New National Cholesterol Education Program III Guidelines for Primary Prevention Lipid-Lowering Drug Therapy
Projected Impact on the Size, Sex, and Age Distribution of the Treatment-Eligible Population
Background— The guidelines in the Third Report of the National Cholesterol Education Program (NCEP III) include absolute risk and lower LDL cholesterol (LDL-C) levels to assess eligibility for lipid-lowering drug therapy. We studied the impact of these changes on the size, sex, and age distribution of the target US population using data from the Third Annual National Health and Nutrition Survey (NHANES III) (1988 to 1994).
Methods and Results— A subsample of NHANES III participants aged 20 to 79 years with known cardiovascular risk factors and LDL-C levels was identified (n=13 589). We assessed their eligibility for drug therapy first using NCEP II guidelines and then using the new NCEP III criteria. We also calculated the number eligible for LDL-C lowering to <100 mg/dL. An estimated 15 million individuals aged 20 to 79 years are eligible for drug therapy under NCEP II; 51% are males, 49% are females, 26% are <45 years old, and 28% are ≥65 years old. Under NCEP III, 36 million would be eligible for treatment; 55% are males, 45% are females, 32% are <45 years old, and 27% are ≥65 years old. This represents a 140% increase in eligibility overall, a 157% increase among males, a 122% increase among females, a 131% increase among those ≥65 years old, and a 201% increase among those <45 years old. Of treatment-eligible individuals, 26% of males, 24% of females, 39% of elderly, and 14% of those <45 years old are targeted for LDL-C lowering to <100 mg/dL.
Conclusions— The NCEP III guidelines will alter the age and sex distributions of the treatment-eligible population, targeting many more younger (<45 years old) and greater numbers of elderly (≥65 years) individuals, particularly for aggressive intervention.
Received October 2, 2001; revision received October 31, 2001; accepted November 1, 2001.
The National Cholesterol Education Program’s (NCEP) recommendations for cholesterol management identify elevations in LDL cholesterol (LDL-C) as the primary rationale for cholesterol- lowering therapy.1 Dietary therapy is the first line of treatment of high blood cholesterol, and drug therapy is reserved for patients at elevated risk for coronary heart disease (CHD). The guidelines consider additional cardiovascular risk factors, such as age, family history of premature CHD, current cigarette smoking, hypertension, low HDL cholesterol (HDL-C; <35 mg/dL), and diabetes mellitus, in the assessment of primary risk. Under the guidelines contained in the Second Report of the NCEP (NCEP II), patients exhibiting an LDL-C level ≥190 mg/dL and who have ≤1 additional risk factor are considered eligible for pharmacological therapy after lifestyle recommendations have failed to alter lipid levels. Patients exhibiting ≥2 of these risk factors and an LDL-C level ≥160 mg/dL are also deemed eligible for drug therapy.1,2
A revised version of the second report of the Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II) has been published recently.3 These revisions represent a major change from NCEP I and NCEP II. The NCEP III primary prevention algorithm now includes absolute risk estimation and lower LDL-C target levels to determine eligibility for drug therapy. Revised Framingham risk functions3 are used for the patient-specific risk computation, which includes lipid levels and other known cardiovascular risk factors. A key question regarding these guideline revisions is their impact on the numbers and demographic distribution of eligible patients and the proportion targeted for the aggressive lipid-lowering regimen.
See p 136
The goal of the present study was to use subjects from the Third Annual National Health and Nutrition Survey (NHANES III)4 and the revised Framingham risk functions to estimate the change in the fraction of the US adult population deemed eligible for primary prevention lipid-lowering drug therapy under the new guidelines. The further impact of these changes on age-sex distributions and target LDL-C levels of eligible patients is also explored.
The analyses in the present study are based on data from >20 000 subjects in the adult sample of the NHANES III survey, conducted from 1988 to 1994.4 One of the goals of the NHANES III survey was to estimate the national prevalence of cardiovascular diseases and associated risk factors. The NHANES III survey also included detailed clinical data, such as serum lipid levels and blood pressure. All risk factor and laboratory data needed for the application of the algorithms contained in the guidelines and risk calculation were thus available for most subjects in NHANES III.
