0
Original Contribution |

Improving the Prediction of Coronary Heart Disease to Aid in the Management of High Cholesterol Levels:  What a Difference a Decade Makes FREE

Andrew L. Avins, MD, MPH; Warren S. Browner, MD, MPH
[+] Author Affiliations

From the General Internal Medicine Section, Veterans Affairs Medical Center, San Francisco, Calif, and the Department of Epidemiology and Biostatistics, University of California, San Francisco.


JAMA. 1998;279(6):445-449. doi:10.1001/jama.279.6.445.
Text Size: A A A
Published online

Context.— A patient's coronary heart disease (CHD) risk must be correctly classified to successfully apply risk-based guidelines for treatment of hypercholesterolemia.

Objective.— To determine the classification accuracy of the National Cholesterol Education Program (NCEP) CHD risk-stratification system and compare it with a simple revised system that gives greater weight to age as a CHD risk factor.

Design.— Modeling of 10-year CHD risk, using equations from the Framingham Heart Study applied to a cross-sectional survey of the US population.

Subjects.— The 3284 subjects aged 20 to 74 years surveyed in the Second National Health and Nutrition Examination Survey (1978-1982) who had fasting lipid levels measured.

Main Outcome Measures.— The area under the receiver operating characteristic curve (AUC) for 10-year CHD risk for the NCEP and revised scales.

Results.— Among all adults with a low-density lipoprotein cholesterol value of at least 4.1 mmol/L (160 mg/dL), the NCEP system showed fairly good discrimination (AUC=0.90), though there was a substantial decline among men 35 to 74 years old and women 55 to 74 years old (AUC=0.81). By contrast, the revised system showed superior performance in all hypercholesterolemic adults (AUC=0.94-0.97) as well as in the subgroup of men 35 to 74 years old and women 55 to 74 years old (AUC=0.94-0.96).

Conclusions.— Simple modifications of the NCEP treatment criteria result in a substantially improved ability to discriminate between higher and lower CHD risk groups. Unlike the NCEP system, this revised system retains its classification ability in all age groups studied.

Figures in this Article

THE NATIONAL Cholesterol Education Program (NCEP) has recognized the importance of risk-based guidelines in its policies regarding the treatment of high blood cholesterol levels for the prevention of coronary heart disease (CHD).1 The logic underlying this strategy is 2-fold.2 First, individuals at lower risk of CHD have less likelihood of benefiting from cholesterol-reducing medical therapy but are at the same risk of potential adverse medication effects as higher-risk patients.3 Second, the cost-effectiveness of lipid-reducing therapy is greater when treatment is targeted toward individuals at higher risk.4

The successful application of risk-based treatment recommendations is greatly dependent on being able to distinguish higher-risk from lower-risk individuals. The updated NCEP guidelines (NCEP II) make this distinction by defining a set of criteria that, if present in sufficient number and proper combination, identify an individual as being at high risk for CHD.1 Because cholesterol screening affects millions of US adults, it is important to examine how well the guidelines succeed in discriminating between higher- and lower-risk groups.

A study of the earlier NCEP criteria suggested that the protocols may not adequately stratify individuals by CHD risk,5 since considerable overlap existed in the possible ranges of risk among those recommended for treatment and those not so advised. This analysis suggested that serious misclassification could exist but did not estimate the extent of misclassification, which depends on the actual distributions of CHD risk in the US population.

One problem with the NCEP guidelines is an inadequate emphasis on the importance of advancing age in the estimation of CHD risk.5 Age is a powerful determinant of risk 6,7 and should play a central role in CHD risk stratification. Therefore, we developed a revised CHD risk-scoring system that places greater emphasis on age. We then compared the performance of this revised system with that of the NCEP II to determine the relative abilities of the scales to discriminate between higher-risk and lower-risk individuals.

Comparing the classification performance of the NCEP II system with a revised scale required estimating the 10-year risk of CHD for each subject in a population-based data set, and generating a measure of discrimination with which to test each scale. We limited our analysis to adults aged 20 to 74 years without a history of CHD.

