Increased Risk of Herpes Zoster Following
Dermatomyositis and Polymyositis
A Nationwide Population-Based Cohort Study
Shin-Yi Tsai, MD, MSc, Cheng-Li Lin, MSc, Ying-Chi Wong, MD, Tse-Yen Yang,
PhD,
Chien-Feng Kuo, MD, Jiung-Mou Cheng, MSc, Jyh-Seng Wang, MD, PhD, and
Chia-Hung Kao, MD
INTRODUCTION
D
ermatomyositis (DM) and polymyositis (PM), systemic autoimmune diseases with clinical features associated with inflammatory myopathies,1 are diagnosed according to thepresence of autoantibodies and intravascular complement deposition and supported by distinct immunohistopathological findings
and responses to immunosuppressive therapies.2 In
addition to shared features of symmetrical weakness of the proximal skeletal muscles and evidence of myositis, DM presents with hallmark cutaneous features including a heliotrope
rash, Gottron papules, a Gottron sign, and poikiloderma in photo-exposed areas (V-sign on the neck and upper chest, shawl sign on the upper back).3 Extramuscular manifestations of DM
or PM, such as malignancy, pulmonary and cardiac involvement, and infections, are the most common causes of death.4
Epidemiological studies have demonstrated an increased susceptibility of viral infections in patients with autoimmune diseases.5 Immunocompromised hosts are particularly vulnerable
to varicella-zoster virus (VZV) dissemination, visceral
involvement and reactivation, and an atypical presentation with considerable morbidity and mortality.6 Herpes zoster (HZ)
infection was reported to be associated with DM or PM and increased risk for DM or PM;7 the correlation benefits healthcare
providers to be attentive to the various indications of primary and secondary VZV infection in the patients with
commencing immunosuppression.
However, the correlation has not been verified by a nationwide population-based retrospective cohort study using the National Health Insurance Research Database. Thus, the study aim is to investigate the interaction relationship between these diseases. In this population-based retrospective cohort study, data from the Taiwan National Health Insurance (NHI) database were used to explore the possible association of an increased subsequent HZ risk in patients with DM or PM according to
demographic characteristics and comorbidities (diabetes, renal disease, obesity, cancer, other autoimmune diseases, and therapeutic drugs).
METHODS
Data Source
The Taiwan NHI program was established in 1995, and
approximately 99% of Taiwan residents are enrolled and 97% of medical providers are under contract (http://www.nhi.gov.tw/ english/index.aspx). The present cohort study used data obtained from the NHI Research Database (NHIRD), part of the NHI electronic records system in Taiwan that contains medical claims data from 1996 to 2011 and is maintained by the National Health Research Institutes (NHRI). The NHIRD includes comprehensive data about the clinical visits of insureds, such as outpatient visits, hospital admissions, prescriptions, disease status, and diagnostic codes in the format used in the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM). Details of the NHIRD have been described in previous studies.8,9
The NHI program allows insureds withmajor diseases, such as malignancies, transplant, or autoimmune diseases, to apply for a Catastrophic Illness Certificate. Application for such a certificate requires cytological or pathological evidence supporting
diagnosis. The NHI database of catastrophic illnesses integrates multipleNHI databases to provide comprehensive information on utilization and enrollment for patients with severe diseases who obtained copayment exemption from the NHI program.
For research purposes, the NHRI released the NHIRD in electronic format, encrypted patient personal information for
privacy protection, and provided researchers with anonymous identification numbers associated with relevant claims information including sex, date of birth, registry of medical services,
and medication prescriptions. Patient consent is not required for accessing the NHIRD or the Longitudinal Health Insurance Database, so the study was approved to exempt from patient consent by the Institutional Review Board of China Medical University in Central Taiwan (CMU-REC-101-012).
