Author + information
- Received May 20, 2014
- Revision received July 27, 2014
- Accepted August 3, 2014
- Published online December 2, 2014.
- Thomas M. Maddox, MD, MSc∗,†∗ (, )
- William B. Borden, MD‡,
- Fengming Tang, MS§,
- Salim S. Virani, MD, PhD‖,
- William J. Oetgen, MD, MBA¶,
- J. Brendan Mullen, BSFS#,
- Paul S. Chan, MD, MSc§,
- Paul N. Casale, MD∗∗,
- Pamela S. Douglas, MD††,
- Fredrick A. Masoudi, MD, MSPH∗,†,
- Steven A. Farmer, MD, PhD‡ and
- John S. Rumsfeld, MD, PhD∗,†
- ∗Cardiology Section, VA Eastern Colorado Health Care System, Denver, Colorado
- †University of Colorado School of Medicine, Aurora, Colorado
- ‡George Washington University, Washington, DC
- §Mid-America Heart Institute, Kansas City, Missouri
- ‖Michael E. DeBakey VA Medical Center, Baylor College of Medicine, Houston, Texas
- ¶American College of Cardiology, Washington, DC
- #National Quality Forum, Washington, DC
- ∗∗Lancaster General Health, Lancaster, Pennsylvania
- ††Duke University Medical Center, Durham, North Carolina
- ↵∗Reprint requests and correspondence:
Dr. Thomas M. Maddox, VA Eastern Colorado Health Care System, Cardiology Section (111B), 1055 Clermont Street, Denver, Colorado 80220.
Background In a significant update, the 2013 American College of Cardiology/American Heart Association (ACC/AHA) cholesterol guidelines recommend fixed-dose statin therapy for those at risk and do not recommend nonstatin therapies or treatment to target low-density lipoprotein cholesterol (LDL-C) levels, limiting the need for repeated LDL-C testing.
Objectives The goal of this study was to examine the impact of the 2013 ACC/AHA cholesterol guidelines on current U.S. cardiovascular practice.
Methods Using the NCDR PINNACLE (National Cardiovascular Data Registry Practice Innovation and Clinical Excellence) registry data from 2008 to 2012, we assessed current practice patterns as a function of the 2013 cholesterol guidelines. Lipid-lowering therapies and LDL-C testing patterns by patient risk group (atherosclerotic cardiovascular disease [ASCVD], diabetes, LDL-C ≥190 mg/dl, or an estimated 10-year ASCVD risk ≥7.5%) were described.
Results Among a cohort of 1,174,545 patients, 1,129,205 (96.1%) were statin-eligible (91.2% ASCVD, 6.6% diabetes, 0.3% off-treatment LDL-C ≥190 mg/dl, 1.9% estimated 10-year ASCVD risk ≥7.5%). There were 377,311 patients (32.4%) not receiving statin therapy and 259,143 (22.6%) receiving nonstatin therapies. During the study period, 20.8% of patients had 2 or more LDL-C assessments, and 7.0% had more than 4.
Conclusions In U.S. cardiovascular practices, 32.4% of statin-eligible patients, as defined by the 2013 ACC/AHA cholesterol guidelines, were not currently receiving statins. In addition, 22.6% were receiving nonstatin lipid-lowering therapies and 20.8% had repeated LDL-C testing. Achieving concordance with the new cholesterol guidelines in patients treated in U.S. cardiovascular practices would result in significant increases in statin use, as well as significant reductions in nonstatin therapies and laboratory testing.
