Frequently Asked Questions

Table of Contents

Explore common questions and answers to the NKF Workgroup for Implementation of Race-Free eGFR for Medication-Related Decisions.

Overarching Questions and Broader Context

What is the difference between estimated creatinine clearance (eCrCL) and estimated glomerular filtration rate (eGFR)?

 

Kidney clearance can be thought of as an equation: glomerular filtration plus tubular secretion minus reabsorption. The Cockcroft-Gault (C-G) eCrCL equation is validated against measured creatinine clearance (mCrCL) using timed urine collections. It was developed as a proxy for GFR at a time when mGFR values were not as available. However, creatinine is filtered by the glomeruli and secreted by the proximal tubule, thus mCrCL is known to overestimate measured GFR (mGFR). Conversely, eGFR equations are validated against mGFR, the current standard for evaluating kidney filtration function. mGFR is determined from the urinary or plasma clearance of an exogenous filtration marker, which is not reabsorbed or secreted by the tubules, such as iothalamate, iohexol, or 51Cr-EDTA (not available in the U.S.). GFR is recognized as the best contemporary measure for assessing kidney filtration and overall kidney function.

 

References:

Why is the Cockcroft-Gault (C-G) eCrCL equation being impacted by the NKF-ASN Task Force’s recommendations when race was never a variable within the C-G equation?

 

There are many factors driving the necessity for change, not just the use of race. While it is true that race was never a variable used within the C-G equation, the patient population used to validate the C-G equation was comprised of 249 white males at a veterans hospital in Canada. Alternatively, the newer NKF-ASN Task Force-recommended eGFR equations were developed and validated in significantly larger, more diverse, and heterogeneous populations.

 

References:

Is it sufficient to use ANY eGFR equation?

 

When switching to an eGFR equation, it is critical to use one of the NKF-ASN Task Force-recommended race-free eGFR equations: 2021 CKD-EPIcr, 2021 CKD-EPIcr-cys, or 2012 CKD-EPIcys. The MDRD and 2009 CKD-EPIcr equations for estimating GFR include race as a variable. Race is a social and not a biological construct and given the rapidly increasing multiracial makeup of the US population, continued use of eGFR equations that include race is not defensible or supported by the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease.

 

If your laboratory or practice has not yet adopted the race-free equation(s), the NKF has developed several resources and guidance documents to assist laboratories with implementing the race-agnostic CKD-EPI eGFR equations into practice.

 

References:

What populations and reference standards were used to develop the different creatinine-based GFR estimating equations?

 

The Cockcroft-Gault equation was developed in 1976, derived from 249 white men (mean age ~57 years, range 18-92), and does not include race as a variable. The reference standard was measured creatinine clearance (mCrCL), which is known to overestimate GFR due to tubular secretion. The equation also predates creatinine assay standardization, which occurred from 2005 to 2010 in the U.S., generating lower serum creatinine (Scr) results on average by ~12% and leading to higher corresponding C-G-based CrCL estimates.

 

The MDRD equation was developed in 1999 (refit in 2007 to accommodate newly standardized creatinine assays) and includes “Black race” as a variable. It was derived from ~1000 patients and validated in another ~550 patients. The reference standard was measured GFR (mGFR), the current standard. The overall study population was 12% Black and 40% female, with a mean age of 50.6±12.7 years and 41.8% being over 55 years. Furthermore, inclusion criteria required an eGFR <60 mL/min/1.73m2 (making it unreliable to calculate eGFR values >60 mL/min/1.73m2). This is the reason most laboratories started reporting eGFR as “>60 mL/min/1.73 m2” rather than providing a specific value for results over 60 mL/min/1.73m2.

 

The 2009 CKD-EPIcr equation was derived from ~8200 patients (32% Black, 5% Hispanic, 1% Asian; 43% female; mean age 47±15 years; 12.4% over 65 years) and validated in another ~3900 patients (10% Black, 2% Hispanic, 2% Asian; 45% female; mean age 50±15 years; 15% over 65 years). The reference standard was measured GFR (mGFR), the current standard. The equation also includes “Black race” as a variable. The study population included individuals with a wider range of eGFR than the MDRD equation and was shown to be more accurate, especially at higher eGFR ranges (i.e., >60 mL/min/1.73 m2). Unfortunately, some laboratories continue to report CKD-EPI results in the same manner as MDRD (i.e. not reporting specific values for results over 60 mL/min/1.73m2), which is not appropriate. This equation was developed and validated using standardized Scr assays.

 

The 2012 CKD-EPIcr-cys equation was derived from ~5300 patients (40% Black; 42% female; mean age 47±15 years; 13% over 65 years) and validated in another ~1100 patients (3% Black; 41% female; mean age 50±17 years; 21% over 65 years). The reference standard was measured GFR (mGFR), the current standard. The equation also includes “Black race” as a variable. This equation was developed and validated using standardized Scr and cystatin C assays.

 

The 2021 CKD-EPIcr and 2021 CKD-EPIcr-cys equations were derived from the same patients as the 2009 CKD-EPIcr & 2012 CKD-EPIcr-cys equations (~8200 and ~5300 patients, respectively) and validated in a new data set with ~4000 patients (14.3% Black; 38.4% female; mean age 57.0±17.4 years; 33.2% over 65 years). Notably, race was removed from the model, and the equation was reformulated. The reference standard was measured GFR (mGFR), the current standard. This equation was developed and validated using standardized Scr assays.

