Systematic Review of the Effect of Adherence to Statin Treatment on Critical Cardiovascular Events and Mortality in Primary Prevention

Abstract

Aim:

Analyze the relative risks of critical cardiovascular outcomes and mortality associated with adherence to statin treatment in a clinical setting in people with no history of prior cardiovascular disease (CVD).

Methods:

A systematic review of the literature was conducted up to December 2016. The outcomes of interest were cardiovascular fatal or nonfatal events and all-cause mortality.

Results:

A total of 17 articles were included in a qualitative synthesis. Four were case–control nested in a retrospective cohort design and the other 11 were a cohort design. Seven studies compared the best adherer patients with the worst adherers. In the 3 studies (317 603 participants) that considered ischemic heart disease in this group, the pooled reduction in risk was 18% (95% confidence interval [CI]: 14%-22%, I² = 0%); for the CVD outcome, 2 studies (131 477 participants) showed a pooled reduction in risk of 47% (95% CI: 36%-56%, I² = 84.7%) with 1 included study showing a much larger reduction than the others; for the cerebrovascular event (CeVD) outcome, 2 studies (155 726 participants) showed a pooled reduction in risk of 26% (95% CI: 18%-34%, I² = 0%); and for mortality, the reduction in risk was 49% (95% CI: 39%-57%, I² = 62.4%). The other 4 studies (147 859 participants) compared the most adherent group with the rest. These showed a pooled risk reduction of CVD of 22% (95% CI: 6%-27%, I² = 0).

Conclusion:

Adherence to statins treatment is shown as a key element for primary prevention, although these are observational data and the risk of bias from confounding cannot be ruled out. Standardization of measures of adherence to treatment would improve comparability between studies. Further research is warranted to design effective interventions to improve patients’ adherence.

Keywords

hydroxymethyl glutaryl coenzyme-A reductase inhibitors cardiovascular diseases adherence primary prevention systematic review meta-analysis

Introduction

Cardiovascular diseases (CVDs) are 1 of the leading causes of death worldwide.¹ Behavioral risk factors such as unhealthy diet, physical inactivity, tobacco use, and harmful use of alcohol are thought to be responsible for about 68% (38 million) of death globally of the world’s 56 million deaths in 2012. More than 40% were premature deaths under age 70 years.² “Hydroxylmethyl glutaryl coenzyme-A reductase inhibitors” (statins) have been shown to be an effective class of pharmaceutical therapy to reduce the incidence of cardiovascular events and mortality in primary^3,4 and secondary prevention.^5,6

Although the literature defines various related concepts, such as persistence, compliance, or discontinuation,^7
–9 the most frequently used and accepted term is adherence. This term implies an active collaboration between the health professional and the patient on the decisions which affect the person’s health.⁹

Various studies have found that adherence is suboptimal in chronic asymptomatic diseases.^10,11 Adherence to long-term cardiovascular and noncardiovascular treatments in the general population is around 50% in developed countries and is much lower in developing countries.⁷ Naderi et al¹¹ found that adherence to statin in primary prevention was 57% (95% confidence interval [CI]: 51%-64%) and 76% (95% CI 70%-82%) in secondary prevention. A meta-analysis of nonadherence to statin therapy found that the percentage of adherent patients at 1 year of follow-up was 49.0% for observational studies and 90.3% for randomized trials.¹² Therefore, it is possible that randomized trials overestimate the benefits of statin that would be obtained in the real-world settings.

While several articles have reported some kind of relationship between statin adherence and clinical outcomes using diverse data and methods,^13

–16 the relationship has not been extensively studied in primary prevention. The objective of this study is to estimate the empirical relationship between statin adherence in primary prevention with critical cardiovascular outcomes and mortality from a systematic review of the literature.

Methods

A systematic review was performed in the MedLine database (through PubMed) and Trip Database was searched up to December 2016.

Search Strategy

In MedLine, the terms used in the search strategies were (“Hydroxymethylglutaryl-CoA Reductase Inhibitors” [Mesh] OR statin* OR hmg-coa) AND (“Primary Prevention” [Mesh] OR “primary therapy”) AND (“Medication Adherence” [Mesh] OR “Patient Compliance” [Mesh] OR “statins adherence” OR “patient adherence” OR “patient cooperation”). Trip Database was consulted using free terms: “statins adherence” and “statins discontinuation.” Finally, a search was conducted of the reference list of the articles identified. Only articles written in English or Spanish were included.

Box 1

Inclusion and Exclusion Criteria.

Inclusion criteria:

Adult men or women in treatment with statins with no recent history of cardiovascular disease but at risk of developing cardiovascular disease (primary prevention, new statin users with no cardiovascular disease at baseline).

The assessment of exposure was the degree of adherence or noncompliance with statins.

The outcomes were relative risks of nonfatal or fatal cardiovascular or cerebrovascular events and overall mortality for adherent versus nonadherent patients.

Exclusion criteria:

Studies conducted only in populations with specific conditions, for example, only smoking participants, only people with obesity, diabetes, and so on, and studies that analyze risk factors associated with adherence.

Studies were excluded which use a comparison group that does not take statin, and studies that analyzed effects in patients who discontinue or are treated intermittently according to their health professionals’ instructions.

Excluded study designs: case reports, case series, review, and simulation and modeling studies.

Data extraction and management

Identification of the articles was carried out by 2 reviewers (E.M.R. and A.O.L.) who selected independently and any discrepancies were agreement with a third person (D.E.). Once the studies had been identified, and duplicate articles excluded, the title and abstract of each review was checked against the inclusion and exclusion criteria before requesting the full text. The review was managed using Reference Manager Software.

Two researchers carried out an independent evaluation of the methodological quality of each study (E.M.R. and A.O.L.) using the Newcastle-Ottawa Scale (NOS)¹⁷ for assessing the quality and risk of bias of nonrandomized studies.

A data extraction form was developed to record bibliographic information, general characteristics of the study included, measure of exposure, and outcome data. The relative risk (RR) measures extracted from the articles were, where possible, those that had been adjusted for observed confounding variables. Relative risks were transformed if necessary so that a RR of less than 1 indicates that better adherence improves CVD outcoms.

Outcomes

Our main end points were CVDs, cardiovascular event, cerebrovascular event, disease-specific mortality, and all-cause mortality. International Classification of Diseases, Ninth Revision codes are diagnosis of chronic heart failure (CHF; codes 398 402 428) or complications of heart disease (codes 429), myocardial infarction or angina (codes 410-414), cerebrovascular events or stroke (codes 430-438), atherosclerosis arterial embolism and thrombosis (codes 440, 443-444), a medical procedure (coronary artery bypass grafting, angiography, angioplasty, stent), and/or coronary revascularizations.

