Statin initiation and treatment non-adherence following a first acute myocardial infarction in patients with inflammatory rheumatic disease versus the general population

People with inflammatory rheumatic diseases (IRD), including rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), systemic sclerosis (SSc), spondyloarthritis (SpA), vasculitis and other chronic inflammatory diseases of the connective tissues, are known to have increased cardiovascular mortality and morbidity compared to the general population [1]-[9]. This phenomenon is thought to be related to chronic systemic inflammation that underlies the progression of typical atherosclerotic disease and may be accelerated in IRD sub-types [10]. Following a myocardial infarction (MI), RA patients have been shown to have increased fatality rates, with adjusted odds (ORs) ranging from 1.2 to 3.0 [11]-[13]. More recently, we have shown that IRD patients experiencing a first myocardial infarction (MI) have significantly increased adjusted odds of death at 30 days (OR 1.51, 95% CI 1.24, 1.84) when compared to the general population [14]. The majority of deaths were found to be due to cardiovascular causes.

Clinical trials have demonstrated the efficacy of secondary pharmacological prevention strategies for reducing cardiovascular morbidity and mortality in the general population [15]-[18]. Based on this evidence, current guidelines advocate the use of secondary prevention medications (anti-thrombotic therapy, oral beta-blockers, angiotensin-converting-enzyme (ACE) inhibitor and statins) following acute MI, where such therapy is not contra-indicated [19]-[21]. Despite this evidence, our previous research suggests that RA patients receive sub-optimal pharmacotherapy after an MI. In a detailed review of the medical records of 90 RA patients admitted with acute MI, we found treatment with β-blockers (OR 0.42, 95% CI 0.18, 0.96) and statins (OR 0.21, 95% CI 0.09, 0.46) to be significantly lower than matched controls without RA at discharge [22]. The reasons for differential treatment were not accounted for by contraindication to therapy in RA patients. As the study included only three hospitals (two tertiary public hospitals and one large metropolitan private hospital in Victoria, Australia) and was focused on in-hospital treatment only, it is unknown if these findings are generalisable to other settings. Subsequently Lindharsen et al conducted a population-based study of national registries in Denmark, which included 877 patients with RA after a first MI. They found RA patients had significantly lower odds of treatment with aspirin (OR 0.80, 95% CI:0.67, 0.96), β-blockers (OR 0.77, 95% CI 0.65, 0.92) and statins (OR 0.69, 95% CI 0.58, 0.82) [23].

As the therapeutic benefit of secondary pharmacological prevention therapies requires adherence, discontinuing treatment can have serious repercussions on patient outcomes. RA patients in one study had adherence rates of 45.4% to statin therapies over a 4-year period [24] with poor adherence increasing the risk of MI by 2% for each 1-month increase in the duration of non-adherence [25]. Poor adherence was also found to be associated with a significantly increased risk of cardiovascular mortality (hazard ratio (HR) 1.60, 95% CI 1.15, 2.23) and all-cause mortality (HR 1.79, 95% CI 1.46, 2.20) compared to RA patients who adhered to treatment [24].

Data on pharmacotherapy initiation and adherence are lacking for patients with IRD sub-types other than RA. Given the known burden of cardiovascular disease, mortality and increased fatality following an MI in IRD [14], the primary aim of the present study was to compare initiation and non-adherence to statin treatment following a first acute MI in patients with IRD, when compared to the general population.

In a cohort of 21,126 people experiencing a first MI over an eight-year period, including 382 people with IRD, we found approximately 50% of IRD patients were not initiated on statins within the first month following MI. This translates to an adjusted OR of 0.67. Among patients who were initiated on statins, the proportion of patients who adhered to treatment was very good (80%) and not significantly different from the general population, after accounting for the higher mortality in IRD patients. Given the high burden of cardiovascular disease and mortality in IRD patients, ensuring the prompt initiation and continuation of evidence-based therapies is critical for ensuring the best outcomes.

One explanation for the lower rates of statin initiation in IRD patients may be the non-traditional presentation of cardiovascular (CV) risk in this group. IRD patients who were already receiving statins prior to their AMI did not have significantly lower rates of continuation when compared to non-IRD patients. This suggests that the issue with statin initiation relates to identifying and treating statin-naïve IRD patients. Previous guidelines have recommended statin treatment be initiated after MI only when total cholesterol is greater than 5 mmol/l [35]. Changes to guidelines and practice started to occur in the early 2000s when statin use was recommended regardless of lipid levels [36], but it may have taken several years for this recommendation to gain widespread adoption by clinicians.

