Synthetic data only. Never enter patient information; nothing is sent or stored. For clinicians. Not FDA cleared.

CPE-003 · version 1.2.0

CKD Pattern Engine: CKD-EPI 2021, chronicity gates and CGA staging, step by step

No CKD label from a single result. Safety gates run first, then chronicity, then CGA.

46 open rules20 reviewed scenariosDeterministicRuns in your browserReviewed by Dr. Vinicius Amaral, D.O.

The Chronic Kidney Disease Pattern Engine follows the order a careful nephrologist reasons in: rule out an acute or emergent process, detect an abnormality, confirm it is chronic, then classify by cause, GFR and albuminuria (CGA).

A single low eGFR or a single high UACR never becomes a CKD diagnosis. The engine says so, tells you what would confirm it, and keeps criteria-based status clearly apart from pattern hypotheses.

Playground

Load a synthetic case or type your own values. The result recomputes as you type, and every rule links to its page.

Live eGFR and CGA map

eGFR (mL/min/1.73 m²)

93.6

CKD-EPI 2021 creatinine, race-free

GFR \ ALBA1<30A230–300A3>300
G1≥90●
G260–89
G3a45–59
G3b30–44
G415–29
G5<15

Classified: CKD G1A2. Colors follow the KDIGO risk heat map: green low, yellow moderately increased, orange high, red very high.

Kidney functionAdd a prior result with dates to test chronicity
Sex for the equation

Switches the engine to the combined eGFRcr-cys equation.

Albuminuria
Chemistry
Kidney findings and red flags
Context
Transient albuminuria confounders
Creatinine may be unreliable
Complications (persistent)
Therapy
Documented history and special populations

Live result · computed in your browser

Detected clinical pattern

Criteria for CKD Met — Classified

Chronicity is documented and the result is classified CKD G1A2. This is criteria-based status, not a probability.

  • Chronicity: Chronicity confirmed (≥3 months)
  • CGA: CKD G1A2

Clinical summary

CKD G1A2: criteria met on documented chronicity (eGFR 93.6 and UACR 89 mg/g). Priority: complete risk stratification, review etiology and kidney-protective management.

Why this pattern was identified

  • Estimated GFR. 93.6 mL/min/1.73 m² (eGFRcr 2021, race-free), category G1 (≥90).
  • Albuminuria. UACR 89 mg/g, category A2 (30–300 mg/g).
  • Persistence. Abnormal results are dated at least three months apart.
  • CKD criteria met. CKD G1A2 on documented chronicity. Criteria-based status, not a probability.

Calculations

eGFRcr (CKD-EPI 2021, race-free)
93.6 mL/min/1.73 m²

Creatinine-based.

Recommended next steps

For physician consideration. Urgent review always comes first.

  1. 1Confirm complete risk stratification (CGA) and review etiology, blood pressure and kidney-protective management.
  2. 2Monitor eGFR and UACR longitudinally at an interval set by the CGA risk category.

Missing information

  • Cystatin C (optional; improves GFR accuracy near decision thresholds).
  • Urine sediment / microscopy (supports etiologic adjudication).

Why it matters

CKD with albuminuria and reduced GFR is associated with higher risks of kidney progression and cardiovascular disease, and the CGA category sets monitoring intensity.

Abnormal results are present on dated measurements separated by at least three months, supporting chronicity.

Evidence and rule trace · 8 rules

DQ-004eGFR calculation assay provenance

Inputs: creatinine assay IDMS traceability not confirmed

eGFR computed with unknown/non-confirmed standardized assay; interpret with assay provenance visible.

GFR-001Compute initial adult race-free creatinine eGFR

Inputs: creatinine 0.78 mg/dL · age 48 · sex female

eGFRcr 2021 = 93.6 mL/min/1.73 m²

DET-001Kidney abnormality detected

Inputs: UACR 89 mg/g (≥30)

kidney_abnormality_detected: determine chronicity and evaluate acute causes before confirming CKD

CHR-001CKD chronicity confirmed through longitudinal evidence

Inputs: UACR ≥30 on 2 dated results spanning 121 days

chronicity_status: confirmed

CGA-001Assign G category from selected indexed eGFR

Inputs: eGFR 93.6 (eGFRcr_2021)

G1 (≥90 mL/min/1.73 m²)

CGA-002Assign A category from UACR

Inputs: UACR 89 mg/g

A2 (30–300 mg/g)

CGA-004Complete CGA classification

Inputs: G1 · A2 · chronicity confirmed

CKD G1A2

KFRE-002KFRE not used in G1–G2 as release decision tool

Inputs: G1

not_eligible_for_primary_KFRE_output

Evidence

  • R1 · KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Disease: Improving Global Outcomes (KDIGO), 2024.
  • R3 · eGFR Equations for Adults: 2021 CKD-EPI creatinine and creatinine-cystatin C race-free equations. National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) / NIH, 2026.
  • R4 · NKF/ASN recommendations implementing race-free CKD-EPI equations. National Kidney Foundation / American Society of Nephrology, 2021.

