The association between nonadherence to medication and morbidity and mortality: the current state of evidence



by S. B. Harrap, Australia

Stephen B. HARRAP, PhD, FRACP
Department of Physiology
University of Melbourne
Victoria, AUSTRALIA




Given the accumulated data of consistent benefits of cardiovascular therapies for hypertension, diabetes, hypercholesterolemia, and related risk factors, it might seem self-evident that nonadherence to prescribed treatment would be bad. Knowing how bad should be part of the discussion between doctors and their patients with the goal of achieving committed and sustained adherence to therapeutic regimes for chronic conditions that are often asymptomatic. Distilling the effect of nonadherence to a specific drug is made difficult by misclassification of adherer status, multimorbidity, the frequent use of complex drug regimens, and the so-called “healthy adherer” effect, where the likelihood of adhering to prescribed medication correlates with general interest in healthy behavior. Nevertheless, there is good evidence that nonadherence comes at a significant cost in terms of increased morbidity, hospitalizations, and death. This is true across all cardiovascular risk factors in primary prevention studies and for secondary prevention of existing coronary artery disease. This brief overview summarizes the methodological approaches and data that have defined the dangers of nonadherence. It is our job as clinicians to share responsibility with our patients to adhere to prescribed treatments and maximize the benefits they deliver.


Chronic disease demands a commitment to consistent care from both physician and patient for the rest of life. Cardiovascular disease and its risk factors, such as hypertension, diabetes and hypercholesterolemia, are not only typical examples of such chronic disease, but also the most common and important causes of death and disability in society today. Effective, safe, and generally readily available modern pharmacological therapies are key elements for reducing cardiovascular risk.

They offer people longer and healthier lives. But only if commitment is maintained and therapy is taken religiously.1 The sad truth is that only about 50%-75% of patients with cardiovascular risk maintain their adherence to prescribed treatments.1,2 The important issue for doctors and patients is the risk that accompanies nonadherence, and this paper examines the available evidence in relation to this question.

Before considering the available data, there are a number of factors that need to be considered related to research designed to determine the impact of nonadherence.

The impact of nonadherence

Defining nonadherence
Consideration of the effects on major cardiovascular outcomes that accompany nonadherence to medications has to take into account the degree of nonadherence. Thus, it is relevant to question: is there a threshold minimum level of adherence needed to obtain benefit? Studies that measure adherence in terms of the proportion of days when patients are assumed to take their medications often define partial adherence for values between 60% and 79% and full adherence for values of ≥80%.3 Early studies demonstrated that reliable blood pressure reductions are achieved only when adherence with antihypertensive drugs was ≥80%.4 This threshold has been validated in studies of statins, β-blockers, and renin-angiotensin system blockers.3 For these reasons, nonadherence is usually defined as taking <80% of the course of prescribed medications.

Measuring nonadherence
The three general approaches to estimating nonadherence offer varying reliability. Least reliable are subjective measurements that depend on information from patients, family, or friends. The factors that might distort such estimates, such as giving the benefit of any doubt, are obvious. More objective methods include pill counting, dispensing records, and medication-related insurance claims. However, these methods cannot prove that tablets were actually taken. More direct assessment of drug intake includes biochemical measurements of drug or drug metabolite levels in individual patients. However, even biochemistry is a single snapshot that may fall prone to what is sometimes called “white coat adherence,” a phenomenon whereby nonadherent patients, knowing that a test is imminent, will take their medications around that time.2 In reality, a variety of methods are used, sometimes systematized into research instruments, such as the Morisky Medication Adherence Scale-8 (MMAS-8)5 and the Adherence Scale by Culig et al.6

Measurement implications
Nevertheless, whichever measure is used, none would be likely to exaggerate nonadherence. This means that the estimates obtained are likely to underestimate the true magnitude of the problem. In comparisons of outcomes between adherent and nonadherent patients, misclassification of nonadherent patients as adherent would dilute any real differences between the groups and serve to undervalue the effect of nonadherence on health or financial outcomes.

