Prevalence and Factors Associated With Statin Use Among a Nationally Representative Sample of US Adults: National Health and Nutrition Examination Survey, 2011–2012

Abstract

The 2013 American College of Cardiology/American Heart Association guidelines recommend statins for adults age ≤75 years who have clinical atherosclerotic cardiovascular disease (IA) and adults age 40 to 75 years with diabetes mellitus and LDL‐C 70–189 mg/dl (IA). Our aim was to estimate the prevalence and likelihood of statin use among selected statin benefit groups. Using data from the National Health and Nutrition Examination Survey (NHANES) 2011–2012, we examined 5319 adults age ≥20 years. We estimated weighted frequencies and prevalence of statin use for adults with diabetes mellitus and dyslipidemia (or low‐density lipoprotein cholesterol ≥70 mg/dL), defined as statin benefit group 1 (SBG1); and for adults with atherosclerotic cardiovascular disease, defined as statin benefit group 2 (SBG2). We constructed a logistic regression model to estimate odds of statin use in SBG1. Overall, an estimated 38.6 million Americans are on a statin. In adjusted models, uninsured and Hispanic adults were less likely to be on a statin compared with white adults; 59.5% (95% confidence interval [CI]: 53.0‐66.1) of all adults in SBG1, 58.8% (95% CI: 51.5‐66.1) of adults age 40 to 75 in SBG1, and 63.5% (95% CI: 55.6‐71.4) of all adults in SBG2 were on a statin. Although the prevalence of statin use has increased over time, Hispanic ethnicity and lack of insurance remain barriers to statin use. Black‐white racial disparities were not significant. Our study provides a baseline estimate of statin use in the noninstitutionalized population just prior to introduction of the new guidelines and provides a reference for evaluating the impact of the new guidelines on statin utilization.

1. INTRODUCTION

Multiple randomized controlled trials have shown statins to be effective in lowering blood levels of low‐density lipoprotein cholesterol (LDL‐C) as well as reducing cardiovascular disease (CVD) risk1, 2, 3 through inhibition and regression of coronary atherosclerosis,4 anti‐inflammatory properties,5, 6 and atherosclerotic plaque stabilization.7 Their widespread use in clinical practice has resulted in decreased rates of CVD complications and mortality.8

In 2011, statin use was estimated to cost > $20 billion in the United States.9 This number is likely to increase greatly in future years, based on National Health and Nutrition Examination Survey (NHANES) projections,10 although the exact number cannot be predicted due to the guidelines calling for a risk‐benefit discussion between patient and physician before the initiation of statin therapy in lower‐risk primary‐prevention individuals. On the other hand, more than a third of the US adult population has atherosclerotic cardiovascular disease (ASCVD), which accounts for 35% of all deaths,11 with CVD and stroke generating $312.6 billion in direct and indirect health care costs in 2009.12

In 2013, the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) guidelines for treatment of high blood cholesterol in adults (last published in 2001)13 were revised by the American College of Cardiology/American Heart Association (ACC/AHA). The new guidelines, published in 2014, expand the indications for statin use.14 Recommendations were tailored to reduce the risks of ASCVD events, which include coronary heart disease (CHD), stroke, and peripheral arterial disease.14

The new guidelines recommend the use of statins for 3 high‐risk groups: adults age <75 years who have clinical ASCVD (grade IA recommendation) for secondary prevention of ASCVD; adults age 40 to 75 years with diabetes mellitus (DM) and LDL‐C between 70 and 189 mg/dL (grade IA recommendation) for primary prevention of ASCVD; and those with LDL‐C ≥190 mg/dL.14 The aim of this study was to estimate the prevalence and likelihood of statin use among the first 2 selected statin benefit groups just prior to the release of the new guidelines. This study does not consider the guideline recommendations for determining statin use in asymptomatic, nondiabetic individuals with LDL‐C of 70 to 189 mg/dL.