Patients with SLE, SV and SpA were found to receive less statin treatment compared to the other IRD groups. These patients are known to have a two- to three-fold risk of CV mortality compared to the general population and this risk may be further increased in the presence of traditional CV risk factors [37]-[39]. Statin therapy is recommended due to its lipid-lowering properties and anti-inflammatory effects [37],[40]. However, several reports in the literature have suggested that long-term statin exposure may trigger or aggravate autoimmune diseases, particularly SLE, due to immunomodulatory effects [41]. While the evidence for a causal relationship between statin use and SLE is not well-established, it is unclear if clinicians are hesitant to initiate statin therapy in SLE patients due to this concern. There was also a higher proportion of indigenous Australians in the SLE group (17.7%) [42], and disparities in statin treatment for indigenous populations have been noted previously [43]. A recent study has shown that classic CV risk scores underestimate the risk for patients with PsA when compared with ultrasound assessment of carotid intima thickness, which may contribute to under-treatment in this population [44]. Given the heightened risk of CV disease morbidity and mortality in patients with SLE, SV and SpA, further research is required to understand barriers to statin initiation in these populations.

In their study of initiation and adherence to secondary prevention pharmacotherapies, Linhdardsen et al. found RA patients in Denmark were less likely to be initiated on and adhere to statin therapies following MI [23]. These authors also suggest the added clinical complexity of RA may discourage clinicians from initiating statin treatment, which carries risks of myopathy, rhabdomyolysis, and hepatotoxicity. We also identified lower rates of statin adherence among our IRD cohort, but this was not statistically significant, which may relate to our smaller sample size or to social differences between Denmark and Australia. While prescription of lipid-lowering agents was 79% less (OR 0.21, 95% CI 0.09, 0.46) in our previous study of RA patients over the period from 1995 to 2005 [22], in the present study we found prescription of statins to be 33% lower in IRD patients over the period from 2001 to 2009. Although our data were taken from a different state in Australia, this observation could suggest changing practice over time, as we identified an increasing trend in statin initiation for IRD patients by time period. This trend may coincide with guidelines on the management of acute coronary syndrome published by the Australian National Heart Foundation in 2006, which recommend the use of medications at discharge, including statins, in long-term management after control of the MI [20].

This study included a large, population-based cohort including all incident MIs within the state of Western Australia, however there are several limitations worth noting. It is not known if patients received statins during their hospital admission or had a supply dispensed at the time of discharge, as PBS records do not record hospital drug dispensing. We also assumed that if prescriptions were filled, then the drugs were taken, but this may not always be so. The data on prescription dates were in a month-year format, rather than precise dates. Therefore, we had to make assumptions about the exact duration of scripts and period of adherence. We accepted three months as being an acceptable gap between scripts; however, accepting a full 90-day gap between prescriptions may underestimate non-adherence. The study relied on coded hospital administrative data, which were not designed primarily for research purposes and may therefore contain coding errors. While CV risk factors are known to be under-coded in administrative data, removal of these factors from the multivariate analysis did not change our findings. We chose a five-year look-back period to identify a history of previous MIs. While this may have misclassified some subsequent MIs as the first MI, an individual who experienced an MI during the study period and had a history of previous MI more than five years prior would be likely to receive comparable treatment to individuals with a true first MI [45],[46]. Furthermore, this misclassification is likely to be equal in the IRD and non-IRD groups and therefore not introduce bias. Similarly, the coding of IRD conditions may have inaccuracies and be biased towards only coding the most severe cases. However, we chose a three-year look-back period to capture IRD conditions that may have been coded in a patient’s previous hospital admissions. As our primary analyses examined initiation and non-adherence in IRD patients overall, this may lead to an over-generalisation of findings from one rheumatic disease group to all groups. While we also considered initiation and non-adherence among the IRD sub-types, there was a limited sample size in some populations. The presence of multiple IRD sub-types may be due to coding errors over time or diagnostic uncertainty. As data were de-identified, we were unable to validate diagnoses against a secondary source of information and exclude diagnoses due to potential misclassification.

Despite recent improvements in statin initiation rates, this large population-based study found that patients with IRD were less likely to be initiated on statin therapies following an MI when compared to the general population. However, once initiated, IRD patients were found to adhere to statins at similar rates to non-IRD patients. Given the high risk of CV disease and mortality in IRD populations, increased rates of early initiation on statins following MI as an effective and safe secondary preventative therapy should be encouraged. The underlying factors driving this differential treatment should be examined further to improve care and outcomes of patients with IRD.

MB undertook statistical analysis, interpretation of results and was the lead author of the manuscript. MT participated in the design of the study and undertook statistical analysis. CB, IW and VS participated in the design of the study and interpretation of the statistical analysis. SVD conceived of the study, participated in its design and interpretation of the statistical analysis. All authors were involved in drafting the manuscript and gave final approval of the submitted version of the manuscript. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.