Definition of Exposure

Exposure was the degree of adherence. Broadly, 2 types of comparison are reported in the literature. The first divides the cohort into 4 or more groups according to the degree of adherence (expressed in percentage terms). The worst adherers are taken to be the reference group, and RRs of CVD end points are then estimated for more adherent groups of patients compared to this reference group. In the meta-analysis, we then extract and pool the RRs for the most adherent patients compared to the least adherent, for each end point.

Other studies divide the cohort into 2 groups at some arbitrary cut point of adherence, for example, 60%, 75%, or 80%. The study then reports RRs comparing CVD end points for the more adherent group with the less adherent group. In the meta-analysis, we extract and pool these RRs for each end point.

Heterogeneity between studies was assessed using Higgin I², where I² values of >50% represent a substantial level of heterogeneity.¹⁸ A random-effects model was used to combine study-specific RR. The analyses were performed with Stata.v13 software.

Results

Characteristics of Included Studies

The initial search identified 155 articles from which 113 studies were excluded as duplicates or not meeting the inclusion criteria from title and abstract. The most common reason for exclusion was that in spite of studying aspects relating to the research question, the articles focused on factors associated with adherence, or focused on secondary prevention, or familial hypercholesterolemia. After reading the full text, we excluded a further 30 articles (see Figure 1). Four of these excluded articles were reviews which were used only to identify relevant articles from their reference lists and for background information or discussion.^19

–22 Finally, 12 articles were included from the search plus a further 5 were found via cross references or hand searching. Of the 17 included articles, 4 were case–control nested in a retrospective cohort design and the other 11 had a cohort design.

Figure 1.

Study flow diagram.

Table 1 shows patient and study characteristics of included studies. One study (with results were published across 2 articles) recruited male participants as a follow-on study from a previous RCT.^30,31 A Finnish study published 2 articles,^34,35 one of which only included women.³⁴ The mean age within studies ranged from 51.4 years³² to 66.5 years²⁴. The mean time of follow-up ranged from 1.6 years³⁶ to 5.8 years.²⁸ Several articles required a minimum follow-up time as an inclusion criteria.^{23,26,27,29,33} In 11 of the cohort studies,^{23

–35} the minimum follow-up was 1 year,³³ and in 4 nested case–control studies,^36

–39 the range of follow-up was between 6 months and 6.5 years.^38,39

Table 1.

Characteristics Patients of Studies Included.

Study (year)	Country Setting, Design	Cohort Study	Total Participants		Mean Age (SD), years	Sex (% Women)	Follow-Up (Mean)	Adherence^a (During the First Year)
Study (year)	Country Setting, Design	Cohort Study	Total Participants		Mean Age (SD), years	Sex (% Women)	Follow-Up (Mean)	Definition	Category (Cut Points)	N (%)^a
Corrao et al (2010)²³	Italy, Cohort	New statin users aged ≥18 years, ≥1 year of follow-up, >1 statin prescription^b,c	90 832		61.8 (11.1)	59.3	4.25 years	PDC	≤25%	31 537 (34.7)
									26%-50%	22 544 (24.8)
									51%-75%	18 918 (80.8)
									≥75%	17 833 (19.6)
Degli Esposti et al (2012)²⁴	Italy, Cohort	New statin users aged ≥18 years, PDC >20%, no CV events during 6 months after the enrollment date^b	19 232		66.5 (11.4)	48.9	Media 1.9 years (first 6 months after the enrollment date excluded)	PDC	21%-40%	4427 (23.0)
									41%-60%	3117 (16.2)
									61%-80%	3784 (19.7)
									≥80%	7904 (41.1)
Slejko et al (2014)²⁵	USA, Cohort	New statin users, PDC >80% during first year of statin use and no CV events during first year after the enrollment date^d,e	11 126^f		55.85 (10.3)	49.9	3.5 years	PDC		Second year	Third year
									<20%	676 (6.1)	2303 (21.4)
									20%-80%	2673 (23.9)	2152 (20)
									≥80%	7813 (70.0)	6316 (58.6)
Poluzzi et al (2011)²⁶	Italy, Cohort	Patients who received statin prescription and followed for up to 36 months, PDC ≥80%, use of other cardiovascular drugs and survivors the 3 years follow-up^c	108 073		Unspecific for primary prevention		3 years^e	Amount of statins received during the period (number of tablets purchased)	Adherents patient (at least 300 tablets per year)	Not detailed
									Slightly nonadherent patients: At least 200 tablets in each year, but no adherent through the whole 3-year period
									Strongly nonadherent patients: never covered and at least 1 year with less than 200 tablets
									Highly variable patients: never covered and at least 1 year with less than 200 tablets
Shalev et al (2009)^g,27	Israel, Cohort	New statin users aged ≥18 years^h	136 052		54.8 (12.9)	55.4	4.0 years	PDC	<10%	Not detailed
									10%-19%
									20%-29%
									30%-39%
									40%-49%
									50%-59%
									60%-69%
									70%-79%
									80%-89%
									≥90%
Shalev et al (2012)²⁸	Israel, Cohort	New statin users aged 45 to 75, no CV events or death during first-year follow-up and patients purchased amount of statin ≥85% above the amount required for complete coverage	171 535		57.82 (7.94)	57.7	5.79 years	PDC	<20%	42 017 (25%)
									20%-39%	23 554 (14%)
									40%-59%	25 883 (15%)
									60%-79%	31 756 (19%)
									≥80%	48 325 (28.0%)
Dragomir et al (2010)²⁹	Canada, Cohort	New statin users and 3 years follow-up period^d,i	55 134		PDC<80: 62.5(10); PDC>80: 63.5 (9)	PDC<80%: 60; PDC>80%: 61	3 years	MPR	< 80%	26 585 (48.22)
Dragomir et al (2010)²⁹	Canada, Cohort	New statin users and 3 years follow-up period^d,i	55 134		PDC<80: 62.5(10); PDC>80: 63.5 (9)	PDC<80%: 60; PDC>80%: 61	3 years	MPR	≥ 80%	28 549 (51.78)
WOSCOPS^30,31 (1997; 1999)	Scotland, Cohort	Participants RCT preview: men aged 45 to 64 years and LDL levels were identified by population screening in primary care	6595		Not detailed	0	4.9 years (range 3.5-6.1 years), 4 visits per annum	The records of a subject’s attendance for each trial visit and the issue or not of a supply of trial medication	<75%	Not detailed
									75 <100%
									100%
Rublee et al (2012)³²	USA, Cohort	New atorvastatin users ≥18 years, follow-up >18 months^b,e,m	79 010		PDC<60: 51.4 (11.9); PDC>60: 53.7 (10.7)	PDC<60%: 47.2; PDC>60%: 46.0	Range:13 months after initiation until the end of follow-up (≥36 months)	PDC	<60%	38 014 (48.1)
Rublee et al (2012)³²	USA, Cohort		79 010		PDC<60: 51.4 (11.9); PDC>60: 53.7 (10.7)	PDC<60%: 47.2; PDC>60%: 46.0		PDC	>60%	40 996 (51.9)
Chapman et al (2010)³³	USA, Cohort	Patients ≥18 years at least 1 prescription for SPAA or CCB + statin continuously enrolled for ≥6 months prior and 18 months following treatment initiation, >1 diagnosis of hypertension prior to the index date^j	Total: 19 447 patients: 1537 SPAA, 17 910 CCB + statin		53.5 (7.6) for SPAA patients; 54.8 (8.5) CCB + statin patient	41.4% SPAA; 47.9% CCB + statin patient	1 year	PDC	<80%	Not detailed
Chapman et al (2010)³³	USA, Cohort		Total: 19 447 patients: 1537 SPAA, 17 910 CCB + statin			41.4% SPAA; 47.9% CCB + statin patient	1 year	PDC	≥80%	Not detailed
Lavikainen et al (2016)³⁴	Finland, Cohort	New statin women users aged 45-64 years^k,f,n	42 807		56.2 (4.9)	100	3 years	PDC		first year	second year	Third years
									<80%	47.0%	24.0%	36.2%
									≥80%	53.0%	76.0%	63.8%
Rannanheimo et al (2015)³⁵	Finland, Cohort	New statin users aged 45-75 years^k,f,n	97 575		59	66	3 years	PDC	<40%	21.0%
									40%-79%	26.0%
									≥80%	53.0%
Bouchard (2007)³⁶	Canada, nested case–control	New statin users aged between 50 and 64^d,l	First year of follow-up 20 118 (case 958, control 19 160)	After first year of follow-up 12 180 (Case 580, control 11 600)	58 (5)	Full cohort 63	1.6 years	Percentage of the prescribed doses of medication actually taken by the patient over a specified period	<90%	Not detailed
Bouchard (2007)³⁶	Canada, nested case–control	New statin users aged between 50 and 64^d,l	First year of follow-up 20 118 (case 958, control 19 160)		58 (5)	Full cohort 63	1.6 years		≥90%	Not detailed
Perreault et al (2009) CeVD³⁷	Canada, nested case–control	New statin users aged between 45 and 85^b,i	first year of follow-up 21 791 (case 1366; control 20 425)	After first year of follow-up 41 140 (case 2593; control 38 547)	Full cohort 63 (10). PDC<80% 62.5 (10); PDC>80% 63.5 (9)	Full cohort 59: PDC < 80%: 60; PDC ≥ 80%: 61	2.95 years	MPR	<80%	26 585 (48.22)
Perreault et al (2009) CeVD³⁷	Canada, nested case–control	New statin users aged between 45 and 85^b,i	first year of follow-up 21 791 (case 1366; control 20 425)		Full cohort 63 (10). PDC<80% 62.5 (10); PDC>80% 63.5 (9)	Full cohort 59: PDC < 80%: 60; PDC ≥ 80%: 61	2.95 years	MPR	≥80%	28 549 (51.78)
Perreault et al (2009) CAD³⁸	Canada, nested case–control	New statin users aged 45-85^d,i	first year of follow-up 95.313 (case 5974; control 89 339)	After first year of follow-up 147 601 (case 9294; control 138 3074)	Full cohort 63. PDC<80% 64.6 (9.6) PDC>80% 65.1 (9.2)	Full cohort 59: PDC < 80%: 51; PDC ≥ 80%: 52	Minimum 6 month, maximum 6.5 years	MPR		First year	After first year
									<80%	26.0%	47.0%
									≥80%	74.0%	53.0%
Perrault et al (2008) CHF³⁹	Canada, nested case–control	New statin users aged 45-85^d,i	First year of follow-up 21 831 (case 1369; control 20 462)	After first year of follow-up 46 601 (case 2940; control 43 661)	Full cohort: 63	Full cohort: 59	Minimum 6 months, maximum 6.5 years	MPR		First year	After first year
									<80%	26.0%	46.0%
									≥80%	74.0%	54.0%