 

References:

Why did standardization of creatinine assays result in an average 12% decrease in serum creatinine results across labs?

 

Between 2005 and 2010, the serum creatinine assays used across the U.S. were standardized to the Isotope Dilution Mass Spectrometry (IDMS) reference method, the international standard. This improved the accuracy and interlaboratory consistency of results. The average decrease in serum creatinine results is due to standardization removing the positive bias created from non-creatinine substances (e.g., proteins, glucose, ketones, and other chromogens), which interfered with serum creatinine measurement to falsely elevate sCr results. This decrease was especially prominent for laboratories using the Jaffe method.

 

References:

Applying BSA-adjusted eGFR to Dosing Information/Guidance

Can I use eGFRBSAadj for dosing recommendations when the package insert and/or drug compendia resource is providing recommendations based on C-G eCrCL?


If so, is there anything I should take into consideration when doing this?

The NKF supports using eGFRBSAadj in place of C-G eCrCL for medication-related decision-making in adults with stable kidney function. Package inserts and compendia information on older medications do not consider that, on average, Scr has decreased ~12% across the U.S. with the implementation of standardized Scr assays, which may lead to higher corresponding C-G based CrCL estimates as compared to when those dosage guidelines were developed for pharmacokinetic (PK) studies conducted before ~2011. Also, keep in mind that package inserts are not routinely updated as practice standards change, especially for older medications that have lost marketing exclusivity. As such, clinical guidelines and drug information compendia often provide additional context based on updated literature to supplement the information found in prescribing information since product approval to help inform clinical decision-making.

 

Today, it's essential to use estimating equations that have been developed and validated in more diverse populations and with standardized Scr assays. Similarly, it’s essential to avoid additional idiosyncrasies when using estimating equations that lead to widespread inconsistencies across clinicians and practices (e.g., use of alternative weights, rounding up actual Scr values). To calculate eGFRBSAadj for a patient, multiply their standardized eGFR result (in mL/min/1.73m2) by their BSA (m2), then divide by 1.73.

 

Pharmacists and other clinicians who have relied on C-G eCrCL to determine medication dosing adjustments may feel more comfortable easing into using eGFRBSAadj by comparing it with the results of C-G eCrCL. However, it is important to remember that eGFRBSAadj and C-G eCrCL results (using various body weights and/or rounding of Scr) are not directly comparable because there is no consistent application of a single C-G eCrCL method across the U.S. When using eGFR for medication-related decision-making, remember to adjust the value for BSA (convert standardized eGFR to mL/min). If dosing discordance occurs between C-G eCrCL and eGFRBSAadj and a thorough comparative analysis is not available, clinical judgment is advised to determine whether the lower or higher dosage regimen should be initiated based on the patient’s clinical condition(s) and medication-specific characteristics. Adjusting for individual BSA is particularly important at extremes of BSA.

 

Clinicians must remember that results from these equations only provide estimates of kidney function and are not exact (i.e., measured) values for GFR and assume stable kidney function. This is true for both eGFR and eCrCL. Results of all estimating equations using Scr (including C-G eCrCL) show more variability from measured values at weight and age extremes.

 

An mGFR (or mCrCL if mGFR is not easily accessible or available) should be considered when greater accuracy is needed. This is particularly important when measurement of Scr and/or cystatin C may be unreliable due to potential non-GFR determinants, or in high-risk situations, such as use of narrow therapeutic index medications when therapeutic drug monitoring is not available.

 

References:

What if the product labeling recommends dosing based on standardized eGFR (mL/min/1.73m2) - should I still use eGFRBSAadj (mL/min)?

 

The NKF recommends that the adjustment of eGFR for a patient’s BSA should be a standard of practice for medication-related decision-making. eGFR adjusted for BSA (eGFRBSAadj) correlates better with mGFR and medication clearance (both expressed in mL/min), particularly in persons who have a BSA above or below 1.73m2. The practical implications are that, on average, using standardized eGFR (mL/min/1.73 m2) can lead to underdosing (when BSA is over 1.73 m2), overdosing (when BSA is under 1.73 m2), and/or misinformed decisions in both scenarios regarding when to initiate or discontinue certain medications.

 

Although product labeling for many medications recommends dosage adjustments using standardized eGFR (mL/min/1.73 m2), sometimes the body sizes of participants in the PK/clinical trials or retrospective analyses may not represent the wider range of patients seen in practice. Accordingly, we suggest using eGFRBSAadj to determine dosing of medications that are cleared through the kidneys, unless clinical studies evaluating dosing with standardized eGFR (mL/min/1.73 m2) have included adequate numbers of patients across the broader body size spectrum seen in the U.S. Adjusting for individual BSA is particularly important at extremes of BSA.

 

If dosing discordance occurs between standardized eGFR (mL/min/1.73m2) and eGFRBSAadj (mL/min), clinical judgment is advised to determine whether the lower or higher dosage regimen should be initiated based on the patient’s clinical condition(s) and medication-specific characteristics.

 

References:

Should I use the equation that was used in the original PK studies for the medicine I’m needing to dose adjust?


For example, if clinical trials used C-G eCrCL to inform medication dosing, can I use eGFRBSAadj to make dosing recommendations?