To use the Framingham risk functions to compute risk in the NHANES III subjects, it was necessary to apply certain entry criteria for all subjects included in the analysis. Subjects’ data records also had to include complete information on Framingham risk factors. Thus, the study population was restricted to subjects aged 20 to 79 years with no prior evidence of cardiovascular disease. We defined the absence of CHD as a negative response to the question, “Has a doctor ever told you that you had a heart attack?” When these criteria were applied, 14 131 subjects were eligible for the study.
The LDL-C level for each subject was calculated from their total cholesterol, HDL-C, and triglyceride levels with the following formula: LDL-C=total cholesterol−HDL-C−(triglycerides/5). This formula, known as the Friedewald equation,5 was used by NHANES III to calculate LDL levels for respondents who fasted. To increase sample size, we also applied this formula to nonfasting respondents. Because the formula is not accurate for triglyceride levels >400 mg/dL, we excluded subjects with triglyceride levels >400 mg/dL, leaving a study population of 13 589 subjects for the analysis.
Family history of premature CHD was based on reports by respondents that a first-degree blood relative had a heart attack before age 50. Current cigarette use was defined as a positive response to the question, “Do you smoke cigarettes now?” We classified respondents as hypertensive if they reported being told by a physician to take antihypertensive medication or if they had 2 or more observed blood pressure readings of ≥140/90 mm Hg (systolic ≥140 or diastolic ≥90 mm Hg). We defined the presence of diabetes mellitus as a positive response to the question, “Have you ever been told by a doctor that you have diabetes or sugar diabetes?”
The 10-year absolute risk was calculated with sex- and age-specific Framingham score sheets for computing the “hard” risk of myocardial infarction or CHD according to age, smoking, Joint National Committee (JNC-VI) blood pressure categories, and NCEP total serum cholesterol categories.3 Hypertension treatment was defined as reported current drug therapy for hypertension.
We then applied the NCEP II guidelines listed in Table 1 to each predefined NHANES III subject to determine the percentage eligible for drug therapy. A patient with no CHD is considered eligible for treatment under NCEP II guidelines if he/she belongs to 1 of the following 2 risk groups: (1) ≤1 CHD risk factors and LDL-C ≥190 mg/dL or (2) ≥2 CHD risk factors and LDL-C ≥160 mg/dL. We then recalculated the number of eligible subjects by applying the NCEP III guidelines listed in Table 2. A patient is eligible for treatment under the newer recommendations if he/she exhibits one of the following: (1) ≤1 CHD risk factor and LDL-C ≥190 mg/dL (drug-optional level: 160 to 189 mg/dL); (2) ≥2 CHD risk factors and a 10-year hard CHD risk <10% and LDL-C ≥160 mg/dL; (3) ≥2 CHD risk factors and a 10-year hard CHD risk ≥10% to 20% and LDL-C ≥130 mg/dL; (4) 10-year hard CHD risk >20% and LDL-C ≥130 mg/dL (drug-optional levels: 100 to 129 mg/dL); or (5) diabetes and LDL-C ≥130 mg/dL (drug-optional levels: 100 to 129 mg/dL). We also separately calculated the number of individuals who would qualify for the more aggressive therapy (LDL-C target <100 mg/dL). All primary analyses assumed a liberal scenario wherein lifestyle recommendations are assumed to fail and the “drug-optional” (ie, lowest targets) are selected. We also performed a secondary (conservative) analysis wherein all lifestyle recommendations were assumed to succeed and the higher target LDL-C levels were applied. Analyses focused on describing the distribution of treatment-eligible patients by age category and sex.
Tables 3 and 4⇓ show the distribution of the study population aged 20 to 79 years who qualify for primary prevention drug therapy under both NCEP II and NCEP III recommendations by sex and age. An estimated 15 million individuals aged 20 to 79 years were found to be eligible for drug therapy under NCEP II guidelines. Of these, 51% are male, 49% are female, 26% are <45 years old, and 28% are ≥65 years old. Under the new NCEP III guidelines, 36 million would be treatment eligible; of these, 55% would be male, 45% would be female, 32% would be <45 years old, and 27% would be ≥65 years old. This represents a 140% increase in eligibility overall, a 157% increase among males (from 8 to 20 million), a 122% increase among females (from 7 to 16 million), a 131% increase among those ≥65 years old (from 4 to 10 million), and a 201% increase among those <45 years old (from 4 to 12 million). Of those deemed eligible for treatment, 26% of males, 24% of females, 39% of elderly, and 14% of those <45 years old are targeted for LDL-C lowering to <100 mg/dL.
