Kidney Function Assessment
by Creatinine-Based
Estimation Equations

Martin E. Lascano

Emilio D. Poggio

Print this Content

 

Chronic kidney disease (CKD) has been recognized as a common health care problem, and therefore the National Kidney Foundation (NKF) has published the Kidney Disease Outcomes Quality Initiative (KDOQI) Practice Guidelines for the early detection, evaluation, diagnosis, and treatment of this condition (http://www.kidney.org/professionals/kdoqi/guidelines.cfm). As recommended by this initiative, the degree of kidney dysfunction, as assessed by the estimation of the glomerular filtration rate (GFR), is essential for the diagnosis, classification, and staging of CKD; see elsewhere in this section (“Chronic Kidney Disease”). Among the various alternatives available for the assessment of GFR, the NKF, and the National Kidney Disease Education Program (NKDEP) recommend that renal function be estimated by creatinine-based GFR estimation equations, and these are the focus of this chapter.

Evaluation of glomerular filtration rate

Among the many physiologic roles of the renal system, GFR is considered the best indicator of overall kidney function and therefore its assessment has become an important clinical tool in the daily care of patients. GFR cannot be measured directly, but instead it can be assessed by the renal clearance of filtration markers.1 The total kidney GFR is the sum of the filtration rates of all single functional nephrons—that is, it is determined by the total number of functional nephrons. Because of its highly dynamic and adaptive nature, despite initial structural damage to the renal parenchyma (i.e., reduction in functional nephron number), an individual's total GFR may not proportionally decrease because of the compensatory features of the remaining renal units, enabling the kidneys to maintain kidney function temporarily despite the loss of functional tissue. Moreover, the GFR may also be affected in the absence of parenchymal renal disease because of hemodynamic, pharmacologic factors, or both. The clinical assessment of GFR can aid the clinician in measuring the degree of renal dysfunction, progression of established kidney disease, or both; however, it is not informative in the determination of the cause of kidney disease, making it imperative to interpret the GFR in the context of the clinical setting.

The GFR can be determined from the renal clearance of a marker that achieves stable plasma concentration, is inert, and is freely filtered by the glomeruli but not reabsorbed, secreted, or metabolized.1,2 Such an ideal endogenous marker does not exist. Serum creatinine, one of the clinically useful analytes (others are serum urea and, more recently, serum cystatin C), has long been used by clinicians as a marker of renal function. However, it is important to emphasize that often the isolated use of serum creatinine concentration may not reflect the actual degree of kidney function of a particular subject. This is because multiple factors affect the concentration of serum creatinine (Table 1) and that the inverse relation between serum creatinine and GFR is nonlinear, particularly when patients have near-normal renal function (Fig. 1). An alternative to this approach is the measurement of creatinine clearance. This determination does not require highly trained personnel or expensive assays and can be performed by standard laboratories. However, this approach is limited by the difficulties in obtaining accurate urine collections and its potential misinterpretation because of the large biologic variability of creatinine metabolism in various clinical settings, including the unpredictable level of creatinine secretion at different levels of GFR. This method, however, is widely available and familiar to the health care community. On the other hand, more exact methods of GFR measurement, such as the clearance of exogenous markers such as inulin or renally excreted isotopes, are expensive and usually not readily available. More novel serum measurements of analytes, such as cystatin C, are under investigation and are not yet fully validated in all clinical settings; therefore, this approach remains a research tool at present.

Table 1: Factors Affecting Serum Creatinine Concentration
Factor Effect on Serum Creatinine Level Comment
Demographics
Aging Decreased Caused by decline in muscle mass
Female sex Decreased Reduced muscle mass
Ethnicity *
African American Increased Higher average muscle mass in African Americans
Hispanic Decreased
Asian Decreased
Body Habitus
Muscular Increased Increased muscle mass
Muscle wasting, amputation, malnutrition Decreased Reduced muscle mass ± decreased protein intake
Obesity No change No change in muscle mass
Diet
Vegetarian Decreased Decrease in creatinine generation
Ingestion of cooked meats Increased Transient increase in creatinine generation
Medications, Endogenous Substances
Cimetidine, trimethoprim, probenecid, potassium-sparing diuretics Increased Reduced creatinine tubular secretion
Ketoacids, ascorbic acid, glucose, some cephalosporins Increased Interference with alkaline picrate assay (Jaffé reaction) for creatinine
Bilirubin, hemoglobin Decreased Interference with alkaline picrate assay (Jaffé reaction) for creatinine
Flucytosine, praline, hemoglobin Increased Interference with enzymatic assays for creatinine
Metamizole, methyldopa, ethamsylate Decreased Interference with enzymatic assays for creatinine

*Whites used as reference group.

