Background Prognosis of systemic lupus erythematosus (SLE) has been improved during past decades. However, some clinical features which might be associated with poor prognosis persist or increase. This study aimed to identify changes of clinical features and mortality between late and early cohort.

Methods Among 1,448 SLE patients, 621 defined to inception cohort. They divided into early (1998–2007, n=317) and late (2008–2017, n=304) cohort and followed for 10 years until 2008 and 2018, respectively. They were compared with ACR criteria, SLEDAI, Adjusted Mean SLEDAI (AMS), SDI, and mortality. Mortality data were collected by linking with data from the Korean National Statistics Office. Poisson Cox hazard model was used to investigate risk factors for mortality.

Results Mean age at enrollment was 28.5 ± 10.9 years and women were 92.3% in overall inception cohort. Baseline demographic characteristics and the number of ACR criteria during follow-up were not different between two cohorts (all P>0.05). SLEDAI at enrollment (P<0.001) and AMS (P=0.03) were lower in late cohort than early cohort. There was no difference in SDI accrual (P=0.546). However, most common organ damage in early cohort was musculoskeletal (12.3%), followed by neuropsychiatric (8.2%) and renal (7.3%), whereas in late cohort, musculoskeletal (11.5%), followed by pulmonary (6.6%) and skin (5.6%). Renal damage was less in late cohort (7.3% and 2.6%, P=0.013). Mortality was not different between them [n=10 (3.2%) and n=8 (2.6%), P=0.882]. Risk factors for mortality in early cohort was no use of hydroxychloroquine (P=0.017) and neuropsychiatric damage (P=0.024), whereas in late cohort pulmonary damage (P=0.028).

Conclusions Prognosis of late cohort have been improved regarding to disease activity and renal damage. Mortality was not different, but risk factors for mortality in late cohort have been changed from neuropsychiatric to pulmonary damage, which could be a target to improve outcomes for SLE patients diagnosed recently.