Partitioning drug-related effects
The second challenge is to separate the specific effect of not taking prescribed drugs from the effects of other related environment or behaviors that might contribute to risks for cardiovascular problems. Studies of the characteristics of those more likely to be nonadherent to prescribed medications report associations with socioeconomic disadvantage.7 Under these circumstances, the cost of medication and reduced health literacy are direct barriers to adherence. More importantly, disadvantage itself brings with it a broad increase in cardiovascular risks.8

Depression is another comorbidity that is associated with nonadherence, particularly with diabetes.9 The effects of depression extend beyond nonadherence to prescribed medication to factors with a broad health impact, such as missed medical appointments and poor self-care.9 Even marital status has been linked with medication adherence and event-free survival in heart failure.10

Some, but not all, studies have accounted for such confounding in their comparisons of cardiovascular outcomes by adjusting for observed differences in socioeconomic circumstances, comorbidities, etc, between adherent and nonadherent patients.

Confounding by the “healthy adherer” effect
Less easy to detect and account for is a phenomenon known as the “healthy adherer” effect.11 Here, drug adherence reflects an individual’s concern for good health and dedication to a healthy lifestyle in general. It also reflects trust in medical advice. Nonadherent patients are likely to include those for whom a healthy diet and lifestyle are not high priorities. The “healthy adherer” phenomenon has been inferred from the observation in randomized placebo-controlled trials that the outcomes (including mortality) in the placebo group are worse for those who were nonadherent compared with those patients who took their placebos.12 In a meta-analysis of the association between adherence to placebo and mortality in just under 20 000 subjects who took part in 8 cardiovascular or diabetic trials, the average observed effect was a reduction in the risk of mortality by almost half (odds ratio, 0.56; 95% confidence interval [CI], 0.43 to 0.74) with 415 deaths in those who adhered with placebo treatment compared with 581 deaths in those who did not.12





The very existence of a “healthy adherer” effect means it can be difficult to determine the real contribution of taking medication as prescribed. Conversely, the effects of nonadherence are not easy to define, but methods have been devised that look at adherence over time in the same patient and the correlation between periods of lower adherence and the emergence of clinically relevant outcomes. One such method is the “self-controlled case series” approach that allows better control for interpersonal confounders.13 This has been used to follow patients prescribed β-blockers for 1 year after myocardial infarction to examine patterns of recurrent infarction.14 Each individual’s observation time was divided into periods exposed or unexposed to β-blockers (as judged by pharmacy dispensing records) and the relative myocardial infarction incidence rate ratio (IRR) of β-blocker–exposed versus β-blocker–unexposed periods was 0.79 (95% CI, 0.69 to 0.90; P=0.001). In other words, over time adherence to recommended β-blocker therapy was associated with a ≈20% reduction in recurrent myocardial infarction.14 This result was generally consistent with previous research that compared adherent and nonadherent patients, suggesting that at least in studies of the secondary prevention of myocardial infarction, the “healthy adherer” effect might not be a major factor.3

Other studies15 have cast doubt on the “healthy adherer” effect after myocardial infarction. If there were implications for healthy behaviors generally then this might be obvious as reductions in the incidence of other noncardiovascular disease. However, analysis in 31 455 patients over 4 years for lung, prostate, or breast cancer could not define any difference in rates of hospitalization for these cancers between those who were shown to be adherent or nonadherent to cardiovascular medications following myocardial infarction.15

The point is that the drugs themselves make a difference, but they can’t achieve benefit unless the patients take their medication as prescribed. This is true for a number of outcomes for a variety of drugs for cardiovascular disease and diabetes.

Treatment and adherence

Blood pressure treatment
A simple example is the difference in blood pressure in those adherent and nonadherent to antihypertensive medications. A recent meta-analysis16 reported that as many as 45% of 12 603 patients failed to adhere to prescribed antihypertensive treatments as assessed using the MMAS-8 instrument. However, among patients with uncontrolled blood pressure, the prevalence of medication nonadherence was especially high at 84%. Other studies17 have shown high rates of nonadherence (63%) among those classed as having resistant hypertension. This emphasizes that the consequences of nonadherence are not only increased risk for cardiovascular complications of high blood pressure per se, but also the potential exposure to alternative nontrivial treatments of uncertain benefit that might be offered to people with “resistant” hypertension, such as renal denervation.18

