Incidence, Risk Factors, and Severity of Herpesvirus Infections in a Cohort of 121 Patients With Primary Dermatomyositis and Dermatomyositis Associated With a Malignant Neoplasm FREE

Dermatomyositis (DM) is a rare systemic disorder with inflammatory myopathy and typical cutaneous involvement. In adults, DM is usually primary or associated with a malignant neoplasm. It is considered to be associated with a 10% to 30% mortality rate.^1- 2 Death is primarily related to cardiac and lung involvement or associated with malignant neoplasms,^1- 2 but some cases of fatal infections in patients with DM have also been reported.^3- 4 In a study of 47 patients with DM, Viguier et al⁵ reported the incidence of opportunistic infections to be 21%. In a larger series of 156 patients (66 with DM and 90 with polymyositis), Marie et al³ reported a lower but significant rate (11.5%) of such infections. In these studies, the main pathogen microorganisms responsible for opportunistic infections were Pneumocystis jiroveci, Candida albicans, Mycobacterium xenopi, Mycobacterium marinum, or Mycobacterium tuberculosis. The incidence of herpesvirus infections has been rarely assessed. In a series of 12 patients with juvenile DM treated with cyclophosphamide, 3 patients (25%) developed herpes zoster (HZ) infection⁶ but, to our knowledge, no data are available on larger series. The aim of this study was to assess the incidence, risk factors, and severity of herpesvirus infections in a retrospective cohort of patients with DM seen in 2 departments during a 13-year period.

We conducted a retrospective study in all adult patients with DM who were referred to 1 department of dermatology and 1 department of internal medicine between January 1,1995, and December 31, 2007. Patients with an ascertained diagnosis of DM according to the criteria of Bohan and Peter^7- 8 were included in this retrospective cohort. However, only incident cases of DM were retained, and patients who were referred for DM recurrence during the study period were excluded. Patients with DM associated with other connective tissue disorders were also excluded. The diagnosis of DM associated with malignant neoplasm was retained if DM occurred in a context of recently diagnosed (ie, <1 year) neoplasia or if a neoplasia was diagnosed during the 5 years following the diagnosis of DM.

All data reported in this retrospective inception cohort study were based on hospital records. Demographic data (age and sex), clinical data (DM associated or not associated with malignant neoplasms, treatment regimen), and biological data (leukocytes and lymphocytes counts, C-reactive protein, gamma globulins, aspartate aminostransferase [AST], alanine aminotransferase, creatine phosphokinase [CPK], lactate dehydrogenase, antinuclear antibody, and complement levels) obtained at the first visit in our departments were recorded. Biological parameters were obtained before initiation of any therapy. For cytomegalovirus (CMV) or herpes simplex virus (HSV) infections, the clinical diagnoses were confirmed by positive findings from polymerase chain reaction or viral culture. For CMV infections, only a viral load of 4 log or greater was taken into account. Clinical suspicion of Kaposi sarcoma was histologically confirmed. The diagnosis of HZ infection was performed clinically (ie, typical presentation of vesicles and neuralgia along dermatomal distribution), and no biological test was routinely required for confirmation.

To assess risk factors for herpesvirus infections, we chose to consider only incident infections diagnosed during the first year of DM. For this retrospective cohort study, the approval of our institutional review board was not required.

Proportion was used as the descriptive statistic for categorical variables. Continuous variables were described by mean (SD). Groups were compared by the χ² test for categorical variables. The Kaplan-Meier method was used to assess the cumulative incidence rates of herpesvirus infections during the first 5 years of DM. Factors associated with the risk of developing herpesvirus infection during the first year of DM were identified by using Cox proportional hazards models. Each parameter was modeled as a qualitative variable. Variables with a P < .20 in univariate analysis were retained in the model for multivariate analysis. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. All statistical analyses were tested at the significance threshold α = .05 and were 2-tailed. Analyses were performed using SAS statistical software (version 8.2; SAS Institute, Cary, North Carolina).

During the study period, 134 patients with DM were seen in our departments. Ten with DM recurrence, and 3 other patients with associated connective tissue disease were excluded. Thus, 121 patients met the inclusion criteria and were evaluated.

The baseline characteristics of the 121 included patients are reported in Table 1. Ninety-two patients (76%) presented with primary DM (14 with amyopathic) and 29 (24%) with DM associated with a malignant neoplasm (Figure 1). Their mean duration of follow-up was 42 (33) months.

