Systematic underutilisation of secondary preventive drugs in patients with acute coronary syndrome and reduced renal function

Abstract

Aims

The high risk of recurrent events in patients with reduced renal function following an acute coronary syndrome (ACS) may in part be due to suboptimal secondary prevention. We aimed to describe the association between renal dysfunction and the prescription, initiation and persistent use of secondary prevention during the first year after a first ACS.

Methods

We identified all patients admitted to any Swedish coronary care unit for a first ACS between 2005 and 2010 (n = 77,432). In 75,129 patients, creatinine levels were available in order to obtain the estimated glomerular filtration rate (eGFR). Persistent use of prescribed drugs was determined for 1 year using the National Prescription Registry, with complete coverage of all prescribed and dispensed drugs in Sweden.

Results

After adjustment for relative and absolute contraindications, compared to patients with eGFR ≥ 60 mL/min/1.73 m², patients with eGFR 30–59 had higher odds of not being prescribed acetylsalicylic acid (ASA; odds ratio [OR]: 1.56, 95% confidence interval [CI]: 1.47–1.67), statins (OR: 2.94, 95% CI: 2.86–3.13) or β-blockade (OR: 1.25, 95% CI: 1.18–1.32). Patients with eGFR 30–59 were more likely to discontinue treatment with ASA (hazard ratio [HR]: 1.59, 95% CI: 1.42–1.56), statins (HR: 1.35, 95% CI: 1.29–1.41), angiotensin-converting enzyme inhibitors and angiotensin-II receptor blockers (HR: 1.37, 95% CI: 1.31–1.43) or β-blockade (HR: 1.22, 95% CI: 1.18–1.27). Patients with eGFR < 30 showed a similar pattern in both prescription and discontinuation.

Conclusion

High-risk ACS patients with reduced renal function are less likely to be prescribed secondary prevention drugs at discharge, are less likely to initiate treatment when being prescribed these drugs, are less likely to be persistent in the use of these drugs and more often discontinue treatment.

Introduction

Patients with acute coronary syndrome (ACS) have a high risk of recurrent events, and this risk is significantly higher in patients with concurrent renal dysfunction.^1,2 Even a minor renal dysfunction is associated with a higher rate of cardiovascular disease.¹ At the same time, about 40% of all patients with ACS have moderate to severe renal dysfunction.³ There are many pathophysiological mechanisms linking cardiac and renal diseases. However, suboptimal usage of acute percutaneous interventions and poor secondary preventive treatment in chronic kidney disease (CKD) could in part explain the poorer prognosis.^2–5 In general, current guidelines recommend the same treatment for these patients as for patients with normal kidney function, except from minor dose adjustments,^6–8 although the evidence in patients with CKD is limited.⁹ A description of the associations between secondary preventive treatment initiation and compliance in patients with renal dysfunction is still lacking and would be valuable in order to design strategies to improve outcomes in this large group of patients.

We used a cohort of nearly all consecutive ACS patients admitted to a coronary care unit in Sweden between 2005 and 2010. We aimed to describe the association between renal dysfunction stages and the prescription, initiation and persistent use of secondary prevention during the first year after a first ACS in all patients who survived to discharge.

Methods

Study population

All patients who were admitted for their first diagnosed ACS between October 2005 and December 2010 were identified using the nationwide Swedish Web-system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommend Therapies (SWEDEHEART) registry. Patients with a previous diagnosed ACS were identified using the SWEDEHEART registry and the National Inpatient Register and were excluded from this cohort. The SWEDEHEART database comprises all Swedish hospitals (n = 72) and enrols all consecutive patients admitted to a coronary care unit. Patients without available serum creatinine levels at admission and patients who died during their hospitalisation in the coronary care unit were excluded from the study (Figure 1). The study complies with the Declaration of Helsinki and was approved by the regional Human Research Ethics Committee in Uppsala, Sweden.

