Abstract Background/Aims Organ damage remains an important predictor of morbidity and mortality in systemic lupus erythematosus (SLE). The Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index (SDI) is the only validated clinician completed tool for measuring this. Routinely collected electronic health records (EHR) provide a data source about unselected patients. Optimising the SDI for use in this setting provides the chance for studying SLE damage in a larger, more representative context than traditional hospital cohorts. This study aimed to (1) develop an electronic adaptation of the SDI (eSDI) suitable for use within an EHR dataset, the UK Clinical Practice Research Datalink (CPRD), and (2) examine preliminary associations between damage and mortality. Methods A retrospective cohort study was conducted using CPRD GOLD (study period 1990-2020). Adults (≥18 years) with SLE were selected using our previously published algorithm. Damage items were operationalised from SDI definitions using diagnostic, procedural, and medication codes. Associations between individual and overall damage (eSDI 0) and mortality were tested using univariate and multivariable logistic regression adjusting for age, sex, smoking, ethnicity, and baseline comorbidities. Odds ratios (OR) with 95% confidence intervals (CI) were reported. Results 8,363 SLE patients were included (87.5% female, mean age 46.7 ± 16 years); 1,004 deaths occurred during follow-up. Overall, 42.3% developed damage (eSDI 0). The most frequent items were osteoporosis (10.4%), malignancy (8.4%), and cataract (7.4%), with musculoskeletal, ocular, and cardiovascular the most affected domains. Most damage items were associated with increased mortality on univariate analysis. The highest odds were seen for malignancy (OR 6.26, 95% CI 5.27-7.42), pulmonary fibrosis (OR 4.89, 95% CI 3.15-7.48), and significant tissue loss (OR 6.82, 95% CI 2.95-15.60). The presence of any damage (eSDI 0) was strongly associated with mortality (adjusted OR 4.46, 95% CI 3.57-5.57, p 0.001) in the multivariable model. Age at diagnosis (OR 1.06, 95% CI 1.05-1.07), male sex (OR 1.40, 95% CI 1.08-1.80), and smoking (OR 1.62, 95% CI 1.29-2.03) were independent predictors of death. Conclusion Assessment of SLE-related organ damage using EHR data is feasible. Damage was common and independently associated with a fourfold increase in the odds of mortality, even after adjustment for demographic and comorbidities. EHR datasets therefore provides a viable and important opportunity for further study of SLE damage in the ‘real world’ setting; further work interrogating the associations between damage and survival as well as patterns of damage will be valuable. Disclosure J. Ellis: None. N. McHugh: None. J.D. Pauling: Honoraria; JDP has undertaken consultancy work and/or received speaker honoraria from Janssen, Astra Zeneca, Boehringer Ingelheim, IsoMab, Sojournix Pharma and Permeatus Inc. I. Bruce: Consultancies; INB has received consulting fees from AstraZeneca, Eli Lilly, GSK, Takeda, UCB and Dragonfly Therapeutics. Honoraria; INB was a speaker for AstraZeneca, Janssen, GSK and UCB. Grants/research support; INB has received grant support from GSK, Janssen and Astra Zeneca. S. Gor: None. J.H. Humphreys: None. H. Vaughan-Williams: None. E. Korendowych: None. S. Skeoch: None. A. McGrogan: None.
Ellis et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: