Aspirin Use and Misuse for the Primary Prevention of Cardiovascular Diseases

Abstract

Introduction:

Daily aspirin use for primary cardiovascular disease prevention is common among adults. Numerous clinical trials observe reduced cardiovascular disease with regular low-dose aspirin. The U.S. Preventive Services Task Force in 2016 published guidelines for aspirin use, but controversy exists about side effects and overuse or underuse may be common despite the guidelines. Using the Task Force recommendations, this paper describes the prevalence of appropriate aspirin use and physician advice in a population sample.

Methods:

A random sample of men and women (aged 50–69 years) living in the Upper Midwest in 2017–2018 were surveyed, collecting demographic data, health history, and aspirin use. Appropriate primary prevention with aspirin was defined as those with ≥10% cardiovascular disease risk (hypertension, hyperlipidemia, diabetes, smoking) who take aspirin daily or every other day. Those with prevalent cardiovascular disease were labeled secondary prevention.

Results:

A total of 1,352 adults were surveyed (697 women, 655 men). The criteria for secondary prevention were fulfilled in 188 participants and these were eliminated from the analysis. In the remaining group aspirin was indicated in 32.9% (383/1,164). Among those, 46.0% (176/383) were appropriate users, and 54.0% (207/383) were non-users despite indications. Overuse, where not indicated, was common at 26.9% (210/781). Discussion with a physician, though reported in 29% of subjects, was associated with some improvement in appropriate use but also overuse and underuse.

Conclusions:

Aspirin use for primary cardiovascular disease prevention is common. However, many adults are medicating without indication (overuse) or are not using aspirin despite guidelines (underuse).

INTRODUCTION

Cardiovascular disease (CVD) remains among the major causes of death and disability worldwide and their prevention is an attainable goal. Daily aspirin use for CVD prevention is common practice among adults in the U.S.^1–3 Aspirin is known to act by inhibiting platelet aggregation in acute arterial thrombosis, an underlying pathologic process in myocardial infarction and stroke.⁴ Following numerous prospective clinical trials, meta-analyses found reduced rates of first acute myocardial infarction and ischemic stroke with regular low-dose aspirin use for primary CVD prevention.^5,6 These observations led the U.S. Preventive Services Task Force (USPSTF) to make recommendations for aspirin use for primary CVD prevention in those at increased risk in 2002, 2009, and 2016 (Appendix Figures 1 and 2).^7–9

Widespread availability of aspirin over the counter at low cost and substantial industry advertising encourages overuse for primary prevention, as documented in several studies.^10,11 Similarly, there is underuse of aspirin where use is indicated by guidelines.^12,13 The recent release (2018) of 3 large randomized aspirin trials initiated new debate about the appropriateness of widespread aspirin use for primary prevention.^14–17 New meta-analyses, statements from professional groups, and the popular press fueled the controversy.^6,18–23 Concerns include overuse, bleeding side effects, and the balance between benefits and harms.

To understand the use of aspirin for primary CVD prevention in the general population in the context of new data and recommendations, investigators undertook a population survey of 5 Upper Midwestern states. Information on current aspirin use for CVD prevention, health behaviors, cardiovascular risk factors, and demographic characteristics were collected to evaluate prevalence of aspirin use as recommended using the 2016 USPSTF guidelines.

METHODS

Study Population

A cross-sectional telephone survey of non-institutionalized resident adults aged 50–69 years was conducted from August 2017 to March 2018 and analyzed in 2020. The target population lived in the 5-state Upper Midwest area (Iowa, Minnesota, North Dakota, South Dakota, and Wisconsin).

Measures

Lists of addresses with ZIP codes and associated landline telephone numbers were used to generate random samples within states related to state population size.

Randomly selected households were sent a letter describing the study. The option to decline receiving a telephone survey was provided via a call-in site. Only 4% of the sampled population declined participation. The remaining households were called as many as 10 times to administer the survey. Resident adults aged 50–69 years were selected as recommended in the 2016 USPSTF guidelines.⁹ The overall participation rate was 48% after accounting for refusals, address errors, disconnected telephones, and non-responders. This response rate is consistent with the authors’ previous surveys and other telephone surveys.²⁴

Trained interviewers administered a 10-minute telephone survey that included questions about aspirin use, history of CVD, cardiovascular risk factors, reason for aspirin use, physician discussions about aspirin, health behaviors, and demographic characteristics. The questionnaire is available in the Appendix. Because women are more likely to answer the telephone, men were oversampled to obtain an approximately equal male–female ratio.

