Why the Kidney Function Metric Matters
A dose rule can appear to depend on a single number—such as a kidney function value above or below a threshold—but that number is meaningful only when its metric, units, and scale are clear. Measured glomerular filtration rate (GFR), estimated GFR (eGFR), and creatinine clearance are related measures, but they are not interchangeable. They may yield different numerical values for the same person, and a difference that seems small can change a category when the value is near a decision threshold.
The appropriate metric depends on the evidence supporting the dose rule, the metric used as a covariate in the pharmacokinetic model, and how the rule defines its threshold. The goal is not to choose whichever number looks most convenient or to average different estimates. It is to use the measure that matches the decision and to make any necessary unit conversion explicit.
Metric, Units, and Body-Size Scale
Measured GFR can be obtained using an exogenous filtration marker, such as iohexol, when a more direct assessment is warranted. It is not error-free: the result depends on the measurement protocol, sampling, and analytical method. It is also more involved than obtaining a routine serum-creatinine-based estimate.
Many laboratories report eGFR in mL/min/1.73 m2. This is an indexed value: it expresses kidney function relative to a standard BSA. A dose rule that requires an absolute value in mL/min may therefore require de-indexing, provided the rule’s intended metric and scale support that conversion. A commonly used approximation is:
This conversion changes the scale of the reported estimate; it does not turn eGFR into measured GFR or creatinine clearance. BSA can be estimated by the Mosteller formula when height and weight are available:
Creatinine clearance may be measured from a timed urine collection or estimated. Cockcroft–Gault is a commonly used estimate that takes age, body weight, serum creatinine, and a conventional sex coefficient as inputs. In its conventional form, with serum creatinine in mg/dL and weight in kg, it produces an estimate in mL/min. The equation and its conventions should be checked against the protocol or source used for a dose recommendation; it is not an equation for eGFR.
The distinction between indexed and absolute values matters most when patient BSA differs appreciably from 1.73 m2. A value of 50 mL/min/1.73 m2 and a value of 50 mL/min do not describe the same scale. Never compare one with a threshold defined for the other without checking the intended conversion.
Worked Example: De-Indexing eGFR for a Patient With Low BSA
A hypothetical dosing protocol defines its threshold as an absolute GFR of 45 mL/min. A patient is 170 cm tall and weighs 50 kg. The laboratory reports an indexed eGFR of 48 mL/min/1.73 m2. What is the approximate absolute eGFR, and how does it compare with the protocol threshold?
First calculate BSA:
Then de-index the reported eGFR:
The approximate absolute eGFR is 42.65 mL/min, below the protocol’s threshold of 45 mL/min. Comparing the indexed value of 48 directly with an absolute threshold would put the patient on the wrong side of that threshold. This calculation is appropriate only because the hypothetical protocol specifies an absolute value and permits use of eGFR on that scale.
Creatinine Clearance and the Source of Disagreement
Serum-creatinine-based estimates rely on a marker whose concentration is affected by creatinine production and elimination. Age, body composition, diet, and other clinical circumstances can influence how well a given equation estimates kidney function for an individual. A serum creatinine value alone is therefore not a direct reading of filtration.
Creatinine is filtered by the kidney and also secreted into the urine. Consequently, measured creatinine clearance commonly exceeds GFR, although the size of the difference varies. A timed urine collection adds another potential source of error: incomplete or inaccurately timed collection can bias the calculated clearance. Estimated CrCl has different limitations because it depends on the equation and its inputs, including how body weight is handled.
These differences explain why eGFR and CrCl can disagree without either calculation being arithmetically wrong. They estimate related but nonidentical quantities, using different data and assumptions. The method that best supports a particular decision is the one specified by the dose rule or supported by its underlying evidence—not automatically the numerically larger, smaller, or more familiar estimate.
Worked Example: Comparing eGFR and Cockcroft–Gault CrCl
For a hypothetical patient, age is 68 years, weight is 62 kg, and serum creatinine is 1.2 mg/dL. For illustration, use the conventional Cockcroft–Gault equation with the 0.85 coefficient:
Substituting the patient’s values gives:
Suppose the same hypothetical patient has a laboratory-reported indexed eGFR of 52 mL/min/1.73 m2 and a BSA of 1.60 m2. The approximate absolute eGFR is:
The estimates are approximately 43.92 mL/min by Cockcroft–Gault and 48.09 mL/min by de-indexed eGFR. If a hypothetical rule uses an absolute threshold of 45 mL/min, these estimates fall on opposite sides. That disagreement is a reason to verify which metric the rule requires and whether its threshold is absolute or indexed. It is not a reason to average the values: their methods and targets differ, and their arithmetic mean would not become a validated kidney function measure.
Choosing and Checking a Metric
A defensible choice begins with the decision rule rather than with the available calculator. Check the source of the rule, the metric named in it, the units, and the patient population in which the rule was developed or evaluated. If the rule uses CrCl, substituting eGFR simply because a laboratory reports it routinely may change the decision. If the rule uses indexed eGFR, converting it to mL/min without a reason may also be inappropriate.
For a model-based assessment, check how kidney function was defined in the pharmacokinetic model. A model relationship estimated using one metric should not be applied as though a different metric were identical. The covariate’s unit and scale must match those used in the model, including whether it was indexed to BSA. This check is separate from deciding whether a model-based dose recommendation is appropriate for a particular patient.