 

The NKF supports using eGFRBSAadj in place of C-G eCrCL for medication-related decision-making in adults with stable kidney function. Medications that were approved by the FDA and marketed in the US before ~2018 are unlikely to have used a standardized Scr assay in their phase 1 PK studies, as creatinine assay standardization was not fully implemented in the US until ~2010. Typically, the medication approval process from phase 1 through phase 3 clinical trials takes up to 10 years. Remember that on average, Scr has decreased by ~12% across the U.S. with the use of standardized Scr assays (with large variability in impact between laboratories), which led to higher corresponding C-G-based CrCL estimates compared to when those PK studies were conducted. Also, the race-agnostic CKD-EPI eGFR equations were developed and validated using standardized Scr assays and with larger, more diverse populations (e.g., body weight, sex, race, ethnicity) than were used in C-G eCrCL equation development.

 

Pharmaceutical manufacturers have typically used total body weight in the C-G eCrCL equation and in their prescribing information. However, pharmacists generally use a variety of body weights (albeit inconsistently) in the C-G eCrCL equation because it’s been demonstrated that use of different body weights across the weight spectrum increases accuracy of results. Lastly, labeling for kidney-related dosing is based on smaller studies of patients and will not capture all the variability from a larger, more heterogeneous patient population and/or presence of non-GFR determinants of Scr.

 

In many circumstances, the results obtained across GFR estimating equations based on standardized Scr assays will be highly similar to each other, though not always. Clinicians should use their clinical judgement for each patient-specific scenario, applying either a more conservative or more intensive approach depending on which circumstance is most relevant to the patient’s situation.

 

References:

Are there any additional considerations when using eGFRBSAadj to adjust medication doses for narrow therapeutic index agents (e.g. DOACs, chemotherapy)?

 

Clinicians must remember that eGFR values are estimates of kidney function and not exact (i.e., measured) values for GFR. Each estimating equation has an average bias from the measured GFR value and an accuracy value that shows the percentage of individual values in eGFR development or validation datasets that fall outside of +/- 30% of those individuals’ measured GFR values (also known as the P30 value). This is true for both eGFR and eCrCL. Therefore, clinical judgment is always essential when interpreting these results, especially in the context of narrow therapeutic index agents. Factors to consider when deciding whether to implement more intensive or conservative dose adjustments include the clinical relevance of any non-GFR factors affecting Scr (and/or cystatin C) concentrations, risk-benefit ratio for the specific medication, patient’s clinical condition, intended duration of therapy, the turnaround time for obtaining a cystatin C concentration, and availability of therapeutic drug monitoring. In situations where even more accuracy is required, mGFR (or mCrCL if mGFR is not easily accessible or available) may be warranted. Please consult Figure 2 in the NKF consensus report for a pictorial representation of this pragmatic approach.

 

References:

Why can’t we simply switch to using standardized eGFR (mL/min/1.73m2) for simplicity (instead of adjusting for BSA) as Vd (L) and CL (L/min) usually change in the same direction?

 

Some inpatient settings use C-G eCrCL equation results without including the individual’s weight to modify the medication dosing interval based on the expected change in medication half-life.

 

In FDA studies, bioequivalence (comparison of medication exposure) is primarily determined by review of Cmax, Tmax and area under the plasma concentration-time curve (AUC). For the majority of medications, AUC drives the determination of dosage adjustment, and clearance is the primary pharmacokinetic driver of AUC. The 2024 FDA final guidance for industry recommends manufacturers use eGFRBSAadj over C-G eCrCL in pharmacokinetic studies determining drug dosage adjustments moving forward.

 

When it comes to medication dosing in a patient with kidney impairment, it is the patient’s own kidney function that determines the capacity to clear any given drug. While using standardized eGFR (mL/min/1.73m2) for medication dosing is convenient, eGFR adjusted for BSA (eGFRBSAadj) correlates better with mGFR and medication clearance (both expressed in mL/min), particularly in persons who are underweight or overweight. To calculate eGFRBSAadj for a patient, multiply their standardized eGFR result (in mL/min/1,73m2) by their BSA (m2), then divide by 1.73.

 

Pharmacokinetic modeling for four intravenous antibiotics (i.e. aminoglycosides, cefepime, meropenem, vancomycin) show that eGFRBSAadj was better than standardized eGFR (mL/min/1.73m2, analogous to C-G eCrCL without weight adjustment) in predicting medication clearance and/or target AUC attainment. Using standardized eGFR instead of eGFRBSAadj would unnecessarily lead to less accuracy in medication-related decision-making.

 

Two additional studies within the area of oncology practice evaluating dosing strategies for methotrexate and carboplatin also found that eGFRBSAadj (mL/min) more closely correlated with medication clearance and reduced dosing discrepancies respectively, compared to standardized eGFR (mL/min/1.73m2). However, it is worth noting that these two studies predate the race-agnostic 2021 CKD-EPIcr equation, using the 2009 CKD-EPIcr equation instead.

 

References:

Even though eGFRBSAadj (mL/min) is considered better than standardized eGFR (mL/min/1.73m2) for medication dosing, does it justify the extra steps needed to adjust for BSA?

