Heterogeneity in Statin Indications Within the 2013 American College of Cardiology/American Heart Association Guidelines Ravi V. Shah, MDa,b, Melvyn Rubenfire, MDc, Robert D. Brook, MDc, João A.C. Lima, MDd,e, Brahmajee Nallamothu, MD, MPHc, and Venkatesh L. Murthy, MD, PhDc,f,g,* A standard (“core”) implementation of American College of Cardiology/American Heart Association 2013 lipid guidelines (based on 10-year risk) dramatically increases the statin-eligible population in older Americans, raising controversy in the cardiovascular community. The guidelines also endorse a more “comprehensive” risk approach based in part on lifetime risk. The impact of this broader approach on statin eligibility remains unclear. We studied the impact of 2 different implementations of the new guidelines (“core” and “comprehensive”) using the National Health and Nutrition Examination Survey. Although “core” guidelines led to 72.0 million subjects qualifying for statin therapy, the broader “comprehensive” application led to nearly a twofold greater estimate for statin-eligible subjects (121.2 million), with the greatest impact among those aged 21 to 45 years. Subjects indicated for statin therapy under comprehensive guidelines had a greater burden of cardiovascular risk factors and a higher lifetime risk of cardiovascular disease than those not indicated for statins. In particular, men aged 21 to 45 years had a 3.13-fold increased odds of being eligible for statin therapy only under the “comprehensive” guidelines (vs standard “core” guidelines; 95% confidence interval 2.82 to 3.47, p 7.5% (based on the Pooled Cohort Equations [PCE]). However, the guidelines also recognize an optional, broader criterion for treatment in primary prevention, including basing therapeutic decisions on abnormalities in emerging risk factors such as highsensitivity C-reactive protein, coronary artery calcification, family history, and lifetime risk. It is important to understand a Department of Medicine and b Cardiovascular Institute, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; c Division of Cardiovascular Medicine, Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan; Departments of dCardiology and eRadiology, Johns Hopkins Hospital, Baltimore, Maryland; and Divisions of fNuclear Medicine and gCardiothoracic Imaging, Department of Radiology, University of Michigan, Ann Arbor, Michigan. Manuscript received August 16, 2014; revised manuscript received and accepted September 26, 2014. See page 32 for disclosure information. *Corresponding author: Tel: (734) 936-5387; fax: (734) 232-3246. E-mail address: [email protected] (V.L. Murthy).
0002-9149/14/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.09.045
how many subjects may qualify based on this broader reading of the guidelines and what types of patients will be most affected. We sought to investigate the incremental increase in statin prescriptions based on health care providers following the optional “comprehensive” application of the 2013 ACC/AHA guidelines to identify subpopulations in which statin implementation may be emphasized. Methods The National Health and Nutrition Examination Survey (NHANES) is an ongoing research program conducted by the National Center for Health Statistics aimed at assessing the health and nutritional status of noninstitutionalized, civilian adults and children in the United States using a combination of interviews, physical examinations, and diagnostic testing.3 The complex survey design of NHANES assigns sample weights to each participant, allowing estimation of subgroup sizes, proportions, and characteristics among the overall noninstitutionalized United States population. The overall study design was approved by the National Center for Health Statistics’ Institutional Review Board and informed consent was obtained for all participants. We combined data from the 2007 to 2008 and 2009 to 2010 data collection cycles to define the populations indicated for statin therapy under Third Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (ATP3) and under the 2013 ACC/AHA guideline updates. Data from participants in morning sessions (n ¼ 4,910, adults aged 21 years) were used, as fasting lipid www.ajconline.org
The American Journal of Cardiology (www.ajconline.org) 100
Age Neither ATP3 nor AHA 2013
76+
71−75
66−70
61−65
56−60
51−55
46−50
41−45
36−40
21−25
76+
71−75
66−70
61−65
0 56−60
0 51−55
25
46−50
25
41−45
50
36−40
50
31−35
75
26−30
75
Comprehensive
31−35
Core
21−25
% with Indication for Statin
100
26−30
28
Age AHA 2013 Only
Both ATP3 and AHA 2013
ATP3 Only
Figure 1. Distribution of subjects in the United States who are statin eligible based on core (left) and comprehensive (right) criteria, stratified by age.
