Executive Summary
- The 2021 CKD-EPI equations without race coefficients are now standard for GFR estimation using creatinine, cystatin C, or both markers combined
- Creatinine-based eGFR is affected by muscle mass, diet, medications, and tubular secretion, leading to systematic over- or underestimation in specific populations
- Cystatin C-based eGFR is affected by thyroid dysfunction, glucocorticoids, obesity, inflammation, and malignancy
- The combined eGFRcr-cys equation provides the highest accuracy with approximately 91–92% of estimates within 30% of measured GFR
- The eGFRcys/eGFRcr ratio provides prognostic information beyond GFR staging; a ratio <0.7 indicates increased mortality risk
Why Knowing the Estimated GFR Matters
The relationship between serum creatinine and GFR is inverse and nonlinear. A serum creatinine of 1.0 mg/dL may represent a GFR of 120 mL/min/1.73m² in a young muscular male or 50 mL/min/1.73m² in an elderly sarcopenic female. Patients can lose more than 50% of kidney function before serum creatinine rises above the normal reference range—the so-called “creatinine-blind” range.
A “normal” serum creatinine does not exclude significant kidney dysfunction. Always report and interpret eGFR rather than creatinine alone, particularly in populations with altered muscle mass such as the elderly, sarcopenic, or malnourished.
Creatinine-Based GFR Estimation
Creatinine Physiology
Creatinine is a 113-dalton breakdown product of creatine and phosphocreatine, found predominantly in skeletal muscle. It is freely filtered at the glomerulus and undergoes active tubular secretion via organic cation transporters (OCT2) in the proximal tubule. This tubular secretion typically accounts for 10–40% of urinary creatinine excretion at normal GFR, increasing proportionally as GFR declines.
Evolution of Creatinine-Based Equations
The 2021 CKD-EPI creatinine equation removes the race coefficient and was developed using diverse populations. This race-free equation is now recommended by KDIGO 2024 and the NKF-ASN Task Force for routine clinical use.
Table 1: Conditions Causing eGFRcr to Overestimate True GFR
| Condition | Mechanism | Clinical Significance |
|---|---|---|
| Sarcopenia / Low muscle mass | Decreased creatinine generation | Common in elderly, chronic illness; may mask CKD |
| Cachexia / Wasting diseases | Severe reduction in muscle mass | Cancer, HF, COPD; significantly overestimates GFR |
| Amputations | Loss of muscle mass | Consider cystatin C for more accurate assessment |
| Paraplegia / Quadriplegia | Muscle atrophy from disuse | May have very low creatinine despite significant CKD |
| Advanced cirrhosis | Reduced hepatic creatine synthesis; muscle wasting; increased tubular secretion | eGFRcr very unreliable; use cystatin C or measured GFR |
| Vegetarian / Low-protein diet | Reduced dietary creatine intake | May lower creatinine by 10–15% |
| Advanced CKD (GFR <15–20) | Increased tubular secretion (up to 50%); extrarenal clearance | eGFRcr overestimates true GFR in stage 5 CKD |
Table 2: Conditions Causing eGFRcr to Underestimate True GFR
| Condition | Mechanism | Clinical Significance |
|---|---|---|
| High muscle mass / Bodybuilders | Increased creatinine generation | May appear to have CKD when GFR is normal |
| High dietary meat intake | Exogenous creatinine from cooked meat | Acute increase in creatinine post-meal |
| Creatine supplements | Direct precursor to creatinine | Common in athletes; falsely elevates creatinine |
| Trimethoprim | Inhibits tubular creatinine secretion (OCT2) | Reversible increase of 0.2–0.5 mg/dL |
| Cobicistat / Ritonavir | Inhibit tubular creatinine secretion | HIV medications; increase Cr without affecting GFR |
| Dolutegravir | Inhibits OCT2-mediated tubular secretion | Increases creatinine by ~0.1–0.2 mg/dL |
| Diabetic ketoacidosis | Acetoacetate interferes with Jaffe reaction | Falsely elevated creatinine during DKA |
In patients taking medications that inhibit tubular creatinine secretion (trimethoprim, cobicistat, dolutegravir), the initial rise in serum creatinine does not indicate AKI. Establish a new baseline after 2–4 weeks of stable therapy.
Cystatin C-Based GFR Estimation
Cystatin C is a 13-kDa cysteine protease inhibitor produced at a relatively constant rate by all nucleated cells. Unlike creatinine, cystatin C production is independent of muscle mass, diet, and sex. It is freely filtered at the glomerulus and completely reabsorbed and catabolized in the proximal tubule, with no tubular secretion.
The combined eGFRcr-cys equation demonstrates the highest accuracy, with approximately 91–92% of estimates within 30% of measured GFR, compared to 87–88% for creatinine alone.