 

Pharmacokinetic modeling for 4 intravenous antibiotics (i.e. aminoglycosides, cefepime, meropenem, vancomycin) show that eGFRBSAadj was better than standardized eGFR (mL/min/1.73m2) and C-G eCrCL in predicting drug clearance and/or target AUC attainment (vancomycin).

 

As an example, two studies used the AIC (Akaike information criterion, a mathematical method for evaluating how well a model fits the data it was generated from) to evaluate this question. In the vancomycin study, the 2021 CKD-EPIcr eGFRBSAadj demonstrated a better fit than C-G eCrCL (variable dosing weight) and standardized eGFRcr (mL/min/1.73m2). The gentamicin study found a similar result, 2021 CKD-EPIcr eGFRBSAadj was also a better fit than C-G eCrCL (also with variable dosing weight) and standardized eGFRcr (mL/min/1.73m2). Similar studies using a different mathematical method (corrected Bayesian Information Criterion, BICc) with cefepime and meropenem pharmacokinetics showed similar findings - an improvement in model fit with 2021 CKD-EPIcr eGFRBSAadj than standardized eGFR (mL/min/1.73m2) or C-G eCrCL. Please consult Table 4 in the NKF consensus report for additional summaries of these studies and guidance for interpreting AIC and BICc values.

 

While using standardized eGFR (mL/min/1.73m2) for medication dosing is convenient, eGFR adjusted for BSA (eGFRBSAadj) correlates better with mGFR and medication clearance (both expressed in mL/min), particularly in persons who are underweight or overweight.

 

References:

Results from a secondary analysis of the ORBIT-AF II study have discouraged clinicians from using recommended eGFR equations for the purposes of initiating or dose selection of DOACs. How should we interpret the results of this study?

 

The ORBIT-AF II registry was a national, prospective cohort of patients with atrial fibrillation across the United States. This secondary analysis included 8727 patients receiving a direct oral anticoagulant (DOAC) with an available baseline serum creatinine. Patients with an estimated creatinine clearance (eCrCL) <15 mL/min, estimated glomerular filtration rate (eGFR) <15 mL/min/1.73m2, or on dialysis were excluded. Kidney function was estimated using the Cockcroft-Gault (C-G) eCrCL equation (specific body weight parameter not stated) and the standardized MDRD and 2009 CKD-EPIcr equations, which both include Black race as a factor, and are reported in mL/min/1.73m2. There was no referent standard (either measured CrCL or measured GFR) for comparison. 

 

Undertreatment and overtreatment were defined by discordance between C-G eCrCL-based DOAC recommendations and those derived from MDRD or CKD-EPI eGFR. The primary outcome was a composite of major adverse cardiovascular and neurologic events (MACNE), including cardiovascular death, stroke, systemic embolism/transient ischemic attack, and new-onset heart failure or myocardial infarction. Secondary outcomes also included all-cause death, major bleeding events, and all-cause hospitalization. Both primary and secondary events were assessed at 1-year post-DOAC initiation. 

 

The cohort was predominantly white (88% of patients), 26% had chronic kidney disease, and the mean weight was 89 kg (SD 75.0-106.0). The baseline median C-G eCrCL, MDRD eGFR, and 2009 CKD-EPI eGFRcr were 82.6 mL/min, 73.7 mL/min/1.73m2, and 71.3 mL/min/1.73m2, respectively, with the majority taking rivaroxaban (48.9%) or apixaban (44.6%). Agreement in dosing recommendations between C-G eCrCL and eGFR equations was high overall (~88% for rivaroxaban, ~99% for dabigatran and apixaban). For rivaroxaban, ‘undertreatment’ was more likely in the subgroup of patients with eGFR <50 mL/min/1.73m2 (4.40% with MDRD and 6.51% with 2009 CKD-EPIcr), while ‘overtreatment’ was more likely in the subgroup of patients with eGFR ≥30 mL/min/1.73m2 (7.12% with MDRD and 5.53% with 2009 CKD-EPI). 

 

The authors summarized the clinical implications of this research as follows: “These findings highlight the importance of using the C-G eCrCL equation, and not eGFR for dose adjustment in all patients with atrial fibrillation receiving DOACs.” However, this is not supported by the data presented; these data show standardized eGFR results (mL/min/1.73m2) in patients with lower weights were more likely to be overtreated, and those with higher weights were more likely to be undertreated compared to C-G eCrCL (mL/min), which isn’t surprising. The average (SD) weight and height of the population were 89 kg (75-106) and 172 cm (163.0–180.0). This suggests an average BMI and BSA for this cohort were 30.1 (Class 1 obesity) and 2.02m2, respectively. Worth noting, the average weights in the ‘overtreatment’ and ‘undertreatment’ groups were approximately 60 kg (29 kg below group average) and 105 kg (16 kg above group average), respectively. Other key limitations of this study include the lack of a measured kidney function reference standard and the use of older race-based eGFR equations, which are no longer recommended.

 

The researchers’ conclusion noted that the primary outcome (MACNE) was significantly worse with eGFR equations than C-G eCrCL. However, after multivariable adjustment, there was no statistical difference between C-G eCrCL and MDRD or 2009 CKD-EPIcr eGFR for the primary MACNE outcome or first major bleeding event (or in almost all the other numerous secondary outcomes that were evaluated). Unadjusted results should not be used to justify these conclusions because this was not a randomized controlled trial, and there were significant differences in several patient characteristics at baseline (e.g., sex, CHF history, weight, height), which would be expected to affect the outcomes. Overall, the study implications or conclusions are not supported by the evidence presented in this study due to the limitations noted above.

