METHODS: Data were analysed from patients in a multinational longitudinal cohort with known anti-dsDNA results from 2013 to 2021. Patients were categorized based on their anti-dsDNA results as persistently negative, fluctuating or persistently positive. Cox regression models were used to examine longitudinal associations of anti-dsDNA results with flare.

RESULTS: Data from 37 582 visits of 3484 patients were analysed. Of the patients 1029 (29.5%) had persistently positive anti-dsDNA and 1195 (34.3%) had fluctuating results. Anti-dsDNA expressed as a ratio to the normal cut-off was associated with the risk of subsequent flare, including in the persistently positive cohort (adjusted hazard ratio [HR] 1.56; 95% CI: 1.30, 1.87; P 3. Both increases and decreases in anti-dsDNA more than 2-fold compared with the previous visit were associated with increased risk of flare in the fluctuating cohort (adjusted HR 1.33; 95% CI: 1.08, 1.65; P = 0.008) and the persistently positive cohort (adjusted HR 1.36; 95% CI: 1.08, 1.71; P = 0.009).

METHODS: In this prospective, multinational, longitudinal cohort study, we used data from patients with SLE in the Asia Pacific Lupus Collaboration cohort collected between May 1, 2013, and Dec 31, 2020. Eligible patients were adults (aged ≥18 years) who met either the 1997 American College of Rheumatology modified classification criteria for SLE or the 2012 Systemic Lupus International Collaborating Clinics classification criteria. The primary outcome was all-cause mortality, and LLDAS, remission, and variations of remission with lower glucocorticoid thresholds were the primary exposure variables. Survival analyses were used to examine longitudinal associations between these endpoints and risk of mortality. This study is registered with ClinicalTrials.gov, NCT03138941.

FINDINGS: Among a total of 4106 patients in the cohort, 3811 (92·8%) patients were included in the final analysis (median follow-up 2·8 years [IQR 1·0-5·3]; 3509 [92·1%] women and 302 [7·9%] men), of whom 80 died during the observation period (crude mortality rate 6·4 deaths per 1000 person-years). LLDAS was attained at least once in 43 (53·8%) of 80 participants who died and in 3035 (81·3%) of 3731 participants who were alive at the end of the study (p<0·0001); 22 (27·5%) participants who died versus 1966 (52·7%) who were alive at the end of the study attained LLDAS for at least 50% of observed time (p<0·0001). Remission was attained by 32 (40·0%) of 80 who died and in 2403 (64·4%) of 3731 participants who were alive at the end of the study (p<0·0001); 14 (17·5%) participants who died versus 1389 (37·2%) who were alive at the end of the study attained remission for at least 50% of observed time (p<0·0001). LLDAS for at least 50% of observed time (adjusted hazard ratio 0·51 [95% CI 0·31-0·85]; p=0·010) and remission for at least 50% of observed time (0·52 [0·29-0·93]; p=0·027) were associated with reduced risk of mortality. Modifying the remission glucocorticoid threshold (<5·0 mg/day prednisolone) was more protective against mortality than current remission definitions (0·31 [0·12-0·77]; p=0·012), and glucocorticoid-free remission was the most protective (0·13 [0·02-0·96]; p=0·046).

INTERPRETATION: LLDAS significantly reduced the risk of mortality in patients with SLE. Remission did not further reduce the risk of mortality compared with LLDAS, unless lower glucocorticoid thresholds were used.