profiles were not available from afternoon session participants. We used sample weights developed by the National Center for Health Statistics to enable computation of population estimates from only the morning session subjects. Subjects were considered to have diabetes if their fasting glucose was 126 mg/dl, hemoglobin A1c was 6.5%, or if it was reported by history. Prediabetes was considered present among nondiabetics if hemoglobin A1c was 5.7% or fasting glucose was 100 mg/dl. Subjects were considered to have known CVD if they reported being told by a health professional that they had coronary heart disease, angina, heart attack, or stroke. We used the 2004 update of the ATP3 and AHA 2013 recommendations to define which subjects qualified for statin therapy under ATP3, AHA 2013, both, or neither guidelines (Supplementary Figure 1). All patients who self-reported being treated with lipid-lowering medications were presumed to meet indications for both ATP3 and ACC/ AHA 2013 as pretreatment lipid profiles were not available, consistent with previous studies.2 In addition to “core” criteria described in Figure 2 of the ACC/AHA 2013 guidelines, we also applied “comprehensive” criteria described in Figure 4 of the ACC/AHA 2013 guidelines. For the “comprehensive” criteria, we did not include coronary artery calcium score or ankle-brachial indexes, as these data were not collected in the NHANES cycles used for this analysis. Additionally, we considered family history of premature CVD to be history of myocardial infarction in a first degree relative before age 50 to coincide with NHANES data collection. The 10-year risk of myocardial infarction or coronary death and atherosclerotic CVD were computed in accordance with the Framingham Heart Study hard coronary heart disease risk equations4 and the PCE,5 respectively. For the ATP3 risk equation, the continuous risk equations for age, gender, total- and high-density lipoprotein cholesterol, systolic blood pressure, and smoking status were used using coefficients obtained from the Framingham Heart Study website to maximize accuracy and precision of risk estimates.6 Race- and gender-specific PCE were used for nonHispanic whites and non-Hispanic blacks. For Hispanics and all other nonwhite or nonblack racial groups, the genderspecific PCE for non-Hispanic whites was used, as recommended in the guidelines. Subjects were considered indicated under “core” ACC/ AHA 2013 guidelines if they had established CVD, LDL
190 mg/dl, were aged 40 to 75 years with diabetes and LDL 70 mg/dl, were aged 40 to 75 years with PCE risk 7.5% and LDL 70 mg/dl, or were being treated with cholesterol lowering medications. Subjects were considered indicated under “comprehensive” guidelines if they met “core” criteria or if they were aged 40 to 75 years with PCE risk 5% to 7.5% with LDL 70 mg/dl, had LDL 160 mg/dl at any age, had high sensitivity C-reactive protein levels 2.0 mg/dl, were aged 21 to 59 years with high lifetime risk of CVD (defined for this study as 50%), or had early family history of CVD in a first degree relative (defined as myocardial infarction before age 50 for this study). Although ankle-brachial index and coronary artery calcification criteria are also endorsed by the guidelines, these data were not available in the NHANES cycles analyzed for this study. Values are presented as means with 95% confidence intervals [CIs]. All statistical tests were performed using SAS 9.4 (SAS Institute Inc., Cary, North Carolina). Type I error (a) threshold was set at 0.05. Analytic methods accounted for the complex survey design of NHANES. Domain analysis was used to estimate parameters for subgroups and to exclude those aged
0002-9149/14/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjcard.2014.09.045
how many subjects may qualify based on this broader reading of the guidelines and what types of patients will be most affected. We sought to investigate the incremental increase in statin prescriptions based on health care providers following the optional “comprehensive” application of the 2013 ACC/AHA guidelines to identify subpopulations in which statin implementation may be emphasized. Methods The National Health and Nutrition Examination Survey (NHANES) is an ongoing research program conducted by the National Center for Health Statistics aimed at assessing the health and nutritional status of noninstitutionalized, civilian adults and children in the United States using a combination of interviews, physical examinations, and diagnostic testing.3 The complex survey design of NHANES assigns sample weights to each participant, allowing estimation of subgroup sizes, proportions, and characteristics among the overall noninstitutionalized United States population. The overall study design was approved by the National Center for Health Statistics’ Institutional Review Board and informed consent was obtained for all participants. We combined data from the 2007 to 2008 and 2009 to 2010 data collection cycles to define the populations indicated for statin therapy under Third Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults (ATP3) and under the 2013 ACC/AHA guideline updates. Data from participants in morning sessions (n ¼ 4,910, adults aged 21 years) were used, as fasting lipid www.ajconline.org
The American Journal of Cardiology (www.ajconline.org) 100
Age Neither ATP3 nor AHA 2013
76+
71−75
66−70
61−65
56−60
51−55
46−50
41−45
36−40
21−25
76+
71−75
66−70
61−65
0 56−60
0 51−55
25
46−50
25
41−45
50
36−40
50
31−35
75
26−30
75
Comprehensive
31−35
Core
21−25
% with Indication for Statin
100
26−30
28
Age AHA 2013 Only
Both ATP3 and AHA 2013
ATP3 Only
Figure 1. Distribution of subjects in the United States who are statin eligible based on core (left) and comprehensive (right) criteria, stratified by age.