- cardiovascular diseases
- cohort studies
- hydroxymethylglutaryl-coA reductase inhibitors
- primary prevention
- secondary prevention
In 2013, at the request of the National Heart, Lung, and Blood Institute of the National Institutes of Health, the American College of Cardiology (ACC) and American Heart Association (AHA) updated the guidelines on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults (1–3). This update was a significant departure from the previous Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel [ATP] III) guidelines (4), and focused primarily on randomized controlled trials, consistent with the recent recommendations by the Institute of Medicine for guideline development (5,6). The new guidelines redefined patient populations for treatment, targeting those with confirmed atherosclerotic cardiovascular disease (ASCVD), diabetes, native low-density lipoprotein cholesterol (LDL-C) levels ≥190 mg/dl, or 10-year cardiovascular risk ≥7.5%. The new guidelines recommended a “treat to risk” strategy using fixed-dose statin medications, rather than the previous “treat to LDL-C target” strategy; did not recommend use of nonstatin therapies; and did not recommend treatment to target LDL-C lipid levels, thus rendering repeated on-treatment testing unnecessary.
The potential impact of the new guidelines on current U.S. cardiovascular practice is unknown. Because cardiologists typically treat patients at the highest risk for cardiac events, optimizing cholesterol management in light of these new guidelines would be expected to have a significant impact. Although there has been work published on the population impact of these new guidelines (7), important questions remain unanswered. In particular, little is known about current lipid-lowering therapies and LDL-C testing patterns; this knowledge would help quantify expected shifts in care and subsequent implications for statin use, nonstatin use, and LDL-C testing among risk groups.
Accordingly, we examined the implications of the 2013 ACC/AHA cholesterol guidelines on current lipid-lowering therapy and testing patterns in contemporary cardiology practices using data from the NCDR PINNACLE (National Cardiovascular Data Registry Practice Innovation and Clinical Excellence). The PINNACLE registry collects continuous, real-time, clinical information on all patients treated in participating outpatient cardiology practices in the Unites States. As such, current lipid-lowering therapy and LDL-C testing patterns can be assessed, and the shifts expected under the new guidelines can be predicted. This study sought to determine the prevalence in the PINNACLE registry of patients meeting eligibility criteria for statin therapy under the new guidelines and to assess their current therapy and LDL-C testing patterns. Using registry data from 2008 to 2012, patients were classified by risk group (ASCVD, diabetes, LDL-C ≥190 mg/dl, or an estimated 10-year ASCVD risk ≥7.5%) and lipid-lowering therapies and LDL-C testing patterns were described.
The NCDR PINNACLE served as the study data source. Cardiology practices voluntarily participate in and submit data to PINNACLE as part of a national office-based cardiovascular quality improvement program (8). Data are collected on all patients with hypertension, coronary artery disease, heart failure, and/or atrial fibrillation. The data are collected at the point of care through a validated electronic medical record mapping algorithm designed to comprehensively capture required data elements or, in rare cases when electronic health records are not in use, a paper chart abstraction form. These data elements include demographics, insurance status, and detailed clinical information including symptoms, medical conditions, vital signs, medications, and laboratory values. Registry data quality assurance is maintained through rigorous data definitions, standard data collection and transmission, and periodic data quality checks (9,10).
