What are the sensitivity and specificity of estimated glomerular filtration rate (eGFR) for detecting chronic kidney disease (CKD) stage 3 or higher (eGFR < 60 mL/min/1.73 m²)?

Sensitivity and Specificity of eGFR for Detecting CKD Stage 3 or Higher

eGFR is not a diagnostic test with traditional sensitivity and specificity values; rather, it is an estimation method that approximates measured GFR with varying degrees of accuracy depending on the equation used and clinical context. The combined creatinine-cystatin C equation (eGFRcr-cys) achieves 94.9% accuracy within 30% of measured GFR, making it the most reliable estimation method available 1.

Understanding eGFR as an Estimation Tool, Not a Diagnostic Test

eGFR functions fundamentally differently from traditional diagnostic tests:

• eGFR estimates kidney function rather than detecting disease presence or absence, making conventional sensitivity/specificity calculations inappropriate 2
• The accuracy of eGFR is measured by how closely it approximates measured GFR (the gold standard), typically reported as the percentage of estimates within 30% of measured GFR (P30 accuracy) 3
• Creatinine-based eGFR (CKD-EPI) demonstrates 84.1% accuracy (P30) in validation studies, compared to 80.6% for the older MDRD equation 3

Accuracy Metrics for Different eGFR Equations

The performance of eGFR varies substantially by equation type:

• The combined creatinine-cystatin C equation provides the highest accuracy at 94.9% within 30% of measured GFR and should be the definitive value for clinical decision-making when discordance exists between methods 1
• Creatinine-based eGFR alone shows only 50% accuracy in discordant cases, while cystatin C-based eGFR shows 73% accuracy 4
• In patients with creatinine-based eGFR 45-59 mL/min/1.73 m², approximately 23% actually have normal kidney function when confirmed with cystatin C, highlighting the risk of false-positive CKD diagnoses 1

Critical Limitations Affecting eGFR Accuracy

Several factors significantly compromise eGFR reliability:

• eGFR accuracy is suboptimal in patients with normal or near-normal renal function, which is why values >60 mL/min/1.73 m² should be reported as ">60" rather than precise figures 5
• Creatinine-based equations are confounded by muscle mass, age, sex, diet, and race, with interlaboratory variation as high as 20% 5
• The reference interval for serum creatinine includes up to 25% of people (particularly thin, elderly women) who have significantly reduced eGFR <60 mL/min/1.73 m², demonstrating poor sensitivity of creatinine alone 5

Diagnostic Threshold Performance

The <60 mL/min/1.73 m² threshold for CKD Stage 3 has specific performance characteristics:

• An eGFR <60 mL/min/1.73 m² alone is sufficient to diagnose CKD Stages III-V without requiring additional markers of kidney damage 6
• However, healthy adults can have eGFR values >63.5 mL/min/1.73 m², meaning normal values overlap with CKD Stages 1 and 2, so an eGFR >60 mL/min/1.73 m² does not exclude kidney disease 7
• The CKD-EPI equation shows less bias (median difference 2.5 vs. 5.5 mL/min/1.73 m²) and improved precision (IQR 16.6 vs. 18.3 mL/min/1.73 m²) compared to MDRD 3

Recommended Confirmation Strategy for Borderline Cases

For patients with creatinine-based eGFR 45-59 mL/min/1.73 m² who lack other markers of kidney damage, measure cystatin C and calculate the combined equation (eGFRcr-cys) 1, 8:

• If eGFRcr-cys is ≥60 mL/min/1.73 m², CKD diagnosis is not confirmed 1
• If eGFRcr-cys is <60 mL/min/1.73 m², CKD diagnosis is confirmed 1
• This approach prevents misclassification in the 23% of patients who would be false positives with creatinine alone 1

Common Pitfalls to Avoid

• Do not rely solely on creatinine-based eGFR in patients with variable muscle mass, malnutrition, extremes of body composition, eating disorders, extreme exercise, amputations, or spinal cord injuries 8
• Do not interpret eGFR values just below 60 mL/min/1.73 m² without considering clinical context and markers of kidney damage 2
• Recognize that eGFR lags behind acute changes in kidney function and cannot identify early GFR changes, unlike measured GFR 2
• Consider measured GFR when eGFRcr-cys is thought to be inaccurate, though this is more expensive and time-consuming 2