Data sources

Data on baseline characteristics, medication at admission, hospital course variables and drug prescription at discharge were obtained from the SWEDEHEART registry. Information from other relevant registries was obtained using the unique personal identity number that is assigned to each Swedish resident. Data on previous history of diabetes mellitus, hypertension, myocardial infarction (MI) or stroke were also obtained from the National Inpatient Register and death dates were obtained from the Swedish population registry, which includes the vital status of all Swedish citizens. Data on drug dispensations were collected from the National Prescribed Drug Register that records all drugs that have been dispensed in Sweden since July 2005. Four of the established secondary prevention drugs were studied: acetylsalicylic acid (ASA), statins, angiotensin-converting enzyme inhibitors and angiotensin-II receptor blockers (ACEs/ARBs) and β-blockers.

Assessment of renal function

Serum creatinine was collected routinely at the time of patient hospital admission and analysed by either the enzymatic or corrected Jaffe method traceable to isotope dilution mass spectrometry standards. Creatinine measurements performed with non-standardised methods were reduced by 5% prior to being entered into the Chronic Kidney Disease Epidemiology Collaboration version 2009 (CKD-EPI) formula in order to obtain the estimated glomerular filtration rate (eGFR). We lack albuminuria data and therefore describe renal function strata, categorised as mild to normal CKD, stages 1–2 (eGFR ≥ 60 mL/min/1.73 m²), moderate CKD, stage 3 (eGFR 30–59 mL/min/1.73 m²) and severe CKD, stages 4–5 (eGFR < 30 mL/min/1.73 m²).

Assessment of initiation, adherence and persistence

Initiation of treatment was assessed in all patients who survived at least 90 days after the index event. Patients were considered not to initiate treatment if they did not have any dispensed prescription within 90 days.

Adherence was determined as proportions of covered days (PDCs). The PDC is the ratio between the numbers of days covered by the prescription claims of a certain drug divided by the total number of days in the period. A threshold level for PDC of≥80% was used in order to classify patients as adherent or non-adherent.

Persistence in the use of each of the secondary prevention drugs was measured at between 3 and 12 months after the index event. The number of days on which a patient had access to the treatment was estimated according to the quantity of pills and the dosage of each individually dispensed prescription. As daily dosages are recorded as strings in the National Prescribed Drug Register, all textual data were converted to numerical values using a semi-automated supervised script. Drugs from prescription claims exceeding the 365th day after the index event were excluded from the count. Persistence in drug use was measured in each patient for 1 year and patients with a gap of >30 days between the end of dispensed supply and the next dispensed prescription were considered to have discontinued treatment.

Statistical analysis

Patient characteristics were described using the mean and standard deviation for continuous variables and proportions for categorical variables. Logistic regression was used to calculate odds ratios (ORs) for comparisons between renal function groups regarding non-prescription, non-initiation and non-adherence using eGFR > 60 mL/min/1.73 m² as the reference group. We report crude and multivariable adjusted ORs with 95% confidence intervals (CIs). Multivariable adjustment was made using three models. In Model 1, ORs were adjusted for factors that should affect the prescription/compliance to each of the drugs (for ASA: type of MI ST-elevation/non-ST-elevation myocardial infarction [STEMI/NSTEMI], warfarin and/or other antiplatelet treatment at discharge, prior cancer within 3 years and prior bleeding; for statins and ACEs/ARBs: type of MI; for β-blockers: type of MI, AV block II/III, previous peripheral artery disease and chronic obstructive pulmonary disease [COPD]). In Model 2, we additionally adjusted for factors that, besides renal function, could affect prescription/compliance (age, hypertension, diabetes, previous congestive heart failure, peripheral artery disease, previous stroke, COPD, prior cancer within 3 years, atrial fibrillation at discharge, percutaneous coronary intervention performed during hospitalisation and coronary arterial bypass graft surgery). In Model 3, we also adjusted for gender and socioeconomic factors (country of birth, civil status, educational level and income). Hazard ratios (HRs) and 95% CIs for discontinuation of treatment were estimated using Cox proportional hazard regression. Only patients who initiated treatment within 90 days and did not have any early event (death, re-infarction, bleeding or stroke within 90 days) were included. Patients were censored at death, re-infarction, bleeding or any stroke.