METHODS: This multinational cohort study was conducted at 25 sites across 13 Asia-Pacific countries. We included adult patients aged 18 years or older with stable SLE who were receiving routine clinical care, had two or more visits and had attained stable disease at one or more visits. We categorised stable disease into: LLDAS (Systemic Lupus Erythematosus Disease Activity Index 2000 [SLEDAI-2K] score ≤4, Physician Global Assessment [PGA] ≤1, and prednisolone ≤7·5 mg/day); Definitions of Remission in SLE (DORIS) remission (clinical SLEDAI-2K score 0, PGA <0·5, and prednisolone ≤5 mg/day); or complete remission on therapy (SLEDAI-2K score 0, PGA <0·5, and prednisolone ≤5 mg/day). Stable disease categories were mutually exclusive. Tapering was defined as any decrease in dose of corticosteroids or immunosuppressive therapy (mycophenolate mofetil, calcineurin inhibitors, azathioprine, leflunomide, or methotrexate). Using multivariable generalised estimating equations, we compared flares (SELENA-SLEDAI Flare Index) at the subsequent visit after drug tapering. We used generalised estimating equations and Cox proportional hazard models to compare tapering attempts that had begun in LLDAS, remission, and complete remission.

FINDINGS: Between May 1, 2013, and Dec 31, 2020, 4106 patients were recruited to the cohort, 3002 (73·1%) of whom were included in our analysis. 2769 (92·2%) participants were female, 233 (7·8%) were male, and 2636 (88·1%) of 2993 with ethnicity data available were Asian. The median age was 39·5 years (IQR 29·0-50·0). There were 14 808 patient visits for patients in LLDAS, or remission or complete remission, of which 13 140 (88·7%) entered the final multivariable model after excluding missing data. Among the 9863 visits at which patients continued the same therapy, 1121 (11·4%) flared at the next visit, of which 221 (19·7%) were severe flares. Of the 3277 visits at which a patient received a tapering of therapy, 557 (17·0%) flared at the next visit, of which 120 (21·5%) were severe flares. Tapering was associated with higher odds of flare compared with continuing the same therapy (odds ratio [OR] 1·24 [95% CI 1·10-1·39]; p=0·0005). Of 2095 continuous tapering attempts, 860 (41·1%) were initiated in LLDAS, 596 (28·4%) in remission, and 639 (30·5%) in complete remission. Tapering initiated in LLDAS (OR 1·37 [95% CI 1·03-1·81]; p=0·029) or remission (1·45 [1·08-1·94]; p=0·013) had higher odds of flare in 1 year compared with complete remission. Tapering in LLDAS (hazard ratio 1·24 [95% CI 1·04-1·48]; p=0·016) or remission (1·30 [1·08-1·56]; p=0·0054) had a significantly shorter time to first flare than tapering initiated in complete remission. Attaining sustained LLDAS, remission, or complete remission for at least 6 months just before the time of taper was associated with lower odds of flare at next visit, flares in 1 year, and longer time to flare.

INTERPRETATION: Tapering of corticosteroids or immunosuppressive therapy in patients with stable SLE was associated with excess flares. Our findings suggest that drug tapering should be carefully considered, weighing the risks and benefits, and is best exercised in complete (clinical and serological) remission and after maintaining stable disease for at least 6 months.

METHODS: Data from a 12-country longitudinal SLE cohort, collected prospectively between 2013 and 2020, were analysed. SLE patients with mSACQ defined as the state with serological activity (increased anti-dsDNA and/or hypocomplementemia) but without clinical activity, treated with ≤7.5 mg/day of prednisolone-equivalent GCs and not-considering duration, were studied. The risk of subsequent flare or damage accrual per 1 mg decrease of prednisolone was assessed using Cox proportional hazard models while adjusting for confounders. Observation periods were 2 years and censored if each event occurred.

RESULTS: Data from 1850 mSACQ patients were analysed: 742, 271 and 180 patients experienced overall flare, severe flare and damage accrual, respectively. Tapering GCs by 1 mg/day of prednisolone was not associated with increased risk of overall or severe flare: adjusted HRs 1.02 (95% CI, 0.99 to 1.05) and 0.98 (95% CI, 0.96 to 1.004), respectively. Antimalarial use was associated with decreased flare risk. Tapering GCs was associated with decreased risk of damage accrual (adjusted HR 0.96, 95% CI, 0.93 to 0.99) in the patients whose initial prednisolone dosages were >5 mg/day.