 

This study underscores the importance of adjusting for individual BSA, especially at extremes of BSA. 

 

References:

Cystatin C for Medication-related Decisions

When should I consider using cystatin C to estimate GFR (e.g., eGFRcys or eGFRcr-cys)?

 

Cystatin C can be most helpful in situations when a patient has significant non-GFR determinants that affect their Scr concentration. These clinical situations are featured within Table 2 of the NKF consensus report. Briefly, these include eating disorders, extreme muscle bulk, amputation, muscle wasting conditions, diet considerations, medications that inhibit tubular secretion of creatinine, broad-spectrum antibiotics that decrease non-kidney elimination, etc. However, cystatin C also has non-GFR determinants that must be considered. Furthermore, some patient factors are non-GFR determinants for both creatinine and cystatin C (e.g., obesity, concurrent medical conditions, malnutrition).

 

When evaluating non-GFR determinants, it’s important to emphasize these are not binary (e.g., yes/no, present/absent). Nor are they directly quantifiable (i.e., one non-GFR determinant for Scr does not cancel out one non-GFR determinant for cystatin C). Clinical judgment is imperative in all these situations. Most importantly, if a patient has one or more significant non-GFR determinant(s) for either creatinine or cystatin C that lead you to seriously question its utility, using that biomarker to estimate the patient’s GFR is not appropriate. An mGFR (or mCrCL value if mGFR is not easily accessible or available) should be considered if greater accuracy is needed and the patient has a condition(s) where measurement of Scr and/or cystatin C would be inaccurate or uncertain due to potential non-GFR determinants. An mGFR (or mCrCL) can also be considered in certain high-risk scenarios, such as patients in whom eGFR and C-G eCrCL results are generally less accurate (e.g., those at age and body weight extremes) and/or therapeutic drug monitoring is not available.

 

References:

Within your paper, there is mention of “access to serum cystatin C for timely decisions”; what exactly does “timely decisions” mean?

 

The interpretation of “timely decisions” is highly dependent on clinical context. Cystatin C is likely available in nearly all settings; the question is how long it will take to obtain the result. For example, a test result is likely to be available sooner if the laboratory uses an in-house assay as opposed to needing to send the sample out to a referral lab. In many situations, eGFR information is needed relatively quickly to inform drug decisions. In these circumstances, it would be more helpful to have access to a laboratory with an in-house assay for cystatin C. However, there may be situations where the information is not needed quickly (e.g., medication does not need to be started today, obtaining a baseline GFR before initiating a known nephrotoxin such as chemotherapy). In these cases, having the ability to run the test in-house may not be as relevant.

 

References:

What are some important considerations regarding cost when deciding whether to use cystatin C for medication-related decision making?

 

The cost of the cystatin C test is a factor to consider when deciding whether to order it. Like other lab tests, it should not be ordered unless there is an intention to use the resultant information. Ordering unnecessary tests contributes to increased costs for the healthcare system - both direct (e.g., bioassays, supplies) and indirect (e.g., time to draw blood, run the test, process results). Currently, coverage for cystatin C tests is relatively limited to very specific situations and usually not within the context of dosing medications. Therefore, it is critical to also consider out-of-pocket costs for the patient and direct costs to the lab and/or health system. However, when there are non-GFR determinants limiting the reliability of Scr, ordering cystatin C is recommended for appropriate decision making.

Since serum creatinine (Scr) is the biomarker predominantly used in PK studies, is it okay to use cystatin C instead of (or in addition to) Scr for medication-related decision making?

 

Yes, there is a gap in data when applying eGFRBSAadj based on 2021 CKD-EPIcr-cys or 2012 CKD-EPIcys to medication dosing. Most (if not all) product labeling is derived from creatinine-based estimates (either eCrCL or eGFRcr). The decision regarding whether to use cystatin C instead of (or in addition to) creatinine will be driven largely by how unreliable a patient’s Scr concentration is, based on non-GFR determinants. The most important element is using the best possible estimate of kidney filtration function available for that specific patient and applying that result to medication dosing decisions (in addition to clinical judgment).

 

References:

Since cystatin C has a shorter half-life, does that make it more helpful to estimate GFR for medication-related decisions in the acute care setting?

 

Use of any GFR estimating equation (either C-G eCrCL or eGFR) in the acute care setting can be problematic, especially for patients with unstable GFR. The studies featured in Table 4 within the NKF consensus report all included adults in the acute care setting and still found that the combined 2021 CKD-EPIcr-cys equation predicts drug clearance better than either 2021 CKD-EPIcr, 2012 CKD-EPIcys, or C-G eCrCL. As such, this may be a situation where the combination equation is justified if mCrCL or mGFR is not feasible.

 

An eGFR concordance study in over 17,700 hospitalized adults showed a discrepancy between 2021 CKD-EPIcr and 2012 CKD-EPIcys in a large portion of acute care patients. Unfortunately, since plasma drug concentrations were not available for most of the medications and mGFR and mCrCL were not used as a relative standard, it's unknown which biomarker may have been best. It’s important to emphasize that these equations were not developed for use in the acute care setting.