Abbreviations: CAD, coronary artery disease; CCB, calcium-channel blocker; CeVD, cerebrovascular events; CHF, chronic heart failure; CVD, cardiovascular disease; LDL, Low Density Lipoprotein ;MPR, medication possession rate; PDC, proportion of days covered; SD, standard deviation; SPAA, single-pill amlodipine/atorvastatin; WOSCOPS, West of Scotland Coronary Prevention Study.

^aData referred to the first year if there is no express indication.

^bNew users defined no prior statin prescription or lipid-lowering drugs filled during the 2 years preceding the enrollment date.

^cNo history of CV events in 2 years previous to index date.

^dNew users defined no prior statin prescription or lipid-lowering drugs filled during the 1 year preceding the enrollment date.

^eNo history of CV events in 1 year previous to index date.

^fSample size decreased after 2 years due to noncontinuous health plan enrollment.

^gOnly patients surviving the 3-year follow-up period were included in the logistic analysis.

^hNew users defined no prior statin prescription or lipid-lowering drugs filled during the 3 years preceding the enrollment date.

ⁱNo history of CV events in 5 years previous to index date.

^jNo claims for the index prescriptions for 6 months

^kNew users defined no prior statin prescription filled during the 7 years preceding (when de register was established) and lipid-lowering drugs filled during the 3 year preceding the enrollment date. Patients who died or were institutionalized long term within 3 years prior to statin or had an outcome event within the first year after statin initiation were excluded.

^lNo history of CV events in 3 years previous to index date.

^mOnly information about primary prevention.

ⁿOnly patients who initiated statin therapy for primary prevention of CVD.

Four case control studies nested in a cohort were found,^36

–39 and all of them were performed by the same Canadian research team (Table 1). In the nested case–control study, cases of a disease that occur in a defined cohort are identified and, for each, a specified number of matched controls is selected from among those in the cohort who have not developed the disease by the time of disease occurrence in the case. The cases were defined as those patients who have CeVD,³⁷ coronary artery disease (CAD),^36,38 or CHF.³⁹ Meanwhile, all controls (15 controls^37
–39-20 controls³⁶ per case) were randomly selected from the source population (cohort) of the same age and follow-up using density sampling. All cohorts were included in the analysis as this study design should lead to independent sampling with little risk of duplication of cases or controls.

