Increased risk of coronary heart disease among patients with idiopathic inflammatory myositis: a nationwide population study in Taiwan

Abstract

Objective

To evaluate the risk of incident coronary heart disease (CHD) among patients with DM and PM in a general population context.

Methods

We conducted a retrospective cohort study using the Taiwan National Health Insurance Research Database containing records covering the years from 2000 to 2010. DM and PM were confined for the purposes of this study to those aged ⩾18 years who were eligible for the Taiwan catastrophic illness certificate. The diagnoses, CHD outcomes and cardiovascular risk factors were identified from electronic claims data. We conducted two cohort analyses: CHD and DM, and CHD and PM, excluding for each analysis individuals with CHD already identified at baseline. Data for the comparison group was obtained from the Longitudinal Health Insurance database, comprising 1 million persons randomly sampled from the total beneficiaries during 2000. We estimated hazard ratios comparing myositis with comparison cohorts, adjusting for potential cardiovascular risk factors.

Results

A total of 1145 patients with idiopathic myositis were identified, along with 732 723 control patients aged ⩾18 years. The incidence rates of CHD were 15.1 in DM and 30.1 in PM per 1000 person-years, vs 8.4 and 10.5 per 1000 person-years in the comparison cohort. The adjusted hazard ratios for CHD in patients with idiopathic myositis were 2.21 (95% CI 1.64, 2.99) for DM and 3.73 (95% CI 2.83, 4.90) for PM.

Conclusion

Results of this general population-based cohort study suggest that DM and PM are associated with an increased risk of CHD.

Introduction

Coronary heart disease (CHD) is a major health problem in developed countries [1, 2], and acute myocardial infarction (AMI) is a potentially lethal complication. There is growing consensus regarding the significance of links between inflammation and CHD, and increasing recognition that the innate and adaptive immune systems play an important role in the initiation and progression of atherosclerosis [3, 4].

DM and PM are idiopathic inflammatory myopathies, characterized by the shared features of symmetric proximal skeletal muscle weakness and evidence of chronic muscle inflammation [5–7]. DM, unlike PM, is associated with a range of characteristic skin manifestations including Gottron’s papules and the heliotrope eruption [8]. Both DM and PM are relatively rare systemic autoimmune diseases [9, 10]. In a group of patients with primary PM/DM, cardiovascular involvement was the most common cause of death [11–13]. However, whether patients with DM or PM have higher risk of developing CHD when compared with those without is not certain due to the lack of controlled studies.

There is evidence to suggest that a T_h1 immune response, involving activated T cells, antigen presenting cells, cytokines and systemic inflammation markers such as CRP, is important to the development of atherosclerosis and ultimately, CHD [14–18]. The objective of our study was to determine using a population-based cohort whether forms of idiopathic inflammatory myositis, including DM and PM, are identifiable risk factors for CHD, based on established theory for atherosclerotic inflammation.

Methods

Patients and data sources

Taiwan launched a single-payer National Health Insurance (NHI) programme on 1 March 1995. In 2014, 99.9% of Taiwan’s population was enrolled [19]. The NHI Research Database (NHIRD) has been accessible to researchers in an electronically encrypted form since 1999. The NHIRD registry of files for catastrophic illness patients was used to establish the cohort of idiopathic myositis in our study. To avoid severe financial hardship for families coping with major illnesses, the NHI designates 31 categories of catastrophic illness, with DM and PM included. To obtain a catastrophic illness certificate, the attending physician of a patient diagnosed as DM or PM is required to provide relevant clinical and laboratory information as part of the application for review. The review committee then assesses the application based on the definite Bohan and Peter criteria for each diagnosis, and subsequent approval indicates eligibility for the catastrophic illness certificate and exemption from co-payment [20]. Therefore, the catastrophic illness file has very high validity.

The comparison group was compiled from the Longitudinal Health Insurance Database 2000 [21]. The Longitudinal Health Insurance Database 2000 contains the entire original claims data for 1 000 000 beneficiaries randomly sampled from the entire population of NHI beneficiaries in the year 2000. We used Taiwan’s Longitudinal Health Insurance Database 2000 data, covering the 10-year period ending 31 December 2010 as our comparison cohort. The large sample sizes and reliability of catastrophic illness-related diagnoses in the claims data have ensured that this data set provides a valid source to estimate the incidence of CHD among patients with idiopathic inflammatory myositis (IIM).

All data in this study were anonymous. This study was approved by the Institutional Review Board (A-EX-107-047) of National Cheng-Kung University Hospital, Tainan, Taiwan.

Definition of idiopathic myositis and coronary heart disease

The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), was used for coding the diagnosis of disease. A retrospective, population-based cohort in Taiwan of patients with DM (ICD-9-CM 710.3) or PM (ICD-9-CM 710.4) was confined to individuals aged ⩾18 years between 2000 and 2010, and who had been approved for the catastrophic illness certificate as a result of their DM or PM.

For solid outcome, incident CHD cases appearing in our study included AMI (ICD-9-CM 410) and ischaemic heart disease (ICD-9-CM 411), these having been retrieved from the inpatient diagnosis claims data as the principle diagnosis. The validity of AMI diagnosis coding in Taiwan NHIRD has been done and the positive predictive value for AMI was 0.88. The positive predictive value increased to 0.93 when using only the principal diagnosis in the NHIRD [22].