Participants
Figure 1 shows the selection process of the participants in the 2 study cohorts. From the Catastrophic Illness and NHI databases, we identified patients with newly diagnosed with DM or PM (ICD-9-CM 710.3 and 710.4) between January 1, 1998 and December 31, 2011 (exposure cohort). The date of diagnosis for DM or PM was defined as the index date. Patients with HZ before the index date were excluded.
The control cohort used same exclusion criteria and was
frequency matched with the exposure cohort at an approximately 4:1 ratio for sex, age (every 5 years) and comorbidities of
diabetes, renal disease, obesity, malignancy, rheumatoid arthritis, immunodeficiency virus infection, autoimmune disease not
elsewhere classified, mixed connective tissue disease, or vasculitis, and index year.
Finally, we recruited 2023 patients with DM or PM and 7409 unexposed participants in this study.
Outcome Measurement and Comorbidities
Every participant in both cohorts was followed up until a diagnosis of HZ (ICD-9-CM 053) was made, or until the patients were censored because of loss to follow-up, death, withdrawal from the database, or December 31, 2011. For other baseline comorbid conditions, we searched for diagnoses of diabetes CM 250), renal disease CM 580–589), obesity CM 278.0), malignancy (ICD-9-CM 140–208, 230–235), rheumatoid arthritis (ICD-9-(ICD-9-CM 714.0–714.3), immunodeficiency virus infection (ICD-9-CM 041–044), autoimmune disease not elsewhere classified (ICD-9-CM 279.4), mixed connective tissue disease (ICD-(ICD-9-CM 710.0–710.2, 710.8, 710.9), or vasculitis (ICD-9-CM 446,
447.6). Medication included immunosuppressants and corticosteroids, the primary medications use in treating DM and PM.
Statistical Analysis
The distribution of categorical sociodemographic characteristics and comorbidities were compared between the 2
cohorts, and the differences were examined using the Chisquare test. Follow-up person-years were used to estimate the
incidence rate of each variable. Cox proportional hazards regression analysis was applied to assess the risk of HZ with DM or PM and the interaction of DM or PM with the comorbidities associated with the relative risk of HZ. The cohortspecific survival curves were plotted after adjusting for age, sex, and comorbidities in the Cox model. A P<0.05 was considered statistically significant. All statistical analyses were performed using the SAS statistical software (Version 9.3 for Windows; SAS Institute, Inc., Cary, NC).
RESULTS
Table 1 summarizes the baseline characteristics of the 2023 patients with DM or PM and the 7409 matched controls. Women accounted for 67.2% of the study population. The mean age of the study participants was 45.9_17.9 years for the control cohort and 45.9_18.0 years for the exposure cohort; 30.2% of the participants were aged between 35 and 49 years. The baseline comorbidities of the cohorts are shown in
Table 1. The exposure cohort was more likely to have rheumatoid
arthritis (1.52% versus 2.17%, P¼0.04), autoimmune disease not elsewhere classified (3.40% versus 6.13%,
P¼0.0001), mixed connective tissue disease (32.3% versus 35.2%, P¼0.01), and vasculitis (2.27% versus 3.61%, P¼0.001) than was the control cohort. Medication use was more prevalent in the exposure cohort than in the control cohort at baseline (P<0.0001).
Table 2 shows theHZincidence densities in both cohorts and the HRs for the HZ. During the follow-up period, 324 patients in the control cohort (incidence rate¼7.01 per 1000 person-years) and 338 patients in the exposure cohort (incidence rate¼35.8 per 1000 person-years) developed an HZ infection. The exposure cohort had a significantly higher risk of HZ than did the control
cohort (crude hazard ratio [HR]¼4.96, 95% confidence interval [CI]¼4.26–5.78; adjusted HR¼3.90, 95% CI¼3.18–4.77) (Table 2). The results of the cumulative incidence curve for the HZ (Figure 2) indicate a significantly higher risk of HZ in patients with DM or PM.
