Creatinine Artifact — Creatine Supplementation

What this model detects

Creatine monohydrate supplementation causes a pharmacokinetic artifact that mimics kidney dysfunction on standard clinical chemistry: serum creatinine rises 15–30% above pre-supplementation baseline, while true glomerular filtration rate is unchanged.

This artifact affects:

  • Serum creatinine (direct measurement)
  • Creatinine-based eGFR (all standard equations: CKD-EPI, MDRD, Cockcroft-Gault)
  • BUN:creatinine ratio (shifts artifactually)
  • Urine creatinine (elevated excretion rate)
  • Apple Watch estimated GFR (uses creatinine-based equations)

What features mislead

FeatureDirection of artifactMagnitudeReliability
Serum creatinine↑ 15–30%Dose/duration dependentHigh — well-characterized
Creatinine-based eGFR↓ artifactually (10–30 mL/min)VariableHigh — direct consequence
Apple Watch eGFR↓ artifactuallySame magnitudeHigh — same equation
Cystatin CUnchangedNoneHigh — not affected by creatine
BUNNo meaningful changeMinimalHigh
True GFR (iohexol/51Cr-EDTA)UnchangedNoneGold standard

Body composition artifacts during loading:

  • BIA impedance: Falsely lower body fat %, falsely higher lean mass — ICW shifts increase conductive mass
  • DEXA: Falsely higher lean mass — water registered as lean tissue
  • Scale weight: +0.5–2.0 kg during loading (water), +0.5–1 kg long-term (sustained ICW)

Certainty limits

High certainty:

  • Creatinine rises are a direct pharmacokinetic consequence of creatine → creatinine conversion (~1–2% of total body creatine pool per day)
  • True GFR is unaffected in healthy adults
  • Cystatin C is not affected by creatine supplementation
  • The artifact is fully reversible upon cessation (creatinine normalizes within ~7–14 days of stopping)

Moderate certainty:

  • The artifact magnitude scales roughly with dose (loading > maintenance) but individual variation is substantial
  • Body composition BIA/DEXA artifacts during loading are well-characterized qualitatively; precise magnitude varies

Low certainty:

  • Non-responders (~20–30% of users) may show minimal creatinine elevation — absence of artifact does not rule out creatine use

Route / dose / timing caveats

  • Loading phase (20 g/day × 5–7 days): Artifact is largest — expect near-maximum creatinine elevation
  • Maintenance phase (3–5 g/day): Artifact is smaller but persistent and dose-proportional — eGFR still confounded at steady state
  • Timing after cessation: Creatinine normalizes within 7–14 days of stopping; eGFR interpretation should wait ≥14 days post-cessation for clinical decisions
  • Hydration status: Dehydration elevates serum creatinine independently — compounding with creatine artifact can produce clinically misleading values
  • Age 65+: The artifact is the same magnitude but absolute eGFR values are already lower; the artifact can push apparent eGFR across clinical thresholds

What to use instead

For kidney function monitoring in creatine users:

  1. Cystatin C-based eGFR — not confounded by muscle mass, diet, or creatine. Requires clinical blood draw (not available on consumer wearables).
  2. Measured GFR (iohexol or 51Cr-EDTA clearance) — gold standard research method; not clinically practical.
  3. Apple Watch eGFR is not usable during creatine supplementation. Flag and suppress any kidney function alerts from wearable-derived eGFR.

Practical Vitals protocol:

WHILE user is on creatine supplementation:
  SUPPRESS: Apple Watch eGFR alerts
  SUPPRESS: serum creatinine trending as kidney concern
  REQUEST: cystatin C-based eGFR on next blood panel
  NOTE: Body composition BIA/DEXA confounded for 4–6 weeks post-initiation
  CADENCE: re-evaluate cystatin C every 6 months or per standard wellness

Relationship to this vault

  • Creatine — the parent substance note (hub); this detection model is a satellite
  • Sarcopenia Detection — links to this detection model as a kidney monitoring caveat for muscle health panels
  • Muscle Health Biomarkers — mentions cystatin C as the preferred kidney marker; this note provides the mechanistic basis