During the first year of evolution, 81 patients (66%) received systemic corticosteroids, alone (n = 55) or in combination with methotrexate (n = 6) or with intravenous immunoglobulins (n = 20). Fourteen patients (11%) received only intravenous immunoglobulins, and 26 patients (23%) were treated only with chloroquine diphosphate or received no systemic therapy. For the 81 patients treated with corticosteroids, the mean (SD) baseline dosage was 63 (22) mg/d, and the mean dosages at month 6 and month 12 were 35 (22) and 16 (13) mg/d, respectively. Patients with amyopathic DM were less likely to be treated with systemic corticosteroids than other patients with DM (6 of 14 vs 75 of 101; P = .04).

During follow-up, 20 patients developed a total of 22 herpesvirus infections. The incidence rate was 49 episodes per 1000 patient-years. A 68-year-old woman was diagnosed as having extensive HSV 2 vulvovaginitis. One patient developed a HSV 1 pneumonitis with a fatal outcome. Two patients developed a CMV infection soon after the onset of corticosteroid therapy (retinitis in one, pneumonitis with a fatal outcome in the other). A visceral or cutaneous Kaposi sarcoma was diagnosed in 2 other patients who were not infected with the human immunodeficiency virus (HIV) and who were seronegative. Last, 14 patients were diagnosed as having HZ infection (multimetameric HZ in 3 patients, recurrent HZ in 2, and ophthalmic HZ in 2; no patients had disseminated disease). The incidence rate for HZ was 33 episodes per 1000 patient-years. The median steroid dose at the time of herpesvirus infection was 20 mg/d (range, 0-80 mg/d). Only 1 of these patients received methotrexate in combination with corticosteroids when herpesvirus infection occurred. Seven of these 14 patients with HZ were hospitalized for antiviral therapy initiation.

The cumulative incidence rate of herpesvirus infection in our cohort was 13% (95% CI, 7%-19%) at year 1, 20% (95% CI, 11%-29%) at year 3, and 24% (95% CI, 12%-36%) at year 5 (Figure 2). The cumulative incidence rate of HZ infection was 9% (95% CI, 4%-13%) at year 1, 14% (95% CI, 6%-22%) at year 3, and 18% (95% CI, 9%-29%) at year 5.

Univariate analysis showed that the risk of herpesvirus infection during the first year of DM was increased in patients with high AST and CPK levels and tended to be higher in patients with low lymphocyte or leukocyte counts and in patients who received systemic corticosteroids (Table 2). In multivariate analysis (Table 3), 3 factors were significantly associated with the risk of herpesvirus infection: therapy with systemic corticosteroids (HR, 3.71 [95% CI, 1.02-13.41]; P = .04), a lymphocyte count lower than 6000 μL (to convert to a proportion of 1.0, multiply by 0.01) (HR, 3.55 [95% CI, 1.00-12.65]; P = .05), and a CPK level higher than 300 U/L (HR, 4.81 [95% CI, 1.28-18.06]; P = .02). Moreover, DMs associated with malignant neoplasms tended to be negatively associated with the risk of herpesvirus infection (HR, 0.16 [95% CI, 0.02-1.29]; P = .08).

The results of the present study emphasize the high incidence of herpesvirus infections in patients with DM. Indeed, the incidence of herpesvirus infection was 13% at year 1 and 24% at year 5. Herpes zoster represented 70% of these herpesvirus infections. Patients who developed herpesvirus infection during the first year of DM were more likely to have been treated with systemic corticosteroids and to have a low lymphocyte count and a high CPK level. Moreover, DMs associated with malignant neoplasms tended to be negatively associated with the risk of herpesvirus infection.

The frequency of opportunistic infections in patients with DM has rarely been reported. Focusing on opportunistic infections other than herpesviridae (pneumocystosis, candidiasis, and mycobacterial infections), a prevalence rate of 15% to 21% has been reported in small series.^3,5 In a pooled analysis of patients with DM and polymyositis (PM), opportunistic infections tended to be more frequent in patients with an associated malignant neoplasm and low baseline lymphocyte count.³ In these series, the analysis did not focus on the risk of herpesvirus infections.

Over 90% of adults in the United States have serologic evidence of varicella–zoster virus infection and are at risk for HZ infection.⁹ The annualized incidence of HZ is about 1.5 to 2.5 cases per 1000 persons.^10- 11 The incidence of HZ is higher in immunocompromised patients: rates of HZ per 1000 person-years have been calculated to be 29.4 cases in HIV-seropositive patients,¹² 16 to 32 in patients with systemic lupus erythematosus (SLE),^13- 15 and 9.8 to 14.5 in those with rheumatoid arthritis (RA).^16- 18 Herpes zoster infection in RA was associated with corticosteroids used with an HR of 1.50 (95% CI, 1.20-1.80) to 2.51 (95% CI, 2.05-3.06).^16- 17 Except for a small study¹⁹ of 22 patients with inflammatory myopathy published in 1990, no such data were available for DM. Although the mean age of patients was different, the incidence of HZ in our study is in line with the rates reported in series of patients with SLE^13- 15 and is greater than those reported in patients with RA.^16- 18

