Comparative performance of lupus disease-activity indices during pregnancy to predict adverse obstetric outcomes: A colombian SLE cohort

Abstract

Objective

To compare the performance of five lupus disease-activity indices measured at a single retrospective pregnancy timepoint for discriminating a composite adverse obstetric outcome (AO), and to perform exploratory clinically interpretable categorical and renal-item analyses across indices.

Methods

Retrospective cohort at a tertiary center in Bogotá, Colombia (2012–2024). For each pregnancy, one retrospective assessment with complete clinical and laboratory information required to compute all indices was selected. AO was defined as any of: preterm birth, miscarriage, stillbirth/early fetal death, preeclampsia/eclampsia, premature rupture of membranes, postpartum hemorrhage, thromboembolism, placental abruption, or maternal death. We calculated SLEPDAI, Lupus Activity Index in Pregnancy (LAI-P), modified SLAM (m-SLAM), SLE Disease Activity Score (SLE-DAS), and BILAG2004-P. Because the dataset was retrospective and selected on complete data, we treated categorical and renal-item analyses as exploratory.

Results

Twenty-nine pregnancies (28 women) were included; AO occurred in 18/29 (62.1%). The original continuous-score analyses showed overlapping confidence intervals across indices. Exploratory clinically interpretable analyses showed AO in 15/18 (83.3%) pregnancies with any BILAG2004-P A/B domain, 16/20 (80.0%) with SLEPDAI ≥3, and 13/14 (92.9%) with SLE-DAS >7.64. Exploratory renal analyses showed the clearest signals for BILAG2004-P renal A/B (13/14 vs 5/15; OR 26.0, 95% CI 2.61–259.31), SLEPDAI renal-item score >0 (13/13 vs 5/16; OR 56.45, 95% CI 2.81–1133.97), and the LAI-P renal item >0 (12/13 vs 6/16; OR 20.00, 95% CI 2.05–195.01).

Conclusion

In this small single-center cohort, active lupus was associated with adverse obstetric outcomes across disease-activity instruments. However, the study is underpowered to detect meaningful differences in discrimination between indices. Renal activity remained the most clinically relevant organ-specific signal in exploratory analyses, but these findings should be interpreted cautiously and confirmed in larger longitudinal cohorts.