Abstract
Background The current commonly used definitions of flare may not be able to capture patients with a persistently active disease (PAD) course. This study sought to identity the frequency and determinants of flare and PAD in an Asia-Pacific cohort.
Methods Data from Asia-Pacific SLE patients collected between 2013 and 2020 were included. Flare was assessed using the SELENA-SLEDAI flare index (SFI) and PAD was defined as a SLEDAI-2K score of >4, excluding serology, on >2 consecutive visits. Data from 2013 to 2015 were used to model flare and PAD in 2016 through logistic regression and model properties were tested for prediction of flare and PAD in 2020 using the data from 2017 to 2019.
Results During median 2.5 (1.0–5.1) years, 53.1% (2180/4106) of patients experienced at least one episode of flare (flare incidence 0.49 per patient-year). 1786 (43.5%) patients experienced PAD including 368 patients (9.0%) who did not achieve the definition of flare. In the predictive model for flare, being from a country with GDP<$20,000, current smoking, prior mucocutaneous involvement, arthritis, nephritis and low complements were risk factors, and achieving low disease activity state (LLDAS) for ≥50% of follow-up time during the previous three years was a protective factor. Prior nephritis and higher time-adjusted SLEDAI score in the previous three years were predictors for subsequent PAD while spending ≥50% of follow-up time in LLDAS during the previous three years was protectively associated with PAD (table 1). The two models gave 72% and 83.8% correct prediction of flare and PAD in 2020, respectively.
Conclusions Both flare and PAD were common disease activity patterns in SLE, with 9% of patients having PAD that was not captured by the SFI definition. Our predictive models may help identify patients at high risk of flare or PAD and enable targeted interventions to achieve better outcomes.
