Background/Purpose: Ustekinumab (UST), an anti IL12/IL23 monoclonal antibody and apremilast, an inhibitor of phosphodiesterase 4 are two relatively new agents used for the treatment of psoriasis (PSO) and psoriatic arthritis (PSA). Potential differences in their risk profiles with respect to serious infection risk compared to tumor necrosis factor inhibitors (TNFi) have been suggested but minimally evaluated in real-world settings.

Methods: We conducted a retrospective, new user design cohort study of PSO or PSA patients initiating apremilast or one of the target biologic DMARDS using 2010-2014 Market scan and 2006-2014 Medicare data. To be included in the cohort, patients have to have at least 12 consecutive months of observable (Enrolled in Medical and pharmacy) before the initiation date (index date). Patients were excluded if they have any diagnosis code for inflammatory disease other than PSO or PSA, malignancy or HIV or organ transplant. Follow up started at index date and ended at earliest of: discontinue of the index drug (90 days after days of supply) or switch to another drug, the outcome, lose of full insurance coverage, death, and end of study. Patients contributed to only one episode of one specific drug but could contribute to multiple drugs. Hospitalized infection was identified using ICD-9 diagnosis code from hospital discharge (Primary or secondary). Sensitivity analysis was conducted by limiting outcome to primary discharge diagnosis code.  Incidence rates (IR) were calculated by dividing number of outcome by person year exposed. Crude and multi-variable adjusted hazard ratios (HR) were calculated with COX proportional hazard regression with robust estimate. 

Results: The final cohort included 50,125 initiations of therapies of interest (19,922 from Medicare, 30,203 from Market scan) represented by 40,501 unique PsO/PsA patients. Patient’s demographic characteristics were: age 50.4 (14.9) in years, women 54.4%. A total of 2,191 hospital infections occurred in 41,650 person years, resulting in an overall IR of 5.26 (5.04, 5.49) per 100 person years.  IRs ranged from 4.36 (3.74, 5.08) for ustekinumab to 8.78 (8.00, 9.64) for infliximab (Table). Compared to adalimumab, the multivariable adjusted HR were 0.77 (0.43, 1.38) for apremilast, and 0.99 (0.83, 1.17) for ustekinumab (Table 1). Infliximab was associated with a significantly increased risk for hospital infection compared to adalimumab.

Conclusion : In patients with psoriasis and psoriatic arthritis, the comparative rates of hospitalized infection were generally comparable between TNFi and ustekinmab, although rates were highest for infliximab. Apremilast was associated with a numerically lower risk for serious infection compared to adalimumab. Table: Incidence rate and hazard ratio for hospital infection associated with medications for psoriasis and psoriatic arthritis

<>IR: Incidence rate; CI: Confidence interval; HR: Hazard ratio <>*Adjusted for age, gender, diabetes hyperlipidemia, hypertension, obesity, chronic kidney disease, chronic obstructive pulmonary disease, pneumonia, heart failure, sepsis, peptic ulcer disease, fracture,  skin ulcer, glucocorticoid use, prior biologic use, smoking status, prostate specific antigen screen, mammography, pap smear, hospitalization in baseline, number physician visit.