Background Hydroxychloroquine (HCQ) is a cornerstone treatment of systemic lupus erythematosus (SLE). We compared time to flare in SLE patients discontinuing/reducing HCQ versus those maintaining their dose, and identified factors associated with time to flare.

Methods We analyzed prospective data from the Systemic Lupus International Collaborating Clinics (SLICC) cohort, which includes SLE patients from 33 sites in Europe, Asia, and North America, enrolled within 15 months of diagnosis and followed annually (1999-2019). We identified patients with HCQ reduction/discontinuation, regardless of disease activity. We evaluated person-time that patients contributed on their initial dose (‘maintenance’), comparing this to person-time contributed after a first dose reduction, and person-time after a first HCQ discontinuation. We estimated time to first flare, defined as either subsequent need for therapy augmentation (steroids or other immunomodulators), increase of ≥4 points in the SLE Disease Activity Index-2000 (SLEDAI-2K) or hospitalization for SLE. We estimated crude flare rates for each sub-cohort and hazard ratios and 95% confidence intervals (CIs) for various demographic and clinical factors potentially associated with flare risk in the reduction and discontinuation sub-cohorts, as well as comparator maintenance sub-cohorts (matched for time on HCQ to the reduction and discontinuation sub-cohorts).

Results We studied 1460 SLE patients (90% women, 52% Caucasian) on HCQ. Of these, 592 subsequently reduced HCQ at any point, while 407 discontinued HCQ at any point. The crude flare rate for the HCQ reduction sub-cohort was 42.3 per 100 person-years (95% CI 38.6, 46.4), versus 35.6 (95% CI 32.4, 39.1) in the matched maintenance sub-cohort. In the discontinuation sub-cohort, the crude flare rate was 43.1 (95% CI 38.3, 48.4), versus 34.2 (95% CI 30.6, 38.2) in the matched maintenance sub-cohort. Table 1 shows the factors associated with time to flare within each sub-cohort. The hazard ratios are adjusted by all variables in the table. Prednisone or immunosuppressive use at time-zero was associated with higher flare risk in all analyses. Lower education was associated with higher risk of SLE flares among patients who discontinued HCQ. There was a trend across sub-cohorts for lower flare risk among patients from Asia, versus North America.

Conclusions Compared to HCQ maintenance, crude flare rates were numerically higher after HCQ taper/discontinuation. SLE patients on prednisone or immunosuppressives were at higher risk for flare in all groups. The association between lower education and higher SLE flare risk was most clearly seen upon discontinuation of HCQ, suggesting this as a particularly vulnerable group.

Acknowledgements This research was supported by the Canadian Institutes of Health Research (CIHR).