Missing data

Data were missing for some baseline variables that were used for adjustment, including MI classification as STEMI or NSTEMI (12.4%), atrial fibrillation at discharge (2.2%), education level (6.8%) and income level (5.1%). We used multiple imputation with logistic regression to handle missing data. Both baseline variables and outcome variables (medications at discharge and 1-year mortality) were used as predictors in the model. We imputed five datasets, and all multivariable analyses were performed on the imputed data for baseline variables. SPSS 22 (IBM) was used for all data management and statistical analyses.

Results

Study population and baseline characteristics

From the SWEDEHART registry, 81,023 patients were identified as being admitted for their first diagnosed ACS between October 2005 and December 2010. After exclusion of patients who died during hospitalisation and patients with unknown serum creatinine levels at their time of admission, 75,129 patients remained, on whom the analyses were based (Figure 1). Among these, 74.5% had an eGFR ≥ 60 mL/min/1.73 m², 21.8% had an eGFR 30–59 and 3.8% had an eGFR < 30. The baseline characteristics are presented in Table 1. Patients with kidney dysfunction were older, more often female and had more comorbidities such as diabetes, hypertension and heart failure. Baseline characteristics for the patients included in the analysis of initiation are presented in had been treated with ACEs/ARBs and 21.0% were on statin treatment.

Prescription

At the time of discharge, 93.2% of patients were prescribed ASA, 5.8% received an oral anticoagulant, 84.4% received statins, 68.1% received an ACE/ARB and 88.6% a β-blocker. The distribution of non-prescription and corresponding ORs for non-prescription for each level of renal function are shown in Table 2. Compared to patients with an eGFR ≥ 60 mL/min/1.73 m², patients with an eGFR 30–59 were equally likely to receive ACEs, but less likely to receive treatment with ASA, statins and β-blockers. Patients with an eGFR < 30 were less likely to receive each of these drugs. These findings were also consistent after adjustment for those factors that should affect the prescription of each of these drugs (Model 1).

Initiation

After exclusion of patients who died within 90 days of their index event, 72,395 patients remained, on whom analyses of initiation were based. The distribution of non-initiation and corresponding ORs for each level of renal function are shown in Table 3 for all patients and in Table S6 (appendix) for those patients who were prescribed the drug at time of discharge (intention-to-treat group). Among patients who were prescribed the drug at discharge, compared to patients with an eGFR ≥ 60 mL/min/1.73 m², patients with an eGFR 30–59 were equally likely to initiate treatment with statins after adjustment for factors that could influence initiation. However, patients with an eGFR 30–59 were less likely to initiate ASA, ACEs/ARBs and β-blockers. Patients with an eGFR < 30 were less likely to receive each of these drugs.

Adherence

Compared to patients with an eGFR ≥ 60 mL/min/1.73 m², patients with an eGFR 30–59 were less likely to reach an adherence level above 80% for each of the four drugs (Table 4). Patients with an eGFR < 30 were even less likely to reach the threshold for adherence. A similar pattern concerning the differences in adherence between the renal function groups was seen in the subgroup of patients who were prescribed the drug at discharge (intention-to-treat group).

Persistence

Out of the patients who initiated treatment and who did not have any early event, those with an eGFR < 30 mL/min/1.73 m² were most likely to discontinue treatment during the first year (Table S8 [appendix]). Patients with an eGFR 30–59 were more likely to discontinue the treatment of each drug than patients with an eGFR ≥ 60. These results were consistent when both unadjusted and adjusted in all three models. Figure 2 shows persistence curves after adjustment according to Model 1.