All patients aged 18 years or older with clinical encounters in the PINNACLE registry (version 1.2) between January 1, 2008, and December 31, 2012, were identified. Patients with insufficient data to determine their risk group, as outlined in the 2013 cholesterol guidelines, were excluded. Patients were then categorized into 5 mutually exclusive risk groups on the basis of the 2013 cholesterol guidelines: ASCVD; diabetes (without ASCVD); off-treatment LDL-C ≥190 mg/dl (without ASCVD or diabetes); 10-year ASCVD risk ≥7.5% (without ASCVD, diabetes, or off-treatment LDL-C ≥190 mg/dl); or no risk criteria. ASCVD criteria included: medical record documentation of coronary artery disease; peripheral arterial disease; previous stroke or transient ischemic attack; unstable angina; stable angina; myocardial infarction; coronary artery bypass grafting; and/or percutaneous coronary intervention. Next, patients without evidence of ASCVD, ages 40 to 75 years, and with medical record documentation of diabetes were identified. Patients without evidence of ASCVD or diabetes, on no lipid-lowering therapies, and with ≥1 LDL-C value ≥190 mg/dl were then identified. Patients without ASCVD or diabetes or an off-treatment LDL-C ≥190 mg/dl and ages 40 to 75 years were next identified and had their estimated 10-year ASCVD risk calculated using the pooled risk calculator referenced in the 2013 cholesterol guidelines (1). Those with an estimated 10-year risk ≥7.5% were included in the fourth group. Finally, those patients who did not meet the criteria for any of the 4 risk groups were classified as “no risk criteria.” To determine the final risk group classification for patients with multiple risk group qualifications over time, all clinic visits recorded in PINNACLE for each patient were examined for each risk group classification in a sequential, hierarchical fashion. For example, if 2 years after his or her diabetes diagnosis, a diabetic patient were noted to develop ASCVD, that patient would be placed in the ASCVD group for analysis. If none of the risk group criteria were met by the final recorded visit in PINNACLE, then that patient was placed in the “no risk criteria” group. The visit where the final risk group classification criteria were met was designated as the index visit.
The PINNACLE registry included data on 1,711,326 patients treated in 111 U.S. cardiovascular practice clinics from 2008 to 2012 (Figure 1). In total, 536,781 patients were excluded from the study cohort: 6,561 patients were excluded because they were <18 years of age; and 530,220 patients were excluded because they could not be mapped to 1 of the 5 patient groups (465,234 [87.7%] were excluded due to missing LDL-C values, 50,431 [9.5%] due to missing race, 4,704 [0.9%]) due to missing tobacco use, and 9,851 [1.9%] for other miscellaneous causes). The final study cohort comprised 1,174,545 patients.
Clinical outcomes and characteristics
The primary study outcomes were lipid-lowering therapies and LDL-C testing patterns among the 5 patient risk groups. To determine therapy patterns, the 5 patient risk groups outlined were stratified into 4 categories: those receiving no lipid-lowering therapy; those receiving nonstatin lipid-lowering therapies only; those receiving statins as their sole lipid-lowering therapy; and those receiving both statin and nonstatin therapies. Examples of nonstatin lipid-lowering therapies include fibrates, cholestyramine, and nicotinic acid. Receipt of therapy was determined by examining all clinic visits recorded after the index visit, as we have already defined. For example, if statin therapy was noted at any visit after the index visit and nonstatin therapy was never noted at any visit after the index visit, then the patient was classified in the statin therapy–only group. We also excluded patients with documented reasons for nonprescription of lipid-lowering therapies from our analyses to ensure that our calculations represented prescription rates among eligible patients. Specifically, patients with documented nonprescription reasons for statins, nonstatins, or both were excluded from the no lipid-lowering therapy and both statin and nonstatin therapy groups. Patients with nonprescription reasons for statins only were excluded from the statin-only group. Patients with nonprescription reasons for nonstatins only were excluded from the non-statin-only group.
To determine LDL-C testing patterns, we collected the number of LDL-C assessments obtained per patient, both overall and by each patient group. All assessments collected at the index visit and those at subsequent visits were counted. For those patients in the “no risk criteria” group, all LDL-C assessments collected at any visit recorded in PINNACLE were counted. For those patients with >1 assessment on a particular day, only 1 was tabulated, as this likely indicated double counting.
Descriptive analyses were performed for the 5 patient risk groups. Demographic, clinical, and medication characteristics were described and compared. Variables were compared using 1-way analyses of variances. Adjustments for multiple comparisons were not made. Next, among those patients with an indication for statin therapy under the new guidelines, those receiving no therapies, those receiving nonstatin lipid-lowering therapies, those receiving statin therapies, and those receiving both statin and nonstatin therapies were calculated. Therapy rates by patient risk group were also calculated. Finally, LDL-C assessments were calculated and grouped by ascending number.