Defining Scale Criteria

The NCEP II algorithm for intervention is detailed in Table 1. 1 In this system, men younger than 35 years and premenopausal women are recommended for treatment if their low-density lipoprotein (LDL) level is at least 5.7 mmol/L (220 mg/dL) or if their LDL cholesterol level is at least 4.9 mmol/L (190 mg/dL) and at least 2 other cardiac risk factors are present. The same strategy is recommended for older persons, with LDL cholesterol thresholds reduced by 0.8 mmol/L (30 mg/dL).

Table Graphic Jump LocationTable 1.—NCEP II Treatment Criteria*

For the revised scale (Table 2), we substituted the following rule for the NCEP age criteria: men are assigned 1 point for every decade of age, beginning at 35 years, and women acquire age points beginning at 45 years. Thus, a woman is assigned 1 point if she is 45 to 54 years old, 2 points if she is 55 to 64 years old, and so on. One additional point is added if the LDL cholesterol level is at least 5.2 mmol/L (200 mg/dL). The other risk factors (hypertension, diabetes, smoking, and low high-density lipoprotein [HDL] cholesterol) are retained as specified in the NCEP II guidelines, as is the "negative risk factor" of a high HDL cholesterol level. For example, a 60-year-old woman (+2 points) with hypertension (+1 point), an LDL cholesterol level of 4.1 mmol/L (160 mg/dL) (no additional points), and an HDL cholesterol level of 1.8 mmol/L (70 mg/dL) (−1 point) would receive a total of 2 points.

Table Graphic Jump LocationTable 2.—Revised Treatment Criteria*

Under the revised system, treatment of elevated cholesterol levels is indicated for persons with an LDL cholesterol level of at least 4.1 mmol/L (160 mg/dL) and a specified number of high-risk points. We performed separate analyses of the revised scale for "higher-risk" cutoffs of 2, 3, or 4 points.

Assigning CHD Risk to Population-Based Data

To estimate the distribution of CHD risk in US adults aged 20 to 74 years, we used data from the Second National Health and Nutrition Examination Survey (NHANES II), a population-based survey of the US noninstitutionalized civilian population, carried out between 1978 and 1982.8 We included the 3284 subjects who had full measurements of lipids following at least a 9-hour fast. Because data from the more recent NHANES III had not yet been released at the time of this analysis, we used published summary estimates from NHANES III to adjust total cholesterol and HDL values for changes that have occurred since NHANES II was performed.9 Expansion weights were used in all calculations and 1996 US Bureau of the Census population projections were used to convert percentages of the population into absolute numbers.10

We used the Framingham equations6 to estimate the 10-year CHD risk for each respondent in the NHANES II data set. This value was calculated by the method of Anderson et al6 as follows: 1 − exp(−exp[(ln(10) − µ)/σ]), where µ = 4.4284+m, and σ = exp(−0.3171 − 0.2825 × m). For men, m = a − 1.6346 × ln(age) −0.2082 × diabetes; for women, m = a − 6.5306 + 2.1059 × (ln[age/74])2 − 0.4055 × diabetes. For both men and women, a = 11.0938 − 0.8670 × ln(diastolic blood pressure) − 0.2789 × smoking − 0.7142 × ln(total cholesterol/HDL cholesterol) − 0.7195 × ECG-LVH, where "smoking" is defined as current smoker and "ECG-LVH" is the presence of left ventricular hypertrophy on the electrocardiogram (set to 0 since this variable was not present in the NHANES II data set).

Measurement of Discrimination

We used the area under the receiver operating characteristic (ROC) curve as the primary measure of the classification performance of the scales.11,12 Unlike typical ROC curves, which plot sensitivity against 1−specificity, these curves plot the positive predictive value against 1−negative predictive value. The ROC curves were constructed by first ranking all NHANES II respondents according to their CHD risk. As the cut point separating "higher-risk" from "lower-risk" persons was incrementally increased from a 10-year CHD risk of 0% to 100%, positive and negative predictive values were calculated. For example, if the cut point between higher- and lower-risk individuals is chosen as a 10-year risk of 10%, the positive predictive value of the NCEP system is 0.993 while the 1−negative predictive value is 0.638. If the cut point is increased to a 10-year risk of 11%, the corresponding values are 0.989 and 0.611. Thus, these curves incorporate the changing prevalence of "higher-risk" individuals as the cut point of CHD risk increases. The area under the curve (AUC) was calculated with the trapezoidal rule. All analyses were performed with Stata statistical software, version 5.0.13