Table 3: Conditions Causing eGFRcys to Overestimate True GFR
| Condition | Mechanism | Significance |
|---|---|---|
| Hypothyroidism | Decreased cellular production of cystatin C | eGFRcys increases when hypothyroidism is treated |
| Some immunosuppressive therapy | May suppress cystatin C production | Variable effect; monitor with creatinine also |
Table 4: Conditions Causing eGFRcys to Underestimate True GFR
| Condition | Mechanism | Significance |
|---|---|---|
| Hyperthyroidism | Increased metabolic rate and cystatin C production | 62% of hyperthyroid patients have elevated cystatin C |
| Glucocorticoid therapy | Stimulates cystatin C gene expression | Dose-dependent; may persist with chronic use |
| Obesity (Class II–III) | Increased production from adipose tissue | Graded association with BMI; may overdiagnose CKD |
| Systemic inflammation / High CRP | Inflammatory cytokines stimulate production | Sepsis, autoimmune disease, critical illness |
| Malignancy | Increased cell turnover | Particularly hematologic malignancies |
| Heart failure | Elevated beyond GFR effects | Predicts mortality independent of eGFR |
| Diabetes mellitus | Mildly elevated independent of GFR | Inflammatory component |
| Smoking | Chronic low-grade inflammation | Associated with elevated cystatin C levels |
Cystatin C is particularly valuable when creatinine is unreliable: sarcopenia, cirrhosis, amputees, and extremes of muscle mass. However, avoid relying solely on cystatin C in patients receiving glucocorticoids, with uncontrolled hyperthyroidism, or with active malignancy.
The eGFRcys/eGFRcr Ratio: Clinical Interpretation
When both creatinine and cystatin C are measured, the ratio provides valuable diagnostic and prognostic information beyond either marker alone. In a systematic review of 1,489 studies, nearly 30% of participants had discordance between eGFRcys and eGFRcr.
Table 5: Interpretation of the eGFRcys/eGFRcr Ratio
| Ratio | Interpretation | Possible Causes | Clinical Action |
|---|---|---|---|
| >1.0–1.15 | eGFRcys higher than eGFRcr | High muscle mass, meat intake, creatine supplements, trimethoprim, hypothyroidism | eGFRcys may be more accurate; evaluate for confounders |
| 0.85–1.0 | Normal concordance | Neither marker disproportionately affected | Either eGFR likely reliable; eGFRcr-cys for best accuracy |
| 0.70–0.84 | Mild discordance | Mild sarcopenia, low-dose glucocorticoids, subclinical inflammation | Assess for sarcopenia; consider eGFRcys or eGFRcr-cys |
| <0.70 | Marked discordance | Sarcopenia, glucocorticoids, hyperthyroidism, shrunken pore syndrome, malignancy | Evaluate for confounders; if absent, consider SPS; increased mortality risk |
| <0.60 | Severe discordance | Shrunken pore syndrome most likely if confounders excluded | High mortality risk independent of GFR; close monitoring warranted |
Shrunken Pore Syndrome (SPS)
First described in 2015, SPS is defined by an eGFRcys/eGFRcr ratio below 0.60–0.70 in the absence of known confounders. It reflects selective impairment of filtration of molecules in the 10–30 kDa range while preserving filtration of smaller molecules. SPS has been associated with markedly increased mortality risk, with hazard ratios of 3.0–7.3.
A low eGFRcys/eGFRcr ratio (<0.7) is not merely marker discordance but independently predicts mortality, even in patients with normal measured GFR. After excluding sarcopenia and glucocorticoid use, consider shrunken pore syndrome and heightened cardiovascular monitoring.
Measured GFR Validation: Stockholm Iohexol Cohort
The Stockholm Iohexol Cohort (Fu et al., JASN 2023) included 6,185 adults with 9,404 concurrent measurements of creatinine, cystatin C, and iohexol plasma clearance, including subgroups with cardiovascular disease, heart failure, diabetes, liver disease, and cancer.
Key Finding: When Markers Disagree, Creatinine Has the Larger Error
In 47% of samples where eGFRcys was >20% lower than eGFRcr, creatinine's overestimation (+15.0 mL/min/1.73m²) was nearly twice the magnitude of cystatin C's underestimation (−8.5 mL/min/1.73m²).
