Lupus & antiphospholipid syndrome
Neutrophil to lymphocyte ratio (NLR) outperforms platelet to lymphocyte ratio (PLR) in SLE diagnosis and prognosis: an artificial intelligence (AI)-enhanced meta-analysis of 12 850 patients with ethnicity-specific cut-offs
Authors' conclusion
NLR outperforms PLR in SLE diagnosis and risk stratification, with validated cut-offs that vary significantly by ethnicity.
Study context
- Population
- Adults (≥18 years) meeting 2019 EULAR/ACR classification criteria for SLE, global SLE registries (Asia, Europe, North America, Africa); comparators healthy controls and non-SLE autoimmune diseases
- Participants
- 12,850 adults50 studies
- Comparison
- NLR vs PLR as diagnostic and prognostic biomarkers in SLE
- Outcome
- Diagnostic accuracy (AUC, sensitivity, specificity) for active SLE (SLEDAI ≥6)
Methodological detailstwo-phase mixed-methods, PRISMA-compliant, XGBoost model with SHAP explainability, multicentre
- Two-phase mixed-methods
- PRISMA-compliant
- XGBoost model with SHAP explainability
- Multicentre
Pooled result
As reported in the abstract
Diagnostic accuracy for active SLE
| Group | Pooled AUC | Pooled sensitivity | Pooled specificity |
|---|---|---|---|
| NLR | 0.85 | 78% | 82% |
| PLR | 0.78 | 70% | 75% |
Pooled AUC
NLR0.85
PLR0.78
Pooled sensitivity
NLR78%
PLR70%
Pooled specificity
NLR82%
PLR75%
- NLR vs PLR: P = 0.02
Further pooled results
| Outcome | NLR | PLR |
|---|---|---|
| Prognosis: risk with elevated NLR (>3.5) and PLR (>185) | ||
| HR lupus nephritis | 2.1 | 1.8 |
| HR cardiovascular events | 2.3 | 1.9 |
| HR mortality | 3.1 | 2.1 |
- Prognosis: risk with elevated NLR (>3.5) and PLR (>185): all p<0.01
Optimal NLR cut-offs by ethnicity
- Asian
- >3.1
- Caucasian
- >2.8
- African
- >3.4
- Significant ethnic variations
AI model performance
- AUC in training
- 0.87
- AUC in validation
- 0.82
- NLR top predictive feature, SHAP score
- 0.25
Original publication
NLR outperforms PLR in SLE diagnosis and prognosis: an AI-enhanced meta-analysis of 12 850 patients with ethnicity-specific cut-offs. Lupus Sci Med. Published 30 September 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →