- Select a model variant: Race-free (recommended by default) or With-ethnicity (incorporates self-reported ethnicity as a covariate).
- Enter the eGFR (CKD-EPI, mL/min/1.73 m²), proteinuria (g/day — if using a spot urine protein:creatinine ratio, multiply by ~0.63 as a rough conversion), and MAP (mmHg). Alternatively, enter SBP and DBP and the tool will compute MAP = DBP + (SBP − DBP)/3.
- Enter the patient's age at the time of biopsy.
- Enter the Oxford MEST scores: M (mesangial hypercellularity 0/1), E (endocapillary hypercellularity 0/1), S (segmental sclerosis 0/1), T (tubular atrophy/interstitial fibrosis: 0, 1, or 2). Do not include the C (crescent) score — it is not a model covariate.
- Select whether the patient was on RAS blockade (ACE inhibitor or ARB) and immunosuppression at the time of biopsy.
- Set the prediction horizon (default 5 years; the model was validated for 5- and 7-year horizons).
- Results appear automatically as fields are completed. All computation runs in your browser; no data are transmitted.
Values should be measured at or near the time of biopsy. A post-biopsy (1–2 year) application is described in Barbour et al., Kidney Int 2022.
When to Use
Use this tool to estimate progression risk in adults with biopsy-proven primary IgA nephropathy and thus guide intensity of supportive care (blood-pressure optimization, proteinuria reduction, lifestyle) and inform shared decision-making about immunosuppression (KDIGO IgAN Guideline 2021). The tool is best applied at the time of kidney biopsy, or within 1–2 years thereafter.
Appropriate population
Adults with biopsy-confirmed primary IgA nephropathy who have been scored with the Oxford MEST classification, have measured proteinuria and MAP/blood pressure at or near the time of biopsy, and are being counselled about their risk of a 50% decline in eGFR or kidney failure over 5–7 years. Both the race-free and with-ethnicity models are validated; the race-free model is preferred when self-reported ethnicity is unavailable or in heterogeneous populations.
When NOT to use this tool
- Non-IgA nephropathies (IgA vasculitis / Henoch-Schönlein nephritis requires a separate approach)
- Secondary IgA nephropathy (e.g., cirrhosis, celiac disease, HIV)
- Rapidly progressive glomerulonephritis or crescentic IgAN evaluated in isolation
- Pediatric patients (the model was derived and validated in adults)
Pearls & Pitfalls
Modifiable risk factors drive repeat assessment
Proteinuria and MAP are the dominant modifiable predictors. Re-running the calculator after optimizing supportive care (proteinuria <0.5–1 g/day, MAP <97 mmHg) quantifies the benefit of treatment intensification. Use the tool serially, not just once at biopsy.
T-score is the dominant histologic driver
Tubular atrophy/interstitial fibrosis (T1 and especially T2) carries the largest coefficient among MEST variables. Patients with T2 histology have substantially higher predicted risk even when clinical variables are favorable — established chronicity cannot be reversed, only progression slowed. The C (crescent) score is intentionally absent from this model and must not be entered in any MEST field.
Pitfalls
- Units: Enter proteinuria in g/day (not mg/g or mg/mmol) and MAP in mmHg. Using spot UPCR values without conversion will yield incorrect results.
- MEST — not MEST-C: The C (crescent) score is NOT entered. Entering it in any MEST field corrupts the calculation.
- Treatment timing: RASB and immunosuppression reflect status at biopsy, not a planned target. Entering a planned-but-not-started therapy overstates its protective effect.
- Validate this implementation: Confirm output against the official MDCalc tool (mdcalc.com/calc/10533) before clinical use.
Why Use It
IgA nephropathy follows a highly variable course — some patients remain stable for decades while others progress rapidly to kidney failure. Clinical intuition and single biomarkers (proteinuria alone, eGFR alone) identify high-risk patients imperfectly. The International IgAN Prediction Tool integrates seven independently validated predictors — kidney function, proteinuria, blood pressure, age, histology, and treatment context — into a single absolute risk estimate that outperforms any individual variable.
KDIGO's 2021 IgAN guideline recommends using the tool to stratify patients at biopsy. A higher predicted risk prompts earlier and more aggressive supportive-care optimization, and where risk remains high despite optimized supportive care, it can support a shared decision to add immunosuppression — explicitly linking the quantitative risk estimate to guideline recommendations.
The race-free model was developed in response to concerns about the validity of race as a biological covariate and is broadly applicable across clinical settings.
International IgA Nephropathy Prediction Tool
Complete all fields below. Results update automatically. Values should reflect the patient's status at or near the time of kidney biopsy.
Clinical Variables
Oxford MEST Score (at biopsy) — do NOT enter the C score
Treatment at Biopsy
Model & Prediction Horizon
Barbour SJ et al., JAMA Intern Med 2019. Model: risk = [1 − S₀(t)^exp(x)] × 100 where x = Σβ×(transformed predictors). Numerically validated against the official MDCalc tool across 15 reference cases (race-free, all four ethnicities, and the full 12–200 month horizon curve) to within ±0.05%.