These AUCs can be interpreted as follows: if 1 individual is randomly selected from the group labeled "higher-risk" and another is randomly chosen from the group labeled "lower-risk, " the AUC is the probability that the higher-risk individual has a CHD risk greater than that of the lower-risk individual. This interpretation was verified by Monte Carlo simulations. A scale with perfect discrimination would have an AUC of 1.0; a scale with no discrimination ability would have an AUC of 0.5.14 Ninety-five percent confidence intervals (CIs) for the AUCs were calculated by bootstrap methods, which make no distributional assumptions about the variables.15

Levels of CHD risk for groups of individuals are expressed as the 10th to 90th percentiles of 10-year risk of a CHD event. The full range of risk is not presented since all scales include a few very low-risk and very high-risk persons, obscuring differences in the general levels of risk associated with the application of the different systems.

These analyses were performed for 2 groups: all adults aged 20 to 74 years with an LDL cholesterol level of at least 4.1 mmol/L (160 mg/dL), and all men 35 to 74 years of age and women 55 to 74 years of age with an elevated LDL cholesterol level (≥4.1 mmol/L [160 mg/dL]).

Validation

In a separate analysis, we used data from the Lipid Research Clinics Prevalence and Follow-up Study (LRC)16 to test the performance of the revised and NCEP systems. This investigation was a prospective cohort study of risk factors for CHD death, beginning in 1977. The available LRC data set contained data on 2919 hypercholesterolemic (LDL, ≥4.1 mmol/L [160 mg/dL]) subjects who were at least 30 years old and free of CHD at baseline, and who were followed for an average of 12.2 years. Ascertainment of deaths was 99% complete; 131 CHD deaths were identified during the follow-up period.

In these analyses, each subject in the LRC data set was designated "higher-risk" or "lower-risk" according to the NCEP and revised criteria (using 3 points as the cutoff for the latter). We then calculated the sensitivity and specificity of each scale for the outcome of definite or suspected CHD death (note that sensitivity and specificity, rather than predictive values, were calculated since this data set contained actual clinical outcomes, not simply estimated CHD risk). Comparison of the sensitivities, specificities, and numbers of subjects labeled "higher-risk" was performed with McNemar test for matched pairs.

Among all adults aged 20 to 74 years, the NCEP II criteria showed good classification performance, with an AUC of 0.90 (Table 3). In large part, this was due to the correct classification of most young adults as lower risk. Among men aged 35 to 74 years and women aged 55 to 74 years, the scale performed less well, with the AUC falling to 0.81.

Table Graphic Jump LocationTable 3.—Area Under the ROC Curve, Number of Persons Recommended for Lipid-Lowering Treatment, and Levels of CHD Risk Under Alternative Risk-Stratification Scales*

The revised scales performed substantially better than the NCEP criteria (Table 3). Depending on the number of points used as the criterion for "higher risk, " the AUCs ranged from 0.94 to 0.97 for all adults and 0.94 to 0.96 for the older subgroup. In nearly all cases, there was no overlap in the CIs of the 2 classification systems. Visual comparison of the ROC curves generated by using the NCEP II system and the revised system (using the 3-point cutoff) for the older hypercholesterolemic subgroup demonstrates a substantially greater AUC for the revised system (Figure 1).

Graphic Jump Location
Figure 1.—Receiver operating characteristic (ROC) curves for the updated National Cholesterol Education Program guidelines (NCEP II) and the revised system (3-point cutoff), for the subgroup of men at least 35 years old and women at least 55 years old with a low-density lipoprotein cholesterol level of at least 4.1 mmol/L (160 mg/dL) in the NHANES II (Second National Health and Nutrition Examination Survey) data set. Dashed line represents lower limit of a ROC curve where the area under the curve equals 0.5.

Reflecting the explicit incorporation of advancing age as a CHD risk factor, those individuals identified as "higher-risk" tended to be older under the revised system than under the NCEP II system (Figure 2). Note that there are fewer treated individuals in the oldest age group (ages 65-74 years) than in the next younger group only because there are fewer persons in the older age group in the United States.

Graphic Jump Location
Figure 2.—Age distributions of those designated "higher-risk" by the updated National Cholesterol Education Program guidelines (NCEP II) and the revised system (3-point cutoff) among persons 20 to 74 years old with a low-density lipoprotein cholesterol level of at least 4.1 mmol/L (160 mg/dL) in the NHANES II (Second National Health and Nutrition Examination Survey) data set.