Statin Adherence

Studies were heterogeneous in the definition of adherence and cut points used to quantify the degree of adherence (Table 1). The most common definition of adherence was proportion of days covered (PDC).^{23
–25,27,28,32

–35} This is calculated by dividing the quantity of statin dispensed by the total time interval from index date to events, deaths or cardiovascular events, whichever occurred earlier. Another indicator used was medication possession rate (MPR).^29,37
–39 This is defined as number of days of medication supplied within the refill interval divided by number of days in refill interval. In the West of Scotland Coronary Prevention Study (WOSCOPS) studies,^30,31 adherence was measured by whether a subject attended a scheduled trial visit (1 visit every 3 months) and was issued or not with a further supply of trial medication according to the doctor’s indication. Bouchard et al³⁶ used the percentage of the prescribed doses of medication actually taken by the patient over a specified period, and finally, Poluzzi et al²⁶ used the amount of statins received during the period (number of tablets purchased). The majority of studies, the denominator period used to calculate adherence was the total follow-up of the cohort, except for Chapman et al³³ and Rublee et al,³² which categorized the patient into adherent or not over the first 6 months or used a “anniversary model”: “a 90-day period at the end of the 1-year assessment period was used as the measure of discontinuation, and patients who did not have pill coverage during these 90 days were considered nonadherent.”^32p.25 Lavikainen et al³⁴ and Rannanheimo et al³⁵ exposure was defined as baseline and time-varying cumulative adherence, and statins adherence was measured from as PDC at first-year intervals since statin initiation. Four studies required a minimum level of adherence to statin^28,34,36,33 in order to include the patient in the analysis: PDC >20%,²⁶ PDC >80% during first year,^24,25 or patients purchases amount of statins ≥85%.²⁸ Table 1 shows adherence rates as measured in each study.

In 4 case–control studies nested in a cohort, adherence for cases was calculated from the cohort entry date to the date of cardiovascular event, while for the controls it was from the cohort entry date to the date of selection.

Risk of Bias in Included Studies

In general, we judged the individual studies to have a low risk of bias, with similar scores in the NOS scales (Table 2). The WOSCOPS study^30,31 recruited patients as a follow-on study from a previous RCT, and this study was judged as the higher risk of bias in terms of patient selection. Most studies included patients with no history of CVD events in given period prior to the index date, except in 4 studies.^24,27,30,31 None of studies reported the number of participants lost to follow-up. With the exception of 2 studies,^30,31 all included studies had more than 19 000 participants and used a pharmacy claims database as the source of data on adherence and outcome.

Table 2.

Methodological Quality Summary.

	Selection				Comparability	Outcome^a\Exposure^b
Cohorts Studies	1	2	3	4	5	6	7	8
Chapman et al³³
Corrao et al²³
Degli Esposti et al²⁴
Dragomir et al²⁹
Lavikainen et al³⁴
Poluzzi et al²⁶
Rannanheimo et al³⁵
Rublee et al³²
Shalev et al²⁷
Shalev et al²⁸
Slejko et al²⁵
WOSCOPS^30,31
Nested case–control studies	9	10	11	12	13	14	15	16
Bouchard et al³⁶
Perreault et al³⁷
Perreault et al³⁸
Perreault et al³⁹

Abbreviation: WOSCOPS, West of Scotland Coronary Prevention Study.

^aCohort Studies: (1) representativeness of the exposed cohort; (2) selection of nonexposed cohort; (3) ascertainment of exposure; (4) outcome not present at beginning; (5) comparability of cohorts; (6) assessment of outcome; (7) follow-up long enough; (8) adequacy of follow-up.

^bNested case–control studies: (9) is the case definition adequate? (10) Representativeness of the cases; (11) selection of the controls; (12) definition of the controls; (13) comparability of cases and controls; (14) ascertain of exposure; (15) same method of ascertain for cases and control; (16) nonresponse rate.

Effects of Adherence on Risk of Developing Critical Cardiovascular Outcomes and Mortality

Tables 3 and 4 show the effects of adherence to statin on clinical outcome (heart diseases, cardiovascular events, and mortality).

Table 3.

Effects of Adherence on Critical Cardiovascular Outcomes and Mortality: Cohort Studies.