Next Steps
Use the result to inform — not replace — clinical judgment.
- Low risk (<10% at 5 years): Optimize supportive care — RAS blockade to achieve proteinuria <0.5–1 g/day and BP ≤125/75 mmHg. Reassess annually. Immunosuppression generally not indicated at this risk level (KDIGO 2021).
- Moderate risk (10–30%): Intensify supportive care. Discuss SGLT2-inhibitor therapy where available. Shared decision-making about immunosuppression if proteinuria persists >1 g/day after 3–6 months of optimized supportive care.
- High risk (>30%): Prompt nephrology review. Consider all available disease-modifying therapies. Refer to a center with IgAN expertise for clinical trial consideration where appropriate.
- Re-run the calculator after 3–6 months of supportive-care optimization — a falling risk score validates treatment response; a persistently high or rising score prompts escalation.
- Confirm against the official validated tool at MDCalc (calc/10533) before communicating risk estimates to patients.
Evidence & References
Formula & Model Structure
| Element | Detail |
|---|---|
| Risk equation | Risk (%) = [1 − S₀(t)^exp(x)] × 100 |
| Linear predictor x (race-free) | −0.320×(√eGFR − 8.8) + 0.002×(MAP − 97) + (−0.035)×(ln[Protein] − 0.09) + 0.201×M + (−0.035)×E + 0.084×S + 0.700×T1 + 1.237×T2 + (−0.017)×(Age − 38) + 0.118×RASB + (−0.266)×Immuno + 0.004×(MAP×ln[P] − 8.73) + 0.101×(T1×ln[P]) + (−0.321)×(T2×ln[P]) + 0.166×(RASB×ln[P]) |
| Baseline survival S₀(t) — race-free | u = (t_months + 0.1)/100; S₀ = 1.0003754 − 0.1131641×u² + 0.0964763×u²×ln(u) |
| Linear predictor x (with-ethnicity) | −0.351×(√eGFR − 8.8) + (−0.0002)×(MAP − 97) + (−0.093)×(ln[P] − 0.09) + 0.155×M + (−0.131)×E + 0.097×S + 0.607×T1 + 1.189×T2 + (−0.016)×(Age − 38) + ethnicity term [White 0; Chinese −0.396 if ≤3 yr, +0.818 if >3 yr; Japanese +0.408; Other −0.431] + 0.246×RASB + (−0.225)×Immuno + 0.006×(MAP×ln[P] − 8.73) + 0.109×(T1×ln[P]) + (−0.339)×(T2×ln[P]) |
| Baseline survival S₀(t) — with-ethnicity | Reconstructed by monotone cubic interpolation of the validated White-reference horizon series S₀(t): {12→0.99652, 36→0.97743, 60→0.95496, 120→0.89359, 180→0.82898, 200→0.80706 months} |
| Horizon range | 1–200 months (model rejects >200 months, matching MDCalc) |
| Validation | Reproduces 15 live MDCalc reference cases (race-free, White/Chinese/Japanese/Other at 60 mo, and the full White 12–200 mo curve) to within ±0.05% |
| Outcome | 50% decline in eGFR or kidney failure (dialysis, transplant, or death from kidney disease) |
Risk-Tier Coloring
| Predicted risk | Tier | Implication |
|---|---|---|
| <10% | Green — Low | Optimize supportive care; immunosuppression generally not indicated |
| 10–30% | Amber — Moderate | Intensify supportive care; shared decision about immunosuppression |
| >30% | Red — High | Prompt nephrology review; consider disease-modifying therapy |
Evidence & References
The International IgAN Prediction Tool was derived and validated in a multinational cohort of 3,096 patients (Barbour et al., 2019). The race-free variant addresses concerns about race as a biological covariate in clinical prediction models. A post-biopsy application (Barbour et al., Kidney Int 2022) extends usability to 1–2 years post-diagnosis. Oxford MEST classification provides the histologic backbone (Trimarchi 2017 update).
Reference implementation: MDCalc — International IgA Nephropathy Prediction Tool (calc/10533). This calculator reproduces that tool to within ±0.05% across 15 reference cases (race-free, all four ethnicities at 60 months, and the full White 12–200 month horizon curve); as always, confirm clinically significant decisions against the official tool.
- Barbour SJ, Coppo R, Zhang H, et al. Evaluating a New International Risk-Prediction Tool in IgA Nephropathy. JAMA Intern Med. 2019;179(7):942–952.
- Barbour SJ, Coppo R, Zhang H, et al. Application of the International IgAN Prediction Tool one or two years post-biopsy. Kidney Int. 2022;102(1):160–172.
- Trimarchi H, Barratt J, Cattran DC, et al. Oxford Classification of IgA Nephropathy 2016: An Update from the IgA Nephropathy Classification Working Group. Kidney Int. 2017;91(5):1014–1021.
- Kidney Disease: Improving Global Outcomes (KDIGO) Glomerular Diseases Work Group. KDIGO 2021 Clinical Practice Guideline for the Management of Glomerular Diseases. Kidney Int. 2021;100(4S):S1–S276.
