Estimating Glomerular Filtration Rate

The normal serum creatinine reference interval does not necessarily reflect a normal GFR for a patient. Because mild and moderate kidney injury is poorly inferred from serum creatinine alone, NKDEP strongly encourages clinical laboratories to routinely estimate glomerular filtration rate (GFR) and report the value when serum creatinine is measured for patients 18 and older, when appropriate and feasible. An estimated GFR (eGFR) calculated from serum creatinine using an isotope dilution mass spectrometry (IDMS) traceable equation is a simple and effective way in which laboratories can help health care providers detect CKD among those with risk factors—diabetes, hypertension, cardiovascular disease, or family history of kidney disease. Assessment of kidney function through eGFR is essential once albuminuria is discovered. Providers also may use eGFR to monitor patients already diagnosed with CKD.

IDMS Traceable Equations

To reduce interlaboratory variation in creatinine assay calibration and enable more accurate eGFR results, all major manufacturers have calibrated their serum creatinine measurement procedures to be traceable to IDMS. Because creatinine results that are calibrated to IDMS may differ by 5 to 30% compared to uncalibrated results,1 use of a non-IDMS traceable equation with IDMS calibrated results will yield an inaccurate eGFR. Therefore, all laboratories should use an IDMS traceable equation when estimating and reporting GFR.

Read more about creatinine standardization.

Selecting an Equation

The Modification of Diet in Renal Disease (MDRD) Study equation and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation are the most widely used IDMS traceable equations for estimating GFR in patients age 18 and over. For estimating GFR from serum creatinine in patients under age 18 (including infants, toddlers, children, and teens), the Bedside Schwartz equation should be used.

Both the MDRD Study and CKD-EPI equations include variables for age, gender, and race, which may allow providers to observe that CKD is present despite a serum creatinine concentration that appears to fall within or just above the normal reference interval. Direct comparison of the MDRD and CKD-EPI equations to other equations such as Cockcroft-Gault2, 3 and to creatinine clearance measured from 24-hour urine collections has demonstrated this superiority.4

Note that creatinine clearance should be considered for assessing kidney function when the patient's basal creatinine production is very abnormal. This may be the case with patients of extreme body size or muscle mass (e.g., obese, severely malnourished, amputees, paraplegics, or other muscle-wasting diseases), or with unusual dietary intake (e.g., vegetarian, creatine supplements).

The MDRD Equation

The following is the IDMS-traceable MDRD Study equation (for creatinine methods calibrated to an IDMS reference method)

GFR (mL/min/1.73 m2) = 175 × (Scr)-1.154 × (Age)-0.203 × (0.742 if female) × (1.212 if African American)

The equation does not require weight or height variables because the results are reported normalized to 1.73 m2 body surface area, which is an accepted average adult surface area.

The equation has been validated extensively in Caucasian and African American populations between the ages of 18 and 70* with impaired kidney function (eGFR < 60 mL/min/1.73 m2) and has shown good performance for patients with all common causes of kidney disease.2

*The equation has not been validated in patients older than 70, but an MDRD-derived eGFR may still be a useful tool for providers caring for patients older than 70.

The CKD-EPI Equation

The CKD-EPI equation uses a 2-slope "spline" to model the relationship between GFR and serum creatinine, age, sex, and race. The equation is given in the following table for creatinine in mg/dL (see Appendix for creatinine in µmol/L). The equation can be expressed in a single equation (see table legend) or as a series of equations for different race, sex, and creatinine conditions (see table rows).

Table 1: CKD EPI Equation for Estimating GFR Expressed for Specified Race, Sex and Serum Creatinine in mg/dL (From Ann Intern Med 2009;150:604–612, used with permission)

Race Sex Serum Creatinine,
Scr (mg/dL)
Equation (age in years for ≥ 18)
Black Female ≤ 0.7 GFR = 166 × (Scr/0.7)-0.329 × (0.993)Age
Black Female > 0.7 GFR = 166 × (Scr/0.7)-1.209 × (0.993)Age
Black Male ≤ 0.9 GFR = 163 × (Scr/0.9)-0.411 × (0.993)Age
Black Male > 0.9 GFR = 163 × (Scr/0.9)-1.209 × (0.993)Age
White or other Female ≤ 0.7 GFR = 144 × (Scr/0.7)-0.329 × (0.993)Age
White or other Female > 0.7 GFR = 144 × (Scr/0.7)-1.209 × (0.993)Age
White or other Male ≤ 0.9 GFR = 141 × (Scr/0.9)-0.411 × (0.993)Age
White or other Male > 0.9 GFR = 141 × (Scr/0.9)-1.209 × (0.993)Age

CKD-EPI equation expressed as a single equation:

GFR = 141 × min (Scr /κ, 1)α × max(Scr /κ, 1)-1.209 × 0.993Age × 1.018 [if female] × 1.159 [if black]
Scr is serum creatinine in mg/dL,
κ is 0.7 for females and 0.9 for males,
α is -0.329 for females and -0.411 for males,
min indicates the minimum of Scr /κ or 1, and
max indicates the maximum of Scr /κ or 1.

A laboratory that reports eGFR numeric values > 60 mL/min/1.73 m2 should use the CKD-EPI equation, because the CKD-EPI equation is more accurate for values > 60 mL/min/1.73 m2 than is the MDRD Study equation. However, the influence of imprecision of creatinine assays on the uncertainty of an eGFR value is greater at higher eGFR values and should be considered when determining the highest eGFR value to report.