The sex-specific American College of Cardiology/American Heart Association Pooled Cohort Risk Equation (PCE) 10-year CVD risk was calculated from age, and self-reported hypertension, hyperlipidemia, diabetes, and current smoking. This value was estimated using an algorithm derived from the Minnesota Heart Survey in which there were both clinical measures (measured systolic blood pressure, anti-hypertensives, measured total cholesterol, high-density lipoprotein cholesterol, and a biochemical measure of current smoking) and self-reported history of hypertension, hypercholesterolemia, and current smoking.^25,26 The equations take the binary self-reported risk factors from the survey and weight them appropriately to derive the PCE, which is then used to classify participants into risk categories used in the 2016 USPSTF recommendations.⁹

In a previously published report using the same telephone survey, the validity of self-reported aspirin use was tested with blood levels of thromboxane B2, which is very sensitive to aspirin use.²⁷ Sensitivity and specificity >90%, indicating that this questionnaire and the telephone approach were valid for assessing regular aspirin use.

The validity of the self-reported history of hypertension and high blood cholesterol was also tested against measured values in a population sample, finding a high concordance between the interview and actual measures.²⁵

A primary prevention candidate was defined as a person with no self-reported history of myocardial infarction, stroke, peripheral artery disease, or revascularization procedure. A secondary prevention candidate was defined as an individual with a self-reported history of myocardial infarction, stroke, peripheral artery disease, or revascularization procedure. They were eliminated from this analysis. Regular aspirin use was defined as aspirin use every day or every other day to prevent a heart attack or stroke. Aspirin indicated was defined as no self-reported contraindications (gastrointestinal bleed/peptic ulcer, aspirin allergy, or anticoagulant use), and a CVD 10-year risk level at or above the USPSTF 10% cut off. Appropriate aspirin use was defined as aspirin indicated and regular aspirin use. Overuse was defined as regular aspirin use where aspirin not indicated by a 10-year risk score <10% or contraindications. Underuse was defined as aspirin indicated but no regular aspirin use. Appropriate aspirin non-use was defined as no aspirin use and not indicated.

Statistical Analysis

Data are presented as n (%) for categorical variables and mean (SD) for continuous variables. All statistics were weighted by the source target population size; thus, percentages may not add to 100% or be calculated directly from numerators and denominators. Associations between aspirin use and sex, age group, number of risk factors, and aspirin discussions were performed using multivariable logistic regression.

Analyses were performed in Stata, version 16.1. The IRB of the University of Minnesota approved this study and all participants provided verbal consent for the interview.

RESULTS

The survey included 1,352 participants. Demographic characteristics were similar to the 2010 Census data (Table 1).²⁸

Table 1.

Characteristics of Primary Prevention Individuals

Characteristics	Women	Men
n	638	526
Demographics
Age, years, mean ± SD	60.8 ± 5.5	61.1 ± 5.7
50–59, n (%)	252 (39.3)	196 (38.0)
60–69, n (%)	386 (60.7)	330 (62.0)
Education, n (%)
Less than high school	7 (1.0)	17 (3.0)
High school graduate	151 (25.3)	136 (24.0)
Some college	247 (33.0)	178 (30.7)
College graduate	230 (40.4)	190 (41.8)
Other	3 (0.4)	5 (0.5)
Marital status, n (%)
Married	459 (71.5)	376 (67.8)
Single	59 (10.2)	79 (18.0)
Divorced	60 (9.3)	48 (9.1)
Separated	4 (0.3)	1 (0.4)
Widowed	52 (8.0)	19 (3.9)
Other	4 (0.7)	3 (0.8)
Regular health provider, n (%)	609 (96.2)	480 (92.7)
Risk factors, n (%)
Hypertension	245 (41.7)	237 (45.7)
Hyperlipidemia	240 (36.4)	220 (42.8)
Diabetes	83 (12.6)	82 (17.2)
Current smoker	76 (10.7)	61 (11.8)
Estimated 10-year CVD risk, n (%)
<10%	549 (86.4)	165 (32.3)
≥10%	89 (13.6)	361 (67.7)
Aspirin contraindications, n (%)
Gastrointestinal bleed/peptic ulcer	31 (4.4)	35 (7.4)
Aspirin allergy	7 (0.7)	10 (1.4)
Anticoagulant use	14 (2.9)	34 (8.6)
Any contraindications	50 (7.5)	72 (16.4)