Patients with DM or PM was significantly associated with increased risk of HZ than the subjects without DM or PM for both women and men (HR¼4.32, 95% CI¼3.41–5.47; HR¼2.88, 95% CI¼1.93–4.28). The age-specific exposure cohort to control cohort relative risk for HZ was significant for all age group. Stratified by comorbidity, the relative risk of HZ was higher in the exposure cohort than in the control cohort for both without or with comorbidity (HR¼4.74, 95% CI¼3.33– 6.74; HR¼3.55, 95% CI¼2.77–4.54).
We analyzed the association between comorbidities and DM or PM with the risk of HZ (Table 3). Relative to the control cohort without comorbidity, the DM or PM patients with comorbidity were at a higher risk of HZ (HR¼6.03, 95% CI¼4.60–7.91; interaction P<0.001). Table 4 shows the effects of medications on the risk of HZ. Compared with patients without DM or PM and who received no corticosteroid and immunosuppressant treatment, the patients with DM or PM receiving corticosteroid and immunosuppressant therapy had a significantly higher risk of HZ (HR¼7.43, 95% CI¼6.02–
9.18), followed by patients with DM or PM receiving inhaled corticosteroid or immunosuppressant treatment (HR¼5.89, 95% CI¼4.68–7.43), patients with DM or PM who received no corticosteroid or immunosuppressant treatment (HR¼4.98, 95% CI¼3.62–6.84), and patients with noDMor PM receiving corticosteroid or immunosuppressant treatment (HR¼1.68, 95% CI¼1.35–2.10). Compared with patients with DM or
PM and who received no corticosteroid and immunosuppressant treatment, the patients with DM or PM receiving corticosteroid and immunosuppressant therapy had a significantly higher risk of HZ (HR¼1.45, 95% CI¼1.05–1.99).
DISCUSSION
This study presents conclusive evidence supporting the association of DM or PM with a greater risk of subsequent HZ.
In the study population, a higher incidence of HZ was observed in patients with DM or PM who were predominantly female, aged older than 50 years, and had one or more comorbidities, including diabetes,10 renal disease,11 obesity,12 cancer,13 other
autoimmune diseases,14 mixed connective tissue disease,15
vasculitis,16 or those who underwent medical treatment with
immunosuppressants or corticosteroids.17
The Taiwan NHI database provides complete and valid
information on the demographic characteristics of patients with DMor PM because of the distinctive diagnosis ofDMor PM and HZ. The incidence ofHZwas significantly higher in patients with DM or PM than in those without DM or PM. The findings of our
study corresponded with previous epidemiological data:18,19 more women were diagnosed
with DM or PM (67.2%) and the
adjusted HR was greatest in a female population with DMor PM. Moreover, VZV-specific cellular immunity declines with an
increasing age.20 Our findings were consistent with a retrospective
review conducted by Chen et al,21 who identified lateDMor
PM onset (at 52.3 years of age) as a predictor of infectious complications. Patients aged 50 years or older had a high HZ incidence in our study. The mean age of patients withDMor PM in our study was 45.9_18.0 years, reflecting a peak incidence of DM or PM at 40 to 50 years.22 The adjusted HR appeared to be
higher in the older age group and the exposure cohort was more likely to have coexisting comorbidities and a higherHZincidence than was the control cohort.
Current treatment of DM or PM is based on immunosuppression. The core therapeutic approach is daily administration of
a high dose of oral corticosteroid therapy along with immunosuppressive agents.23 In our study, the coexistence of comorbidities
and concurrent treatment with immunosuppressants and corticosteroids is independently associated with an increased incidence of subsequent HZ. As documented by Schmader and Oxman,20 an immunosuppressed patient has a 20- to 100-fold
greater risk of developing HZ than does an immunocompetent individual of the same age.
