Varicella vaccine safety and immunogenicity in patients with juvenile rheumatic diseases receiving methotrexate and corticosteroids

Varicella has been a common rite of passage among children and is typically considered to be a mild disease. However, even among healthy individuals, varicella can cause significant morbidity related to secondary bacterial infections of the excoriated lesions, complications of bacterial sepsis, and virally mediated complications, including pneumonia, encephalitis, and hemorrhagic conditions, that may result in death (1, 2). In 1995, this led to the addition of varicella vaccine (Varivax; Merck) into the routine immunization schedule for children in the US (3). Varicella vaccines are now available throughout the world in both single-antigen and combination (measles, mumps, rubella, and varicella [MMRV]) vaccine formulations, and a number of other countries, including Taiwan, Australia, Canada, Germany, and Uruguay, include varicella vaccine in their routine childhood vaccination program. Globally distributed varicella vaccines, all using the Oka vaccine strain, include Varivax, which is the only varicella vaccine licensed in the US, Okavax (Biken; Sanofi Pasteur), and Varilrix (GlaxoSmithKline).

Varicella-zoster virus (VZV) is a highly cell-associated virus. Because of this, cellular immunity is critical for host defense and recovery from wild-type infection and is also likely to be important in the immune response to vaccination. Therefore, people with altered immune systems, particularly conditions affecting lymphocytes, are at increased risk of severe varicella disease. The prototypical group is children with acute lymphocytic leukemia receiving chemotherapeutic agents. The high morbidity and substantial mortality of varicella among immunocompromised children was the impetus to the development of varicella vaccine in Japan in the 1970s (4). Japanese researchers initially demonstrated the safety and efficacy of the live attenuated varicella vaccine in children with underlying diseases, including nephrotic syndrome and asthma, and then extended their studies to children with leukemia (5). In a large multisite study in the US and Canada, varicella vaccine was administered to children with leukemia in remission. Among the children, 511 had maintenance chemotherapy suspended 2 weeks prior to vaccination and 64 children had completed chemotherapy. A concern raised about the study was that nearly 50% of children in the maintenance chemotherapy group developed a rash postvaccination, some severe enough to require intravenous acyclovir. However, the advantages of avoiding severe, and possibly fatal, wild-type disease were thought to be worth the risk (6). Initial experience showed that approximately 15% of the children with leukemia did not develop antibodies (as measured by the fluorescent antibody to membrane antigen [FAMA] test) after a single dose of vaccine. Therefore, subsequent studies administered 2 doses of varicella vaccine with children receiving the second dose 3 months after the first, resulting in 95% developing antibodies after the second dose. These children were followed closely for varicella exposures and varicella disease over the next 10 years. Of the 123 children who reported a household varicella exposure after receiving 2 doses of varicella vaccine, approximately 85% were protected against developing varicella disease. Based on these studies, although varicella vaccine is not licensed in the US for use in children with leukemia, the vaccine was originally available for use in children with acute lymphocytic leukemia under a special protocol. Current recommendations in the US permit the use of live attenuated vaccines, including varicella vaccine (Varivax), in patients with leukemia, lymphoma, or other malignancies whose disease is in remission, providing their chemotherapy has been terminated for at least 3 months (3). Okavax and Varilrix vaccines are both licensed for use in children with acute lymphocytic leukemia, provided that they are in remission and meet other clinical criteria for vaccination.

Subsequent studies conducted in the US, Japan, and Europe have demonstrated the safety, immunogenicity (ranging from 60–100% using different tests), and in a limited number of studies, effectiveness of varicella vaccine administered to other groups of patients at high risk of developing unusually severe and progressive disease following infection with wild-type varicella. These groups include children with human immunodeficiency virus (HIV) with varying levels of immunocompromise, solid tumors, renal failure, and other conditions (5). In the US, these studies have resulted in the consideration of using the varicella vaccine in children with HIV and CD4+ T lymphocyte counts ≥15% of normal for their age (3).