Notes: Estimated sex-specific 10-year CVD risk was calculated from the four risk factors and age, using the algorithm described in reference 25.

CVD, cardiovascular disease.

The primary prevention group had 1,164 participants with 638 women and 526 men (Table 1). A history of hypertension, hypercholesterolemia, diabetes, or current cigarette smoking was common and was higher for men than women. Self-reported contraindications to aspirin use (aspirin allergy, history of gastrointestinal bleeding, or anticoagulant use) were modest. Most participants reported a regular healthcare provider (>90%).

The secondary prevention group included 188 participants and was not included in this analysis of primary prevention.

Regular aspirin use for primary prevention was 31.0% overall, with 25.4% for women and 38.3% for men. Table 2 shows the prevalence of aspirin use according to sex, age group, and the number of risk factors. Also shown are the mean and 95% CI of the estimated PCE for each subgroup. Aspirin use for primary prevention increased with the number of risk factors and PCE score in both women and men. In a logistic regression model for aspirin use including sex, age group, and number of risk factors, men used aspirin more than women (OR=1.7, 95% CI=1.21, 2.43) and those in the older age group (60–69 years) used aspirin more than those in the younger age group (OR=1.8, 95% CI=1.2, 2.6). Those with 1, 2, and 3 or 4 risk factors used aspirin more than those with 0 risk factors (OR=1.8, 95% CI=1.1, 2.8; OR=3.5, 95% CI=2.1, 5.6; OR=5.0, 95% CI=2.6, 9.6).

Table 2.

Regular Aspirin Use in Upper Midwest by Sex, Age Group, Number of Risk Factors and PCE 10-year CVD risk

Sex/Age group/n	Number of risk factors	PCE (95% CI) (10 year CVD risk, %)	Aspirin use (%)
Women
50–59 years
120	0	1.5 (1.4, 1.6)	3.9
74	1	2.8 (2.4, 3.2)	16.3
41	2	3.9 (3.3, 4.5)	39.4
17	3/4	9.7 (8.1, 11.6)	52.5
60–69
115	0	5.0 (4.6, 5.4)	26.1
146	1	6.2 (5.7, 6.7)	27.0
90	2	9.1 (8.3, 10.0)	34.7
35	3/4	17.1 (14.9, 19.7)	55.2
Men
50–59
75	0	5.6 (5.2, 6.0)	17.5
70	1	7.8 (6.8, 8.8)	32.8
40	2	10.5 (9.0, 12.3)	41.7
11	3/4	20.5 (16.8, 25.0)	33.8
60–69
91	0	12.4 (11.6, 13.3)	24.6
103	1	15.9 (14.9, 16.9)	38.4
100	2	19.9 (18.3, 21.7)	59.3
36	3/4	32.2 (29.0, 35.8)	60.4

Notes: Values for PCE are mean (95% CI) and for aspirin use are prevalence (%). Risk factors are hypertension, hypercholesterolemia, diabetes and current smoking. Those with 3 or 4 risk factors were combined due to small numbers of subjects with all 4 risk factors.

PCE, pooled cohort equations; CVD, cardiovascular disease.

Self-reported smokers reported lower aspirin use than those who were not current smokers (26% vs 32%, p=0.29). Those with hypertension reported 44% aspirin use whereas those without hypertension reported 21% aspirin use (p<0.001). Individuals with hyperlipidemia used aspirin more than those without (41% vs 25%, p<0.001). Aspirin use was substantially higher in those with diabetes than those without diabetes (55% vs 27%, p<0.001).