Several secondary analyses were conducted to support or clarify our primary analyses. First, we excluded all patients with heart failure symptoms of New York Heart Association functional class III or IV severity because the guidelines do not recommend statin therapy in this patient population (1). Second, we restricted the categorization of those in the 10-year estimated ASCVD risk to those with a risk ≥10% to address concerns about the potential overestimation of risk of the pooled risk calculator (11). Third, we expanded the categorization of those in the 10-year estimated ASCVD risk group to those with a risk ≥5% because the guidelines indicate that statin therapy confers a net absolute benefit in this group of patients, thus warranting its consideration. Fourth, we measured LDL-C levels <100 mg/dl and <70 mg/dl among those patients without any lipid-lowering therapies because providers may have justifiably elected to not prescribe lipid-lowering therapies for these patients, in line with ATP-3 guidelines. All analyses were performed with SAS (version 9.3, SAS Institute, Cary, North Carolina).
Patient risk groups and characteristics
In the study cohort of 1,174,545 patients, 1,129,205 (96.1%) patients met criteria for statin therapy: 1,029,633 (91.2%) patients had ASCVD; 75,046 (6.6%) were diabetic patients without ASCVD; 3,176 (0.3%) had an off-treatment LDL-C ≥190 mg/dl without ASCVD or diabetes; and 21,350 (1.9%) had an estimated 10-year ASCVD risk ≥7.5% without ASCVD, diabetes, or LDL ≥190 mg/dl. Demographic and clinical characteristics of each patient group are listed in Table 1. In general, more men than women qualified for the ASCVD or estimated 10-year ASCVD risk ≥7.5% risk groups, whereas more women than men qualified for the diabetes, LDL-C ≥190 mg/dl, or no risk criteria risk groups. Among the ASCVD risk group, 25.3% had diabetes. The median estimated 10-year ASCVD event risk was 20.6% for the ASCVD group, 16.2% for the diabetes group, 12.5% for the LDL group, 13.8% for the estimated risk ≥7.5% group, and 15.6% for the no risk criteria group (the median score is 15.6% because, in line with guideline recommendations, patients over 75 years of age are not included in the estimated risk ≥7.5% group).
Treatment categories by patient group
Patients’ lipid-lowering therapies were assessed, both overall and by risk group. As described in the methods section, to determine rates of therapy use among eligible patients, we excluded patients with contraindications to lipid-lowering therapies. Overall, 341,669 eligible patients (29.3%) were not receiving any lipid-lowering therapy, 35,642 (3.1%) were receiving nonstatin therapies, 563,685 (48.7%) were receiving statin therapies, and 223,501 (19.5%) were receiving both statin and nonstatin therapies (Figure 2). Treatment patterns by patient risk group were also assessed. Among the ASCVD risk group, 285,211 eligible patients (27.9%) were not receiving any lipid-lowering therapy, 28,887 (2.9%) were receiving nonstatin therapies, 506,009 (49.9%) were receiving statin therapies, and 200,789 (20.0%) were receiving both statin and nonstatin therapies. Among the diabetic risk group, 26,827 eligible patients (35.9%) were not receiving any lipid-lowering therapy, 3,540 (4.7%) were receiving nonstatin therapies, 32,392 (43.5%) were receiving statin therapies, and 11,927 (16.1%) were receiving both statin and nonstatin therapies. Among the off-treatment LDL-C ≥190 mg/dl risk group, 916 eligible patients (29.3%) were not receiving any lipid-lowering therapy, 194 (6.2%) were receiving nonstatin therapies, 1,384 (45.1%) were receiving statin therapies, and 630 (20.5%) were receiving both statin and nonstatin therapies. Among the estimated 10-year ASCVD risk ≥7.5% risk group, 7,491 eligible patients (35.5%) were not receiving any lipid-lowering therapy, 1,092 (5.2%) were receiving nonstatin therapies, 8,699 (41.5%) were receiving statin therapies, and 3,799 (18.2%) were receiving both statin and nonstatin therapies. Among the no risk criteria group, 21,224 eligible patients (47.5%) were not receiving any lipid-lowering therapy, 1,929 (4.3%) were receiving nonstatin therapies, 15,201 (34.2%) were receiving statin therapies, and 6,356 (14.4%) were receiving both statin and nonstatin therapies.
