Factors Affecting Creatinine, eGFR, and Cystatin C
Multiple demographic, clinical, and physiological factors significantly influence serum creatinine, creatinine-based eGFR, and cystatin C levels beyond actual kidney function, requiring careful interpretation when assessing GFR.
Factors Affecting Serum Creatinine and Creatinine-Based eGFR
Demographic and Body Composition Factors
- Age: Creatinine production decreases with age due to declining muscle mass 1
- Sex: Males generate more creatinine due to higher muscle mass 1
- Muscle mass: Lower muscle mass artificially elevates eGFRcr above true GFR 2, 3
- Weight and body habitus: Obesity and higher body fat affect creatinine generation 4, 5
Clinical and Pathological Factors
- Diet: Meat consumption can transiently increase creatinine levels 2
- Glomerular disease activity: Disease flares and proteinuria significantly alter creatinine generation 5
- Edema and fluid status: Can dilute or concentrate creatinine measurements 5
Medication Effects
- Corticosteroids: Increase muscle catabolism, affecting creatinine production 5
- Immunosuppressants: Alter creatinine generation in glomerular disease patients 5
Critical pitfall: Creatinine-based equations systematically overestimate GFR in patients with low muscle mass, neurogenic bladder disorders, and certain glomerular diseases 5, 6. The median overestimation can be 13-17 mL/min/1.73 m² in vulnerable populations 6.
Factors Affecting Cystatin C and Cystatin C-Based eGFR
Demographic Factors
- Age: Older age increases cystatin C levels 4
- Sex: Males have higher cystatin C levels 4
- Ethnicity: South Asian ethnicity associated with lower eGFRcys 4
Body Composition
- Obesity: Higher waist circumference and body fat artificially lower eGFRcys 2, 4, 7
- Weight status: Significant predictor of cystatin C variability 4
Inflammatory and Endocrine Conditions
- Thyroid dysfunction: Strongly affects cystatin C production 2, 4
- Chronic inflammatory diseases: Increase cystatin C production 2, 4, 7
- C-reactive protein elevation: Associated with higher cystatin C along non-GFR pathways 4, 7
- Malignancy: Can alter cystatin C levels 2
Lifestyle Factors
- Smoking: Current smokers have significantly different cystatin C levels (strong predictor of lower eGFRcys) 4
Medication Effects
Metabolic Factors
- Hypertension: Associated with cystatin C changes along non-GFR pathways 7
- Diabetes: Affects cystatin C through non-GFR mechanisms 7
- Hypertriglyceridemia: Influences cystatin C levels 7
- Hyperuricemia: Associated with cystatin C alterations 7
Kidney-Specific Factors
- Proteinuria/albuminuria: Particularly urinary albumin-creatinine ratio >300 mg/g strongly predicts lower eGFRcys 4, 7
- Shrunken pore syndrome: Decreased cystatin C clearance relative to creatinine clearance 3
Understanding Creatinine-Cystatin C Discordance
The difference between eGFRcys and eGFRcr (eGFRdiff) is often large—occurring in 25-30% of patients—and carries prognostic significance 3, 4. However, the pathophysiology remains poorly understood 3.
Key Clinical Implications:
When eGFRcys is >15 mL/min/1.73 m² lower than eGFRcr (25% of patients), suspect:
- Low muscle mass (creatinine underproduction)
- Obesity, inflammation, thyroid dysfunction, steroid use (cystatin C overproduction)
- South Asian ethnicity, current smoking, chronic inflammatory disease 4
When eGFRcr is >15 mL/min/1.73 m² lower than eGFRcys (5% of patients), the predictors are largely inverse 4.
Measurement Variability
Cystatin C shows higher biological variability (coefficient of variation 10.7%) compared to creatinine (6.4%) 7, meaning cystatin C measurements fluctuate more over time even when true GFR is stable.
Practical Clinical Approach
According to the 2022 NKF-ASN Task Force guidelines, when cystatin C is available, use the combined creatinine-cystatin C equation (CKD-EPIcr-cys) 1. This approach:
- Minimizes bias (0.63 mL/min/1.73 m² vs -4.95 for creatinine alone and +5.01 for cystatin C alone) 8
- Achieves superior accuracy (P30 of 84-88%) 8, 6
- Partially compensates for the non-GFR determinants of each biomarker 7
When interpreting discordant values, models incorporating all known factors explain only a minority of the variation 3. Until better understanding emerges, favor the combined equation estimate rather than attempting to determine which single biomarker is more accurate 3.
Special populations requiring caution:
- Neurogenic bladder: Creatinine equations overestimate by 13-17 mL/min/1.73 m²; prefer cystatin C-based equations 6
- Glomerular disease: Creatinine generation is highly dynamic with disease activity; combined equations recommended 5
- Extremes of muscle mass: Combined equations perform better than either alone 2, 4