2. METHODS

2.1. Study Design

NHANES is an annual survey of the noninstitutionalized civilian resident population of the United States. It is a stratified multistage probability survey that provides annual data on a nationally representative sample of US residents. It oversamples Hispanics, non‐Hispanic blacks, low‐income whites, and those age ≥80 years to increase the reliability and precision of estimates of health‐status indicators for these population subgroups. Details of the methodology of the NHANES surveys have been previously published.15 All participants gave written informed consent for the interview, examination, and specimen collection and storage. The survey protocol was reviewed and approved by National Center for Health Statistics (NCHS) Research Ethics Review Board.15

2.2. Study Population

We obtained the 2011–2012 NHANES data,16 the most recent survey cycle available from the US Centers for Disease Control and Prevention (CDC). We examined 5319 adults age ≥20 years who completed the interview, were examined, and had blood samples collected.

2.3. Variables

Our primary independent variable (predictor variable) was a diagnosis of DM and LDL‐C between 70 and 189 mg/dL, which we defined as statin benefit group 1 (SBG1). This group was created based on a self‐reported diagnosis of DM and (1) participant has a measured serum LDL‐C value ≥70 mg/dL or a self‐reported diagnosis of dyslipidemia (“ever been told by a doctor or other health professional that your blood cholesterol level was high?”) or (2) participant has been told to take a prescription to lower cholesterol levels (“To lower your blood cholesterol, have you ever been told by a doctor or other health professional to take prescribed medicine?”). We also created a second category of patients: persons with clinical ASCVD, defined as statin benefit group 2 (SBG2). This group was created based on self‐reported history of heart attack, angina, CHD, or stroke.

Our outcome of interest (dependent variable) was current statin use, defined as statin use in the 30 days prior to the survey. Survey participants were asked if they had taken prescription medications in the past 30 days. If “yes,” they were asked to provide medication containers, and the medication name was documented from the container. If no container was available, they were asked to verbally report the name of the medication.17 If any of the medications reported by a participant in the 30 days prior to the survey was a statin, they were classified as “Yes”; if otherwise, they were classified as “No.” Statin medications reported by study participants included simvastatin, lovastatin, atorvastatin, rosuvastatin, pravastatin, and pitavastatin.

Other independent variables or covariates of interest included age, sex, race/ethnicity, insurance status (includes health insurance obtained through employment or purchased directly, as well as government programs like Medicare and Medicaid), educational level, waist circumference, CHD, stroke, hypertension, smoking, and family history of premature CHD (“Including living and deceased, were any of your close blood relatives, including father, mother, sisters, or brothers, ever told by a health professional that they had a heart attack or angina before the age of 50?”).

2.4. Statistical Analysis

We estimated the total number of US adults who reported using a statin using sample weights provided by the CDC. We calculated the proportion of persons on a statin for the entire study population and among a subpopulation of adults age 40 to 75 years and generated 95% confidence interval (CI) limits for all estimates. We performed bivariate analyses to assess the relationship between covariates and our primary predictor variable using standard parametric tests including χ² and t tests. We then constructed a multivariable logistic regression model to estimate the odds of statin use among persons in SBG1 while controlling for potential confounders. All variables were systematically assessed for inclusion in the final model using a backward elimination approach. Missing variables were excluded from analysis. All analyses were conducted using SAS survey procedures (SAS Institute, Inc., Cary, NC) for complex sample data and data evaluated for significance at α = 0.05.

3. RESULTS

We examined a total of 5319 adults representing an estimated 224 million US adults age ≥20 years in 2011–2012. Overall, 48% were male, 66.4% were non‐Hispanic whites, 11.5% were African American, and 14.3% were Hispanic. Persons in SBG1 had a mean age of 61 years, and 59.5%—an estimated 10.6 million persons—were taking a statin (Table 1). A subpopulation analysis among persons age 40 to 75 years showed statin use to be 58.8% for persons in SBG1 (Table 1). Persons in SBG2 had a mean age of 65 years, and 63.5%—an estimated 10.8 million persons—were on a statin (Table 1).

Table 1.