LDL-C testing patterns by patient group
LDL-C assessments were calculated, both overall and by patient group (Figure 3). Overall, 594,771 patients (50.6%) had no assessments, 335,233 (28.5%) had 1 assessment, and the remaining 244,541 (20.8%) had ≥2 assessments. Of the 244,541 patients with ≥2 assessments, 82,037 (7.0%) had >4 assessments. LDL-C assessments by patient risk group were also calculated. Among the ASCVD risk group, 540,937 patients (52.5%) had no assessments, 267,686 (26.0%) had 1 assessment, and the remaining 221,010 (21.5%) had ≥2 assessments. Among the diabetic group, 53,500 patients (71.3%) had no assessments, 16,462 (21.9%) had 1 assessment, and the remaining 5,084 (6.8%) had ≥2 assessments. Among the off-treatment LDL ≥190 mg/dl group, 46 patients (1.4%) had no assessments, 2,327 (73.3%) had 1 assessment, and the remaining 803 (25.3%) had ≥2 assessments. Among the estimated 10-year ASCVD risk ≥7.5% group, 152 patients (0.7%) had no assessments, 15,759 (73.8%) had 1 assessment, and the remaining 5,439 (25.5%) had ≥2 assessments. Among the no risk criteria group, 136 patients (0.3%) had no assessments, 32,999 (72.8%) had 1 assessment, and the remaining 12,205 (26.9%) had ≥2 assessments.
Exclusion of patients with New York Heart Association functional class III to IV
In our study cohort, 14,478 patients (1.2%) had heart failure symptoms of New York Heart Association functional class III to IV severity. Lipid-lowering therapy rates did not significantly vary after their exclusion from the cohort (Figure 4). There were 338,005 eligible patients (29.4%) not receiving any lipid-lowering therapy, 35,185 (3.1%) receiving nonstatin therapies, 558,484 (48.7%) receiving statin therapies, and 218,537 (19.3%) receiving both statin and nonstatin therapies.
Estimated 10-year ASCVD risk ≥10% threshold
When we used an estimated 10-year ASCVD risk ≥10% threshold for the primary prevention risk calculation, 5,344 (0.5%) of our cohort moved from the estimated 10-year ASCVD risk ≥7.5% prevention group to the no risk criteria group. The rates of lipid-lowering therapies among these groups did not significantly change from the primary analysis (Figure 5). Among the estimated 10-year ASCVD risk ≥10% group, 5,473 eligible patients (34.7%) were not receiving any lipid-lowering therapy, 833 (5.3%) were receiving nonstatin therapies, 6,574 (41.8%) were receiving statin therapies, and 2,904 (18.6%) were receiving both statin and nonstatin therapies. Among the no risk criteria group, 23,164 eligible patients (46.4%) were not receiving any lipid-lowering therapy, 2,183 (4.4%) were receiving nonstatin therapies, 17,277 (34.8%) were receiving statin therapies, and 7,290 (14.8%) were receiving both statin and nonstatin therapies.
Estimated 10-year ASCVD risk ≥5% threshold
When we used an estimated 10-year ASCVD risk ≥5% threshold for the primary prevention risk calculation, 6,608 of our cohort patients (0.6%) moved from the no risk criteria group to the estimated 10-year ASCVD risk ≥5% prevention group. The rates of lipid-lowering therapies among these groups did not change significantly from the primary analysis (Figure 6). Among the estimated 10-year ASCVD risk ≥5% group, 10,209 eligible patients (37.0%) were not receiving any lipid-lowering therapy, 1,366 (5.0%) were receiving nonstatin therapies, 11,114 (40.4%) were receiving statin therapies, and 4,923 (18.0%) were receiving both statin and nonstatin therapies. Among the no risk criteria group, 18,639 eligible patients (48.7%) were not receiving any lipid-lowering therapy, 1,671 (4.4%) were receiving nonstatin therapies, 12,808 (34.8%) were receiving statin therapies, and 5,193 (13.7%) were receiving both statin and nonstatin therapies.