Physician adjudication required. Clinical decision support for physician review. Criteria-based status is not a probability. Findings require physician adjudication and correlation with the complete clinical context. Validation-only release: not for production clinical use.

What it does

  • Computes the 2021 CKD-EPI race-free eGFR from creatinine, and the combined creatinine–cystatin C equation when cystatin C is available.
  • Runs AKI, severe hyperkalemia, active glomerular disease and obstruction gates before any CKD classification.
  • Confirms chronicity only from dated evidence spanning at least three months, or documented external evidence.
  • Assigns G and A categories and reports a complete CGA classification such as CKD G3bA3.
  • Flags possible etiologic patterns, complications, medication safety, progression and referral triggers.

What it never does

  • Confirm CKD from one abnormal eGFR or UACR.
  • Say dialysis, transplant, an ESA, a phosphate binder or calcitriol is required.
  • Show a probability for criteria-based CKD status.
  • Compute a KFRE risk from unsourced coefficients.

Safety gates

These gates run before the pattern is shown, and urgent findings are displayed first.

  • A creatinine rise ≥0.3 mg/dL in 48 hours, or ≥1.5× baseline in 7 days, defers chronic staging and raises an urgent review.
  • Potassium ≥6.0 mmol/L raises an urgent alert independent of any CKD label.
  • RBC casts, dysmorphic RBCs with heavy proteinuria, or persistent hematuria with rapid decline raise an active glomerular disease alert.
  • Reduced function with hydronephrosis, retention or outlet obstruction raises an obstruction alert.

Inputs

  • Age, sex, serum creatinine (with dates)
  • Cystatin C, optional
  • UACR (with dates and collection type)
  • Potassium, urine sediment, imaging, medications
  • Context: diabetes, hypertension, hematuria

Scope

In scope

  • Adults aged 18 and over
  • Pre-dialysis CKD detection and classification

Routed out

  • Pediatrics
  • Pregnancy
  • Kidney transplant recipients
  • Dialysis or kidney replacement therapy
  • Kidney donor evaluation

The full rule catalog

46 active rules. Each has its own page with conditions, outputs and evidence. 45 of 46 active rules are exercised by at least one synthetic case.

Download the catalog as JSON

Eligibility

Data quality

Global safety gates

Abnormality detection

Chronicity

Albuminuria

GFR estimation

CGA classification

Etiology patterns

Progression

Risk prediction (KFRE)

Complications

Medication safety

Referral

Synthetic cases

Every case is synthetic. Open one to see the inputs, the full result and the rule trace; then change the values.

CKD-FX-001

Persistent moderately increased albuminuria with preserved GFR

Persistent moderately increased albuminuria with preserved GFR

CKD-FX-002

Persistent severely increased albuminuria with mildly decreased GFR

Persistent severely increased albuminuria with mildly decreased GFR

CKD-FX-003

Single reduced eGFR without chronicity

Single reduced eGFR without chronicity

CKD-FX-004

Acute creatinine rise before CKD staging

Acute creatinine rise before CKD staging

CKD-FX-005

Isolated suspected G3a requiring cystatin C confirmation support

Isolated suspected G3a requiring cystatin C confirmation support

CKD-FX-006

Combined eGFR reclassification support

Combined eGFR reclassification support

CKD-FX-007

Confirmed CKD G3bA3 with diabetes

Confirmed CKD G3bA3 with diabetes

CKD-FX-008

Active glomerular disease red flag

Active glomerular disease red flag

CKD-FX-009

Anemia in CKD context without overdiagnosing EPO deficiency

Anemia in CKD context without overdiagnosing EPO deficiency

CKD-FX-010

Vitamin D low alone does not establish CKD-MBD

Vitamin D low alone does not establish CKD-MBD

CKD-FX-011

Severe potassium elevation with eGFR unknown

Severe potassium elevation with eGFR unknown

CKD-FX-012

Metformin contraindication threshold

Metformin contraindication threshold

CKD-FX-013

Kidney failure range does not automatically mandate dialysis

Kidney failure range does not automatically mandate dialysis

CKD-FX-014

Possible monoclonal protein-related kidney disease

Possible monoclonal protein-related kidney disease

CKD-FX-015

ACE inhibitor monitoring rather than automatic cessation

ACE inhibitor monitoring rather than automatic cessation

CKD-FX-016

ACE inhibitor creatinine rise requiring evaluation

ACE inhibitor creatinine rise requiring evaluation

CKD-FX-017

UACR positive random sample requires confirmation

UACR positive random sample requires confirmation

CKD-FX-018

Transient UACR confounder

Transient UACR confounder

CKD-FX-019

Possible obstructive kidney process

Possible obstructive kidney process

CKD-FX-020

SGLT2i early eGFR dip

SGLT2i early eGFR dip

CKD-EX-01 · extended

A 12-year-old is outside the adult pathway

Pediatric patients are routed out; no classification is produced.

CKD-EX-02 · extended

Dialysis patients are routed to a specialized pathway

On maintenance dialysis, so the general CKD classification does not apply.