Based on the 2016 USPSTF recommendations for primary prevention, 383/1,164 (32.9%) of the participants had indications for primary prevention aspirin use (Figure 1). Among those indicated for primary prevention aspirin use, 46.0% (176/383) were appropriate users, whereas 54.0% (207/383) were underusers (Figure 1). Men were more often underusers than women. In the 781 adults without indication for aspirin, 26.9% (210/781) were regularly using aspirin (overuse). Women were more often overusers than men.

Figure 1.

Pie charts showing use of aspirin in women and men, according to USPSTF 2016 guidelines.

USPSTF, U.S. Preventive Services Task Force.

Table 3 shows the prevalence of overuse and underuse of aspirin, stratified by sex and whether or not the participant had been involved in discussions with their clinician regarding the use of aspirin for prevention of CVD. Men, (32.4%, 173/526) were somewhat more likely to have this discussion whereas women were less likely to have this discussion (28%, 168/638). In models including sex, age group, and aspirin discussion between clinician and participant and their interactions, men who did not engage in discussions about aspirin were 8.8 (95% CI=5.0, 15.5) times more often underusers than women. Discussions were significantly inversely associated with underuse (OR=0.27, 95% CI=0.15, 0.49), and older participants were more often underusers (OR=5.8, 95% CI=3.3, 10.1). Women were more often overusers than men (OR=1.7, 95% CI=1.1, 2.6), and discussions were more often associated with overuse in both men and women (OR=5.9, 95% CI=3.8, 9.0).

Table 3.

Associations Between Discussion Between Participant and Healthcare Provider and Aspirin Use

Variable	Women (n=638)		Men (n=526)
	Yes	No	Yes	No
Any discussion	168/638 (28)	470/638 (72)	173/526 (34)	353/526 (66)
Aspirin use
Appropriate	26/168 (10.8)	19/470 (3.2)	79/173 (40.9)	52/353 (14.5)
Overuse	73/168 (43.6)	64/470 (10.9)	47/173 (29.4)	26/353 (7.5)
Aspirin non-use
Appropriate	60/168 (39.3)	357/470 (78.2)	22/173 (16.7)	132/353 (39.3)
Underuse	9/168 (6.3)	30/470 (7.7)	25/173 (13.0)	143/353 (38.7)

Notes: Values are n/N (%). Aspirin indicated: No self-reported contraindications (gastrointestinal (GI) bleed/peptic ulcer, aspirin allergy, or anticoagulant use), and a CVD 10-year risk level at or above the USPSTF 10% cutoff. Appropriate aspirin use: Aspirin indicated and regular aspirin use. Overuse: Regular aspirin use where aspirin not indicated by a 10-year risk score <10% or contraindications. Underuse: Aspirin indicated but no regular aspirin use. Appropriate aspirin non-use: No aspirin use and not indicated.

CVD, cardiovascular disease; USPSTF, U.S. Preventive Services Task Force.

DISCUSSION

The use of aspirin for secondary prevention of CVD is widely accepted.²⁹ However, aspirin use for primary prevention is debated despite large clinical trials and guidelines supporting the benefits.^{5–9,14–16} Concerns have been raised about bleeding side effects and overuse in individuals who are not at increased CVD risk. Recommendations for aspirin use in primary prevention have varied with the most recent USPSTF (2016) limiting the recommended use to ages 50–69 years and an increased 10-year cardiovascular risk.⁹ Compared with the 2009 USPSTF recommendations (Appendix Figure 1), aspirin use was recommended in fewer patients in 2016 (Appendix Figure 2), particularly among women. Despite these changing guidelines restricting indications, aspirin use for prevention of CVD is widely utilized in the U.S. population and has continued to grow in recent decades.^1–3