LDL-C targets <100 mg/dl and <70 mg/dl
In our primary cohort, 341,669 eligible patients (29.3%) were not receiving any lipid-lowering therapies. Among these patients, 73,604 (21.5%) had an LDL-C assessment, of which 43,097 (12.6%) had an LDL-C value <100 mg/dl and 7,029 (2.1%) had an LDL-C value <70 mg/dl. Therefore, 12.6% of eligible patients not on lipid-lowering therapies appeared to be treated in line with the ATP-3 guidelines.
We assessed the potential impact of the new 2013 cholesterol guidelines on current lipid-lowering treatment and LDL-C testing patterns in a large, contemporary clinical registry analysis of U.S. cardiology practices (Central Illustration). We demonstrated that the majority of patients in the PINNACLE registry qualify for statin therapy under the new guidelines, primarily for secondary prevention of cardiac events. However, 32.4% of these patients were not receiving statin therapy. In addition, 22.6% were receiving nonstatin therapies, including 3.1% receiving nonstatins as their sole lipid-lowering therapy. Moreover, repeated LDL-C testing occurred in 20.8% of patients, which may no longer be needed given the shift away from a “treat to LDL target” strategy (1). Overall, these findings suggest the existence of significant gaps in secondary prevention for contemporary cardiac patients, and the implementation of the 2013 cholesterol guidelines will prompt significant changes in their current management.
The ATP III guidelines, published in 2001, focused on LDL-C cholesterol lowering as a treatment strategy in managing patients either with or at high risk for coronary heart disease (4). They did not specifically call for statins as a lipid-lowering therapy, and many clinicians employed multiple nonstatin lipid-lowering therapies in an attempt to achieve the recommended LDL-C targets. This “treat to LDL-C target” strategy also necessitated frequent LDL-C assessments. In 2008, the National Heart, Lung, and Blood Institute convened the ATP IV to review and update the cholesterol guidelines to reduce cardiovascular risk. In 2013, the full panel transitioned to the ACC/AHA Expert Panel to complete the guideline update (2,3). In accordance with best practices advocated by the National Heart, Lung, and Blood Institute Systematic Evidence Review, the panel considered randomized controlled trials as the most appropriate source of evidence on which to base their recommendations (5,6). Only 1 approach to reducing cardiovascular events through cholesterol treatment, the use of fixed-dose lipid-lowering drugs, had been evaluated with randomized controlled trials, and almost every trial used statin medications exclusively. Importantly, none of these trials directly evaluated the effect of adjusting statin or other lipid-lowering therapies to achieve specific LDL-C targets, but they demonstrated that fixed-dose statin therapy provides consistent relative risk reduction among all patient subgroups. Accordingly, the panel determined that 4 patient risk groups warranted moderate-dose to high-dose statin therapy: those with ASCVD; diabetes; LDL ≥190 mg/dl; and/or a 10-year estimated ASCVD risk ≥7.5%. The panel did not find evidence to support the use of nonstatin lipid-lowering therapies (such as fibrates or niacin), and thus did not recommend their use. Finally, though the panel indicated that LDL-C assessments should be used to gauge initial ASCVD risk and could be used to check for patient adherence to statin therapy, its routine assessment to adjust lipid-lowering dosing to specific LDL-C targets is no longer needed.