 

Please refer to the 2025 JACCP article referenced below for an analysis of one academic medical center’s experience with integrating cystatin C for medication dosing into their acute care clinical practice.

 

References:

What steps should I take if I identify a significant discrepancy between a patient’s eGFRcr and eGFRcys?

 

In scenarios where 2021 CKD-EPIcr eGFR and 2012 CKD-EPIcys eGFR values are discordant (e.g., results are >20% higher or lower than the other measure), 2021 CKD-EPIcr-cys is generally more accurate than either 2021 CKD-EPIcr or 2012 CKD-EPIcys. However, in patients with significant non-GFR factors that affect only Scr (e.g., muscular atrophy, creatine supplementation in a bodybuilder), 2012 CKD-EPIcys will likely produce more accurate results than 2021 CKD-EPIcr-cys. Similarly, in patients with significant non-GFR factors that affect only cystatin C (e.g., effects of smoking, uncontrolled inflammatory condition), 2021 CKD-EPIcr will likely produce more accurate results than 2021 CKD-EPIcr-cys. Please see Table 2 in the NKF consensus report for further guidance.

 

Therefore, it is important to evaluate the presence and magnitude of various non-GFR determinants for each biomarker a patient may have to identify which may be more appropriate for their situation. In situations where even more accuracy is required for medication-related decision-making (e.g., dosing of narrow therapeutic index medications), mGFR(or mCrCL) and/or therapeutic drug monitoring may be indicated.

 

References:

Is there an easy way to remember some of the non-GFR determinants of cystatin C?

 

Things that affect cellular metabolism rates can affect cystatin C concentrations - from a medication perspective, corticosteroids, growth hormones, and other agents that increase cell turnover may increase cystatin C concentrations. From the perspective of comorbidities, systemic inflammatory states (sepsis, active infection), hypercellular states (active malignancy), smoking, or hyperthyroidism can all increase cystatin C concentrations, while diseases associated with decreased cellular metabolism (e.g., uncontrolled hypothyroidism) may decrease cystatin C. Obesity may also increase cystatin C concentrations because of a greater number of nucleated cells. It is unclear to what degree these factors influence cystatin C (e.g., increase by 5% or 50%), but they should be considered when interpreting a cystatin C value, especially when it is significantly discordant from Scr.

 

References:

Practical Considerations, Tools, and Resources

What tools are available to clinical staff to transition to race-agnostic eGFR for medication-related decision making?

 

The NKF has developed several resources and guidance documents to assist clinicians and practices with implementing the race-free CKD-EPI eGFR equations into practice. Clinicians can also refer to the NKF consensus report for guidance on best practices for incorporating eGFRBSAadj into medication-related decision making. Furthermore, the workgroup has developed this FAQ document to help provide further guidance for some of the most commonly encountered questions. Other tools can be found on the NKF Race-agnostic eGFR for Medication-related Decisions resource page, including a templated slide presentation, a series of case-based self-assessment questions, and a best practices resource with recommendations for implementing these changes within the electronic health record (EHR). If your EHR system cannot report eGFRBSAadj, consider embedding a hyperlink to the NKF eGFR calculator in the screens and menus wherever eGFRBSAadj should be used for medication-related decision-making.

 

References:

What are some factors I should consider when interpreting an eGFRBSAadj value?

 

In addition to the usual factors you would consider when interpreting an eGFR (e.g. stable kidney function, presence/absence of non-GFR determinants, concurrent conditions, inpatient vs. outpatient setting), some additional factors that are important to consider when interpreting an eGFRBSAadj include:

  • How current/accurate are the height and weight used to determine the BSA? Both measures within the electronic health record (EHR) are highly prone to data entry errors or result from estimation. When interpreting eGFRBSAadj, use a current, valid BSA for the patient (informed by height and weight). This becomes especially important if the EHR is automatically performing these calculations for you.
  • How recent was the biomarker (e.g. Scr, cystatin C) measured and is that still clinically relevant given the patient’s situation? Having a good understanding of your health system’s parameters for reporting eGFR with each biomarker will be helpful (e.g. “lookback times”, reporting requirements).

To help individual institutions implement these recommendations into their EHR systems, the NKF has developed a best practices resource with recommendations for leveraging the EHR to safely automate reporting of eGFRBSAadj for informing medication decisions.

Are there specific populations where the transition from C-G eCrCL to eGFRBSAadj for medication-related decision making might have the greatest impact?

 

The eGFR estimating equations are derived from more diverse data sets than the original C-G eCrCL equation in several ways (e.g. sex, race, body size). Therefore, it is likely to be more reflective of the general patient population seen in current clinical practice. Patients who are at either end of the body size spectrum (i.e., underweight, obesity) are also likely to be greatly impacted if the transition is done incorrectly and standardized eGFR is used for dosing decisions instead of eGFR adjusted for BSA.