MDRD and CKD-EPI Equation Performance

As shown in the figure below, the CKD-EPI equation and the MDRD Study equation were equally accurate in a subgroup with estimated GFR (eGFR) less than 60 mL/min/1.73 m2. However, the CKD-EPI equation was more accurate in a subgroup with eGFR between 60 and 120 mL/min/1.73 m2. The receiver operator curves (ROC) for detecting GFR categories less than 90, 75, 60, 45, 30 and 15 mL/min per 1.73 m2 did not differ between the CKD-EPI and MDRD Study equations.1, 2

A graphic of a speedometer-like dial that depicts GFR results of 0 to 15 as kidney failure, 15 to 60 as kidney disease, and 60 to 120 as normal.

Figure 1. Accuracy of the CKD-EPI and MDRD equations to estimate GFR for the validation data set (N=3896). Both panels show the difference between measured and estimated (y-axis) vs. estimated GFR (x-axis). A smoothed regression line is shown with the 95% CI for the distribution of results, using quantile regression, excluding the lowest and highest 2.5% of estimated GFR. From Ann Intern Med 2009;150:604-612, used with permission.

Reduce Rounding Errors

NKDEP recommends using serum creatinine values in mg/dL to two decimal places (e.g., 0.95 mg/dL) OR values in µmol/L to the nearest whole number (e.g., 84 µmol/L) when calculating eGFR using the MDRD Study or CKD-EPI equation. This practice will reduce rounding errors that may contribute to imprecision in the eGFR value.

When Not to Use Creatinine-based Estimating Equations

Creatinine-based estimating equations may not be suitable for all populations. Creatinine-based estimates of kidney function are only useful when renal function is stable; serum creatinine values obtained while kidney function is changing will not provide accurate estimates of kidney function.

Creatinine-based estimating equations are not recommended for use with:

  • Individuals with unstable creatinine concentrations. This includes pregnant women; patients with serious co-morbid conditions; and hospitalized patients, particularly those with acute renal failure. Creatinine-based estimating equations should be used only for patients with stable creatinine concentrations.
  • Persons with extremes in muscle mass and diet. This includes, but is not limited to, individuals who are amputees, paraplegics, bodybuilders, or obese; patients who have a muscle-wasting disease or a neuromuscular disorder; and those suffering from malnutrition, eating a vegetarian or low-meat diet, or taking creatine dietary supplements.

Application of the equation to these patient groups may lead to errors in GFR estimation.5 GFR estimating equations have poorer agreement with measured GFR for ill hospitalized patients6 than for community-dwelling patients.

As noted above, providers should exercise judgment regarding clinical status when presented with an MDRD Study- or CKD-EPI-derived eGFR for a patient with an unstable creatinine level or other condition for which the equation is not suitable. Providers may not understand that estimating equations like the MDRD and CKD-EPI are derived from large populations of patients and provide the best estimate of mean GFR for a group of people of a certain age, race, gender, and serum creatinine value. Thus, the reported eGFR is the best estimate of a patient's GFR; it is not the patient's actual GFR.

Limitations of the CKD-EPI and MDRD Equations

  • Limitations using creatinine as a filtration marker: both the MDRD study and CKD-EPI equations are based on serum creatinine. Despite modest reduction in bias with the CKD-EPI equation, estimates remain imprecise, with some people showing large differences between the measured and estimated GFR. Like all other creatinine-based estimation equations, they suffer from physiologic limitations of creatinine as a filtration marker.4, 7 The terms for age, sex, and race in both equations only capture some of the non-GFR determinants of creatinine concentration in blood plasma, and the coefficients represent average effects observed in the population used to develop the equations.

    All estimates of GFR based on serum creatinine will be less accurate for patients at the extremes of muscle mass (including frail elderly, critically ill, or cancer patients), those with unusual diets, and those with conditions associated with reduced secretion or extra-renal elimination of creatinine. Confirmatory tests with exogenous measured GFR or measured creatinine clearance should be performed for people in whom estimates based on serum/plasma/blood creatinine alone may be inaccurate.
  • Populations not well represented in the development or validation cohorts: Elderly people and blacks with higher levels of GFR, racial and ethnic minorities other than blacks.
  • The influence of creatinine measurement imprecision at low creatinine concentrations (high eGFR) has not been carefully studied but has likely contributed to the variability at higher eGFR values.


Table 2: CKD EPI Equation for Estimating GFR Expressed for Specified Race, Sex and Serum Creatinine in µmol/L (Adapted from Ann Intern Med 2009;150:604-612, used with permission)

Race Sex Serum
Scr µmol/L
Equation (age in years for ≥ 18)
Black Female ≤ 61.9 GFR = 166 × (Scr/61.9)-0.329 × (0.993)Age
Black Female > 61.9 GFR = 166 × (Scr/61.9)-1.209 × (0.993)Age
Black Male ≤ 79.6 GFR = 163 × (Scr/79.6)-0.411 × (0.993)Age
Black Male > 79.6 GFR = 163 × (Scr/79.6)-1.209 × (0.993)Age
White or other Female ≤ 61.9 GFR = 144 × (Scr/61.9)-0.329 × (0.993)Age
White or other Female > 61.9 GFR = 144 × (Scr/61.9)-1.209 × (0.993)Age
White or other Male ≤ 79.6 GFR = 141 × (Scr/79.6)-0.411 × (0.993)Age
White or other Male > 79.6 GFR = 141 × (Scr/79.6)-1.209 × (0.993)Age

GFR = 141 × min (Scr /κ, 1)α × max(Scr /κ, 1)-1.209 × 0.993Age × 1.018 [if female] × 1.159 [if black]
Scr is serum creatinine in µmol/L,
κ is 61.9 for females and 79.6 for males,
α is -0.329 for females and -0.411 for males,
min indicates the minimum of Scr /κ or 1,
and max indicates the maximum of Scr /κ or 1.