Associations of improvement in laboratory tests with clinical outcomes in patients with active systemic lupus erythematosus: a multinational longitudinal cohort study

Background: The selection and categorisation of laboratory tests in disease activity measures used within systemic lupus erythematosus (SLE) trial endpoints lack strong evidence. We aimed to determine whether longitudinal improvements in routinely measured laboratory tests are associated with measures of clinical improvement in patients with baseline active SLE. Methods: We included patients from a multicentre longitudinal cohort (recruited between May 1, 2013, and Dec 31, 2019) with active SLE (SLEDAI-2K ≥6) coinciding with an abnormality in at least one of 13 routine laboratory tests, at a visit designated as baseline. At 12 months, we analysed associations between thresholds of improvement in individual laboratory test results, measured as continuous variables, and five clinical outcomes using logistic regression. Primary outcomes were damage accrual and lupus low disease activity state (LLDAS), and secondary outcomes were modified SLE responder index (mSRI), physician global assessment (PGA) improvement of at least 0·3, and flare. Findings: We included 1525 patients (1415 [93%] women and 110 [7%] men, 1328 [87%] Asian ethnicity) in separate subsets for each laboratory test. The strongest associations with LLDAS and damage protection were seen with improvements in proteinuria (complete response: adjusted odds ratio [OR] 62·48, 95% CI 18·79–208·31 for LLDAS, OR 0·22, 95% CI 0·10–0·49 for damage accrual), albumin (complete response: adjusted OR 6·46, 95% CI 2·20–18·98 for LLDAS, OR 0·42, 95% CI 0·20–1·22 for damage accrual), haemoglobin (complete response: adjusted OR 1·97, 95% CI 1·09–3·53 for LLDAS, OR 0·33, 95% CI 0·15–0·71 for damage accrual), erythrocyte sedimentation rate (complete response: adjusted OR 1·71, 95% CI 1·10–2·67 for LLDAS, OR 0·53, 95% CI 0·30–0·94 for damage accrual), and platelets (complete response: adjusted OR 4·82, 95% CI 1·54–15·07 for LLDAS, OR 0·49, 95% CI 0·20–1·19 for damage accrual). Improvement in serological tests were mainly associated with PGA and mSRI. White cell and lymphocyte count improvements were least predictive. Interpretation: Improvements in several routine laboratory tests correspond with clinical outcomes in SLE over 12 months. Tests with the strongest associations were discrepant with laboratory tests included in current trial endpoints, and associations were observed across a range of improvement thresholds including incomplete resolution. These findings suggest the need to revise the use of laboratory test results in SLE trial endpoints. Funding: Abbvie.