The safety of tofacitinib in psoriatic arthritis (PsA) and rheumatoid arthritis (RA) has been demonstrated in clinical studies of≤ 4 and 9.5years, respectively. Post-marketing surveillance (PMS) data for tofacitinib from spontaneous and voluntary adverse event (AE) reports have been published for RA, but not PsA. To inform the real-world safety profile of tofacitinib in PsA, we evaluated AE reports submitted to the Pfizer safety database (including RA data for context). Endpoints included AEs, serious AEs (SAEs), AEs of special interest (AESIs; serious infections, herpes zoster, cardiovascular events, malignancies, venous thromboembolism), and fatal cases. Exposure was estimated using IQVIA global commercial sales data. Number, frequency, and reporting rates (RRs; number of events/100 patient-years' [PY] exposure) were summarized by indication and formulation (immediate release [IR] 5 or 10mg twice daily], modified release [MR] 11mg once daily, or all tofacitinib). The data-collection period differed by indication (PsA: 14December2017 [US approval, IR/MR]to6November2021; RA: 6November2012 [US approval, IR]to6November2021; MR approval, 24February2016). A total of 73,525 case reports were reviewed (PsA = 5394/RA = 68,131), with 20,706/439,370 PY (PsA/RA) ofexposure. More AEs were reported for IR versus MR (IR/MR: PsA = 8349/7602; RA = 137,476/82,153). RRs for AEs (IR/MR: PsA = 59.6/113.4; RA = 44.0/64.8) and SAEs (PsA = 8.1/13.6; RA = 8.0/9.5) were higher with MR versus IR. AE RRs (RA) in the first 4years after IR approval were 95.9 (IR; 49,439 PY) and 147.0 (MR; 2000 PY). Frequency of SAEs, AESIs, and fatal cases was mostly similar across formulations and indications. The most frequently-reported AE Preferred Terms (PsA/RA) included drug ineffective (20.0%/17.8%), pain (9.7%/10.6%), condition aggravated (9.9%/10.5%), headache (8.8%/7.9%) and, for PsA, off-label use (10.5%/3.4%). Tofacitinib PMS safety data from submitted AE reports were consistent between PsA and RA, and aligned with its known safety profile. Exposure data (lower MR versus IR; estimation from commercial sales data), reporting bias, reporter identity, and regional differences in formulation use limit interpretation.