 

Another group of people who are likely to be impacted are older adults (e.g. over 70). This patient population was underrepresented in the development of all the equations used to estimate kidney filtration function (i.e. CKD-EPI eGFR, C-G eCrCL). In older adults, it is important to consider the age-related decline in muscle mass within the context of normal age-related declines in kidney function. In a simulation study of adults aged 70 years, C-G eCrCL (using total body weight) provided lower values than CKD-EPI eGFRBSAadj, especially for patients with lower Scr or lower BMI. Alternatively, the C-G CrCL-based estimate was found to be higher than eGFRBSAadj in patients with higher BMI. There is often a temptation among clinicians to prefer the lower value, especially in older adults, but that does not mean the lower result is always more accurate or safer. In certain situations (e.g., antibiotics, antivirals, chemotherapeutics), automatically deferring to the lower value can cause significant harm by limiting efficacy.

 

Unfortunately, no estimating equation does a great job (including C-G eCrCL) in assessing kidney filtration function in older persons across the weight spectrum. This is a population where clinical judgment is essential for balancing the risks and benefits of treatment decisions within the context of each individual patient’s clinical situation. More data is needed in older persons (across weight strata) for current eGFR equations that compare against mGFR, the current standard for assessing kidney filtration function.

 

Assuming minimal non-GFR determinants for cystatin C, clinicians and pharmacists can consider the eGFRcr-cys (adjusted for BSA) equation for routine GFR evaluation in older adults. For critical treatment decisions, consider using mGFR (or mCrCL if mGFR is not easily accessible or available).

 

References:

What are the potential impacts of the transition from C-G eCrCL to eGFRBSAadj on drug dosing decisions?

 

The most notable impact of this transition will likely be felt around dose thresholds (i.e., the point at which a different dose is recommended based on the patient’s eGFRBSAadj value). In situations where there is conflicting information or discordance (e.g., comparing eGFRBSAadj to package insert or compendia providing guidance based on C-G eCrCL), clinical judgment is imperative. Recall that eGFRBSAadj has been shown in several studies to predict clearance and/or area under the plasma concentration-time curve (AUC) for several medications better than C-G eCrCL using either total body or various dosing weights.

 

Another potential impact includes when certain medications may be indicated and/or recommended for discontinuation at a certain eGFR/eCrCL threshold. If a patient’s eGFR is not adjusted for their BSA, a medication may be discontinued prematurely due to the perception that the patient’s kidney clearance of medications is less than it really is.

What data monitoring or research initiatives would be beneficial for evaluating the impact of this change?

 

While this workgroup does not feel further research is needed to switch from C-G eCrCL to race-agnostic eGFRBSAadj for medication-related decision making, ongoing research is always helpful to assuage concerns clinicians may have. The impact of this change can be evaluated in several ways, including:

Prospective Studies: 

  • Conduct prospective studies to monitor the implementation of the recommendations and assess their impact on medication dosing accuracy and patient outcomes. When conducting these studies, it is important to compare equation performance against mGFR and/or plasma drug concentrations as the current standards (rather than comparing equations to each other in dosing discordance studies).

Retrospective Data Analyses: 

  • Analyze retrospective data to compare medication dosing practices and adverse effects or effectiveness metrics before and after the implementation of these recommendations. This can help identify any significant issues with changes in dosing patterns in various patient populations.
  • Analyze retrospective drug-concentration data sets in medications of interest and apply pharmacokinetic modeling approaches to predict drug clearance or area under the plasma concentration-time curve (AUC) with C-G eCrCL (various weights) compared to recommended CKD-EPI eGFRBSAadj equations.

Clinical Workflow Assessments: 

  • Evaluate the workflow in pharmacies and health-system practices to determine the impact of these recommendations on daily operations, workflows, alignment across disciplines, and clinical outcomes, and identify opportunities for increased efficiency (if necessary).
Is it possible to modify our electronic health record (EHR) setup to incorporate automatic reporting of eGFRBSAadj to avoid the added step of further calculations?

 

The most important place to start is to ensure your electronic health record (EHR) is using the NKF-ASN Task Force-recommended race-agnostic equations for reporting estimated GFR. Review the Clinical Chemistry article referenced below for guidance on how clinical laboratories can implement the recommended eGFR equations.

Once you’ve confirmed your EHR is using the race-agnostic equation(s), it will be important to engage several entities within your organization to identify the most appropriate way(s) to report eGFRBSAadj in your EHR. Some recommendations for how to start such an initiative include:

  • Evaluate for potential information systems (IS) and institutional barriers and facilitators for designing and piloting the change early. Barriers may include the complexity of integrating new eGFR calculation methods into existing systems and obtaining institutional buy-in for the changes. Facilitators may include identifying all the places where C-G eCrCL is currently reported/displayed within your EHR, as these are the areas where replacing with eGFRBSAadj will provide the most utility. Focusing here can help make the scope of the project more manageable 
  • Identify key stakeholders. This involves determining who will be affected by the change, such as clinicians involved in medication decision-making and dose verification, laboratory staff, and IT personnel. 
  • Identify institutional constraints and limitations. This includes assessing workforce collaboration between the laboratory and EHR teams to ensure data integration.
  • Create a detailed implementation plan to outline the steps needed to update the EHR, assign responsibilities, and set timelines for each phase of the project.
  • Train clinical staff and other impacted personnel/departments to ensure all involved in medication decision making and dose verification comprehend this new approach to medication-related decision making and their implications for patient care.
  • Continuous Plan-Do-Study-Act (PDSA) cycles are necessary to review the effects of the change. Regularly monitoring and evaluating the implementation process allows for adjustments and improvements to be made as needed.