Rapid estimation of GFR by using creatinine-based mathematical equations is an attractive alternative to the clinician. These models rely on the inverse relation of serum creatinine with GFR, along with adjustment factors for measurable determinants of serum creatinine concentration (e.g., age, sex, body size, race).

Back to Top

Creatinine-based estimation equations

Numerous kidney function estimation equations, used to estimate creatinine clearance or GFR, have been published over the past few decades. However, the most commonly used formulas, the Cockcroft-Gault formula and the modification of diet in renal disease (MDRD) study equation, have been recommended by the KDOQI practice guidelines to be used for the estimation of GFR. These equations rely heavily on the reciprocal relation between serum creatinine and GFR but also incorporate demographic and anthropometric variables to complement the GFR estimation derived from the use of serum creatinine alone. It is important to emphasize that these mathematical models will reflect the clinical setting used to originate them.3

Cockcroft-Gault Formula

The Cockcroft-Gault formula was developed in 1976 with data from 249 men, primarily in an inpatient setting, with a wide range of renal function.4 It uses age, the inverse of serum creatinine, and lean body weight to estimate creatinine clearance in milliliters per minute (Box 1); it was not originally intended to be adjusted for body surface area (BSA). The inclusion of the weight factor is intended to adjust for muscle mass, a determinant of serum creatinine concentration. This implies that in clinical situations in which a change in weight is not the result of a similar change in muscle mass (e.g., edematous states, pregnancy, third spacing, overweight, obesity), the weight factor will adversely affect the performance of this formula. Because the original mathematical model was derived from data obtained predominantly in a male population, an arbitrary adjustment for female sex by a factor of 0.85 was incorporated. This equation has become popular because of its simple mathematical formulation and bedside applicability. It is important to note that this formula estimates creatinine clearance; this is known to overestimate GFR because of tubular secretion of creatinine, which is not adjustable.

Box 1: Creatinine-based Glomerular Filtration Rate (GFR) Estimation Equations
  • Cockcroft-Gault formula = ([140 − age] × weight)/(72 × SCr)*
  • Abbreviated MDRD equation = 186 × (SCr−1.154) × (age−0.203)
  • Re-expressed MDRD equation = 175 × (SCr−1.154) × (age−0.203)

Note: Weight in kilograms, age in years. *Multiply by 0.85 if female; expressed in mL/min. Multiply by 0.742 if female, by 1.212 if African American; expressed in mL/min/1.73 m2. To be used with assays that have been calibrated to be traceable to an isotope dilution mass spectrometry (IDMS) reference method (gold standard), as specified by the National Institute of Standards and Technology (NIST).
MDRD, modification of diet in renal disease; SCr, serum creatinine level, expressed in mg/dL.


Modification of Diet in Renal Disease Formula

The MDRD equation was developed in 1999 using data from 1628 patients with established chronic kidney disease.5 Whereas the Cockcroft-Gault formula estimates creatinine clearance, the set of equations developed from the data derived from the MDRD study are aimed at estimating GFR as measured by 125I-iothalamate urinary clearance, the reference method used by this study. The base population for these equations was outpatients with established CKD. Various MDRD equations have been published; however, the most widely used equation by the health care community is the abbreviated (four-variable) MDRD equation, which has been reformulated to be used with a standardized serum creatinine assay (see Box 1). It uses age, the inverse of serum creatinine, gender, and race (African American versus non–African American). In contrast to the Cockcroft-Gault formula, this model accounts for the biologic relation of creatinine metabolism observed in African Americans, but there is no adjustment for other ethnicities. This equation directly relates the accounted variables (e.g., serum creatinine, age, gender, race) to GFR adjusted for BSA—that is, the determinants of body size are prepackaged in the equation and thus additional adjustment is not required. A relative limitation of this equation is the need for a calculator. The NKF and the NKEDP now recommend using this equation, rather than the Cockcroft-Gault formula, to estimate kidney function.

Serum Creatinine Assay Calibration

Serum creatinine measurements are susceptible to calibration bias. This refers to a systematic absolute difference in measured serum creatinine concentrations throughout the whole range of creatinine values among laboratories because of variations in the assay calibration. This issue is critical in the application of estimation equations, especially at the normal levels of serum creatinine values.2,3,6 Data derived from the College of American Pathologists have suggested that there is a bias of up to 0.37 mg/dL in measured serum creatinine levels among U.S. laboratories. The importance of this difference is exemplified by the fact that a serum creatinine level of 0.8 mg/dL in one laboratory could represent a value of 1.2 mg/dL in a different one, with both falling within the “normal” range. Assuming that this sample belongs to a 60-year-old white woman, the estimated GFR could range from 78 to 49 mL/min/1.73 m2, clearly indicating the possibility of decreased kidney function.

