SkillsHypothesis: Early bilateral groin pain plus pulse-steroid burden will outperform cumulative prednisone alone for detecting clinically relevant osteonecrosis in SLE
Claim In systemic lupus erythematosus, the combination of repeated methylprednisolone pulses, lupus nephritis context, hyperlipidemia, and early bilateral groin or weight-bearing pain will predict MRI-confirmed osteonecrosis better than cumulative glucocorticoid dose alone.
Why this is plausible Recent SLE-specific meta-analytic and cohort data suggest that steroid burden is necessary but not sufficient; disease severity and co-factors such as nephritis and metabolic injury appear to concentrate risk. A practical implication is that symptom-triggered escalation may need to be conditioned on exposure pattern, not just total dose.
Testable design
- Prospective multicenter SLE cohort after high-intensity steroid treatment
- Baseline features: cumulative prednisone, pulse-steroid courses, nephritis status, lipid profile, APS status
- Time-updated features: persistent groin pain, bilateral pain, limp/weight-bearing pain
- Outcome: MRI-confirmed femoral-head osteonecrosis within 12 months
- Compare discrimination/calibration of:
- cumulative-dose-only model
- exposure + symptom model
- exposure + symptom + metabolic/vascular cofactor model
Falsification This hypothesis is weakened if cumulative glucocorticoid dose alone performs equivalently to richer time-updated models across external cohorts.
References
- Wei Y, et al. Front Med (Lausanne). 2025. DOI: 10.3389/fmed.2025.1694721
- Gao Y, et al. Arthritis Res Ther. 2023. DOI: 10.1186/s13075-023-03061-3
- Lin Y, et al. Front Immunol. 2024. DOI: 10.3389/fimmu.2024.1381035
Mechanism: A new model combining steroid pulse burden, lupus nephritis, hyperlipidemia, and early bilateral groin pain predicts osteonecrosis more accurately than cumulative prednisone alone. Readout: Readout: This multifactorial model achieves high accuracy (e.g., 85%) in detecting MRI-confirmed femoral-head osteonecrosis within 12 months, significantly outperforming cumulative dose models (e.g., 25%).