To help individual institutions with implementing these recommendations into their EHR systems, the NKF workgroup has developed a best practices resource with recommendations for leveraging the EHR to safely automate reporting of eGFRBSAadj to inform medication decisions. 

 

References: 

The studies referenced in the paper are all examples from inpatient practice. How does this practice apply to medication-related decisions in the outpatient setting?

 

The studies referenced in the NKF consensus report were used because they evaluated actual clearance (or other relevant pharmacokinetic parameter) of medications that are highly cleared by the kidneys against race-agnostic eGFR & C-G eCrCl. Unfortunately, these types of studies are seldom done in the outpatient setting, so there were no outpatient models to reference. However, the same principles from these inpatient studies would still apply to medications in the outpatient setting. In situations where there is a concern for over/underdosing with medications that have a low therapeutic index, plasma drug concentrations, patient symptoms, and/or effectiveness markers (e.g. INR, blood glucose concentrations) can be monitored. Furthermore, pharmacokinetic/dynamic modeling approaches can be employed to compare recommended eGFR equations as predictive factors for drug clearance and/or AUC.

Potential Downstream Effects

How might these recommendations affect insurance coverage decisions for treatments?

 

The impact of these recommendations on insurance coverage decisions for medications is difficult to ascertain. At a minimum, it is not likely to have a negative impact, as most determinations are based on a medication’s indication rather than specific dosing recommendations. However, there may be situations where a patient’s standardized eGFR might recommend against the use of a medication while the more patient-specific eGFRBSAadj suggests the opposite (especially for patients with a BSA greater than 1.73 m2). In these situations, the references utilized in the NKF consensus report and supplemental resources such as this document may help provide supporting documentation for appeals.

 

Regarding the use of cystatin C, the NKF-ASN Task Force on Reassessing the Inclusion of Race in Diagnosing Kidney Diseases recommends national efforts to facilitate increased, routine, and timely use of cystatin C, especially to confirm eGFR in clinical decision-making. Current coverage determinations for cystatin C focus primarily on the diagnosis and staging of CKD. With increased use and research investigating the use of cystatin C for medication-related decision-making, this may improve insurance coverage for cystatin C testing.

 

Additionally, the NKF Laboratory Engagement Initiative Workgroup is collaborating with US Pathology and Laboratory Society leadership to develop tools to support the implementation of cystatin C testing. This collaboration includes creating content and guidelines to help laboratories integrate cystatin C into their testing protocols.

 

References: 

What impact might this transition have on various protocols, practice standards, and collaborative practice agreements?

 

This will depend on how these items are set up in your practice. The NKF recommends clinicians review their protocols, practice standards, and collaborative practice agreements, paying special attention to any references to C-G eCrCL and replacing them with [race-agnostic] eGFRBSAadj in any areas referencing medication dosing. Pharmacists and other healthcare professionals may also consider incorporating cystatin C into these critical practice documents, including any site-specific guidance for how and when cystatin C testing should be incorporated for medication-related decision-making. Consider using the scenarios presented in Table 2 of the NKF consensus recommendations to help inform those conversations. 

 

References:

What are the implications of this change for population health CKD screening programs?

 

The scope of the NKF recommendations related to this work is on medication-related decision-making, so it is distinct from CKD screening and monitoring. However, removal of race from the equation(s) used to estimate GFR is recommended in all aspects of care for CKD, including CKD population health and quality improvement programs to provide more equitable estimates for all individuals. When testing for CKD, recall the 2021 NKF-ASN task force recommends using standardized eGFR (in mL/min/1.73 m2) as a population-level measure to screen, diagnose, and stage CKD.

 

Once a person has a diagnosis of CKD, the eGFRBSAadj can provide a more tailored assessment of GFR to inform medication-related decisions for that person. With a more precise eGFR, clinicians involved in medication decision-making and dose verification can mitigate the risk of inappropriate medication dosing and medication selection, leading to better patient outcomes and overall population health improvements.

 

References: 

Glossary

  • AUC: area under the plasma-concentration time curve
  • BSA: body surface area
  • C-G: Cockcroft-Gault
  • CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration
  • 2021 CKD-EPIcr: 2021 CKD-EPI eGFR equation with creatinine
  • 2021 CKD-EPIcr-cys: 2021 CKD-EPI eGFR equation with creatinine and cystatin C
  • 2012 CKD-EPIcys: 2012 CKD-EPI eGFR equation with cystatin C
  • eCrCL: estimated creatinine clearance
  • eGFR: estimated glomerular filtration rate
  • eGFRBSAadj: estimated glomerular filtration rate adjusted for body surface area (units: mL/min)
  • mCrCL: measured creatinine clearance
  • mGFR: measured glomerular filtration rate
  • NKF: National Kidney Foundation
  • Scr: serum creatinine

Acknowledgements

Authors:

  • Andrew Bzowyckyj, PharmD, BCPS, CDCES
  • Venita Le Schandorf, DNP, ARNP, FNP-BC
  • Andrew Webb, PharmD, BCCCP

Reviewers:

  • Tracy Anderson-Haag, PharmD, BCPS, BCTXP
  • Joanna Hudson, PharmD, BCPS
  • W. Greg Miller, PhD
  • Tom Nolin, PharmD, PhD
  • Wendy St. Peter, PharmD
  • Joseph Vassalotti, MD