Discussion

This nationwide, population-based study included all patients who, during a period of over 5 years, were admitted for a first episode of acute coronary heart disease. Our main findings were that these high-risk patients with reduced renal function were less likely to be prescribed guideline-recommended secondary prevention drugs at discharge, were less likely to initiate treatment when being prescribed these drugs, were less likely to be persistent in the use of these drugs and more often discontinued treatment.

Current coronary heart disease guidelines recommend the same secondary preventive drugs for patients with renal dysfunction as for all patients, except for dose adjustment of those drugs that are eliminated by the kidney.^6,8 Here this may apply to ACEs/ARBs, some β-blockers and statins in more severe renal dysfunction. However, since patients with renal dysfunction have often been excluded from randomised cardiovascular studies, most recommendations concerning these patients are based on extrapolation of data from the whole population.⁹ This is supported by a few smaller studies – most of them being registry studies – that have shown a prognostic benefit of the individual secondary preventive drugs in patients with renal dysfunction, as detailed below. At the same time, several studies have shown that adherence rates to guideline recommendations are low both in general^10–14 and in particular in patients with renal dysfunction.^5,15–18 Poor compliance to treatment could be explained by comorbidities, the need to take many pills, side effects or frailty.¹⁹ One study found no difference in (the low) adherence to cardiovascular medications between patients with and without CKD.²⁰ However, a more recent study of adherence over 36 months showed that patients with renal dysfunction had significantly lower long-term ACE/ARB and β-blocker adherence rates, whereas long-term statin adherence did not vary by kidney function.²¹ Although patient-related factors might be important, a study of 3000 patients with ACS revealed that, according to these patients, most stopped their cardiovascular medication based on a physician’s decision related to clinical indications or side effects, while spontaneous self-reported discontinuation was infrequent.²² Our study avoids the issues with physician- and patient-reported medications, as it uses collected prescriptions. In Sweden, all prescribed drugs are part of the government-subsidised programme for all citizens, and there are no economic incentives to ordering drugs outside of this system. Each such prescription should be made to cover a period of 3 months, and this is why we used iterated collected prescriptions. It is thus very likely that the data in the present study truly reflect actual drug usage. Our data confirm that patients with moderately and severely reduced renal function are more likely to discontinue treatment with each of the four studied drugs within 1 year of their index event. This pattern is also seen after adjustment for comorbidities and socioeconomic factors that could affect compliance to treatment.

Registry studies have shown a positive impact of statin therapy on long-term prognosis in acute coronary heart disease patients with CKD.^23,24 The absolute benefit that resulted from the use of pravastatin was greater in patients with CKD than in those without.²⁵ Keough-Ryan et al. showed that lipid-lowering therapy was significantly less likely to be used in patients with CKD.¹⁸ In our cohort, patients with moderately reduced renal function are almost three times less likely to receive statins. The reason for this could be a lack of evidence, since current studies show conflicting results,⁸ and there is a prevalent fear of side effects.²⁶ For statins, reported rates of muscle symptoms are invariably higher in observational studies when compared to blinded randomised controlled trials.²⁷ This could be explained by the fact that patients with comorbidities predisposing to myopathy, such as CKD patients, are excluded from these trials. However, a meta-analysis recently found no evidence of difference in myopathy risk between statin treatment and placebo, even in older patients.²⁸ In order to avoid discontinuation, more focus should be on identifying true statin-associated myopathy.²⁷ Zhang et al. showed that 90% of patients reporting myopathy with statin treatment were able to tolerate an alternative statin with continued use after 12 months.²⁹ In contrast to the study of Chang et al.,²¹ our study also shows that 1-year compliance varied according to baseline kidney function.

Observational studies of elderly patients with reduced left ventricular function after MI have shown that ACEs and β-blockers were associated with greater benefit in patients with renal insufficiency than in patients with preserved renal function.³⁰ We have recently shown in a registry study that prescription of ACEs/ARBs following MI was also associated with improved long-term survival in patients with reduced renal function and was associated with few adverse renal events.⁴ Our results from the present study are consistent with previous studies showing that ACEs are used more frequently in patients with mild renal insufficiency than in those with normal kidney function.^15,17,18 However, patients with an eGFR < 30 mL/minute are less likely to be prescribed ACEs.