To avoid inclusion of high-risk patients, we estimated the risk of CHD only in patients with DM or PM who had no previous admission for coronary artery disease (ICD-9-CM 410, 411, 412, 413, 414) 3 years prior to the first admission for CHD.

Study design

A retrospective, population-based cohort was established using data from the Taiwan NHIRD and used to identify the incidence of CHD among individuals with idiopathic myositis (myositis cohorts), as compared with individuals without myositis (comparison groups). Our study spanned the period from 1 January 2000 through to 31 December 2010. With the myositis cohorts, follow-up commenced from the date of catastrophic illness certificate issuance for DM or PM, this date also being matched for the comparison groups (as the index date). When patients are surely diagnosed of DM or PM, their attending physician would apply for a catastrophic illness certificate for them as soon as possible to obtain exemption from co-payment. As a result, the issuance date of catastrophic illness certificate for the myositis patients is very close to the date when disease diagnosis is confirmed.

We excluded patients younger than 18 years, and those who had incomplete information on age or sex. During the follow-up period we linked participants to the admission claims data to identify the first episode of CHD (ICD-9-CM code 410 or 411). From the date of cohort entry, all patients were followed until the development of CHD, withdrawal from the insurance system, death or 31 December 2010, whichever point came first. The control subjects for DM or PM cases were time-matched and randomly selected, then adjusted for age, gender, traditional risks of CHD including diabetes mellitus (ICD-9-CM 648, 250, 249), hypertension (ICD-9-CM 401), dyslipidaemia (ICD-9-CM 272), renal failure (ICD-9-CM 586, 585.9), atherosclerosis (ICD-9-CM 440) and CHD-related medications.

Statistical analysis

Data processing and statistical analysis were performed using SAS statistical software (Version 9.4.1; SAS Institute, Cary, NC, USA). Cox proportional hazards regression models were used to calculate the hazard ratio (HR) of CHD for DM and PM compared with the control group. The Kaplan–Meier method was used to compare the cumulative incidences of CHD in patients with and without myositis. A P-value of <0.05 was considered significant. Additionally, we performed the propensity score method with 5-to-1 greedy matching technique to create more homogeneous groups for comparisons. The propensity score was derived from logistic regression based on known baseline characteristics listed in the above section.

Results

Baseline characteristics

A total of 1145 patients with idiopathic myositis and 732 723 control patients aged ⩾18 years were identified between 2003 and 2010. The demographic background, baseline traditional risk factors and related medication of CHD are shown in Table 1. There were 640 patients with DM and 505 patients with PM in our study. In terms of gender, the comparison group of patients was well balanced with an almost equal number of males and females (49.8% females, numbering 365 039), but female patients were greater in number (at around 70%) among patients with DM or PM, the myositis group in our study. The peak incidences of DM and PM occurred between the ages of 51 and 65 years. There were more DM than PM patients in Taiwan.

The myositis cohort had higher frequencies of traditional risk factors at baseline, including diabetes mellitus, hypertension, dyslipidaemia, renal failure and atherosclerosis; the patients within this cohort also had more frequent exposure to related forms of medication including steroids, antidiabetics, diuretics, beta-blockers, calcium channel blockers, lipid lowing agents, aspirin and NSAIDs. However, after propensity score matching, the demographic characteristics were balanced between DM/PM and control groups (Table 2).

The cohort was followed from 1 January 2003 until 31 December 2010, and the total mean follow-up time was 4 years. During this period, there were 31 (6.1%) and 40 (10.6%) CHD-related hospitalizations among patients with DM and PM, respectively (Table 3). The incidence rate of CHD was higher in the myositis patients than in the comparison group (20.5 vs10.8 per 1000 person-years). Patients with PM had a higher incidence rate of CHD than those with DM (30.0 vs15.1 per 1000 person-years).

Risks of coronary heart disease

DM and PM were both associated with increased risk of CHD. Relative to the comparison cohorts, the crude HRs for CHD were 4.44 (95% CI 3.29, 5.99) for DM and 7.60 (95% CI 5.78, 9.97) for PM (Table 4). After adjusting for traditional risk factors and related medications, the HRs for CHD were reduced slightly to 2.21 (95% CI 1.64, 2.99) and 3.73 (95% CI 2.83, 4.90), respectively, for DM and PM (Table 4). We further used propensity score matched control groups to compare the CHD risks. The results were similar to adjusted HRs, with the P-values all <0.05.

Discussion

The findings of this study suggest that the incidence of hospitalized CHD among patients with DM and/or PM is increased compared with the non-myositis comparison cohort. Cardiovascular manifestations constitute a major cause of death in patients with myositis [11–13, 23, 24], and CHD is known to be the number one cause of death in adults within the general population [1, 2, 25]. However, the risk of CHD in patients with individual DM or PM remains undefined as large population-based epidemiological studies are scarce due to small sample size of the patients. Although DM and PM are idiopathic inflammatory myopathies, they are characterized by the shared features of proximal skeletal muscle weakness and by evidence of muscle inflammation [5–7]. However, the clinical and serologic features of DM and PM vary depending upon immunogenetic and possibly other genetic factors [26].