When two reasonable estimates disagree, use a structured review:
Read the source definition rather than relying on the label “kidney function.” Determine whether it requires measured GFR, eGFR, or CrCl.
Confirm whether values are in mL/min or mL/min/1.73 m2. Convert only when the dose rule and intended scale justify conversion.
Check serum creatinine units, height, weight, age, the equation used, and—if relevant—the timing and completeness of a urine collection.
Determine whether estimates lead to different categories under the specified rule. If the choice would materially change a decision, seek clarification or consider whether a more direct assessment is warranted under the applicable clinical process.
Agreement should be assessed in the context of the decision, not inferred from correlation alone. Two methods can rank people similarly while differing systematically in their numerical values. Near a threshold, even modest disagreement can change classification. Record the estimate, method, units, date, and any conversion so that the basis for the decision remains clear.
Agreement Near a Threshold
A threshold comparison has two parts: first, establish a value on the same scale and using the same metric definition as the threshold; second, compare that value with the threshold. A threshold written as “50” is incomplete unless its units and metric are known. Nor does a value just above a threshold mean that kidney function is known precisely to be above it. Measurement and estimation have uncertainty, and the practical importance of a small difference depends on the decision rule.
Worked Example: Three Estimates Near a Decision Boundary
Consider a hypothetical case with a dose-category threshold of 50 mL/min on an absolute scale. A measured GFR is 48 mL/min. The laboratory reports indexed eGFR of 42 mL/min/1.73 m2, and the patient’s BSA is 1.95 m2. Cockcroft–Gault CrCl is reported as 54 mL/min.
Convert indexed eGFR to an approximate absolute value:
The resulting values are 48 mL/min by measured GFR, approximately 47.34 mL/min by de-indexed eGFR, and 54 mL/min by Cockcroft–Gault. Relative to the hypothetical 50 mL/min threshold, the first two are below it and the CrCl estimate is above it. If the rule explicitly specifies CrCl, the relevant comparison is the CrCl estimate, subject to appropriate application of the rule. If the rule specifies measured GFR, the CrCl value does not replace that measurement. If the rule is unclear, the discrepancy is a prompt to resolve the metric definition—not to take a majority vote among estimates.
A measured value is not automatically preferable for every decision: a dose rule may have been developed around another metric. Conversely, availability of a routine estimate does not make it interchangeable with a more direct measurement when the decision specifically calls for measured GFR. Near a consequential threshold, confirm the intended metric and consider the uncertainty and clinical context rather than treating a rounded estimate as an exact boundary.
Comparison, Limitations, and Common Pitfalls
| Metric | How it is obtained | Key interpretation check | Important limitation |
|---|---|---|---|
| Measured GFR | Clearance of an exogenous filtration marker | Confirm measurement method and units | More involved to obtain; protocol and measurement error remain relevant |
| eGFR | Equation using laboratory and patient characteristics | Check equation, units, and whether it is indexed to 1.73 m2 | Estimate may be less representative for an individual in circumstances that affect its inputs |
| Creatinine clearance | Timed urine measurement or an estimating equation such as Cockcroft–Gault | Check whether measured or estimated, plus equation, inputs, and units | Creatinine secretion, collection quality, or equation conventions can affect the result |
A frequent mistake is to see an indexed eGFR of 48 mL/min/1.73 m2, compare it directly with an absolute threshold of 45 mL/min, and conclude that the patient is above the threshold. That reasoning ignores both the unit difference and the patient’s BSA. The correction is to check what the rule requires and, if appropriate, de-index using the patient’s BSA before comparison. A different but equally common mistake is to convert automatically, even when the threshold is explicitly indexed; conversion is not a universal first step.
Another pitfall is treating a numerical mismatch as proof that one method has failed. The values may differ because the methods estimate different quantities or use different inputs. First verify the calculations and definitions. Then determine which measure is relevant to the particular dose rule or model. These metrics describe kidney function; they do not, by themselves, establish how much a particular drug’s clearance depends on renal function.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A hypothetical protocol uses an absolute eGFR threshold of 60 mL/min. A patient’s indexed eGFR is 55 mL/min/1.73 m2 and BSA is 1.90 m2. What is the approximate absolute eGFR, and which side of the threshold is it on?
- Why should an indexed eGFR not automatically be compared with a CrCl threshold?
- A measured GFR and a Cockcroft–Gault estimate disagree near a threshold. Should their arithmetic mean be used? Explain.
- Name two checks to make before using a reported kidney function value in a dosing rule.
Solutions. (1) De-index: \(55\times(1.90/1.73)=60.40\) mL/min, approximately. This is just above the hypothetical absolute threshold of 60 mL/min. (2) They are different metrics, and the eGFR is indexed to 1.73 m2 while a CrCl threshold is commonly expressed in mL/min. The threshold’s metric and scale must match the value used. (3) No. The mean is not thereby a validated estimate; first establish which metric the rule specifies and investigate the methods, inputs, and units. (4) Examples include checking the metric and equation, the units and BSA indexing, the patient inputs, and whether the reported value is measured or estimated.