In both males and females, we observed large increases in the number of the treatment-eligible patients among those <40 years of age and those >70 years of age. For example, eligibility increased by 321% among females aged 30 to 39 years and 157% among males aged 20 to 29 years. Overall, those aged 20 to 29 and 30 to 39 years increased their eligibility by 186% and 194%, respectively. Among males aged 70 to 79 years, eligibility increased by 140%.
Tables 5 and 6⇓ describe the differences in distribution of treatment-eligible patients under the liberal (ie, lifestyle recommendations fail) versus conservative (ie, lifestyle recommendations succeed) scenarios. On average, 71% of males and 62% of females would be eligible for treatment under the conservative scenario compared with the liberal scenario. Also as shown, younger patients are less likely to be treated than older patients under a conservative scenario. For example, only 44% of males aged 20 to 29 years of age would be treated conservatively compared with 83% of males aged 50 to 59 years.
As shown in Table 6, the age-adjusted sex distribution (% females) of treatment-eligible patients differs from the NCEP II guidelines under both the liberal and conservative scenarios. Previously, 49% of females were eligible for treatment. Under NCEP III, this ratio changed to 45% and 42% when liberal and conservative criteria were applied, respectively.
Table 7 describes the number of patients who would be deemed eligible for treatment on the basis of the sole criterion of LDL-C ≥160 mg/dL. Overall, 71% of treatment-eligible patients exhibited LDL-C ≥160 mg/dL, which alone qualifies them for optional drug treatment if liberal criteria are applied. Proportionally more females (79%) than males (64%) qualify for therapy under this scenario.
The present study demonstrates that under a liberal treatment scenario, 140% more individuals will be eligible for lipid-lowering therapy under the new guidelines than with the previous guidelines and that more than one fourth of these individuals will also be targeted for more aggressive treatment. In contrast to the NCEP II recommendations, proportionally many more younger (<45 years) and elderly (≥65) patients will be treated. With regard to sex, proportionally more males than females will be treated under the new guidelines, particularly if the conservative criteria are applied.
The finding of increased eligibility for younger patients is driven by the age-cholesterol and age-smoking weights provided in the Framingham score sheets.3 For example, young (20 to 34 years) patients with elevated total cholesterol are assigned higher scores than older (≥35 years) patients with similar cholesterol levels. This is also true for the sex-smoking relationship. It seems counterintuitive that younger persons with comorbid conditions such as smoking and hypercholesterolemia should be assigned greater risk scores than older persons with these same comorbid conditions. Using the previously published Framingham risk functions,6 we obtained risk estimates of 5.8% and 9.7% for NHANES III smoking males aged 20 to 29 years and 30 to 39 years, respectively. Similarly, we obtained risk estimates of 8% and 9.4% for hypercholesterolemic (total cholesterol ≥280 mg/dL) NHANES III males aged 20 to 29 years and 30 to 39 years, respectively. It is unclear why the interaction of these covariates modeled in the score sheets would reverse this relationship.
For patients ≥65 years old, the increase in eligibility is due to the age-risk relationship inherent in the Framingham risk functions and is reflected in the age score weights. For both males and females, the risk score rises linearly with age.
Many patients who fall outside the range of the major statin primary prevention trial participants will now be eligible for more aggressive lipid-lowering treatment. For example, patients included in the West of Scotland Coronary Prevention Study were 45 to 64 years of age,7 and patients enrolled in the Air Force/Texas Coronary Atherosclerosis Prevention Study were 45 to 73 years old.8 Clearly, the age distributions of patients in these studies differs from the age distributions of the proposed treatment-eligible population (20 to 79 years). We found that 40% of the eligible patients fall outside the age range for these trials. Furthermore, the Framingham risk for the younger (<45 years) patients (1.3% for females and 6.2% for males) is much less than half that of the patients aged 45 to 73 years (6.3% for females and 16.7% for males). Given the anticipated relative risk reduction associated with primary prevention statin therapy (≈30% to 40%),7,8 the actual benefit in terms of absolute risk reduction will be modest (0.3% for females and 2.5% for males) in the younger age group. Although trials are under way to determine the benefit of lipid lowering in elderly patient populations, no formal studies are ongoing for younger patients to quantify this benefit. The generalizability of the primary prevention trial results, particular to younger patients, must be explored.