HZ development is determined by factors influencing the host–virus relationship, including old age, cellular immune
dysfunction, the female sex, physical trauma to the affected dermatome, and interleukin 10 (IL-10) gene polymorphisms.24
Hassan et al25 proposed that VZV increased tropismfor activated
memory T-cells, which are infiltrated in the muscles affected by DM or PM, leading to a latent HZ viral infection. Cho et al reported a cytokine imbalance, specifically higher levels of IL-10, in patients with DM or PM attributable to the presence of autoantibodies.26 IL-10 is an immunomodulatory cytokine that
suppresses cell-mediated immunity by inhibiting the functions of antigen-presenting cells and cytokine production through activated T-cells.26 In a study conducted by Cho et al,27 polymorphisms
of the IL-10 promoter gene were reported to be possibly
associated with increased susceptibility toHZ. Other nonimmune processes of DM or PM include the endoplasmic reticulum (ER) stress response.28,29 The ER contributes to the synthesis, folding,
and assembly of viral proteins, which can be provoked by a
variety of conditions including ischemia in the tissue microenvironment of DM or PM.30 The ER overload response follows a
viral infection, thereby fuelling the inflammatory cascade observed in patients with DM or PM.31 The DM or PM tissue
microenvironment including endomysial hypoperfusion, proinflammatory cascade, and the production of a variety of antibodies
against immune or stressor molecules contributes to a complex yet efficiently regulated infectious process that plays a crucial role in HZ pathogenesis.7
Several limitations exist in this study. First, because of the retrospective nature of this study, the database analysis did not include addressing information such as disease severity and history. However, the NHI database of Taiwan provides complete and valid information regarding the demographic characteristics of patients in both the case and control groups. Despite
the shortfall which database analysis did not address information such as HZ severity and history, due to the distinctive
diagnoses of DM or PM and HZ, recorded data should be
comparable and representative of HZ severity requiring medical attention in both case and control groups. Nevertheless, a study reported that HZ prevalence was independent of disease status and clinical features.5 Second, unavailable patient information
such as smoking or personal habits was replaced with coexisting comorbidities. Our retrospective cohort study revealed a
positive association between DM or PM and subsequent HZ viral infection. A prospective cohort study including individual information obtained from chart review may enable to determine a causal relationship between the 2 distinct disease
processes. Third, the actual HZ incidence possibly was underestimated because of an underreporting bias. Fourth, the study
was limited by the lack of knowledge on whether HZ is a pathogenic or an opportunistic infection in patients with DM or PM. Patients with DM or PM may be susceptible to viral
infections other than HZ;7 therefore, further studies may
identify potentially threatening pathogens.
HZ treatment can be effectively provided in the form of antiviral agents and reimbursed by the NHI in Australia. Community-based data have supported the success of zoster vaccination.32 According to the statistical results of this study,
prophylaxis in the form of a live attenuated vaccine should be considered for immunocompetent older adults who have higher incident rate of HZ infection and come in contact with female patients with DM or PM and have at least 1 condition of
comorbidities to reduce the risk of transmitting HZ and wildtype VZV to susceptible immunosuppressed contacts.20 Live
attenuated VZV vaccines are effective in healthy individuals; however, vaccine delivery is debatable in immunocompromised patients because of a potential risk of viremia.33 Furthermore,
the involvement of cytokine imbalance in DM or PM and HZ should be considered in future therapy development.
Controlling DM or PM disease activity with immunosuppressive therapy and preventing opportunistic infections are
challenging. Currently, identifying potential individuals with high incidence is essential for reducing the incidence of HZ among patients with DM or PM. These patients should be advised to seek prompt medical attention if clinical symptoms
associated with HZ develop.34 Physicians should anticipate diverse clinical manifestations
and be prepared to start early
treatment if necessary after understanding the atypical presentations of HZ in immunocompromised hosts. The study finding
is crucial to provide insights for identifying high-risk patients likely to suffer from HZ infection, and to open a new avenue of research on the intrinsic defects in DM or PM patients which precipitate HZ infection. Therefore, the health policy to reduce incidence of HZ infection and consequently to improve the patients’ quality of life is worth promoting.
