We identified 44,115 eligible subjects classified as having at least one of the autoimmune diseases of interest who were at least 50 years old at the start of observation. At baseline, their average age was 58.1 (standard deviation (SD), 6.7) years and 27,443 (62.2%) were female; 19,326 (43.8%) had RA, 867 (2.0%) had psoriatic arthritis (PsA), 10,712 (24.3%) had psoriasis (PsO), 633 (1.4%) had AS, 8,639 (19.6%) had IBD, and 3,938 (8.9%) had two or more of these diseases. During follow-up, 551 (1.2%) subjects received zoster vaccine. The earliest claim for vaccination occurred in August 2006; subsequently, vaccine use increased continuously over time (Figure ). The distribution of patients' characteristics at baseline by vaccination status is presented in Table .
Figure 1 Incidence proportion for each 6-month period of zoster vaccination by calendar year from June 1, 2006, to December 31, 2009. The proportion is expressed as the proportion of unvaccinated individuals under observation and vaccinated in each 6-month period. (more ...)
Baseline patient characteristics by vaccination status
Among the 551 subjects who received zoster vaccine, complete claims history for at least 30 days after vaccination was available for 514 subjects. At the time of vaccination, 32 (6.2%) subjects were using anti-TNF therapy, 34 (6.6%) were using methotrexate, and 33 (6.4%) were using oral glucocorticoids (Table ). In the 30 days after vaccination, 40 (7.8%) subjects used anti-TNF therapy, 45 (8.8%) used methotrexate, and 48 (9.3%) used oral glucocorticoids. A total of 47 patients used biologics at some time within 30 days before and after vaccination. None of the 47 subjects developed HZ in the 30 days after vaccination. Their average age was 60 years (standard deviation, 5); 64% were women; and 70% were not exposed to oral glucocorticoids, 28% to a daily average dose of no more than 20 mg, and 2% to a daily average dose of 20 mg or more.
Biologic and nonbiologic DMARDs use 30 days before and after zoster vaccination (n = 514)
Patients using anti-TNF therapies were less likely to receive zoster vaccine than those who were not using anti-TNF agents (hazard ratio (HR), 0.47; 95% CI, 0.33 to 0.67) (Table ). Patients who were using other biologics (HR, 0.52; 95% CI, 0.19 to 1.40) or high-dose oral glucocorticoids (HR, 0.46; 95% CI, 0.15 to 1.45) were only half as likely as were non-users to receive zoster vaccine, but these associations did not reach statistical significance. Patients aged 60 to 64 years were most likely to receive zoster vaccine; those with fewer co-morbidities and those without hospitalization within the past 6 months were more likely to receive zoster vaccine. Patients with recent, but not remote, history of HZ were more likely to be vaccinated. Finally, patients who had more physician visits in outpatient settings and those who had received influenza or pneumococcal vaccine in the past year were more likely to be vaccinated. Results of the analysis that was restricted to individuals aged 60 years or older were similar to the main results (Table ). A majority (> 80%) of the patients received their vaccine from family practice or internal medicine physicians, with less than 5% from dermatologists, gastroenterologists, or rheumatologists, both in the overall cohort and in patients who were exposed to biologics at the time of vaccination.
Fixed and time-varyinga patient characteristics associated with vaccination
During 88,354 observed person-years, 761 cases of HZ occurred (incidence rate (IR), 8.6 per 1,000 person-years). The IR increased with age from 8.14 per 1,000 person-years among those aged 50 to 54 years to 15.30 per 1,000 person-years among those 90 years or older. Five cases of HZ occurred during vaccinated person-time (crude IR, 9.97 per 1,000 person-years), and 756 HR occurred during unvaccinated person-time (crude IR, 8.61 per 1,000 person-years). The age- and sex-standardized IR (expected IR) for the vaccinated was 10.06 per 1,000 person-years, resulting in a standardized IR ratio (vaccinated to unvaccinated) of 0.99 (95% CI, 0.29 to 3.43).
The five cases of HZ in the vaccinated patients occurred 7, 131, 201, 214, and 667 days after vaccination. Within 90 days before vaccination, the only filled prescription for an immunosuppressive agent for the patient who developed HZ on day 7 after vaccination was for a 15-day supply of prednisone approximately 2 months before receipt of zoster vaccination. None of the five patients was hospitalized.
Results from sensitivity analyses restricted to individuals under observation continuously from the time zoster vaccine was approved in 2006 had similar results (data not shown). When requiring only an HZ diagnosis code, the result was consistent, in that the crude incidence rates were the same between vaccinated (14.4 per 1,000 person-years) and unvaccinated (13.1 per 1,000 person-years) patients. With Poisson regression, the incidence rate ratio was 0.99 (95% CI, 0.35 to 2.82).