Large differences were seen in the number and sex distributions of patients eligible for treatment under the conservative versus the liberal scenario. As shown in Table 7, most patients, especially the younger patients, would qualify for drug therapy under the liberal scenario because of LDL-C levels alone. Over 71% of treatment-eligible patients exhibited LDL-C ≥160 mg/dL. Among women, nearly 80% would be treated on the basis of an LDL-C level ≥160 mg/dL. Therefore, most of the increases in treatment eligibility under the liberal scenario will be due to revised LDL-C target levels and not to the Framingham risk estimate. This approach differs considerably from that previously adopted by the European Guidelines Committee.9 That algorithm foregoes drug therapy in patients with lower absolute risk (<20%) despite elevated LDL-C levels (≥190 mg/dL). Under the current NCEP III guidelines, many low-risk patients are still eligible for treatment if LDL-C levels exceed 160 mg/dL.
When conservative criteria are applied (ie, higher target LDL-C), the absolute risk estimate provides greater weight, which also explains the sex differences in eligibility. Under NCEP II, treatment was indicated in roughly equal numbers of males and females, whereas under NCEP III, proportionally more men than women are eligible. Although these differences in risk are factual, this runs counter to the current opinion that cardiovascular risk should be regarded as equally relevant in females and males. This finding may have important health policy implications.
With regard to the application of liberal versus conservative criteria, the “real world” estimate for the numbers of patients actually treated lies somewhere in between. Assuming that lifestyle recommendations would not be followed, it is anticipated that the increase in the number of treated patients would reflect the liberal rather than the conservative scenario. Although consideration is also given to the number and prognostic strength of known cardiovascular risk factors, our analysis demonstrates that most treatment decisions will be premised primarily on the basis of LDL-C levels and not on absolute risk. However, it is laudable that the committee does emphasize the importance of the impact that lifestyle modification will have on all modifiable risk factors. To the extent that these changes affect LDL-C, they might reduce the number of treatment-eligible patients. Therapy to lower blood pressure and prevent diabetes will also have an impact on treatment eligibility but to a lesser extent than LDL-C because of the construct of the algorithm.
There are some limitations to the use of NHANES III data for population projections. The survey was conducted between 1988 and 1994. The prevalence and frequency of cardiovascular risk factors and hyperlipidemias may have changed since the time of the survey. In many patients, the risk factor categories (smoking, hypertension, hypertension treatment, and diabetes) were self-reported and not confirmed. There was no information available on estrogen replacement therapy or premature menopause, which would affect the NCEP II eligibility estimates. Also, the categorical age of first-degree relatives with premature CHD (<65 years for females, <55 years for males) could not be determined precisely. We therefore used the category that was based on respondents reporting a first-degree blood relative having a heart attack before age 50. Finally, our estimates may differ from those derived from other population or trial sources owing to differences in risk factor definitions and projection methodology.
In conclusion, the present study suggests that not only the frequency but also the age distributions of treatment-eligible patients will be greatly altered by the adoption of the new NCEP III guidelines. Furthermore, aggressive therapy will be initiated in patients not previously considered for this type of treatment, and many elderly and younger patients will be so treated. The new guidelines will also alter the sex distribution of treatment-eligible patients. Consideration of the health policy implications of these new guidelines should be addressed concurrently with their adoption.
Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. 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; 285: 2486–2497.
Vital and health statistics, plan and operation of the third national health and nutrition and examination survey, 1988–1994. Rockville, MD: US Dept of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Health Statistics; 1994:4–485. DHHS publication PHS 94-1308, series 1, No. 32.
Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972; 18: 499–502.
Wilson PWF, D’Agostino R, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation. 1998; 97: 1837–1847.
Task Force Report. Prevention of coronary heart disease in clinical practice: recommendations of the second joint task force of European and JOURNAL cardiovascular societies on coronary prevention. Eur Heart J. 1998; 19: 1434–1503.