In the subgroup of all adults with an LDL cholesterol level of at least 4.1 mmol/L (160 mg/dL), slightly fewer individuals would be treated using the revised system (with a 3-point cutoff) relative to that of the NCEP, with a corresponding increase in levels of CHD risk (Table 3). For the revised system, a substantially greater number of individuals would be candidates for drug treatment if a cutoff of 2 points were chosen; many fewer would be so eligible if a cutoff of 4 points were used.

Other modifications of the revised system are possible. Finer gradations of change would occur in the levels of risk and number of individuals designated "higher-risk" if the number of risk factors used as the cutoff is varied from 2 to 4 and the starting age cut point is lowered or raised by 5 years (Table 4). For example, if the revised system is used with a 3-point cutoff and an age criteria starting at 40 years for men and 50 years for women, the AUC for this scale would be 0.96. The overall levels of risk found in the lower-risk (untreated) and higher-risk (treated) groups would increase because of the older age criteria, and there would be a corresponding decline in the number of persons recommended for hypolipidemic therapy. Regardless of how the revised system is modified, the degree of overlap in CHD risk between the lower-risk group and the higher-risk group is substantially less for the revised system than for the NCEP II system.

Table Graphic Jump LocationTable 4.—Performance of NCEP II System Compared With Variations in Point Cutoff and Age Criteria for Revised System*

In the validation analyses, the revised system showed significantly greater sensitivity and specificity than the NCEP system(Table 5). Using a 3-point cutoff, the revised system identified a greater proportion of those who ultimately died from CHD, while labeling significantly fewer individuals as "higher-risk."

Table Graphic Jump LocationTable 5.—Comparison of NCEP II and Revised Systems (Using 3-Point Cutoff) for Subjects in Lipid Research Clinics Prevalence and Follow-up Study*

The compelling logic of risk-based treatment guidelines cannot be fully exploited without an adequate method of separating higher-risk individuals from those at lower risk. If serious misclassification is present, many higher-risk individuals would not be identified, denying them appropriate therapy, and many lower-risk individuals would be subject to overtreatment with expensive drugs having uncertain long-term safety.

For example, using the NCEP II criteria, a 25-year-old woman with an LDL cholesterol level of 5.7 mmol/L (220 mg/dL) and no other cardiac risk factors would be recommended to receive cholesterol-lowering interventions, while a 60-year-old man with an LDL cholesterol level of 4.1 mmol/L (160 mg/dL) and no other cardiac risk factors would not. Yet the young woman has a predicted 10-year CHD risk of about 0.5%, while the older man's risk is approximately 16%, more than 30 times greater. If such inconsistencies were common in the application of the NCEP II system, treatment of hypercholesterolemia would be seriously misdirected.

We found that simple modifications of the NCEP II classification system substantially improved the ability to distinguish higher-risk from lower-risk individuals. For example, the AUC for the NCEP II scale applied to men at least 35 years old and women at least 55 years old was only 0.81, nearly midway between a perfect scale and one that has no discrimination ability. This observation is particularly disturbing, since it is in these age groups that hypolipidemic drug therapy is most commonly considered.1 Among these same individuals, however, use of the revised scale greatly improved the classification ability, with AUCs around 0.95. This absolute improvement of 0.14 is substantial, since the values for the AUC range only from 0.5 to 1.0.

Use of the revised system leads to some interesting and, perhaps, counterintuitive outcomes. For example, using the 3-point cutoff implies that all men at least 55 years old with an LDL cholesterol level of at least 4.1 mmol/L (160 mg/dL) and an HDL cholesterol level of less than 1.6 mmol/L (60 mg/dL) should be offered cholesterol-lowering therapy, a more aggressive recommendation than that of the NCEP. Though these men have an average 10-year CHD risk of 19%, one sixth of them (those with no other cardiac risk factors) are designated "lower-risk" by the NCEP II classification system and are not recommended to receive cholesterol-reducing treatment.

