Kidney failure risk calculator

How this calculator works

This page uses the original Canadian calibration, not the later pooled or regionally recalibrated models. Use a measured numeric ACR; an albuminuria category cannot replace it. The eight-variable option adds albumin, bicarbonate, calcium and phosphorus. Outputs estimate treatment with dialysis or transplantation over the stated horizon, not the date a kidney will fail. Calibration varies by population and treatment era. The model is outside scope in acute kidney injury, children, dialysis and transplant recipients. The result is an estimate for information only and is not medical advice.

Formula

Risk(t) = 1 - S0(t)^exp(LP). Centred variables: A = age/10 - 7.036; M = male(1 or 0) - 0.5642; G = eGFR/5 - 7.222; L = ln(ACR in mg/g) - 5.137. Four-variable LP = -0.2201A + 0.2467M - 0.5567G + 0.4510L; S0 = 0.9750 at 2 years and 0.9240 at 5 years. Eight-variable LP = -0.1992A + 0.1602M - 0.4919G + 0.3364L - 0.3441(albumin g/dL - 3.997) + 0.2604(phosphorus mg/dL - 3.916) - 0.07354(bicarbonate mmol/L - 25.57) - 0.2228(calcium mg/dL - 9.355); S0 = 0.9780 and 0.9301. ACR mg/mmol is divided by 0.11312 to obtain mg/g.

Worked example

At the eight-variable centring values, use age 70.36 years, eGFR 36.11, ACR exp(5.137) mg/g and the listed centred lab values. A man has LP = 0.1602 × (1 - 0.5642) = 0.06981516, giving about 2.357% at 2 years and 7.476% at 5 years.

Frequently asked questions

Sources

• Tangri et al. (2011): original kidney failure prediction models
• Kang et al. (2020): original equations reproduced in Table 7
• Primary CKD clinic study: eight-variable equation, Supplemental Table 2