These new cholesterol guidelines have significant implications for both cardiac patients and the general medical population. Using a national sample of the U.S. population, Pencina et al. (7) found that an additional 12.8 million U.S. adults (11.1%) ages 40 to 75 years of age would qualify for statin therapy under these new guidelines. Our findings complement this work by examining the implications of the guidelines for patients seen in U.S. cardiology practices. As expected, the vast majority of these patients has confirmed cardiovascular disease and thus is at the highest risk for recurrent cardiac events. Despite this need, one-third of these patients were not currently treated with a statin. Furthermore, almost one-quarter of patients were receiving nonstatin lipid-lowering therapies, which are no longer indicated for cardiac event prevention under the new guidelines. Finally, 20.8% of patients received ≥2 LDL-C assessments, and 7.0% had ≥4 assessments, testing that may no longer be as necessary in a “treat to risk,” rather than a “treat to LDL-C target” environment. The secondary analyses excluding severe heart failure patients and varying the primary prevention risk threshold for treatment did not significantly alter these primary findings. Given the controversy surrounding the guideline risk model and its potential for risk overestimation (11,12), it is particularly important that our findings were not significantly altered by varying risk thresholds. In addition, the finding that 12.6% of patients on no lipid-lowering therapy had LDL-C values <100 mg/dl underscores the differences between ATP-3 and ATP-4 and illustrates how these new guidelines will expand the indications for statin therapy.
These findings have several implications for cardiac patients. As the provider and patient community move toward adopting these guidelines, increased statin use, decreased nonstatin use, and decreasing LDL-C testing are expected. Accordingly, if the 377,311 eligible patients in the PINNACLE population not currently receiving statin therapy receive them as a result of these guideline changes, then those patients would have a 25% reduction, on average, in cardiovascular events (13). This benefit would be especially pronounced among those with diabetes or an estimated ASCVD risk ≥7.5%, who were noted to have the highest rates of nonstatin use (36.2% and 36.6%, respectively) in this population. Next, as the number of nonstatin lipid-lowering therapies is reduced, the side effects of nonstatin lipid-lowering therapies (especially myopathic symptoms, which have a higher incidence among patients treated with both statins and either fibrates or niacin) would be expected to decline. In addition, among those patients with no risk criteria, 48.6% were currently receiving statin therapy, which may no longer be indicated under the new guidelines. Finally, the cost and inconvenience of repeated LDL-C testing to titrate statin medication to specific LDL-C targets would be reduced. The overall net cost effect of the new guidelines is complex, as the cost of statin use must be considered against the savings from reduced cardiovascular event occurrence, reduced use of nonstatin therapies, reduced use of statins among patients without indication, and reduced testing.
First, the study population is restricted to cardiology practices that voluntarily participate in the PINNACLE registry. Therefore, the patients in our study may have greater cardiovascular risk and higher lipid-lowering therapy rates than the general population. In addition, the PINNACLE registry participating practices may represent providers with a particular interest in quality improvement efforts. As a result, they may not be representative of all cardiology or primary care practices in the United States. However, we demonstrate significant opportunities for optimizing treatment among PINNACLE practices, suggesting that treatment gaps in the broader cardiology and general medical communities may be even larger. Second, the PINNACLE registry extracts data directly from clinical electronic health records, and lipid-lowering therapies and lipid panel assessments may be incompletely captured. However, PINNACLE conducts periodic data quality assessments to mitigate this potential issue. Third, the current PINNACLE registry does not contain information on statin type or dosing, which could provide additional insight into expected changes from the updated guidelines. In addition, renal function data is not available in the current registry, which prevents us from understanding the number of patients on hemodialysis who would not qualify for statin therapy. The next version of PINNACLE data will collect information on both statin type and dosing as well as renal function, and future analyses will provide insights into those questions. Fourth, if patients were prescribed therapies before their index visit and the electronic medical record failed to carry those prescriptions forward, the method used to categorize lipid-lowering therapy for each patient could have underestimated current lipid-lowering therapy rates. This misclassification would, if anything, result in an even larger gap between current practice and guideline-concordant care than this analysis indicates, further underscoring the need for action. Fifth, LDL-C assessments are often conducted in venues other than cardiology practices, such as primary care and neurology practices. As such, the analysis of repeated LDL-C assessment may underestimate the number of repeated assessments. Finally, many LDL-C levels in the cohort were measured in the setting of lipid-lowering therapy, which means that the estimated risk ≥7.5% patient risk group may be underestimated and the no risk criterion groups may be overestimated. This was partially accounted for in the sensitivity analysis that reduced the risk threshold to ≥5% and (as with the misclassification possibility) continues to highlight that actual practice may have even more significant gaps in therapy than are indicated by this analysis.