In a large cohort of more than 130,000 elderly subjects, patients with normal kidney function received aspirin and β-blockers 20% more often.¹⁵ Further studies have confirmed that patients with renal dysfunction are less likely to immediately receive evidence-based therapies, including aspirin and β-blockers.^16–18 This is consistent with our results showing that patients with reduced renal function are less likely to be prescribed ASA and β-blockers. The higher prevalence of atrial fibrillation and COPD in patients with renal dysfunction could affect the prescription rate of ASA and β-blockers, respectively. However, this pattern remains after adjustment for absolute and/or relative contraindications for ASA, such as treatment with warfarin or other antiplatelet therapy, and for β-blockers COPD or significant AV block. The rationale behind increasing adherence is supported by a prospective registry study in which use of ASA and β-blockers was associated with improved outcomes in patients with CKD.⁵

The reason why patients with renal dysfunction are less likely to receive a secondary prevention drugs is insufficiently understood. It could be due to physicians’ concerns regarding doing further harm to the kidneys or causing unwanted side effects. However, this concern might be misguided in the face of these patients’ significantly higher risk and higher benefit in terms of absolute risk reduction from such treatment. This study highlights that underprescrition and adherence is a complex problem with a lack of sufficient evidence, physician resistance, patient–physician interaction and long-term effects of complex comorbidities.

Our study has several limitations. First, it is a registry-based observational study in which we can prove associations, but not causality. However, contrary to clinical trials, these data describe real-life use and include many patients who would likely be excluded from participation in randomised trials. Since we use pharmacy data regarding collected prescriptions, we cannot be absolutely certain whether a patient actually took the medication or not. However, as a drug prescription in Sweden only covers 3 months at a time, if patients do not refill, we can assume that the large majority of patients have discontinued treatment. Finally, our study does not reveal the reasons for non-prescription, discontinuation or low adherence. Gencer et al. have showed that most patients discontinued medications based on their physicians’ decision, while side effects, perceptions that the medication was unnecessary and cost were less common causes of discontinuation.²² Still, most patients are considered to have remained on their treatment, even though low adherence is very common, especially in CKD patients.¹⁹ This might simply be an effect of the large number of pills these patients are prescribed, but other factors, such as frailty or depression, may also be significant.¹⁹

In conclusion, following an ACS, patients with reduced renal function are less likely to be prescribed guideline-recommended secondary prevention drugs at discharge, are less likely to initiate treatment when being prescribed these drugs, are less likely to be persistent in the use of these drugs and more often discontinue treatment. The combined effects of even small differences in each dimension of treatment might have large prognostic implications for patients with moderate renal dysfunction. These findings highlight the necessity of increased awareness from both physicians and patients of suboptimal adherence to secondary prevention guidelines in patients with renal dysfunction. Further studies should focus on optimal and tailored secondary treatment strategies for patients with reduced renal function in order to improve adherence and prognosis in this large group of high-risk patients.

Author contribution

All authors contributed to either the conception or design of the work. MK, TJ and JS contributed to the acquisition of data for the work. MK, KS, J-JC and JS contributed to analysis. All authors contributed to interpretation. MK and JS drafted the manuscript. All authors critically revised the manuscript, gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy.

Acknowledgement

Mir Khedri is acknowledged for substantial contributions to developing the semi-automated supervised script.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: JS has received speaker honoraria from AbbVie, MSD, ResMed, Medtronics and AstraZeneca. TJ has received speaker honoraria and consultant fees from Astra Zeneca, MSD and Aspen.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was a part of the national TOTAL-AMI project, supported by the Swedish Strategic Research Foundation (SSF). The project also received support from the Swedish Heart and Lung Foundation, the Swedish Research Council and the Stockholm County Council (ALF projects). ME and KS acknowledge support from the Stockholm City Council post-doctorate grants for clinical researchers.

References