profiles were not available from afternoon session participants. We used sample weights developed by the National Center for Health Statistics to enable computation of population estimates from only the morning session subjects. Subjects were considered to have diabetes if their fasting glucose was 126 mg/dl, hemoglobin A1c was 6.5%, or if it was reported by history. Prediabetes was considered present among nondiabetics if hemoglobin A1c was 5.7% or fasting glucose was 100 mg/dl. Subjects were considered to have known CVD if they reported being told by a health professional that they had coronary heart disease, angina, heart attack, or stroke. We used the 2004 update of the ATP3 and AHA 2013 recommendations to define which subjects qualified for statin therapy under ATP3, AHA 2013, both, or neither guidelines (Supplementary Figure 1). All patients who self-reported being treated with lipid-lowering medications were presumed to meet indications for both ATP3 and ACC/ AHA 2013 as pretreatment lipid profiles were not available, consistent with previous studies.2 In addition to “core” criteria described in Figure 2 of the ACC/AHA 2013 guidelines, we also applied “comprehensive” criteria described in Figure 4 of the ACC/AHA 2013 guidelines. For the “comprehensive” criteria, we did not include coronary artery calcium score or ankle-brachial indexes, as these data were not collected in the NHANES cycles used for this analysis. Additionally, we considered family history of premature CVD to be history of myocardial infarction in a first degree relative before age 50 to coincide with NHANES data collection. The 10-year risk of myocardial infarction or coronary death and atherosclerotic CVD were computed in accordance with the Framingham Heart Study hard coronary heart disease risk equations4 and the PCE,5 respectively. For the ATP3 risk equation, the continuous risk equations for age, gender, total- and high-density lipoprotein cholesterol, systolic blood pressure, and smoking status were used using coefficients obtained from the Framingham Heart Study website to maximize accuracy and precision of risk estimates.6 Race- and gender-specific PCE were used for nonHispanic whites and non-Hispanic blacks. For Hispanics and all other nonwhite or nonblack racial groups, the genderspecific PCE for non-Hispanic whites was used, as recommended in the guidelines. Subjects were considered indicated under “core” ACC/ AHA 2013 guidelines if they had established CVD, LDL
190 mg/dl, were aged 40 to 75 years with diabetes and LDL 70 mg/dl, were aged 40 to 75 years with PCE risk 7.5% and LDL 70 mg/dl, or were being treated with cholesterol lowering medications. Subjects were considered indicated under “comprehensive” guidelines if they met “core” criteria or if they were aged 40 to 75 years with PCE risk 5% to 7.5% with LDL 70 mg/dl, had LDL 160 mg/dl at any age, had high sensitivity C-reactive protein levels 2.0 mg/dl, were aged 21 to 59 years with high lifetime risk of CVD (defined for this study as 50%), or had early family history of CVD in a first degree relative (defined as myocardial infarction before age 50 for this study). Although ankle-brachial index and coronary artery calcification criteria are also endorsed by the guidelines, these data were not available in the NHANES cycles analyzed for this study. Values are presented as means with 95% confidence intervals [CIs]. All statistical tests were performed using SAS 9.4 (SAS Institute Inc., Cary, North Carolina). Type I error (a) threshold was set at 0.05. Analytic methods accounted for the complex survey design of NHANES. Domain analysis was used to estimate parameters for subgroups and to exclude those aged