To our knowledge, ours is the first study dedicated to examining CHD in DM and in PM. The previous studies have shown that patients with IIM have increased risk for CHD. However, because IIM is rare, there has been little research on the risk of CHD in patients with DM or PM. In our study, the risks of incident CHD saw around a 2-fold increase for DM and a 3-fold rise for PM when set beside their respective comparison groups (Fig. 1). Additionally, our longitudinal study was dedicated to the simultaneous examination of the incidence of CHD in inflammatory myositis after adjusting for risk factors of CHD. The prevalence of traditional CHD risk factors is higher in patients with DM and PM. The rates of diabetes, hypertension, dyslipidaemia and atherosclerosis were all higher in myositis patients at the baseline relative to comparison cohorts in our study. This can be attributed to the larger number of older people in the myositis cohort, since these individuals tend to use more CHD-related medication. In our study, the crude HRs of CHD were higher than adjusted HRs. These associations decreased after adjustment for traditional cardiovascular risk factors and relevant medications. Our findings suggest that DM or PM should be regarded as an independent risk factor with respect to CHD.

Our study outcomes echo previous findings linking autoimmune rheumatic disease such as RA with chronic inflammation, and an increased risk of CHD [27, 28]. In the previous studies, RA was associated with increased rates of cardiovascular disease, which cannot be explained by traditional risk factors. Such elevated levels of risk appear attributable to systemic inflammation, both directly via its deleterious effects on endothelial dysfunction and indirectly by its accentuation of multiple risk pathways including lipid abnormalities [18, 29, 30]. The results of our study are consistent with previous studies linking DM/PM to an increased risks of CHD [31, 32]. However, the previous study may have overestimated the risk of CHD in patients with DM/PM [31] because they only included hospitalized patients. Also, there was no information regarding CHD risk in DM and PM separately. The immune mechanisms and anatomic focus of injury within the muscle tissue in PM and DM appear distinct. In our study, we found an association between DM and CHD, and an association between PM and CHD. Outcomes from this research are in line with a wider identification of systemic inflammation having a potential role as the major driver augmenting vascular comorbidity.

First among the strengths of this study is the nationwide population cohort, made accessible through access to a sizeable database of sufficient power for the study of a rare disease. One of the great strengths of population-based study is that, given the severity of the disease, the loss of follow-up or the bias of non-inclusion are likely to be very low or irrelevant. Further, we had access to data on cardiovascular risk factors, alongside that for related use of medication. Second, the high validity of catastrophic illness certificate-related diagnoses in the claims data has ensured that this data set provides a valuable opportunity to estimate the incidence of CHD among patients with idiopathic myositis. Third, we ruled out all those who had a CHD history traceable within 3 years prior to the follow-up period, ensuring that we excluded recurrent patients from our study, thus avoiding any overestimation of risk regarding CHD.

Some limitations should be noted in this study. There was no laboratory information available in the claims data to which our study referred; we thus could not analyse in detail the predicting factors for CHD. We included only patients with a catastrophic illness certificate for DM and PM; thus, any patients with myositis who had not applied for a catastrophic illness certificate regarding the condition would not have been included in the cohort. However, once patients were approved for a catastrophic illness certificate, they would then have been exempted from co-payment; therefore, numbers of patients with myositis but no catastrophic illness certificate would most likely be low. The competing risks including cancers and interstitial lung disease are also concerns, but the effects would be relatively minor due to the number of patients being very small. The underlying pathophysiological mechanisms that may cause cardiac manifestations of IIM involve myocarditis and coronary artery disease, as well as involvement of the small vessels of the myocardium [33]. Accurate diagnosis of cardiovascular involvement in CTDs remains challenging [34]. Endomyocardial biopsies have been used to confirm myocardial inflammation in patients with PM or DM; however, this invasive method has rarely been used in clinical practice [35]. The misclassification of CHD may potentially have occurred in our study, although the cases are scarce. In using the NHIRD, we did not have access to information regarding patient body weight, exercise activity, smoking and diet, all of which may also have a bearing on the appearance and outcomes of CHD. Lastly, although we adjusted the use of steroid and the results did not change substantially after adjustments, potential confounding effects due to the use of steroid should be considered.

Conclusions

The results of this general population-based study suggest that DM and PM are associated with an increased risk of CHD. Our findings will support efforts to improve CHD prevention and monitoring among myositis patients, as CHD risk modification represents an opportunity to further reduce the morbidity and mortality of idiopathic myositis. Further studies are needed to elucidate the exact mechanisms of CHD in myositis patients.

Acknowledgements

All authors participated in the conception, design and analyses of the study. M.-Y.W. drafted the manuscript. All authors contributed to interpretation of the results. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Funding: This work was supported by grants from the National Cheng Kung University Hospital (NCKUH-10606005). The sponsor of the study had no role in study design, data collection, data analysis, data interpretation or writing of the report.

Disclosure statement: The authors have declared no conflicts of interest.

References