CKD-EX-03 · extended

Oliguria meets an AKI trigger

Urine output below 0.5 mL/kg/h for 8 hours.

CKD-EX-04 · extended

Documented prior CKD supports chronicity

A single eGFR of about 50 in a patient with CKD documented years ago.

CKD-EX-05 · extended

G2 without a marker of kidney damage

A documented CKD label, G2, with a normal UACR and no other marker.

CKD-EX-06 · extended

Low muscle mass distorts creatinine

Serum creatinine 0.5 mg/dL in a frail patient with low muscle mass.

CKD-EX-07 · extended

Longstanding hypertension with moderate albuminuria

Confirmed CKD G3a, UACR 80 mg/g, no active sediment.

CKD-EX-08 · extended

Creatinine rise after an NSAID

Creatinine 1.0 to 1.6 mg/dL over four days with recent ibuprofen use.

CKD-EX-09 · extended

eGFR fell by more than 20% between visits

Confirmed CKD with creatinine 1.2 to 1.7 mg/dL over six months.

CKD-EX-10 · extended

UACR doubled between measurements

UACR 120 to 260 mg/g in confirmed CKD.

CKD-EX-11 · extended

eGFR fell 35% after a kidney-protective drug

A hemodynamically active therapy was started; eGFR dropped 35%.

CKD-EX-12 · extended

Anemia, acidosis and mineral-bone signals in G4

Hemoglobin 10.2, bicarbonate 17 and persistently high phosphate.

CKD-EX-13 · extended

Metformin in the eGFR 30–44 review zone

eGFR 38 on metformin.

Evidence

Guidelines and references the rules cite. Each rule page lists the ones that support it.

  • R1KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Disease: Improving Global Outcomes (KDIGO), 2024.

    Supports: CKD definition and chronicity CGA classification albuminuria confirmation risk prediction referral medication stewardship progression monitoring

  • R2KDOQI US Commentary on the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of CKD. National Kidney Foundation - KDOQI / American Journal of Kidney Diseases, 2025.

    Supports: US clinical implementation context deployment governance

  • R3eGFR Equations for Adults: 2021 CKD-EPI creatinine and creatinine-cystatin C race-free equations. National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) / NIH, 2026.

    Supports: eGFRcr computation eGFRcr-cys computation assay standardization adult eligibility

  • R4NKF/ASN recommendations implementing race-free CKD-EPI equations. National Kidney Foundation / American Society of Nephrology, 2021.

    Supports: race-free eGFR policy US laboratory alignment

  • R5KDIGO 2026 Clinical Practice Guideline for the Management of Anemia in CKD - Executive Summary. KDIGO, 2026.

    Supports: anemia thresholds first-line anemia workup iron deficiency terminology

  • R6KDIGO 2017 Clinical Practice Guideline Update for CKD-Mineral and Bone Disorder. KDIGO, 2017.

    Supports: CKD-MBD serial interpretation avoid automatic calcitriol/binder outputs

  • R7KDIGO Clinical Practice Guideline for Acute Kidney Injury. KDIGO, 2012.

    Supports: AKI creatinine/urine output safety gate

  • R8GLUCOPHAGE / GLUCOPHAGE XR Prescribing Information - Renal Impairment Recommendations. U.S. Food and Drug Administration (FDA) - Drugs@FDA labeling, 2017.

    Supports: metformin medication safety thresholds

  • R9Clinical Decision Support Software Guidance for Industry and Food and Drug Administration Staff. U.S. Food and Drug Administration (FDA), 2026.

    Supports: HCP independent-review design output transparency knowns/unknowns presentation automation-bias risk controls regulatory positioning

Frequently asked questions

Which eGFR equation does the engine use?

The 2021 CKD-EPI race-free creatinine equation for adults, and the 2021 creatinine–cystatin C equation when cystatin C is available. Both are checked against published reference values in the test suite.

When does the engine call something CKD?

Only when a kidney abnormality is documented over at least three months, or externally documented as chronic. A G1 or G2 category additionally needs a marker of kidney damage such as albuminuria.

Why is there no KFRE risk number?

The engine reports KFRE eligibility (confirmed CKD G3–G5 with age, sex, eGFR and UACR) but does not calculate it. A validated, version-controlled implementation is required, and the engine will not hand-code unverified coefficients.

What does the engine do with medications such as metformin or an ACE inhibitor?

It raises monitoring and review flags tied to labeling and guidelines, for example a metformin labeling flag below eGFR 30 or creatinine and potassium follow-up after an ACEi/ARB start. It never says to stop or start a drug.

Does the pattern strength show a score?

No. Etiologic patterns are shown with their supporting and contradicting evidence only. A numeric pattern score is not shown because its weights have not been clinically validated.

Specification status: Specification v1.2.0, validation release. Reviewed by Dr. Vinicius Amaral, D.O. Not for clinical production use.

Clinical review: Dr. Vinicius Amaral, D.O., Medical reviewer, 2026-10-03.

The Clinical Intelligence Layer — transforming fragmented clinical information into structured intelligence for clinician review and action.

Contact Biod