Table 6: Bias by Discordance Pattern (Stockholm Iohexol Cohort)
| Pattern | Prevalence | eGFRcr Bias | eGFRcys Bias | eGFRcr-cys Bias | Most Accurate |
|---|---|---|---|---|---|
| eGFRcys < eGFRcr by >20% | 47% | +15.0 (overestimate) | −8.5 (underestimate) | +0.8 | Combined |
| eGFRcys ≈ eGFRcr (within 20%) | 45% | +3.0 | −2.0 | +0.5 | All similar |
| eGFRcys > eGFRcr by >20% | 8% | −4.5 (underestimate) | +8.4 (overestimate) | +1.4 | Combined |
Table 8: Performance Metrics by Discordance Pattern
| Metric | eGFRcys < eGFRcr (47%) | eGFRcys ≈ eGFRcr (45%) | eGFRcys > eGFRcr (8%) |
|---|---|---|---|
| Median mGFR | 47 | 71 | 74 |
| P30 for eGFRcr | 50% | 88% | 70% |
| P30 for eGFRcys | 73% | 89% | 66% |
| P30 for eGFRcr-cys | 84% | 93% | 84% |
| Correct CKD Stage — eGFRcr | 38% | 68% | 61% |
| Correct CKD Stage — eGFRcr-cys | 62% | 74% | 72% |
When eGFRcys is substantially lower than eGFRcr in a cardiorenal patient, do not assume cystatin C is wrong. mGFR-validated studies show creatinine typically overestimates true GFR by ~15 mL/min while cystatin C underestimates by only ~8.5 mL/min. Use eGFRcr-cys for the most accurate estimate. The patient likely has more kidney disease than creatinine suggests.
Relying on eGFRcr alone in patients with heart failure, diabetes, cancer, or sarcopenia may substantially overestimate kidney function. This could delay recognition of CKD progression, result in inappropriate drug dosing, or miss eligibility for renal-protective therapies. Always obtain cystatin C and calculate eGFRcr-cys in these populations.
Special Populations
Elderly Patients
Age-related sarcopenia reduces creatinine generation, causing eGFRcr to overestimate true GFR. The combined eGFRcr-cys equation is preferred for elderly patients when accuracy is important.
Cirrhosis
Reduced hepatic creatine synthesis, muscle wasting, and increased tubular creatinine secretion all cause eGFRcr to markedly overestimate true GFR. Measured GFR using iohexol clearance is recommended for liver transplant evaluation.
Kidney Transplant Recipients
Altered creatinine metabolism from immunosuppressants and reduced muscle mass from prior illness. Cystatin C may be affected by glucocorticoid therapy. The combined equation is generally preferred.
Acute Kidney Injury
Neither creatinine nor cystatin C equations are validated for GFR estimation in AKI. Both markers lag behind true GFR changes. Interpret absolute marker values and trends rather than calculated eGFR.
Advanced CKD (eGFR <20)
In patients with eGFR <20 mL/min/1.73m², creatinine-based equations may overestimate true GFR by 20–30%. When accurate GFR is needed for critical decisions such as dialysis initiation timing or living donor evaluation, use eGFRcr-cys or measured GFR.
Clinical Approach to GFR Assessment
- Use eGFRcr as initial screening in most patients — appropriate for routine monitoring in stable patients without known confounders
- Add cystatin C (eGFRcr-cys) when: eGFRcr is near decision thresholds (60, 45, 30, 15); concerns about creatinine reliability; clinical findings discordant with eGFRcr; accurate GFR needed for drug dosing
- Calculate and interpret the eGFRcys/eGFRcr ratio when both markers are available. Ratio <0.7 warrants investigation
- Consider measured GFR for living kidney donor evaluation, when both eGFRcr and eGFRcys are unreliable, or when highest accuracy is required
- Always interpret eGFR in clinical context — 10–20% of patients will have significant discrepancies from true GFR
Cystatin C: Cost and Coverage
- Medicare reimburses cystatin C (CPT 82610) at approximately $18.52 vs. $5.12–10.56 for creatinine
- Coverage requires documentation of medical necessity
- To improve coverage, document why creatinine-based eGFR may be inaccurate and how accurate GFR will affect management
References
- Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. PubMed
- Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function. N Engl J Med. 2006;354(23):2473–2483. PubMed
- Inker LA, Eneanya ND, Coresh J, et al. New creatinine- and cystatin C-based equations to estimate GFR without race. N Engl J Med. 2021;385(19):1737–1749. PubMed
- Delgado C, Baweja M, Crews DC, et al. A unifying approach for GFR estimation: recommendations of the NKF-ASN Task Force. J Am Soc Nephrol. 2021;32(12):2994–3015. PubMed
- Levey AS, Coresh J, Tighiouart H, Greene T, Inker LA. Measured and estimated GFR. Nat Rev Nephrol. 2020;16(1):51–64. PubMed
- Shlipak MG, Matsushita K, Arnlov J, et al. Cystatin C versus creatinine in determining risk. N Engl J Med. 2013;369(10):932–943. PubMed
- KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of CKD. Kidney Int. 2024;105(4S):S117–S314. DOI
- Grubb A, et al. Shrunken pore syndrome. Scand J Clin Lab Invest. 2015;75(4):333–340. PubMed Search
- Fu EL, Coresh J, Engstrom G, et al. Accuracy of GFR estimating equations in patients with discordances. J Am Soc Nephrol. 2023;34(9):1574–1586. PubMed Search
- Grams ME, Estrella MM, Coresh J, et al. Discordance in eGFR and clinical outcomes: a meta-analysis. JAMA. 2025. PubMed Search