Rates of varicella-like rash following the first dose of vaccine varied from 5% (in children with HIV and CD4+ T lymphocyte counts ≥25% of normal) to 33% (in children with solid tumors), and likely was related to the level of immune function in the vaccinated patients. Experience of a rash postvaccination led to the recommendation that systemic steroids be withheld, if possible, for at least 2 weeks after vaccination in children with leukemia, and this guidance was later applied to other persons receiving long-term systemic steroid therapy (3, 7).

Another potential benefit of administering varicella vaccine to immunocompromised children is the prevention of herpes zoster, which may be recurrent and severe. In a landmark study, Hardy et al demonstrated a 3-fold lower rate of herpes zoster among immunocompromised children with leukemia matched according to a chemotherapeutic protocol who had received 2 doses of varicella vaccine compared with those who had natural varicella (8). Similar results have been described in children who were vaccinated before undergoing renal transplantation (approximately half the rate of herpes zoster compared with children with varicella before transplantation) (9). The lower risk of herpes zoster is postulated to be due to a lower risk of reactivation in the attenuated vaccine strain virus as well as reduced skin infections following vaccination compared with natural varicella (8).

An article in this issue of Arthritis Care & Research provides the first published data on another immunocompromised group that was administered varicella vaccine: patients with various juvenile rheumatic diseases receiving various medications for treatment (methotrexate, steroids, and other antirheumatic drugs). Although unusual, there are sporadic case reports of patients with rheumatic diseases receiving immunocompromising agents who have had serious and even fatal varicella infections (10–12). In the current issue, Pileggi et al administered a single subcutaneous dose of varicella vaccine (Okavax) to 25 patients between ages 2 and 19 years with juvenile rheumatic diseases and 19 age-matched, healthy controls (13). No patient or control had a preimmunization history of varicella, although the prevaccination serology was indeterminate for 5 patients who were not included in the immunogenicity analysis. At the time of vaccination, all of the patients were receiving methotrexate (mean dosage of 16.4 mg/m²/week) and 13 patients were receiving prednisone ranging from 0.1–0.7 mg/kg/day (mean total dosage of 4.2 mg/day, range 3–20). Five patients were also receiving other disease-modifying antirheumatic drugs (DMARDs), including cyclosporine (n = 3), penicillamine (n = 1), and leflunomide (n = 1). These medication regimens are not likely to result in severe levels of immunocompromise and with the exception of 1 child, all of the patients were receiving steroid doses that did not contraindicate the receipt of single-antigen varicella vaccines. The results of this study with Okavax are likely generalizable to other single-antigen varicella vaccine products. It should be noted, however, that MMRV vaccines have not been studied in immunocompromised populations and these vaccines should not be substituted for varicella vaccine when vaccinating these and other immunocompromised children (3).

The proportion (~20%) of patients in the current study who developed a mild, varicella-like rash within the first 2 weeks after immunization was in the middle of the range (5–50%) described in studies of immunocompromised children (5). Another important finding is that none of the patients reported worsening of their underlying disease in the 3 months after vaccination as compared with the 3 months before vaccination.

