
| 60–69 | 28% | 15% | 18% |
| 70–74 | 11% | 8% | 9% |
| Race/ethnicity (percentage of respondents) ⁎ | |||
| White | 87% | 37% | 47% |
| Black | 11% | 31% | 21% |
| Mexican-American | NA | 28% | 23% |
| Others | 2% | 4% | 8% |
| Body mass index (kg/m 2 ), mean ± SD | 25.5 ± 5.0 | 27.3 ± 5.9 | 28.7 ± 6.6 |
| Body mass index range (percentage of respondents) | |||
| <25 | 52% | 39% | 31% |
| 25–<30 | 33% | 34% | 34% |
| ≥30 | 15% | 26% | 34% |
⁎ Race/ethnicity was classified only as white, black, and others in NHANES II.
Table 2 presents the proportion of adults with total and individual lipid fractions categorized according to National Cholesterol Education Program/Adult Treatment Panel III classifications. The proportion with high TC decreased 9% from NHANES II to NHANES 1999 to 2006, whereas the proportion with high/very high LDL cholesterol (≥160 mg/dl) decreased 8%. The percentage with low HDL cholesterol (<40 mg/dl) increased from NHANES II to NHANES III (21% to 23%, respectively) but decreased to 19% of adult respondents in NHANES 1999 to 2006. The proportion categorized with high HDL cholesterol (≥60 mg/dl) continuously increased over time to 28% of respondents in 1999 to 2006. The proportion of respondents with borderline high, high, or very high TG decreased from NHANES II to NHANES III by 3% (30% to 27%) but then increased from NHANES III to NHANES 1999 to 2006 by 6% (27% to 33%).
| Variable | NHANES II (1976–1980) | NHANES III (1988–1994) | NHANES 1999–2006 |
|---|---|---|---|
| (n = 5,792) | (n = 7,012) | (n = 8,174) | |
| Total cholesterol (mg/dl) (percentage of respondents) | |||
| ≥240 | 25% | 17% | 16% |
| 200–<240 | 30% | 31% | 31% |
| <200 | 45% | 48% | 52% |
| Not available | 4% | 2% | |
| Low-density lipoprotein cholesterol (mg/dl) (percentage of respondents) | |||
| ≥160 | 20% | 16% | 12% |
| 130–<160 | 21% | 25% | 23% |
| 100–<130 | 23% | 30% | 33% |
| <100 | 17% | 23% | 31% |
| Not available | 18% | 5% | 2% |
| High-density lipoprotein cholesterol (mg/dl) (percentage of respondents) | |||
| <40 | 21% | 23% | 19% |
| 40–<60 | 43% | 51% | 52% |
| ≥60 | 18% | 22% | 28% |
| Not available | 18% | 5% | 1% |
| Triglycerides (mg/dl) (percentage of respondents) | |||
| ≥200 | 16% | 14% | 18% |
| 150–<200 | 14% | 13% | 15% |
| <150 | 67% | 68% | 67% |
| Not available | 3% | 5% | 1% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