To generalize the applicability of creatinine-based estimation equations among clinical laboratories, reference materials must be standardized, traceable to a gold standard from the National Institute of Standards and Technology (NIST). Clinical laboratories are expected to recalibrate their serum creatinine assays in the near future. The MDRD formula was re-expressed in 2005 to be used with serum creatinine measurements from clinical laboratories that have recalibrated their assays traceable to a standardized assay. This yields a value approximately 5% lower when compared with the measurements obtained by the original MDRD laboratory (see Box 1). Otherwise, until the transition to the recalibrated serum creatinine assay has been completed, the conventional abbreviated MDRD equation should be used.

Back to Top

Clinical applicability of creatinine-based estimation equations

One of the main limitations of the currently available GFR estimation equations is the lack of universality across the multiple clinical situations encountered by the clinician. However, growing evidence has suggested that the overall performance of the abbreviated MDRD equation is superior to the GFR estimates obtained by use of the Cockcroft-Gault formula, partially because the latter method estimates creatinine clearance and not GFR. Recognition of the limitations of these estimation equations is essential to using the information obtained from them properly.3 In general, the applicability of the MDRD equation is clinically satisfactory in settings that resemble the original population and methods used to develop the model, with expected poor performance in settings that deviate from the original. None of the available equations is applicable in cases of acute kidney injury, mostly because the serum creatinine level itself is not predictive of the acute changes in GFR.

Subjects with Established Chronic Kidney Disease

Several studies have reported on the performance of the Cockcroft-Gault and MDRD formulas in estimating GFR in subjects with established CKD and different levels of kidney function. Direct comparison between studies is challenging because of different methodologies used but, in general, both formulas perform better at lower GFR values (e.g., estimated GFR from 15 to 60 mL/min/1.73 m2), with the MDRD equation providing a lower bias (absolute or relative difference between estimated and measured GFR) and higher accuracy (percentage of estimated GFR within 30% of measured GFR) than the Cockcroft-Gault formula.7,8 Their performance is compromised as the level of GFR increases because of the caveats presented earlier (Fig. 2). Nevertheless, the abbreviated conventional MDRD equation performs better than the Cockcroft-Gault formula in subjects with known CKD, including those with diabetic nephropathy, and it is considered a reliable method to estimate GFR in this particular setting.

Subjects with Normal-Range Glomerular Filtration Rate

One particular area of interest is the validity of GFR estimation equations in subjects with or without CKD but with normal ranges of renal function. Several studies have reported on the performance of these equations in potential kidney donors (considered healthy) or in subjects with known or at high risk for CKD but with normal GFR levels. The two largest studies that analyzed the performance of estimation equations in healthy subjects have reported underestimation of GFR, which can vary anywhere from 5% to 29%, depending on methodologic issues related to the study.7,8 Results obtained from subjects with CKD or at risk for CKD (e.g., subjects with type 1diabetes mellitus, but no established CKD) but with a normal GFR range have provided results similar to those obtained from the healthy population. In this particular setting, current estimation equations are not precise and accurate enough to provide exact estimates of GFR, and can potentially misclassify patients as having a low GFR because of the trend for underestimation of renal function—hence, the importance of interpreting the obtained data in the context of the subject's clinical situation. For example, in the absence of risk factors for CKD such as hypertension or diabetes, absence of evidence of parenchymal renal disease (e.g., abnormal urine analysis results), and a single measurement of the serum creatinine level, an estimated GFR in the range of 60 mL/min/1.73 m2 should be viewed as a possible laboratory error. Because of this limitation of creatinine-based GFR estimation models, the NKEDP has recommended reporting specific values only if the estimated GFR is lower than 60 mL/min/1.73 m2; for higher values, “estimated GFR higher than 60 mL/min/1.73 m2” should be reported instead.

Subjects of Different Races and Ethnic Origins

The current MDRD equation incorporates African American race as a factor to account for the different creatinine metabolism in this population; hence, good performance is expected when applied to an African American population. A deficit of the current MDRD equation is the lack for adjustment, if needed, for Hispanic origin of the target subject. The Hispanic population is the fastest growing minority group in the United States, and is one of the largest populations in certain areas of the country. It is likely that biologic variations of creatinine metabolism, as well as different cultural and social habits (e.g., different diets), affect serum creatinine levels, thus requiring an adjustment. Correction factors may also be needed for other ethnic populations, such as Asians.