However, although to our knowledge, no study on incidence or risk factors for infections with herpes family viruses other than HZ is available, several case reports of non–HZ herpesvirus infections in patients with connective tissue diseases^20- 24 and several cases of symptomatic CMV infection in patients with DM have been reported.^25- 27

Risk factors for HZ in patients with connective tissue diseases are the result of decreased cellular immunity associated with the disease itself, immunosuppressive therapies, disease activity level, and special disease markers. In patients with SLE, immunosuppressive therapies, severe disease with lupus nephritis, and Sm-antibody positivity have been associated with a higher risk of HZ infections.^14- 15 In our study, such links are also noted, with a clinically significant association between HZ and systemic corticosteroid use and with disease activity represented by enzyme levels.

Despite the retrospective nature of our study, we believe that most infectious events were captured in our medical records because our patients with DM were regularly followed up at 2- to 3-month intervals, and most patients would come back to our departments for treatment of any acute illness. However, the incidence of herpesvirus infections in our study population may represent the lower bound of the true incidence owing to underreporting bias. Moreover, because we included all consecutive adult patients with DM who were referred to our departments, our sample is not biased toward patients with more severe involvement or with refractory disease.

We chose to explore risk factors only for early herpesvirus infections (ie, infections within the first year after diagnosis) because some biological parameters were available only at baseline. We thought that because of disease natural history and therapeutic changes, baseline characteristics were less relevant for events beyond the first year. Moreover, treatment regimens were more homogeneous (ie, few patients received immunosuppressive therapy other than corticosteroids) during this short study period. Thus, our findings regarding the link between herpesvirus infections and immunosuppressive therapy may be more relevant.

In healthy individuals, live-attenuated varicella vaccine with enhanced potency has been shown to be effective in reducing the incidence of HZ infections.²⁸ Recently, the American Advisory Committee on Immunization Practices²⁹ recommended that all persons 60 years or older who have no contraindications should receive 1 dose of zoster vaccine. However, zoster vaccine should not be administered to persons receiving immunosuppressive therapy, including high-dose corticosteroids (≥20 mg/d of prednisone or equivalent) lasting 2 or more weeks. The committee considered that short-term corticosteroid therapy (<14 days), a low to moderate dose (<20 mg/d of prednisone or equivalent), long-term alternate-day treatment with low to moderate doses of short-acting systemic corticosteroids, or therapy with low doses of methotrexate (≤0.4 mg/kg/wk) for chronic diseases such as PM were not considered sufficiently immunosuppressive to create vaccine safety concerns and were not contraindications for administration of zoster vaccine.²⁹ To our knowledge, no data are available regarding the risk/benefit and the cost/benefit balances of a prophylaxis with valacyclovir hydrochloride in patients with DM or PM.

Because most patients in our cohort who had herpesvirus infections were hospitalized and received antiviral drugs, the cost of these herpesvirus infections to patients and to the health care system were substantial. To reduce the incidence of herpes-related disease morbidity and mortality in patients with DM, simple measures should be discussed: First, identify the patients with DM at risk of developing symptomatic herpesvirus infection (patients with low baseline leukocytes counts and those who are to be treated with corticosteroids). Second, educate these patients about the risk and the symptoms observed during herpesvirus infection such as HZ, HSV, or CMV and about the necessity of a rapid medical visit whenever pain, localized skin rash, fever, visual disturbances, or unusual pulmonary symptoms occur. Finally, provide recommendations to their general practitioners regarding the necessity of herpesvirus screening in case of atypical or evocative symptoms and the usefulness of an early appropriate treatment (especially in cases of HZ).

Correspondence: Laurence Fardet, MD, PhD, Fédération de Dermatologie, Hopital Saint Louis, 1 Avenue Claude Vellefaux, 75010 Paris, France (laurence.fardet@sls.aphp.fr).

Accepted for Publication: December 31, 2008.

Author Contributions: Dr Fardet had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Fardet, Rybojad, Gain, and Dupuy. Acquisition of data: Fardet, Gain, Cherin, and Dubertret. Analysis and interpretation of data: Fardet, Kettaneh, Bachelez, Lebbe, Morel, and Dupuy. Drafting of the manuscript: Fardet. Critical revision of the manuscript for important intellectual content: Rybojad, Gain, Kettaneh, Cherin, Bachelez, Dubertret, Lebbe, Morel, and Dupuy. Statistical analysis: Fardet, Kettaneh, and Dupuy. Administrative, technical, and material support: Dubertret. Study supervision: Rybojad, Cherin, Dubertret, Lebbe, and Morel.

Financial Disclosure: None reported.