Prevalence of Statin Use Among a Nationally Representative Sample of US Adults

Variable	Raw Frequency	Weighted Frequency	Weighted % (95% CI)
All adults, age ≥20 years
DM and dyslipidemia	332 (n = 551)	10 557 386	59.53 (52.97‐66.10)
ASCVD	289 (n = 489)	10 844 335	63.48 (55.58‐71.38)
Total population	976 (n = 5319)	38 585 232	17.23 (15.05‐19.41)
Adults, age 40 to 75 years
DM and dyslipidemia	255 (n = 426)	8 461 212	58.82 (51.56‐66.07)
ASCVD	191 (n = 321)	7 815 639	67.45 (59.19‐75.72)
Total population	736 (n = 2957)	30 823 698	24.21 (21.36‐27.06)
Race/ethnicitya
Non‐Hispanic white	399 (n = 1949)	28 879 059	19.41 (17.31‐21.51)
DM and dyslipidemia	109 (n = 179)	6 644 365	61.12 (51.29‐70.94)
ASCVD	144 (n = 227)	8 168 824	67.09 (58.72‐75.47)
Non‐Hispanic black	287 (n = 1405)	4 160 979	16.16 (13.39‐18.92)
DM and dyslipidemia	125 (n = 192)	1 788 584	63.42 (55.82‐71.02)
ASCVD	74 (n = 133)	1 063 102	52.40 (41.30‐63.50)
Hispanic	155 (n = 1061)	2 858 506	8.94 (6.37‐11.50)
DM and dyslipidemia	54 (n = 106)	1 062 795	44.66 (30.34‐58.99)
ASCVD	39 (n = 80)	699 598	42.11 (26.83‐57.39)
Other	135 (n = 904)	2 686 687	15.42 (10.10‐20.74)
DM and dyslipidemia	44 (n = 74)	1 061 643	63.89 (54.45‐73.34)
ASCVD	32 (n = 49)	912 812	75.01 (57.61‐92.42)

Abbreviations: ASCVD, atherosclerotic cardiovascular disease; CI, confidence interval; DM, diabetes mellitus; US, United States.

^aAmong all adults age ≥20 years.

We examined persons in SBG2 further by classifying them into those with composite CAD (includes self‐reported history of heart attack, CAD, or angina) and stroke. We found that among persons with composite CAD age ≥20 years, 70.3% (95% CI: 62.3‐78.4) were on a statin; and in the subgroup of persons age 40 to 75 years, 73.9% (95% CI: 65.3‐82.4) were on a statin. When persons with a history of myocardial infarction alone (“heart attack”) were examined, for persons age ≥20 years and those age 40 to 75 years, 74.7% and 76.3%, respectively, were on a statin. Among persons with a history of stroke, however, 48.9% (95% CI: 39.2‐58.5) of those age ≥20 years and 51.4% (95% CI: 41.5‐61.4) of those age 40 to 75 years were on a statin.

Overall, an estimated 38.7 million Americans were on a statin in 2011–2012, an increase from 24 million in 2003–2004 and from 12.5 million in 1999–2000.18

In bivariate analysis, we found CAD, stroke, hypertension, family history of premature CAD, insurance status, educational level, and waist circumference to be significantly associated with SBG1 (Table 2).

Table 2.

Clinical and Demographic Characteristics by Statin Benefit Group

Covariates	SBG1a
	Yes, n (%)	No, n (%)	P Value
Known risk factors for statin use
CAD	114 (21.13)	220 (4.12)	<0.0001
Stroke	50 (8.05)	166 (2.48)	<0.0001
HTN	394 (72.61)	1516 (28.49)	<0.0001
Smoking	278 (49.72)	1999 (43.54)	0.0676
Family history of premature CADb	81 (17.67)	523 (12.17)	0.0302
Demographics
Age, y, mean (SE)	61.06 (0.68)	46.18 (0.85)	<0.0001
Male sex	285 (48.49)	2335 (47.91)	0.8049
Race
Non‐Hispanic white	179 (61.31)	1770 (66.88)	0.1016
Non‐Hispanic black	192 (15.90)	1213 (11.12)
Hispanic	106 (13.42)	955 (14.36)
Other	74 (9.37)	830 (7.64)
Insured	476 (88.19)	3576 (79.36)	0.0002
Education
≤High school	316 (50.65)	2060 (35.76)	0.0006
Some college	139 (28.23)	1463 (32.49)
≥College graduate	95 (21.11)	1242 (31.76)
Waist circumference, cm, mean (SE)	110.97 (1.21)	97.70 (0.56)	<0.0001

Abbreviations: CAD, coronary artery disease; DM, diabetes mellitus; HTN, hypertension; LDL‐C, low‐density lipoprotein cholesterol; SBG, statin benefit group; SE, standard error.