Decisions regarding the appropriate age at which to initiate cholesterol-lowering therapy are difficult. Older individuals have greater absolute risk for CHD, but the association between hypercholesterolemia and CHD is attenuated in the elderly.17 In post hoc subgroup analyses from 3 recent clinical trials, older higher-risk individuals appeared to benefit from treatment with a hydroxymethylglutaryl-CoA reductase inhibitor (statin). At the time of randomization, subjects aged 55 to 64 years,18 60 to 70 years,19 and 60 to 75 years20 showed significant benefits of therapy with statins. In 2 of these trials, however, the magnitude of benefit was attenuated in the older subgroups, 18,19 though in the third study, the opposite was true.20

The efficacy of therapy in older individuals, who are at greater absolute CHD risk, is critical to determining the optimal age at which to begin medical therapy for elevated cholesterol levels, as is consideration of all relevant costs and benefits. Some have argued that hypolipidemic therapy should be provided to younger patients for greater preventive effectiveness.21 However, this strategy requires treating larger numbers of people for much longer periods with costly drugs that currently have limited evidence of long-term safety. Clinical trials currently under way should provide important information regarding the relative and absolute benefits of cholesterol-lowering therapy in older individuals; all strategies for treating high cholesterol levels must be revised as more data become available. At this point, given the paucity of data in individuals older than 75 years, we do not recommend extrapolation of the revised scale to older individuals.

Modifications to the revised scale can be made, depending on one's view regarding the level of CHD risk that warrants hypocholesterolemic drug therapy. For example, if the cut point were increased from 3 to 4 points, the 10th to 90th percentiles of CHD risk in the higher-risk group would increase from 10% to 30% to 14% to 32%, with a corresponding decrease in the number of higher-risk persons from about 15 million to 9.2 million (Table 4). If, instead, the cutoff were kept at 3 points but the age cut points were increased from 35 to 40 years for men and from 45 to 50 years for women, more modest changes would result. In this case, the 10th to 90th percentile of risk in the higher-risk group would increase to about 12% to 32%, with a reduction in the number of higher-risk individuals to approximately 11.4 million (Table 4).

Limitations of these analyses should be noted. First, they are based on the Framingham risk equations. While these formulas were carefully developed, they were derived from a predominantly middle-class white population in 1 region of the United States,22 and may not generalize well to other populations. Given the need for accurate classification, further development and validation of CHD risk-estimation models is urgently needed. Second, these risk-estimation techniques do not incorporate other CHD risk-reduction strategies, such as prophylactic aspirin therapy.2 Without considering these alternative interventions, all variants of these scales may overestimate an individual's true CHD risk. The NHANES II data that we used are now several years old. To address this concern, we adjusted the lipid levels for respondents in NHANES II to reflect recent changes. We did not include family history of premature CHD as a risk factor. This variable was not included in the Framingham risk equations, and it is not clear how much independent CHD risk (beyond that attributed to the other cardiac risk factors) is associated with such a family history.23

In summary, simple modifications of the current NCEP-recommended system for classifying CHD risk result in a substantially improved ability to discriminate higher-risk from lower-risk individuals. This revised system can be easily incorporated into routine clinical practice and can be adapted by individual clinicians to reflect their personal views regarding risk thresholds for the treatment of elevated cholesterol levels. Use of the revised system could result in a more consistent approach to the management of elevated cholesterol levels for the prevention of CHD.