Author (year)	Outcomes	ICD-9 Codes/ICD-10 Codes	n/N	Assessment of Exposure: Adherence		Relative Risk (95% CI)	P
Ischemic heart disease (ICD-9 codes: 410-414)
Corrao et al (2010)²³	Nonfatal ischemic heart disease: acute myocardial infarction, acute and subacute forms of IHD, and/or coronary revascularizations	410, 411, 36.0-36.9 (ICD-9)	1480/90 832	PDC		Reference
					26%-50%	HR: 0.85 (0.72-0.98)
					51%-75%	HR: 0.82 (0.71-0.95)
						HR: 0.81 (0.71-0.94)	.014
Degli Esposti et al (2012)²⁴	AMI	410.x (ICD-9)	265/19 518	PDC	21%-40%	Reference
					41%-60%	HR: 0.99 (0.66-1,47)	.944
					61%-80%	HR: 0.82 (0.56-1.21)	.313
						HR: 0.79 (0.56-1.10)	.166
Dragomir et al (2010)²⁹	Coronary artery disease	410-414 (ICD-9)	7326/55 134	MPR	<80%	Reference
Dragomir et al (2010)²⁹	Coronary artery disease	410-414 (ICD-9)	7326/55 134	MPR	≥80%	OR: 0.93 (0.87-0.99)
WOSCOP (1997)³⁰	Definite coronary heart disease or nonfatal myocardial infarction	Not detailed	1063/6595	Record for visit: issue or not of a supply of trial medication	<75%	Reference
					≥75%	HR: 0.62 (0.50-0.77)	.0001
	Definite or suspect coronary heart disease death or nonfatal myocardial infarction				<75%	Reference
					≥75%	HR: 0.66 (0.54-0.81)	.0001
	Coronary artery bypass graft or percutaneous transluminal coronary angioplasty				<75%	Reference
					≥75%	HR: 0.54 (0.36-0.81)	.0031
Rannanheimo et al (2015)³⁵	Any CHD event	I20-I25 or related medical procedure (ICD-10)	470/97 575	PDC	< 40%	Reference
					40%-79%	HR: 0.86 (0.80-0.93)
					≥80%	HR 0.75 (0.70, 0.80)
	Acute coronary syndrome	I20.0, I21, I22 (ICD-10)			<40%	Reference
					40%-79%	HR: 0.66 (0.56-0.78)
					≥80%	HR: 0.56 (0.49-0.65)
	Chronic CHD event	I20.8, I20.9, I25 or related medical procedure (ICD-10)			<40%	Reference
					40%-79%	HR: 0.74 (0.66-0.83)
					≥80%	HR: 0.67 (0.61-0.74)
Cardiovascular/cerebrovascular events
Degli Esposti et al (2012)²⁴	Stroke	430-438.x (ICD-9)	549/19 385	PDC	21%-40%	Reference
					41%-60%	HR: 0.77 (0.58-1.01)	.060
					61%-80%	HR: 0.65 (0.50-0.84)	.001
					>80%	HR: 0.73 (0.58-0.90)	.004
Shalev et al (2012)²⁸	Mayor coronary events (myocardial infarction or revascularizations) or stroke	36, 410-412, 414 m 429, v15.1, v45.81, v45.82, 36.411, 429, 793, 794, 996, 33.999, 33.510, 92.984, 92.982, 92.980, v12.40, 342, 344, 362, 388, 433-436, 438, 997.02 (ICD-9)	10 159/171 535	PDC	<20%	Reference
					20%-40%	Women: HR = 0.93 (0.82-1.05); men: HR = 0.95 (0.87-1.03)	Women: .237; Men: .182
					40%-60%	Women: HR = 0.71 (0.63-0.80); men: HR = 0.75 (0.69-0.81)	Women: <.001; men: <.001
					60%-80%	Women: HR = 0.57 (0.51-0.65); men: HR = 0.64 (0.59-0.69)	Women: <.001; men: <.001
					>80%	Women: HR = 0.59 (0.53-0.66); men: HR = 0.60 (0.56-0.64)	Women: <.001; men: <.001
Dragomir et al (2010)²⁹	Cerebrovascular events	431, 432, 433, 434, 436, 437 (ICD-9)	2189/55 134	MPR	<80%	Reference
	Cerebrovascular events	431, 432, 433, 434, 436, 437 (ICD-9)	2189/55 134	MPR	≥80%	OR: 0.87 (0.75-0.99)
	Chronic heart failure	398.91, 402.01, 402.11, 402.91, 428.0, 428.1, 428.9 (ICD-9)	2171/55 134	MPR	<80%	Reference
	Chronic heart failure		2171/55 134	MPR	≥80%	OR: 0.87 (0.85-0.99)
Slejko et al (2014)²⁵	Myocardial infarction or stroke or heart failure	410.xx, 430.xx-438.xx, 428.xx (ICD-9)	828/11 126	PDC year 2	<20%	Reference
					20%-80%	HR: 0.48 (0.45-0.51)
					≥80%	HR: 0.44 (0.35-0.56)
				PDC year 3	<20%	Reference
					20%-80%	HR: 0.43 (0.42-0.44)
					≥80%	HR: 0.36 (0.28-0.48)
Rublee et al (2012)³²	Cardiovascular events: hospitalization with the primary diagnosis as heart attack, angina, stroke, fugax, transient ischemic attack, peripheral or central nervous system vascular disease, aortic aneurism, congestive heart failure, heart revascularization, or percutaneous coronary intervention	36.03, 36.06, 36.07, 36.1x, 36.2x, 36.3x, 62.34, 250.7, 410.xx, 411.xx, 412, 413.xx, 414.2x–414.9x, 429.2x, 429.7x, 433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.1x, 435.xx, 433.0, 433.10, 433.20, 433.30, 433.80, 433.90, 437.0, 437.1, 434.0x, 445.xx, 440.xx, 441.xx, 428.x (ICD-9)	1463/79 010	PDC	<60%	Reference
Rublee et al (2012)³²			1463/79 010	PDC	>60%	HR: 0.82 (0.74-0.91)
Poluzzi et al (2011)²⁶	Cardiovascular events: ischemic heart disease, cerebrovascular disease, atherosclerosis, other peripheral vascular disease, arterial embolism, and thrombosis	410-414, 430-438, 440, 443, 444 (ICD-9)	24 238/108 073	Number of tablets purchased	Nonadherent patients	Reference
Poluzzi et al (2011)²⁶		410-414, 430-438, 440, 443, 444 (ICD-9)	24 238/108 073	Number of tablets purchased	Adherent patients	OR: 0.71 (0.69-0.74)
Chapman et al (2010)³³	Myocardial infarction, other ischemic heart disease including unstable angina, stroke/TIA, peripheral vascular disease, angina with hospitalization, and revascularization procedure	410.xx, 412, 411.xx, 414.xx, 427.xx, 433.xx, 434.xx, 435.x, 436, 437.0, 437.1, 438, 440, 440.1, 443.xx, 413.xx, 33503-33545, 35301, 35390, 35901, 92980, 92981, 92973, 92982, 92984, 92995, 92996, 92997, 92998 (ICD-9)	818/19 447	PDC: adherent CCB/statin	<80%	Reference
				PDC: adherent CCB/statin	≥80%	HR: 0.79 (0.66-0.95)	.01
				PDC: adherent SPAA	<80%	Reference
				PDC: adherent SPAA	≥80%	HR: 0.61 (0.39-0.95)	.03
WOSCOP (1999)³¹	Atherosclerotic CHD	Not detailed	N = 6595	Record for visit: issue or not of a supply of trial medication	<75%	Reference
	Atherosclerotic CHD				≥75%	HR: 0.65 (0.53-0.81)	.0001
	All CV causes			Record for visit: issue or not of a supply of trial medication	<75%	Reference
	All CV causes				≥75%	HR: 0.73 (0.62-0.85)	.0001
Rannanheimo et al (2015)³⁵	Any cerebrovascular disease event	I63-I66, I69, and G45 or related medical procedure (ICD-10)	N = 97 575	PDC	< 40%	Reference
					40%-79%	HR: 0.82 (0.72-0.93)
					≥80%	HR: 0.68 (0.60-0.76)
	Acute cerebrovascular disease event	I63, I64, and G45 (ICD-10)			> 40%	Reference
					40%-79%	HR: 0.84 (0.73-0.96)
					≥80%	HR: 0.67 (0.60-0.76)
	Chronic cerebrovascular disease event	I65, I66 and I69 or related medical procedure (ICD-10)			<40%	Reference
					40%-79%	HR: 0.70 (0.52-0.94)
					≥80%	HR: 0.66 (0.50-0.85)
Lavikainen et al (2016)³⁴	Acute cardiovascular events	I20.0, I21-I22, I63 (ICD-10)	470/42 807	PDC	<80%	Reference
Lavikainen et al (2016)³⁴	Acute cardiovascular events	I20.0, I21-I22, I63 (ICD-10)	470/42 807	PDC	≥80%	HR: 0.76 (0.63-0.91)
Study (year)	Outcomes		n/N	Assessment of exposure: Adherence		Relative Risk (95% CI)	P
Mortality
Degli Esposti et al (2012)²⁴	All cause death		765/19 674	PDC	21%-40%	Reference
					41%-60%	HR: 0.81 (0.66-0.99)	.0046
					61%-80%	HR: 0.53 (0.43-0.66)	<.001
					>80%	HR: 0.46 (0.38-0.55)	<.001
Shalev et al (2009)²⁷	All cause death		4259/136 052	PDC	<10%	Reference
					10%-19%	HR: 1.35 (1.22-1.50)
					20%-29%	HR: 1.07 (0.95-1.21)
					30%-39%	HR: 0.88 (0.77-1.00)
					40%-49%	HR: 0.86 (0.75-0.98)
					50%-59%	HR: 0.77 (0.67-0.88)
					60%-69%	HR: 0.63 (0.55-0.73)
					70%-79%	HR: 0.59 (0.51-0.68)
					80%-89%	HR: 0.61 (0.53-0.69)
					≥ 90%	HR: 0.55 (0.49-0.61)
WOSCOP (1997)³⁰	All cause death		241/6595	Record for visit: issue or not of a supply of trial medication	<75%	Reference
	All cause death				≥75%	HR: 0.68 (0.49-0.93)	.015
	Cardiovascular death			Record for visit: issue or not of a supply of trial medication	<75%	Reference
	Cardiovascular death				≥75%	HR: 0.63 (0.42-0.95)	.029
Rannanheimo et al (2015)³⁵	Death		N = 97 575	PDC	> 40%	Reference
					40%-79%	HR: 0.86 (0.74-1.01)
					≥80%	HR: 0.71 (0.62-0.82)
Study (year)	Outcomes	ICD-9	n/N	Assessment of exposure: Adherence	Statin adherence: Cut points	Risk estimation (95% CI)	P
Others outcomes
Degli Esposti et al (2012)²⁴	Stroke or AMI or all cause death	430-438.x, 410.x (ICD-9)	1382/19 232	PDC	21%-40%	Reference
					41%-60%	HR: 0.83 (0.71-0.98)	.026
					61%-80%	HR: 0.60 (0.51-0.70)	<.001
					>80%	HR: 0.61 (0.54-0.71)	<.001
Rannanheimo et al (2015)³⁵	Any CV events or death	I20-25, I63-66, I69; G45, I70, or related medical procedure, ATC Code C01D (ICD-10)	9002/97 575	PDC	> 40%	Reference
					40%-79%	HR: 0.86 (0.81-0.91)
					≥80%	HR: 0.75 (0.71-0.79)