Nonadherence to blood pressure medication in hypertensive subjects has significant associations with the risk of major cardiovascular end points such as stroke, myocardial infarction, or cardiovascular death. The ANBP2 (Australian National Blood Pressure–2) study compared the effects of angiotensin-converting enzyme inhibitors with diuretic medications in hypertensive subjects.19 Those who did not adhere to their medications were significantly more likely to experience a cardiovascular event or death from any cause compared with those who adhered to the treatment regime (hazard ratio [HR], 1.28; 95% CI, 1.04 to 1.57).19 Large studies based on population screening of newly treated hypertensive subjects reveal similar findings. An Italian study of 31 306 patients in Florence taking antihypertensive medications for primary cardiovascular prevention revealed that the risk of all-cause death, stroke, or myocardial infarction was approximately halved in those with excellent adherence compared with those with poor adherence (HR, 0.53; 95% CI, 0.46 to 0.61).20

A recent Australian survey also suggested that over 48 months the risk of death was significantly lower in those who persisted better with their prescribed cardiovascular therapies.21 In this case better adherence was reported in those prescribed a combined tablet of perindopril and amlodipine compared with those taking 2 single tablets of an angiotensin inhibition therapy and a calcium channel blocker. The advantages of fixed-dose combination formulations for adherence are discussed elsewhere in this issue.

So, adherence to blood pressure medications has significant benefits in terms of both control of blood pressure and major cardiovascular complications.

Cholesterol treatment
In a parallel analysis of people and records in Florence, Degli Esposti et al examined the differences in cardiovascular outcomes in relation to adherence to treatment with statin cholesterol- lowering drugs.22 Similar rates of adherence were observed for statins as seen for antihypertensive drugs in this Florentine population. Furthermore, high adherence was significantly associated with decreased risk of all-cause death, acute myocardial infarction, or stroke compared with low adherence (HR, 0.61; 95% CI, 0.54 to 0.71). A systematic review of the impact of adherence to statin treatment found consistent evidence across 28 studies that significantly increased risk of cardiovascular events and death was associated with poor adherence to statin drugs for both primary and secondary prevention studies.23

Diabetes treatment
Diabetes is a condition that demands more active self-management than most other cardiovascular risk factors, yet a systematic review of antidiabetic therapies found a wide range of adherence rates, from 36% to 93%.24 Type 2 diabetes is also commonly associated with other cardiovascular risk factors, including hypertension and hypercholesterolemia; only 29% and 52% of diabetic patients achieve guideline targets for blood pressure and low-density lipoprotein cholesterol.25 Whether as a result of poor glycemic control per se or coincident nonadherence to blood pressure and cholesterol medications, the cardiovascular risk profile of diabetic patients who are nonadherent to diabetic medication is especially concerning. In a prospective cohort of 11 532 patients with diabetes mellitus, Ho et al found that the 21% of patients who were nonadherent to diabetic treatments had significantly higher glycosylated hemoglobin, diastolic blood pressure, and low density lipoprotein cholesterol.25 They were 58% more likely to be hospitalized and 81% more likely to die than adherent patients after adjustment for relevant covariates.25

Antiplatelet therapy
Studies of nonadherence to aspirin in primary prevention trials for cardiovascular disease are sparse.26 Where post hoc analyses have been attempted, no significant impact of aspirin nonadherence on total cardiovascular events or mortality could be demonstrated.27

However, the contrast couldn’t be greater for secondary prevention with antiplatelet agents. For example, following insertion of drug-eluting stents for coronary artery stenosis, those patients who were nonadherent during their prescribed 3 to 6 months of thienopyridine therapy (ticlopidine or clopidogrel) were nine times more likely to die (7.5% vs 0.7%) in the year following stent placement than patients adhering to their treatment.28

Heart failure therapy
Given the prognostic implications of untreated heart failure, it might be surprising that as few as 10% of patients are fully compliant with their pharmacotherapy over a full year and only 80% of heart failure patients fill their discharge prescription for angiotensin-converting enzyme inhibitors by 30 days post-discharge.29 Indeed, in one retrospective cohort, nonadherence (defined as <80% adherence) was associated with a doubling of the risk of the primary outcome of all-cause mortality plus cardiovascular hospitalizations.30

