How reliable is creatinine (Cr) estimated glomerular filtration rate (eGFR) in assessing kidney function?

Reliability of Creatinine-Based eGFR

Creatinine-based eGFR (eGFRcr) is reasonably reliable for initial kidney function assessment in most adults without conditions affecting creatinine levels independent of GFR, but it has important limitations including lack of sufficient accuracy and precision, lag time behind actual GFR changes, and susceptibility to non-GFR determinant confounding. 1

When eGFRcr is Reliable

For the majority of adults without specific confounding conditions, eGFRcr using validated equations (such as the 2021 race-free CKD-EPI equation) provides a reasonable initial assessment of kidney function. 1

• The 2025 KDOQI/KDIGO guidelines recommend using serum creatinine and an estimating equation as the initial test for GFR assessment 1
• eGFRcr is inexpensive, easy to implement, widely available, and easily repeatable compared to measured GFR 1
• When properly calibrated to international reference standards with enzymatic assays (coefficient of variation <2.3%, bias <3.7%), eGFRcr provides adequate screening for most clinical purposes 1

Critical Limitations and Sources of Error

eGFRcr is fundamentally "not sufficiently accurate and precise" compared to measured GFR and is subject to multiple non-GFR determinants that can significantly compromise reliability. 1

Physiologic Factors Affecting Creatinine Generation

• Muscle mass extremes: Very low or very high muscle mass alters creatinine generation independent of kidney function 1
• Dietary intake: Creatinine levels require consideration of dietary protein and meat intake 1
• Age-related decline: Well-described age relationship affects eGFRcr interpretation 2
• Body composition: Obesity class III (BMI >40 kg/m²) reduces eGFRcr accuracy 1

Tubular Secretion Issues

Creatinine undergoes tubular secretion that progressively increases as kidney function declines, causing systematic overestimation of true GFR. 3

• Fractional creatinine secretion varies inversely with actual GFR, averaging 1.64 times the inulin clearance in glomerulopathic patients 3
• This tubular secretion can be blocked by cimetidine, confirming it is a major source of error 3
• As kidney disease worsens, remnant tubules hypersecrete creatinine, resulting in "gross and unpredictable overestimates of GFR" 3

Clinical States Rendering eGFRcr Unreliable

The following conditions specifically compromise eGFRcr accuracy and warrant alternative assessment: 1

• Advanced cirrhosis or cancer with high cell turnover (high catabolism/inflammation) 1
• Severe malnutrition or muscle wasting conditions 1
• Acute illness states where creatinine is not in steady-state 4
• Extremes of body size or composition 5

Non-Steady-State Creatinine

At emergency departments and during acute illness, 28-55% of patients show significant non-steady-state creatinine changes, invalidating eGFRcr calculations that assume constant urinary creatinine excretion. 4

• Changes exceeding 15% decrease or 18% increase between measurements indicate non-steady-state 4
• eGFR should be interpreted with "great caution" in acute care settings 4
• eGFRcr "lags behind changes in GFR" and cannot identify early GFR changes 1

Measurement Uncertainty

At the critical decision threshold of 60 mL/min/1.73 m², the uncertainty of eGFRcr is ±11 mL/min/1.73 m², meaning values between 49-71 mL/min/1.73 m² require cautious interpretation. 2

• This uncertainty stems from both analytical error in creatinine measurement and biological variation 2
• Normal healthy adults can have eGFRcr values as low as 63.5 mL/min/1.73 m², overlapping with CKD stages 1-2 6
• An eGFR >60 mL/min/1.73 m² does not exclude kidney disease 6

When to Use Alternative Methods

When eGFRcr is expected to be inaccurate and GFR affects clinical decision-making, measure cystatin C and calculate eGFRcr-cys (combined creatinine-cystatin C equation). 1

Advantages of Combined eGFRcr-cys

• Improved precision (interquartile range 13.4 vs 15.4 mL/min/1.73 m² for creatinine alone) 7
• Better accuracy (8.5% vs 12.8% of estimates >30% different from measured GFR) 7
• Correctly reclassifies 16.9% of patients with eGFRcr 45-59 mL/min/1.73 m² as having GFR ≥60 mL/min/1.73 m² 7
• Coefficients of variation <10% and bias <0.9% across manufacturers 1

When Even eGFRcr-cys is Insufficient

For critical clinical decisions requiring precise GFR (such as kidney-cleared chemotherapy dosing), measure GFR using plasma or urinary clearance of exogenous filtration markers. 1

• Measured GFR is "accurate for GFR in all situations and across the GFR range" 1
• Consider measured creatinine clearance via timed urine collection if measured GFR unavailable and eGFRcr-cys thought inaccurate 1
• Measured GFR is less influenced by non-GFR determinants and can identify early GFR changes 1

Common Pitfalls to Avoid

• Do not rely on single eGFRcr values near decision thresholds (especially 60 mL/min/1.73 m²) without considering measurement uncertainty 2
• Do not use eGFRcr in acute care settings without verifying steady-state creatinine 4
• Do not ignore clinical context suggesting altered creatinine generation (extremes of muscle mass, dietary intake, catabolic states) 1
• Do not use race-based equations; use the 2021 race-free CKD-EPI equations 1
• Do not assume eGFRcr accuracy when assessing GFR changes over time without considering potential for error 1