Abbreviations: AMI, Acute Miocardial Infarction; ATC, Anatomical Therapeutic Chemical classification system; CCB, calcium-channel blocker; CHD, Coronary Heart Disease; CI, confidence interval; CV, cardiovascular; HR, Hazard Ratio; ICD, International Classification of Diseases; IHD, ischemic heart disease; MPR, medication possession rate; OR, odds ratio; PDC, proportion of days covered; SPAA, single-pill amlodipine/atorvastatin.

Table 4.

Effects of Adherence on Critical Cardiovascular Outcomes: Nested Case–Control in the First Year and After First Year of Follow-Up and Their Controls.

Study: First Author (year)	Outcomes	ICD-9 codes	Assessment of exposure: Adherence	Statin Adherence: Cut points	RR (95% CI)
Study: First Author (year)	Outcomes	ICD-9 codes	Assessment of exposure: Adherence	Statin Adherence: Cut points	Cases Occurring in the First Year of Follow-Up and Their Controls	Cases Occurring After 1 Year of Follow-Up and Their Controls
Perreault et al (2009)³⁷	Cerebrovascular disease	431, 433, 434, 436, 437	MPR	1%-19%	Reference	Reference
				20%-39%	1.06 (0.74-1.53)	0.97 (0.83-1.15)
				40%-59%	1.24 (0.87-1.77)	0.83 (0.70-0.98)
				60%-79%	1.09 (0.77-1.53)	0.82 (0.71-0.96)
				≥ 80%	1.03 (0.76-1.38)	0.74 (0.65-0,84)
Perreault et al (2009)³⁸	Coronary artery disease	410-414	MPR	1%-19%	Reference	Reference
				20%-39%	0.93 (0.78-1.10)	0.91 (0.84-1.01)
				40%-59%	1.00 (0.85-1.18)	0.87 (0.80-1.00)
				60%-79%	0.99 (0.85-1.15)	0.85 (0.78-0.92)
				≥80%	0.88 (0.77-1.01)	0.82 (0.77-0.87)
Bouchard et al (2007)³⁶	Coronary artery disease	410-414	Percentage of the prescribed doses of medication actually by the patient over a specified period	< 90%	Reference	Reference
Bouchard et al (2007)³⁶	Coronary artery disease	410-414		≥ 90%	0.92 (0.82-1.03)	0.81 (0.67-0.97)
Perrault et al (2008)³⁹	Chronic heart failure	398.91, 402.01, 402.11, 402.91, 428.0, 428.1, 428.9	MPR	1%-19%	Reference	Reference
				20%-39%	0,70 (0.46-1.03)	0.90 (0.76-1.07)
				40%-59%	1.01 (0.70-1.45)	0.91 (0.77-1.08)
				60%-79%	0.62 (0.43-0.99)	0.88 (0.75-1.03)
				≥80%	0.72 (0.53-0.98)	0.81 (0.71-0.91)

Abbreviations: CI, confidence interval; ICD, International Classification of Diseases; MPR, medication possession rate; RR, relative risk.

The nested case–control studies stratified their results by outcomes in the first year of statin use and outcomes in subsequent years (Table 4). Figure 2 shows the increase or reduction in CVD risk in nested case–control studies comparing highest and lowest levels of adherence, stratified by time since starting statins. The relationship between the level of adherence and clinical outcome tends to be stronger after the first year for CeVD, and CAD, although not apparently for CHF.

Figure 2.

Adherence and relative risk clinical outcomes: nested case–control in the first year and after first year of follow-up and their controls.

Pooled Comparison of Best Adherer Versus Worse Adherer Patients

Figure 3 shows the forest plots of pooled RRs for outcomes of ischemic heart disease (IHD), cardiovascular events, cerebrovascular events, and mortality for studies that compared the best adherer patients (≥75%) with the worst adherers (<40%). For the 3 studies (n/N = 21 050/317 603) that considered IHD in this group, the pooled reduction in risk was 18% (95% CI: 14%-22%, I² = 0%). This means that best adherers have an 18% lower risk of IHD events than the worst adherents. For the CVD outcome, 2 studies (n/N = 10 987/182 661) showed a pooled reduction in risk of 47% (95% CI: 36%-56%, I² = 84.7%), with 1 study showing a much bigger effect than the other.²⁵ For the CeVD outcome, 2 studies (n/N = 4511/131 477) showed a pooled reduction in risk of 26% (95% CI: 18%-34%, I² = 0%). A pooled risk reduction of 49% (95% CI: 39%-57%, I² = 62.4%) was observed for mortality in 2 studies (n/N = 5024/155 726).

Figure 3.

Meta-analysis 1: Comparison of relative risk of cardiovascular events between best adherent group and worst adherent group.

Pooled Comparison of Best Adherers Versus All Other Patients

Figure 4 shows a pooled risk reduction of 22% (95% CI: 16%-27%, I² = 0%) for all CVD events, for 4 studies (n/N = 5474/147 859) that compared the most adherent group (≥60%) with the rest (<60%).

Figure 4.

Meta-analysis 2: Comparison of relative risk of cardiovascular events between patients above a given cut point of adherence with those below.

Discussion

The results of this review show the benefits of good adherence to statin medication on primary prevention of CVD. On average over all end points, the most adherent patients had 32% fewer mortality, IHD, stroke, or cardiovascular events than the least adherent, and the most adherent had 22% fewer CVD events when compared with the rest of the sample.