Specific drug effects
Are there priorities for pursuing adherence to particular drugs over others? One might imagine that the prognosis of the underlying condition and the estimated magnitude of benefit for a particular drug or drug class might be relevant here. However, it is becoming increasingly difficult to accurately define the beneficial contributions of a specific drug in clinical trials. For example, multiple drugs from different classes are used together commonly for the treatment of hypertension, diabetes, or heart failure. The situation is further complicated by the frequent associations of these conditions. The corollary is that where nonadherence exists, it is unlikely to target a specific drug, so it is difficult to separate the effects of lack of adherence to one drug from another.

The situation is exacerbated in the case of secondary prevention for cardiovascular disease, where the exposure to drugs is greater again with frequent combinations of antiplatelet, β-blocker, statins, and antihypertensive medications. Where this has been studied, it also appears that the effects of nonadherence to individual drugs do not seem to be additive when patients are nonadherent to two or more drugs. For example, in one study of secondary prevention following myocardial infarction,3 adherence to a single agent (statin, β-blocker, or renin-angiotensin system blocker) resulted in an average 23% reduction in risk of a major vascular event or revascularization compared with nonadherent patients, while the comparison of those who were adherent or nonadherent to two of the three classes of drugs showed an average risk reduction of 29%, with only a 35% reduction in risk comparing those adherent and nonadherent for all 3 drug classes. This makes it hard to design targeted approaches to encourage drug adherence towards particular treatments for which nonadherence might have the greatest individual detrimental effects.

Conclusions

Better health, fewer deaths, and reduced costs to the community are all proven benefits of adherence with prescribed cardiovascular medications. The means by which we can attain better adherence are examined elsewhere in this issue. However, our patients need to understand that they cannot afford to miss their cardiovascular medications more than one day per week and should aim for a perfect score. Positive feedback and meaningful involvement of patients in their own care will reinforce and sustain commitment. Above all, we should encourage them to become “healthy adherers” and take advantage of the specific cardiovascular and more general health benefits that follow. ■


