Factors associated with damage accrual in patients with systemic lupus erythematosus with no clinical or serological disease activity: a multicentre cohort study

Background: Evaluating the contribution of glucocorticoid use to organ damage in systemic lupus erythematosus is confounded by glucocorticoid use in active disease. We sought to determine the independence of the contribution of glucocorticoid use to damage accrual from associations with disease activity by analysing patients without measurable disease activity. Methods: Patients (age >18 years) who met the criteria for systemic lupus erythematosus were recruited from 13 centres in Australia, Indonesia, Japan, Malaysia, the Philippines, Singapore, Taiwan, and Thailand, and followed longitudinally. Disease activity (Systemic Lupus Erythematosus Disease Activity Index 2000 [SLEDAI-2K] and Physician Global Assessment [PGA] scores) and treatment details were recorded at each visit (at least once every 6 months), and organ damage measured annually according to the Systemic Lupus International Collaborating Clinics Damage Index (SDI). Glucocorticoid use during the study period was recorded as any exposure to prednisolone, cumulative prednisolone exposure, and time-adjusted mean daily prednisolone dose. Multivariate survival analyses were used to examine time-dependent associations of glucocorticoid use with damage accrual (defined as an increase of ≥1 on SDI). A SLEDAI-2K score of 0 was taken to indicate the absence of clinical and serological disease activity; a subset of patients without disease activity during the study were defined by a time-adjusted mean SLEDAI-2K (AMS) score of 0. Findings: Between May 1, 2013, and Dec 31, 2016, 1707 patients were recruited. Over a median observation period of 2·2 years (IQR 1·5–3·0), damage accrual events were observed in 255 (14·9%) patients. 1405 (82·3%) of patients were exposed to prednisolone, with a median time-adjusted mean prednisolone dose of 5·0 mg/day (IQR 1·9–8·8). As SLEDAI-2K and PGA scores were highly correlated, two multivariable models were set, each including one of these two variables. In the model including AMS score, baseline SDI damage (SDI >0) was independently associated with damage accrual (HR 1·32 [95% CI 1·01–1·73], p=0·0427). In the other model, time-adjusted mean PGA score was independently associated with damage accrual (1·05 [1·02–1·08], p=0·0012). In both models, factors independently associated with damage accrual included time-adjusted mean prednisolone dose, age at enrolment, and ethnicity (Asian vs non-Asians). 157 (9·2%) patients had an AMS score of 0 (no disease activity), among whom 103 (65·6%) had glucocorticoid exposure and the median time-adjusted mean prednisolone dose was 2·0 mg/day (IQR 0·0–5·0). Accrual of irreversible organ damage occurred in 21 (13·4%) of these patients and was independently associated with time-adjusted mean prednisolone dose (HR 1·14 [95% CI 1·03–1·26], p=0·0117), time-adjusted mean PGA score (1·13 [1·03–1·23], p=0·0144), and age at enrolment (1·04 [1·01–1·07], p=0·0061), but not baseline SDI damage (0·94 [0·43–2·06], p=0·8675). Interpretation: Glucocorticoid use contributes to damage accrual in systemic lupus erythematosus independently of the presence of clinical or serological disease activity. Funding: UCB Pharma, GlaxoSmithKline, Janssen, Bristol-Myers Squibb, and AstraZeneca (to the Asia-Pacific Lupus Collaboration).