Other studies have reviewed adherence to statins in primary and secondary prevention,^{19
–21,40,41} but as far as we know, this is the first systematic review conducted in primary prevention that summarized results using meta-analysis. We only included articles in English and Spanish, and we did not search the “grey” literature (ie, non-peer-reviewed articles in unconventional information channels). We may therefore have missed some relevant studies, though the risk of bias is likely to be small.

There was considerable heterogeneity in the definition of the variables across the included studies, in terms of the measures of adherence used, the cut points, the study design, and the CVD outcome. Most studies classified adherence according to behavior during the entire study prior to the CVD event, and only 2 studies classified patients according to their level of adherence prior to the index date.^27,36 We controlled for this heterogeneity to some extent by stratifying the analysis by CVD end point, type of study design, and cut point. While the quality of these studies was considered generally reasonable, these are observational data and hence have all the corresponding caveats and confounders. For example, patients who adhere to treatment also may be more likely to adhere to other aspects of healthy lifestyle. Under these conditions, the observed effect would be picking up the impact of healthy lifestyle on CVD, not just adherence to statin therapy. The majority of the studies performed a Cox regression to adjust for differences in observed variables at baseline, which controls at least for observed confounders. Other factors such as intensity of statin,^42,43 which we were unable to take account of in this study, may influence both the effectiveness of statins and the degree of adherence.

The articles included in this review evaluated the degree of adherence by indirect methods based on the analysis of the number of units of statin dispensed (the majority of them based on refill claims data over long time periods). This is more properly described as “persistence” than adherence. The advantage is that these data can be collected objectively at low cost and represent the “real-world” settings. However, the measure lacks the patient’s perspective as it does not take account the patient’s acceptance of the physician’s advice nor compliance with the dosing, timing, and so on. Most studies categorized patients as adherent according to PDC or MPR calculated from the start of follow-up to the time of a cardiovascular events, disenrollment from health system, death, or end of time of follow-up, whichever occurred first,^{23

–29,36
–38} and used this as the “exposure” variable. An exception was Slejko et al²⁵ who estimated adherence in yearly segments from the index statin prescription. Furthermore, Lavikainen et al³⁴ and Rannanheimo et al³⁵ calculated the statins adherence in intervals of 1 year from statin initiation, then adherence models were used in order to estimate each individual’s probability of having the observed adherence at 3 time points: 12, 24, and 36 months from statin treatment initiation. Another difference observed in the measurement of adherence was some studies have suggested that the best results are obtained after a full year at good levels of adherence, and they only included the results are obtained before this time.^{29,39
–41,44} Despite the different definitions, and cut points used in the measure of adherence to statins, our meta-analytic approach, by stratifying for these factors, may be considered a reasonable approximation.⁴¹ These results reinforce the recommendation about the need for standardization of measures of adherence and the common data elements involved in key cardiovascular end point events for future research.^42,43

The primary motivation for this study was to estimate how effective statins might be for primary prevention of CVD in the real-world settings. Self-selection into RCTs may end up recruiting patients who are most likely to be adherent, and this may imply that RCTs overstate the benefits of statins.^45,46 Furthermore, patient adherence in the real world is not as easily monitored or evaluated.⁴⁴ Investigation is needed into interventions that improve adherence in the real world. Improved adherence might increase pharmaceutical costs to payers (if more statins are dispensed) and may imply the need for more intensive follow-up in primary care for some patients but should also improve health outcomes and could be highly cost-effective,^29,47,48 though there are unlikely to be easy solutions.^48,49

This study at least highlights the need for health professionals to identify and work with patients at risk of poor adherence.^15,40 Health professionals themselves may need to develop or to be trained in this area. World Health Organization⁷ promote an “adherence counseling toolkit” adaptable to different socioeconomic settings and covering patients’ information needs, clinical decision making, and behavioral tools for health professional.

Footnotes

Acknowledgments

The authors are grateful to Simon Thompson, Stephen Kaptoge, Daysi Yoe Ling Chang Chan for advice and support, and Camila Higueras, librarian EASP.

Author Contributions

Martin-Ruiz and Antonio Olry-de-Labry Lima contributed to design; acquisition, analysis, and interpretation; drafted manuscript; critically revised manuscript; gave final approval; and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Ricardo Ocaña-Riola contributed to analysis and interpretation, critically revised manuscript, gave final approval, and agrees to be accountable for all aspects of work ensuring integrity and accuracy. Epstein, D contributed to conception and design; acquisition, analysis, and interpretation; drafted manuscript; critically revised manuscript; gave final approval; and agrees to be accountable for all aspects of work ensuring integrity and accuracy.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was financed by the EPIC-CVD project under the European Commission. Directorate General for research & Innovation FP 7 Health 2011. Grant Agreement number 279233.

References

The top ten causes of death. 2017. http://www.who.int/mediacentre/factsheets/fs310/en/. (accessed 29 November 2017).

World Health Organization. Global Status Report on Noncommunicable Diseases 2014. Geneva, Switzerland: World Health Organization; 2014.

Robinson

. Starting primary prevention earlier with statins. Am J Cardiol. 2014;114(9):1437–1442.

Taylor

Huffman

Macedo

. Statins for the primary prevention of cardiovascular disease. Cochrane Database Syst Rev. 2013;31(1):CD004816. doi:10.1002/14651858.CD004816.pub5.

Reiner

De Bacquer

Kotseva

. Treatment potential for dyslipidemia management in patients with coronary heart disease across Europe: findings from the EUROASPIRE III survey. Atherosclerosis. 2013;231(2):300–307.

Desai

Martin

Blumenthal

. Non-cardiovascular effects associated with statins. BMJ. 2014;349:g3743.

Adherence to Long-Term Therapies—Evidence for Action. World Health Organization; 2003. Available at: http://apps.who.int/medicinedocs/en/d/Js4883e/

Cramer

Roy

Burrell

. Medication compliance and persistence: terminology and definitions. Value Health. 2008;11(1):44–47.

López

. Treatment adhesion, treatment adherence or treatment compliance? Rev Clin Esp. 2008;206(8):414.

10.

Yeaw

Benner

Walt

. Comparing adherence and persistence across 6 chronic medication classes. J Manag Care Pharm. 2009;15(9):728–740.

11.

Naderi

Bestwick

Wald

Sian

Smith

. Adherence to drugs that prevent cardiovascular disease: meta-analysis on 376,162 patients. Am J Med. 2012;125(9):882–887.e1.

12.

Lemstra

Blackburn

Crawley

. Proportion and risk indicators of nonadherence to statin therapy: a meta-analysis. Can J Cardiol. 2012;28(5):574–580.

13.

Kronish

. Adherence to cardiovascular medications: lessons learned and future directions. Prog Cardiovasc Dis. 2013;55(6):590–600.

14.

Cheung

Lauder

Lau

Kumana

. Meta-analysis of large randomized controlled trials to evaluate the impact of statins on cardiovascular outcomes. Br J Clin Pharmacol. 2004;57(5):640-651.