References
1. Brown MT, Bussell JK. Medication adherence: WHO cares? Mayo Clin Proc. 2011;86:304-314.
2. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353: 487-497.
3. Choudhry NK, Glynn RJ, Avorn J, et al. Untangling the relationship between medication adherence and post-myocardial infarction outcomes: medication adherence and clinical outcomes. Am Heart J. 2014;167:51-58.
4. Fletcher SW, Appel FA, Bourgeois MA. Management of hypertension. Effect of improving patient compliance for follow-up care. JAMA. 1975;233:242-244.
5. Morisky DE, Ang A, Krousel-Wood M, Ward HJ. Predictive validity of a medication adherence for hypertension control. J Hypertens. 2008;10:348-354.
6. Culig J, Leppée M, Boskovic J, Eric M. Determining the difference in medication compliance between general patient population and patients receiving antihypertensive therapy: a case study. Arch Pharm Res. 2011;34:1143-1152.
7. Kripalani S, Goggins K, Nwosu S, et al; Vanderbilt Inpatient Cohort Study. Medication nonadherence before hospitalization for acute cardiac events. J Health Commun. 2015;20(suppl 2):34-42.
8. Chow CK, Lock K, Teo K, Subramanian S, McKee M, Yusuf S. Environmental and societal influences acting on cardiovascular risk factors and disease at a population level: a review. Int J Epidemiol. 2009;38:1580-1594.
9. Gonzalez JS, Peyrot M, McCarl LA, et al. Depression and diabetes treatment nonadherence. Diabetes Care. 2008;31:2398-2403.
10. Wu JR, Lennie TA, Chung ML, et al. Medication adherence mediates the relationship between marital status and cardiac event-free survival in patients with heart failure. Heart Lung. 2012;41:107-114.
11. Ladova K, Vlcek J, Vytrisalova M, Maly J. Healthy adherer effect – the pitfall in the interpretation of the effect of medication adherence on health outcomes. J Eval Clin Pract. 2014;20:111-116.
12. Simpson SH, Eurich DT, Majumdar SR, et al. A meta-analysis of the association between adherence to drug therapy and mortality. BMJ. 2006;333:15. doi:10.1136/bmj.38875.675486.55.
13. Whitaker HJ, Farrington CP, Spiessens B, Musonda P. Tutorial in biostatistics: the self-controlled case series method. Stat Med. 2006;25:1768-1797.
14. Di Bartolomeo S, Marino M, Guastaroba P, Valent F, De Palma R. Self-controlled case-series study to verify the effect of adherence to beta-blockers in secondary prevention of myocardial infarction. J Am Heart Assoc. 2015;4:e001575.
15. Rasmussen JN, Chong A, Alter DA. Relationship between adherence to evidence- based pharmacotherapy and long-term mortality after acute myocardial infarction. JAMA. 2007;297:177-186.
16. Abegaz TM, Shehab A, Gebreyohannes EA, Bhagavathula AS, Elnour AA. Nonadherence to antihypertensive drugs: a systematic review and meta-analysis. Medicine (Baltimore). 2017;96:e5641.
17. de Oliveira-Filho AD, Costa FA, Neves SJ, de Lyra Junior DP, Morisky DE. Pseudoresistant hypertension due to poor medication adherence. Int J Cardiol. 2014;172:e309-e310.
18. Bhatt DL, Kandzari DE, O’Neill WW, et al; SYMPLICITY HTN-3 Investigators. A controlled trial of renal denervation for resistant hypertension. N Engl J Med. 2014;370:1393-1401.
19. Nelson MR, Reid CM, Ryan P, Willson K, Yelland L. Self-reported adherence with medication and cardiovascular disease outcomes in the Second Australian National Blood Pressure Study (ANBP2). Med J Aust. 2006;185:487-489.
20. Degli Esposti L, Saragoni S, Benemei S, et al. Adherence to antihypertensive medications and health outcomes among newly treated hypertensive patients. Clinicoecon Outcomes Res. 2011;3:47-54.
21. Simons LA, Chung E, Ortiz M. Long-term persistence with single-pill, fixed-dose combination therapy versus two pills of amlodipine and perindopril for hypertension: Australian experience. Curr Med Res Opin. 2017;33(10):1783-1787.
22. Degli Esposti L, Saragoni S, Batacchi P, et al. Adherence to statin treatment and health outcomes in an Italian cohort of newly treated patients: results from an administrative database analysis. Clin Ther. 2012;34:190-199.
23. De Vera MA, Bhole V, Burns LC, Lacaille D. Impact of statin adherence on cardiovascular disease and mortality outcomes: a systematic review. Br J Clin Pharmacol. 2014;78:684-698.
24. Cramer JA. A systematic review of adherence with medications for diabetes. Diabetes Care. 2004;27:1218-1224.
25. Ho PM, Rumsfeld JS, Masoudi FA, et al. Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med. 2006;166:1836-1841.
26. Naderi SH, Bestwick JP, Wald DS. Adherence to drugs that prevent cardiovascular disease: meta-analysis on 376,162 patients. Am J Med. 2012;125:882- 887.
27. Cook NR, Cole SR, Buring JE. Aspirin in the primary prevention of cardiovascular disease in the Women’s Health Study: effect of noncompliance. Eur J Epidemiol. 2012;27:431-438.
28. Spertus JA, Kettelkamp R, Vance C, et al. Prevalence, predictors, and outcomes of premature discontinuation of thienopyridine therapy after drug-eluting stent placement: results from the PREMIER registry. Circulation. 2006;113:2803- 2809.
29. Gandhi J, McCue A, Cole R. Nonadherence in the advanced heart failure population. Curr Heart Fail Rep. 2016;13:77-85.
30. Fitzgerald AA, Powers JD, Ho PM, et al. Impact of medication nonadherence on hospitalizations and mortality in heart failure. J Card Fail. 2011;17:664-669.


Keywords: adherence; cardiovascular disease; cholesterol; compliance; coronary disease; diabetes; hypertension; thrombosis