^aSBG1 represents patients with DM and LDL‐C values >70 mg/dL or a diagnosis of dyslipidemia.

^bFamily history of premature CAD (age of close blood relative with CAD <50 years).

In multivariate models, standard clinical indications for statin use were found to be significantly associated with being on a statin. Persons in SBG1 were as likely as persons with CAD to be on a statin (odds ratios [ORs]: 4.15 and 4.96, respectively; Table 3). However, uninsured and Hispanic persons were significantly less likely to be on a statin medication compared with white persons (ORs: 0.33 and 0.70, respectively; Table 3). Black‐white differences were not statistically significant after controlling for other covariates.

Table 3.

ORs for Statin Use Among a Nationally Representative Sample of US Adults: NHANES 2011–2012

Covariates	Unadjusted OR (95% CI)	Adjusted OR (95% CI)
SBG1a	9.36 (7.13‐12.27)	4.15 (2.76‐6.23)
Other known risk factors for statin use
CAD	14.39 (10.34‐20.02)	4.96 (2.72‐9.06)
Stroke	4.92 (3.51‐6.89)	—
HTN	5.93 (4.42‐7.94)	1.97 (1.27‐3.04)
Smoking	1.56 (1.27‐1.93)	—
Family history of premature CADb	2.05 (1.44‐2.92)	1.78 (1.18‐2.68)
Demographics
Age, y, mean (SE)c	1.08 (1.08‐1.09)	1.07 (1.06‐1.08)
Male sex	1.03 (0.87‐1.23)	—
Race
Non‐Hispanic white	Ref	Ref
Non‐Hispanic black	0.80 (0.65‐0.99)	1.01 (0.74‐1.37)
Hispanic	0.41 (0.31‐0.54)	0.70 (0.54‐0.93)
Other	0.76 (0.56‐1.03)	1.17 (0.91‐1.50)
Uninsured	0.17 (0.12‐0.24)	0.33 (0.20‐0.53)
Education
≤High school	1.45 (1.11‐1.90)	—
Some college	1.11 (0.80‐1.53)	—
≥College graduate	Ref	—
Waist circumference, cm, mean (SE)c	1.02 (1.02‐1.03)	1.01 (1.00‐1.02)

Abbreviations: CAD, coronary artery disease; CI, confidence interval; DM, diabetes mellitus; HTN, hypertension; LDL‐C, low‐density lipoprotein cholesterol; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio; Ref, reference; SBG, statin benefit group; SE, standard error; US, United States.

All variables listed in the table were adjusted for as potential confounders.

^aSBG1 represents patients with DM and LDL‐C values >70 mg/dL or a diagnosis of dyslipidemia.

^bFamily history of premature CAD (age of close blood relative with CAD <50 years).

^cEffect estimates calculated for a 1‐unit change.

4. DISCUSSION

Our study provides information on the baseline prevalence of statin use and, indirectly, physician practice just prior to the advent of the new guidelines. We found that statin use had essentially tripled over the last decade (from 1999–2000 to 2011–2012). While this may have health care cost implications, a number of studies have shown statins to be cost effective.19, 20 An even larger number of persons are estimated to meet the new criteria for statin use based on the new ACC/AHA guidelines.

Our estimates demonstrate that prior to the advent of the new guidelines, statin use was suboptimal even among patient groups previously identified to benefit significantly from statin therapy, such as persons with established CAD. It also appears that persons with a history of stroke had a much lower prevalence of statin use compared with persons with CAD. This identifies a potential target group of patients for interventions directed at improving statin coverage.