Renal Transplant Recipients

Various factors in kidney transplant patients may affect the metabolism of creatinine that will then translate into varying performance of estimation equations. Nevertheless, different estimation equations, including the Cockcroft-Gault and the MDRD formulas, have been used by different investigators for this purpose, along with the Nankivell equation, which was derived from mostly white renal transplant recipients who underwent 99Tc-diethylenetriaminepenta-acetic acid (DTPA) clearances.3 Several studies using different methodologies have compared the performance of these equations, with varying results. In one study, the MDRD formula was found to be superior to the Nankivell and Cockcroft-Gault formulas in an American population that included African American subjects and used the same GFR analytical reference method as the MDRD study. In contrast, other authors have reported better performance of the Nankivell formula over the MDRD equation in a Canadian population, in which GFR was measured by clearance of 99Tc-DTPA, the same method used to develop the Nankivell formula. In this study, serum cystatin C-based estimation equations were clearly superior to all creatinine-based models.

Other Considerations

Because of the steady aging of the population and the increase in illness severity of hospitalized patients, estimation of renal function is often needed for drug dosing or patient care in these settings. In older adults, the strength of the association between age and GFR may be overestimated by the Cockcroft-Gault formula; however, this varies among studies. It is not clear whether any of these formulas apply to older adults, but they are currently the best available alternative to assess kidney function quickly. In sick hospitalized patients, both the MDRD and the Cockcroft-Gault formulas significantly overestimate GFR and their poor performance is not clinically acceptable.9 By introducing other variables that may adjust for severity of illness, performance of the six-variable MDRD equation, which includes corrections for albumin and blood urea nitrogen levels, is partially improved, suggesting that future models should incorporate multiple surrogate markers of GFR to improve the estimation.

Back to Top

Summary

  • Estimation of glomerular filtration rate (GFR) to assess kidney function facilitates the detection, evaluation, and management of kidney disease.
  • Estimation equation results should be interpreted in the context of the clinical setting being applied.
  • An important aspect of creatinine-based GFR estimation equations is the recognition and acknowledgment of their limitations in each clinical setting.
  • The applicability of these estimation equations is satisfactory in settings that resemble the original population and methods used to develop the model, such as patients with established chronic kidney disease, and are an invaluable tool for the assessment of patients with kidney disease.

Back to Top

References

  1. Stevens LA, Levey AS. Measurement of kidney function. Med Clin North Am. 2005, 89: 457-473.
  2. Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function—measured and estimated glomerular filtration rate. N Engl J Med. 2006, 354: 2473-2483.
  3. Poggio ED, Hall PM. Estimation of glomerular filtration rate by creatinine-based formulas: Any room for improvement? NephSAP. 2006, 5: 131-140.
  4. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976, 16: 31-41.
  5. Levey AS, Bosch JP, Lewis JB, et al: A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999, 130: 461-470.
  6. Coresh J, Astor BC, McQuillan G, et al: Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate. Am J Kidney Dis. 2002, 39: 920-929.
  7. Poggio ED, Wang X, Greene T, et al: Performance of the modification of diet in renal disease and Cockcroft-Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol. 2005, 16: 459-466.
  8. Rule AD, Larson TS, Bergstralh EJ, et al: Using serum creatinine to estimate glomerular filtration rate: Accuracy in good health and in chronic kidney disease. Ann Intern Med. 2004, 141: 929-937.
  9. Poggio ED, Nef PC, Wang X, et al: Performance of the Cockcroft-Gault and modification of diet in renal disease equations in estimating GFR in ill hospitalized patients. Am J Kidney Dis. 2005, 46: 242-252.

Back to Top

Suggested Readings

  • Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976, 16: 31-41.
  • Coresh J, Astor BC, McQuillan G, et al: Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate. Am J Kidney Dis. 2002, 39: 920-929.
  • Levey AS, Bosch JP, Lewis JB, et al: A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999, 130: 461-470.
  • Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function—measured and estimated glomerular filtration rate. N Engl J Med. 2006, 354: 2473-2483.
  • Stevens LA, Levey AS. Measurement of kidney function. Med Clin North Am. 2005, 89: 457-473.
  • Poggio ED, Hall PM. Estimation of glomerular filtration rate by creatinine-based formulas: Any room for improvement? NephSAP. 2006, 5: 131-140.
  • Poggio ED, Nef PC, Wang X, et al: Performance of the Cockcroft-Gault and modification of diet in renal disease equations in estimating GFR in ill hospitalized patients. Am J Kidney Dis. 2005, 46: 242-252.
  • Poggio ED, Wang X, Greene T, et al: Performance of the modification of diet in renal disease and Cockcroft-Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol. 2005, 16: 459-466.
  • Rule AD, Larson TS, Bergstralh EJ, et al: Using serum creatinine to estimate glomerular filtration rate: Accuracy in good health and in chronic kidney disease. Ann Intern Med. 2004, 141: 929-937.