Adult-onset Still’s disease (AOSD) is a rare systemic autoinflammatory disorder characterized by marked clinical heterogeneity, variable disease course, and potentially severe complications such as macrophage activation syndrome (MAS). Real-world data from single-center cohorts remain valuable for clarifying disease phenotypes, treatment patterns, and long-term outcomes.

We retrospectively reviewed 48 patients with AOSD treated at a single center. Clinical characteristics, laboratory findings, PET/CT data, treatments, and outcomes were collected from medical records. Patients with and without MAS were compared descriptively. Exploratory analyses were performed to explore factors associated with 3-month remission, and Kaplan–Meier analysis was used to evaluate time to sustained long-term remission.

A total of 48 patients with AOSD were included in this single-center retrospective study. Females accounted for 83.3% of the cohort, and the median age was 44 years (IQR 29–59). Thirty-six patients (75.0%) were newly diagnosed and treatment-naïve at presentation. Among the 46 patients with evaluable longitudinal follow-up, 22 (45.8%) had a single-phase pattern, 14 (29.1%) had an intermittent pattern, and 10 (20.8%) had a chronic pattern; two patients were lost to follow-up and could not be classified. Overall, the cohort showed substantial clinical heterogeneity, with systemic inflammatory manifestations predominating.

MAS occurred in 7 of 48 patients (14.6%). Compared with patients without MAS, those with MAS exhibited a markedly higher inflammatory burden, as reflected by increased neutrophil counts and higher CRP, ESR, ferritin, IL-6, and LDH levels, with CRP and IL-6 showing the most prominent elevations. PET/CT metabolic activity in the enlarged spleen, lymph nodes, and bone marrow did not show clear between-group differences. Given the limited number of MAS events, these findings should be interpreted as descriptive and exploratory.

Conventional immunosuppressive agents were used in 79.5% of patients, with methotrexate being the most frequently prescribed drug, followed by cyclosporine. Among biologic agents, tofacitinib and tocilizumab were the most commonly used, accounting for 12 (25.0%) and 11 (22.9%) of the overall cohort, respectively. Ruxolitinib was used only in patients with MAS (3/7, 43%), consistent with its use in severe hyperinflammatory/MAS settings. Biologic switching was observed in 10/48 patients (21%) overall, including 9/41 (22%) in the non-MAS group and 1/7 (14%) in the MAS group.

The remission rate was 60.0% at 3 months and 79.0% at 6 months. Among patients with follow-up longer than 6 months, 67.4% achieved sustained long-term remission. In exploratory analyses, a single-phase disease pattern, lower LDH levels, and a higher initial glucocorticoid dose were associated with remission at 3 months. In exploratory logistic analysis, higher LDH was associated with 3-month non-remission (OR 1.01, 95% CI 1.00–1.01, P = 0.030), whereas a higher glucocorticoid dose showed a borderline inverse association (OR 0.987, 95% CI 0.968–1.000, P = 0.089). These findings should be interpreted with caution because of the limited sample size and incomplete follow-up in some patients.

By the end of follow-up, two deaths had occurred. One patient developed MAS at 2 months after disease onset and subsequently experienced a sudden subarachnoid hemorrhage. The other discontinued treatment at 3 months after disease onset because of severe type I respiratory failure complicated by infection. Severe complications included myocarditis (3/48), neoplasms (3/48), pulmonary involvement (3/48), and single cases of nephritis, epilepsy, and severe cutaneous vasculitis. Some of these complications occurred in patients with MAS, suggesting a more severe systemic inflammatory phenotype in this subgroup. Kaplan–Meier analysis showed that the probability of remaining without sustained long-term remission declined most markedly during the early follow-up period, particularly within the first 40 months, and then gradually plateaued. The median time to sustained long-term remission was 32 months; however, estimates in the late follow-up period should be interpreted cautiously because the number of patients at risk decreased over time.

In this single-center retrospective cohort, a single-phase disease pattern was associated with a more favorable early remission profile in AOSD. MAS is a rare phenotype in AOSD, and our cohort provides a descriptive characterization of this subgroup. Elevated LDH may serve as a practical marker of poorer short-term prognosis, whereas early and adequate glucocorticoid therapy appeared to favor remission. By providing additional longitudinal follow-up data on disease course and sustained long-term remission, this study contributes real-world evidence for clinical phenotyping and prognostic assessment in AOSD.