We were unable to separate persons with transient ischemic attack from those with stroke or to differentiate between persons who had a hemorrhagic stroke vs ischemic stroke, which may have contributed to the low prevalence seen among persons with stroke. The NHANES questionnaire did not include questions about reasons for not being on statin therapy, so we were unable to exclude persons who were not on a statin either due to intolerance or patient preference.

Medication nonadherence has also limited the projected benefits of statin use.9 Medication cost is a known factor, given improved use with insurance. Other factors include the side‐effect profile of statins and also the more universal challenge of continuing adherence to medication for an asymptomatic risk factor. Observational studies have shown statin adherence to drop to 50% at 6 months and to 25% adherence after 1 year.21 These estimates were obtained from a largely elderly population (age ≥65 years) enrolled in state Medicaid and pharmaceutical‐assistance programs. They found that persons on statin therapy for secondary prevention (stroke or severe CHD) were more likely to be adherent with statin therapy. Even in clinical trials where participants are more likely to adhere to therapy due to volunteer bias, statin discontinuation is reported to be as high as 33% in primary prevention trials and 18% in secondary prevention trials.21 Other factors shown to be predictors of statin nonadherence include black ethnicity, household income < $50 000, and muscle‐related side effects.21

Consistent with other studies that have described national trends in statin use using the NHANES, National Hospital Ambulatory Medical Care Survey (NHAMCS), and the National Ambulatory Medical Care Survey (NAMCS) datasets,18, 22 our study shows that ethnic disparities remain a barrier to statin use, with Hispanic adults being less likely to be on a statin compared with non‐Hispanic white adults. However, in contrast to these studies, our results showed that African Americans were just as likely as non‐Hispanic whites to be on a statin. This may represent progress toward reducing racial disparities over time.

Previous research has demonstrated a paradoxical increase in mortality disparities for conditions for which treatments or preventive strategies are rapidly improving. Racially or socioeconomically advantaged segments of the population access and adopt lifesaving innovations more rapidly,23, 24 resulting in unequal diffusion of innovation.25 However, our analysis demonstrates a recent reduction in the black‐white treatment gap. There is increasing evidence that disparities are not inevitable and that some communities are moving more rapidly toward equality of outcomes.26, 27, 28

Uninsured persons were also less likely to be on a statin. Subsequently, to the extent that recent health‐reform legislation reduces the number of uninsured, statin use may increase; but this will depend not only on health insurance coverage defined broadly, but also specifically on the extent to which insurance plans cover pharmaceutical expenses. This is an area where Medicaid expansion or nonexpansion may have a significant impact, bringing health insurance and pharmacy benefits to the population of previously uninsured adults who are in a statin benefit group (eg, cardiovascular risk), but who would have previously been categorically ineligible for Medicaid until they had a disabling stroke or heart attack. Medicaid coverage also impacts disparities; for example, studies show that previously documented black‐white disparities in antiretroviral therapy for human immunodeficiency virus patients have been completely eliminated in the Medicaid population, even as disparities persist in Hispanic and Latino segments of the Medicaid population.29, 30, 31

A major strength of this study is our ability to produce estimates from a nationally representative sample systematically designed to estimate the entire US population. Limitations of this study include use of patient‐reported diagnoses and family history, which may be subject to recall bias.

5. CONCLUSION

Our study provides national estimates of statin use and, indirectly, physician practice just prior to the introduction of the new guidelines. This is useful in assessing treatment practices, provides a baseline measure for evaluating the impact of the new treatment guidelines, and serves as a national reference by which programs can evaluate compliance with the ASCVD prevention guidelines. Improving compliance with statin‐use recommendations and addressing remaining disparities will be an important step in achieving the projected benefits of statin use.

Adedinsewo D, Taka N, Agasthi P, Sachdeva R, Rust G, Onwuanyi A. Prevalence and Factors Associated With Statin Use Among a Nationally Representative Sample of US Adults: National Health and Nutrition Examination Survey, 2011–2012, Clin Cardiol 2016. DOI: 10.1002/clc.22577.