Adult Treatment Panel II.  Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults.  JAMA.1993;269:3015-3023.
Avins AL, Browner WS. Lowering risk without lowering cholesterol: implications for national cholesterol policy.  Ann Intern Med.1996;125:502-506.
Davey-Smith G, Egger M. Who benefits from medical interventions?  BMJ.1994;308:72-74.
Goldman L, Gordon DJ, Rifkind BM.  et al.  Cost and health implications of cholesterol lowering.  Circulation.1992;85:1960-1968.
McIsaac WJ, Naylor CD, Basinski A. Mismatch of coronary risk and treatment intensity under the National Cholesterol Education Program guidelines.  J Gen Intern Med.1991;6:518-523.
Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professionals.  Circulation.1991;83:356-362.
Garber AM, Browner WS, Hulley SB. Cholesterol screening in asymptomatic adults, revisited.  Ann Intern Med.1996;124:518-531.
McDowell A, Engle A, Massey JT, Maurer K. Plan and Operation of the Second National Health and Nutrition Examination Survey 1976-1980.  Washington, DC: US Government Printing Office; 1981.
Johnson CL, Rifkind BM, Sempos CT.  et al.  Declining serum total cholesterol levels among US adults: the National Health and Nutrition Examination Surveys.  JAMA.1993;269:3002-3008.
US Bureau of the Census.  United States Population Estimates, by Age, Sex, Race, and Hispanic Origin, 1996 to 2000.  Washington, DC: US Bureau of the Census; 1996.
Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine.  Clin Chem.1993;39:561-577.
Swets JA. Measuring the accuracy of diagnostic systems.  Science.1988;240:1285-1293.
StataCorp.  Stata Statistical Software: Release 5.0.  College Station, Tex: Stata Corp; 1996.
Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve.  Radiology.1982;143:29-36.
Mossman D. Resampling techniques in the analysis of non-binormal ROC data.  Med Decis Making.1995;15:358-366.
Jacobs DR, Mebane IL, Bangdiwala SI, Criqui MH, Tyroler HA. High density lipoprotein cholesterol as a predictor of cardiovascular disease mortality in men and women: the follow-up study of the Lipid Research Clinics Prevalence Study.  Am J Epidemiol.1990;131:32-47.
Manolio TA, Pearson TA, Wenger NK, Barrett-Connor E, Payne GH, Harlan WR. Cholesterol and heart disease in older persons and women.  Ann Epidemiol.1992;2:161-176.
Shepherd J, Cobbe SM, Ford I.  et al.  Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia: West of Scotland Coronary Prevention Study Group.  N Engl J Med.1995;333:1301-1307.
Pedersen TR, Kjekshus J, Olsson A.  et al.  Scandinavian Simvistatin Study (4S).  Lancet.1994;344:1765-1768.
Sacks FM, Pfeffer MA, Moye LA.  et al.  The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels.  N Engl J Med.1996;335:1001-1009.
Forrester JS, Shah PK. Using serum cholesterol as a screening test for preventing coronary heart disease: the five fundamental flaws of the American College of Physicians guidelines.  Am J Cardiol.1997;79:790-792.
Glick H, Heyse JF, Thompson D, Epstein RS, Smith ME, Oster G. A model for evaluating the cost-effectiveness of cholesterol-lowering treatment.  Int J Technol Assess Health Care.1992;8:719-734.
Perkins KA. Family history of coronary heart disease: is it an independent risk factor?  Am J Epidemiol.1986;124:182-194.

Figures

Graphic Jump Location
Figure 1.—Receiver operating characteristic (ROC) curves for the updated National Cholesterol Education Program guidelines (NCEP II) and the revised system (3-point cutoff), for the subgroup of men at least 35 years old and women at least 55 years old with a low-density lipoprotein cholesterol level of at least 4.1 mmol/L (160 mg/dL) in the NHANES II (Second National Health and Nutrition Examination Survey) data set. Dashed line represents lower limit of a ROC curve where the area under the curve equals 0.5.
Graphic Jump Location
Figure 2.—Age distributions of those designated "higher-risk" by the updated National Cholesterol Education Program guidelines (NCEP II) and the revised system (3-point cutoff) among persons 20 to 74 years old with a low-density lipoprotein cholesterol level of at least 4.1 mmol/L (160 mg/dL) in the NHANES II (Second National Health and Nutrition Examination Survey) data set.

Tables

Table Graphic Jump LocationTable 1.—NCEP II Treatment Criteria*
Table Graphic Jump LocationTable 2.—Revised Treatment Criteria*
Table Graphic Jump LocationTable 3.—Area Under the ROC Curve, Number of Persons Recommended for Lipid-Lowering Treatment, and Levels of CHD Risk Under Alternative Risk-Stratification Scales*
Table Graphic Jump LocationTable 4.—Performance of NCEP II System Compared With Variations in Point Cutoff and Age Criteria for Revised System*
Table Graphic Jump LocationTable 5.—Comparison of NCEP II and Revised Systems (Using 3-Point Cutoff) for Subjects in Lipid Research Clinics Prevalence and Follow-up Study*