Of the 20 evaluable patients and 18 controls in the current study, only 10 (50.0%) of 20 patients and 13 (72.2%) of 18 controls had detectable antibodies 4–6 weeks after vaccination. Although the authors concluded that the vaccine response rates did not differ between patients with juvenile rheumatic diseases and controls, the lack of statistical significance may be a function of small sample size and low power of the study to detect a true difference. The scatter plots indicate a higher mean and a higher upper range of antibody response among control subjects. The type of test used to measure the immune response should also be taken into account. The FAMA test for detecting VZV IgG antibodies is typically considered the gold standard for gauging immunity to varicella, with a result of >1:4 considered to represent protection against infection (5). However, the test requires fresh cells infected with the VZV and cannot be automated (14). As part of the clinical testing of the varicella vaccine in the US, Merck developed an in-house enzyme-linked immunosorbent assay (ELISA) that used VZV glycoproteins, and studies suggest that a glycoprotein ELISA titer of ≥5 units/ml is an approximation of protection (3). However, there are no direct comparisons between FAMA and glycoprotein ELISA results (14). The intricacies of the FAMA test performance and the fact that the glycoprotein ELISA test is not available commercially has resulted in the response to varicella vaccine often being measured, as in the current study, using a commercially available ELISA test. A significant limitation of such tests is their low sensitivity compared with the FAMA test and the lack of a serologic correlate of immunity. The 50% reported response rate in the children with juvenile rheumatic diseases is in contrast to studies reporting 85–90% seroresponse (glycoprotein ELISA titer of ≥5 units/ml) 6 weeks after immunization in healthy children (3). Although suggestive of a lower antibody response, since commercial ELISA tests do not detect low levels of vaccine-induced antibodies, the results from this study cannot be directly compared with glycoprotein ELISA and FAMA results. A more comparable study in terms of laboratory methods is a study in Spain of 34 children receiving either renal dialysis or posttransplantation, where 85% had antibodies detected by ELISA after one dose and 10% developed a rash (15). Compared with this study, the response in the children with juvenile rheumatic diseases may be lower. Some studies in healthy children and adults have demonstrated clinical protection following VZV exposure despite the absence of antibodies, suggesting the important role that cell-mediated immunity may play in VZV immunity. However, 1 of the 2 patients in the current study who was receiving anti–tumor necrosis factor therapy at the time of exposure did not seroconvert following vaccination and developed severe varicella with pneumonia and probable macrophage activation syndrome, which is suggestive of primary vaccine failure. In the current study, the 25% attack rate in 8 vaccinated patients with juvenile rheumatic diseases following household or classroom exposure suggests vaccine effectiveness (protection against disease when exposed) of approximately 70% over the 40 weeks of active surveillance. Although this point estimate is somewhat lower than that described in an epidemiologic study of healthy children that demonstrated a vaccine effectiveness of approximately 86% for the first 2 years after vaccination and remaining at 85% 6 years after vaccination, it is within the range of vaccine effectiveness estimates (44–100%, median 85%) described in the >20 studies conducted in healthy children (3, 16). None of the 25 patients developed herpes zoster during their median followup period of 32 months.

To improve rates of individual immune response and protection and population varicella prevention and control, in 2006, the Advisory Committee on Immunization Practices and the American Academy of Pediatrics in the US recommended that all healthy children receive 2 doses of varicella vaccine, with the first dose at age 12–15 months and the second dose at age 4–6 years or at least 3 months after the first dose (3). Data are not yet available on the impact of this 2-dose vaccine policy. Based on the results of the current study, if a decision was made to immunize children with rheumatic diseases against varicella, a 2-dose regimen would be preferable. Of note, in countries using varicella vaccine (1 - or 2-dose policy) in their universal childhood immunization program, most children will receive varicella vaccine before developing juvenile rheumatic diseases.

Although the current study provides important data, there are limitations that need to be considered. First, only 25 patients were immunized and only 20 were eligible for immunogenicity evaluation. Second, the age of study subjects ranged from 2–19 years, and response to the vaccine may vary by age. Third, the medications the patients were receiving varied greatly. While all of the subjects were receiving a rather consistent dose of methotrexate per kilogram of body weight, 13 of 25 patients were receiving varying doses of prednisone (although with the exception of 1 child receiving 20 mg/day, these doses are not a contraindication to vaccination) and 5 subjects were also receiving other DMARDs, all of which could have affected the response to the vaccine. Finally, patients primarily had juvenile idiopathic arthritis; some carried the diagnosis of dermatomyositis, scleroderma, or vasculitis, again potentially altering the response to vaccination. Despite these limitations, Pileggi et al were able to demonstrate that in a highly selected group of patients, administration of a single dose of live attenuated varicella vaccine was safe, relatively immunogenic, and approximately 70% effective in preventing varicella disease following household exposure (13). They should be commended for their work. Based on the safety and effectiveness found in this small study, expanding the investigation to a broader group of children with rheumatic diseases receiving medications that may be more immunocompromising may be warranted. This new scientific evidence will be useful to countries considering or reevaluating varicella vaccine policy.