This study found that in U.S. cardiovascular practices participating in the PINNACLE registry, 32.4% of patients with an indication for statins under the 2013 ACC/AHA cholesterol guideline were not currently receiving them. In addition, 22.6% were receiving nonstatin therapies. Moreover, repeated lipid testing occurred in 20.8% of patients. Achieving concordance with the new cholesterol guidelines would result in significant increases in statin use and might also lead to significant reductions in nonstatin therapies and laboratory testing.
COMPETENCY IN MEDICAL KNOWLEDGE: The 2013 blood cholesterol guidelines recommend statin treatment for patients with ASCVD, diabetes, LDL-C levels ≥190 mg/dl, or a 10-year ASCVD risk ≥7.5%. LDL-C assessments to titrate statin therapy to specific LDL-C goals and nonstatin treatments are no longer recommended.
COMPETENCY IN SYSTEMS-BASED PRACTICE: Physicians and other healthcare professionals should examine their cholesterol treatment and testing patterns to assess alignment with current guidelines.
TRANSLATIONAL OUTLOOK: Clinical outcomes studies are needed to gauge the effect of application of blood cholesterol on cardiac events in well-defined practice-based patient populations.
The authors acknowledge Ms. Megan Petrich, MPH, for her editorial assistance.
Dr. Maddox is supported with a VA Health Services Research and Development career development award. Dr. Oetgen is the Executive Vice President for Science, Education, and Quality of the American College of Cardiology. Dr. Masoudi is the Senior Medical Officer for the National Cardiovascular Data Registry; and has a contract with the American College of Cardiology. Dr. Rumsfeld is the Chief Science Officer for the National Cardiovascular Data Registry. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- American College of Cardiology
- American Heart Association
- atherosclerotic cardiovascular disease
- Adult Treatment Panel
- low-density lipoprotein cholesterol
- Received May 20, 2014.
- Revision received July 27, 2014.
- Accepted August 3, 2014.
- American College of Cardiology Foundation
- Stone N.J.,
- Robinson J.G.,
- Lichtenstein A.H.,
- et al.,
- for the American College of Cardiology/American Heart Association Task Force on Practice Guidelines
- Gibbons G.H.,
- Shurin S.B.,
- Mensah G.A.,
- Lauer M.S.
- Gibbons G.H.,
- Harold J.G.,
- Jessup M.,
- Robertson R.M.,
- Oetgen W.J.
- Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults
- ↵Eden J, Levit L, Berg A, et al., for the Institute of Medicine, editors. Finding What Works in Health Care: Standards for Systematic Reviews. Washington, DC: National Academies Press, 2011.
- Graham R, Wolman DM, Greenfield S, et al., for the Institute of Medicine, editors. Clinical Practice Guidelines We Can Trust. Washington, DC: National Academies Press, 2011.
- Messenger J.C.,
- Ho K.K.,
- Young C.H.,
- et al.,
- for the NCDR Science and Quality Oversight Committee Data Quality Workgroup
- Masoudi F.A.,
- Ponirakis A.,
- Yeh R.W.,
- et al.
- Lloyd-Jones D.M.,
- Goff D.,
- Stone N.J.