References

Adult Treatment Panel II.  Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults.  JAMA.1993;269:3015-3023.
Avins AL, Browner WS. Lowering risk without lowering cholesterol: implications for national cholesterol policy.  Ann Intern Med.1996;125:502-506.
Davey-Smith G, Egger M. Who benefits from medical interventions?  BMJ.1994;308:72-74.
Goldman L, Gordon DJ, Rifkind BM.  et al.  Cost and health implications of cholesterol lowering.  Circulation.1992;85:1960-1968.
McIsaac WJ, Naylor CD, Basinski A. Mismatch of coronary risk and treatment intensity under the National Cholesterol Education Program guidelines.  J Gen Intern Med.1991;6:518-523.
Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professionals.  Circulation.1991;83:356-362.
Garber AM, Browner WS, Hulley SB. Cholesterol screening in asymptomatic adults, revisited.  Ann Intern Med.1996;124:518-531.
McDowell A, Engle A, Massey JT, Maurer K. Plan and Operation of the Second National Health and Nutrition Examination Survey 1976-1980.  Washington, DC: US Government Printing Office; 1981.
Johnson CL, Rifkind BM, Sempos CT.  et al.  Declining serum total cholesterol levels among US adults: the National Health and Nutrition Examination Surveys.  JAMA.1993;269:3002-3008.
US Bureau of the Census.  United States Population Estimates, by Age, Sex, Race, and Hispanic Origin, 1996 to 2000.  Washington, DC: US Bureau of the Census; 1996.
Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine.  Clin Chem.1993;39:561-577.
Swets JA. Measuring the accuracy of diagnostic systems.  Science.1988;240:1285-1293.
StataCorp.  Stata Statistical Software: Release 5.0.  College Station, Tex: Stata Corp; 1996.
Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve.  Radiology.1982;143:29-36.
Mossman D. Resampling techniques in the analysis of non-binormal ROC data.  Med Decis Making.1995;15:358-366.
Jacobs DR, Mebane IL, Bangdiwala SI, Criqui MH, Tyroler HA. High density lipoprotein cholesterol as a predictor of cardiovascular disease mortality in men and women: the follow-up study of the Lipid Research Clinics Prevalence Study.  Am J Epidemiol.1990;131:32-47.
Manolio TA, Pearson TA, Wenger NK, Barrett-Connor E, Payne GH, Harlan WR. Cholesterol and heart disease in older persons and women.  Ann Epidemiol.1992;2:161-176.
Shepherd J, Cobbe SM, Ford I.  et al.  Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia: West of Scotland Coronary Prevention Study Group.  N Engl J Med.1995;333:1301-1307.
Pedersen TR, Kjekshus J, Olsson A.  et al.  Scandinavian Simvistatin Study (4S).  Lancet.1994;344:1765-1768.
Sacks FM, Pfeffer MA, Moye LA.  et al.  The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels.  N Engl J Med.1996;335:1001-1009.
Forrester JS, Shah PK. Using serum cholesterol as a screening test for preventing coronary heart disease: the five fundamental flaws of the American College of Physicians guidelines.  Am J Cardiol.1997;79:790-792.
Glick H, Heyse JF, Thompson D, Epstein RS, Smith ME, Oster G. A model for evaluating the cost-effectiveness of cholesterol-lowering treatment.  Int J Technol Assess Health Care.1992;8:719-734.
Perkins KA. Family history of coronary heart disease: is it an independent risk factor?  Am J Epidemiol.1986;124:182-194.
CME
Meets CME requirements for:
Browse CME for all U.S. States
Accreditation Information
The American Medical Association is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The AMA designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 CreditTM per course. Physicians should claim only the credit commensurate with the extent of their participation in the activity. Physicians who complete the CME course and score at least 80% correct on the quiz are eligible for AMA PRA Category 1 CreditTM.
Note: You must get at least of the answers correct to pass this quiz.
You have not filled in all the answers to complete this quiz
The following questions were not answered:
Sorry, you have unsuccessfully completed this CME quiz with a score of
The following questions were not answered correctly:
Commitment to Change (optional):
Indicate what change(s) you will implement in your practice, if any, based on this CME course.
Your quiz results:
The filled radio buttons indicate your responses. The preferred responses are highlighted
For CME Course: A Proposed Model for Initial Assessment and Management of Acute Heart Failure Syndromes
Indicate what changes(s) you will implement in your practice, if any, based on this CME course.
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Multimedia

Some tools below are only available to our subscribers or users with an online account.

Web of Science® Times Cited: 26

Related Content

Customize your page view by dragging & repositioning the boxes below.

Articles Related By Topic
Related Topics
PubMed Articles