Three recent clinical trials raise questions about the use of aspirin for primary prevention. Although each demonstrated a decrease in cardiovascular disease outcomes with aspirin, they also revealed more complications.^14–17 In Aspirin in Reducing Events in the Elderly, a study of aspirin in the healthy elderly, the intervention was terminated early owing to increased cancer mortality in the aspirin group.¹⁶ This finding was contrary to previous observations and trials showing reduced cancer risk.^30,31 Cardiovascular events were non-significantly reduced with aspirin, but this was overshadowed by the unexpected cancer mortality.¹⁶ In A Study of Cardiovascular Events in Diabetes, the improvement in cardiovascular disease endpoints was associated with increased bleeding episodes.¹⁵ The Aspirin to Reduce Risk of Initial Vascular Events trial found a non-significant reduction in cardiovascular disease endpoints with the intention-to-treat analysis although the “as treated” analysis showed significant benefit. All 3 of these trials were subject to significant dropout rates and fewer events than anticipated.^14–17

The 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease delivered a mixed message on the use of aspirin for primary prevention based on these 3 trials.²³ The summary recommendations state that “aspirin should be used infrequently in the routine primary prevention of ASCVD [atherosclerotic CVD] because of lack of net benefit” but later in the same document it is recommended that aspirin should be considered in high-risk adults aged 40–70 years who do not have increased bleeding risk.

In this 2017–2018 study of randomly selected adults in the Upper Midwest, a number of observations were made. Use of aspirin for primary prevention was common in men and women. However, many adults who were at increased CVD risk, with indications for aspirin use according to the 2016 guidelines, were not using aspirin (underuse). This was more common in men. Aspirin use in those without indications (overuse) was observed in 14.9% of men and 20.1% of women, putting them at potential increased risk of complications and less benefit. Overuse was more common in women. The excess use among women may be the result of changing USPSTF guidelines between 2009 and 2016. The 2016 guidelines reduced indications for aspirin use among women.

There are several population-based studies on aspirin use providing information regarding prevalence, trends, and associated factors. The Centers for Disease Control and Prevention’s Behavioral Risk Factor Surveillance System population surveys from 1999 to 2003 found increasing aspirin use associated with increased risk factors and moderate use in those with no risk factors.³² The National Health Interview Survey found increasing rates of aspirin use for prevention between 2005 and 2010. Aspirin was more commonly used in men than in women.³ The 2011–2012 National Health and Nutrition Examination Survey data showed high levels of aspirin use but noted that physicians were not prescribing aspirin when indicated in many cases.¹ Rates seen in this study are similar to those observed in a large practice in Wisconsin in 2010–2012 and a national cardiology practice registry.^10,33 A 2016 statement from the European Society of Cardiology does not recommend aspirin use for primary prevention owing to increased risk of bleeding and primary prevention aspirin use is lower in Europe than in the U.S.^34–37

Low-cost, widespread availability and advertising of aspirin may encourage self-diagnosis and aspirin treatment for primary prevention. In this study, discussion with a health professional was associated with more appropriate aspirin use among men and women. But there continues to be significant underuse and overuse despite a discussion with a health professional. This may be due to changing guidelines not reaching clinical practice, professional confusion about guidelines, or discussions before the new guidelines were introduced. The most recent USPSTF recommendations (June 2016) were in place for about 1 year when this survey was performed, but this may be too little time for change in the clinical community.^8,9,38

Limitations

There are limitations to this study that are inherent in most population-based telephone surveys. This study was based on landline phones. This eliminates an increasing proportion of the population who rely only on cell phones. However, landline phones are still more likely to be used in older age groups and those living in non-metropolitan areas and have the advantage of being linked to dwellings in selected geographic areas.³⁹

Participation rates were 48%. This is a level that is similar to other current telephone health surveys but may still introduce bias.²⁴ The study does have the limitation of being in Upper Midwestern states and may not generalize. However, 3 national surveys find similar levels of aspirin use for primary prevention.^1,3,32 Finally, the validity of telephone survey data is always open to question. In a previous study, the authors demonstrated that individuals who respond to this same telephone survey are highly likely to report aspirin use accurately when tested by biochemical marker.²⁷ There is no apparent reason why subjects should prevaricate regarding aspirin use.

CONCLUSIONS

Aspirin continues to be widely used for primary CVD prevention in the general adult population. Many people may be self-prescribing aspirin despite low levels of CVD risk, contrary to guidelines (overuse). There is also significant underuse. A discussion with a physician is not consistently associated with appropriate use.