15.

Whooley

de Jonge

Vittinghoff

. Depressive symptoms, health behaviors, and risk of cardiovascular events in patients with coronary heart disease. JAMA. 2008;300(20):2379–2388.

16.

Rasmussen

Chong

Alter

. Relationship between adherence to evidence-based pharmacotherapy and long-term mortality after acute myocardial infarction. JAMA. 2007;297(2):177–186.

17.

Wells

Shea

O’Connell

. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. (accessed 15 June 2017)

18.

Higgins

Thompson

. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539–1558.

19.

Bitton

Choudhry

Matlin

Swanton

Shrank

. The impact of medication adherence on coronary artery disease costs and outcomes: a systematic review. Am J Med. 2013;126(4):357.e7-e357.e27.

20.

Gomez-Sandoval

Braganza

Daskalopoulou

. Statin discontinuation in high-risk patients: a systematic review of the evidence. Curr Pharm Des. 2011;17(33):3669–3689.

21.

Simpson

Mendys

. The effects of adherence and persistence on clinical outcomes in patients treated with statins: a systematic review. J Clin Lipidol. 2010;4(6):462–471.

22.

Phan

Gomez

Elbaz

. Statin treatment non-adherence and discontinuation: clinical implications and potential solutions. Curr Pharm Des. 2014;20(40):6314–6324.

23.

Corrao

Conti

Merlino

. Results of a retrospective database analysis of adherence to statin therapy and risk of nonfatal ischemic heart disease in daily clinical practice in Italy. Clin Ther. 2010;32(2):300–310.

24.

Degli Esposti

Saragoni

Batacchi

. Adherence to statin treatment and health outcomes in an Italian cohort of newly treated patients: results from an administrative database analysis. Clin Ther. 2012;34(1):190–199.

25.

Slejko

Anderson

Nair

Sullivan

Campbell

. Adherence to statins in primary prevention: yearly adherence changes and outcomes. J Manag Care Pharm. 2014;20(1):51–57.

26.

Poluzzi

Piccinni

Carta

. Cardiovascular events in statin recipients: impact of adherence to treatment in a 3-year record linkage study. Eur J Clin Pharmacol. 2011;67(4):407–414.

27.

Shalev

Chodick

Silber

Kokia

Jan

Heymann

. Continuation of statin treatment and all-cause mortality: a population-based cohort study. Arch Intern Med. 2009;169(3):260–268.

28.

Shalev

Goldshtein

Porath

Weitzman

Shemer

Chodick

. Continuation of statin therapy and primary prevention of nonfatal cardiovascular events. Am J Cardiol. 2012;110(12):1779–1786.

29.

Dragomir

Côté

White

. Relationship between adherence level to statins, clinical issues and health-care costs in real-life clinical setting. Value Health. 2010;13(1):87–94.

30.

Compliance and adverse event withdrawal: their impact on the West of Scotland Coronary Prevention Study. Eur Heart J. 1997;18(11):1718–1724.

31.

The effects of pravastatin on hospital admission in hypercholesterolemic middle-aged men: West of Scotland Coronary Prevention Study. J Am Coll Cardiol. 1999;33(4):909–915.

32.

Rublee

Chen

Mardekian

Rao

Boulanger

. Evaluation of cardiovascular morbidity associated with adherence to atorvastatin therapy. Am J Ther. 2012;19(1):24–32.

33.

Chapman

Yeaw

Roberts

. Association between adherence to calcium-channel blocker and statin medications and likelihood of cardiovascular events among US managed care enrollees. BMC Cardiovasc Disord. 2010;10:29.

34.

Lavikainen

Helin-Salmivaara

Eerola

. Statin adherence and risk of acute cardiovascular events among women: a cohort study accounting for time-dependent confounding affected by previous adherence. BMJ Open. 2016;6(6):e011306. doi:10.1136/bmjopen-2016-011306.

35.

Rannanheimo

Tiittanen

Hartikainen

. Impact of statin adherence on cardiovascular morbidity and all-cause mortality in the primary prevention of cardiovascular disease: a population-based cohort study in Finland. Value Health. 2015;18(6):896–905. doi:10.1016/j.jval.2015.06.002.

36.

Bouchard

Dragomir

Blais

Bérard

Pilon

Perreault

. Impact of adherence to statins on coronary artery disease in primary prevention. Br J Clin Pharmacol. 2007;63(6):698–708.

37.

Perreault

Ellia

Dragomir

. Effect of statin adherence on cerebrovascular disease in primary prevention. Am J Med. 2009;122(7):647–655.

38.

Perreault

Dragomir

Blais

. Impact of better adherence to statin agents in the primary prevention of coronary artery disease. Eur J Clin Pharmacol. 2009;65(10):1013–1024.

39.

Perreault

Dragomir

Blais

Bérard

Lalonde

White

. Impact of adherence to statins on chronic heart failure in primary prevention. Br J Clin Pharmacol. 2008;66(5):706–716.

40.

De Vera

Bhole

Burns

Lacaille

. Impact of statin adherence on cardiovascular disease and mortality outcomes: a systematic review. Br J Clin Pharmacol. 2014;78(4):684–698.

41.

Chowdhury

Khan

Heydon

. Adherence to cardiovascular therapy: a meta-analysis of prevalence and clinical consequences. Eur Heart J. 2013;34(38):2940–2948.

42.

Cardiovascular Disease: Risk Assessment and Reduction, Including Lipid Modification. [CG181]. London, UK: National Institute for Health and Care Excellence (UK); 2014. Updated 2016.

43.

Hicks

Tcheng

Bozkurt

. 2014 ACC/AHA key data elements and definitions for cardiovascular endpoint events in clinical trials: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Data Standards (Writing Committee to Develop Cardiovascular Endpoints Data Standards). J Am Coll Cardiol. 2015;66(4):403–469.

44.

Osterberg

Blaschke

. Adherence to medication. N Engl J Med. 2005;353(5):487–497.

45.

Christensen

. Patient Adherence to Medical Treatment Regimens: Bridging the Gap Between Behavioral Science and Biomedicine. Yale University Press; 2005.

46.

Kennedy-Martin

Curtis

Faries

. A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials. 2015;16:495.

47.

Pittman

Chen

Bowlin

. Adherence to statins, subsequent healthcare costs, and cardiovascular hospitalizations. Am J Cardiol. 2011;107(11):1662–1666.

48.

Corrao

Scotti

Zambon

. Cost-effectiveness of enhancing adherence to therapy with statins in the setting of primary cardiovascular prevention. Evidence from an empirical approach based on administrative databases. Atherosclerosis. 2011;217(2):479–485.

49.

Nieuwlaat

Wilczynski

Navarro

. Intervention for enhancing medication adherence. Cochrane Database Syst Rev. 2014;(11):CD000011. doi:10.1002/14651858.CD000011.pub4.