GFR is defined as the rate of filtration of plasma across the glomerular capillary wall into the Bowman space and is the product of the average of single-nephron GFR (SNGFR) times the number of nephrons. SNGFR is governed by Starling forces (net filtration pressure, ΔP) and the properties of the glomerular capillary wall (ultrafiltration coefficient, K _F ). For further discussion of the physiology of glomerular ultrafiltration, see Chapter 3.

The kidney filters an average of approximately 180 L of plasma per day, equivalent to 125 mL/min, with a wide range. GFR is indexed by body surface area (BSA) because kidney size is proportional to body size; this practice reduces variability in GFR among healthy individuals and allows comparisons to normative values. The normal value of GFR (125 mL/min/1.73 m ² ) was derived from older studies, when 1.73 m ² was the average value for BSA in young adults. Studies of kidney donor candidates in the current era report a lower mean GFR, approximately 105 mL/min/1.73 m ² . The source of differences between recent and older studies is not well understood. GFR >90 mL/min/1.73 m ² is considered normal for children >2 years of age and young adults, and GFR <60 mL/min/1.73 m ² is one of the criteria for AKD and CKD at all ages.

Reduced GFR can be caused by decreased SNGFR due to alteration in Starling forces or decreased number of nephrons. Apart from AKD and CKD, conditions that lead to alterations in Starling forces include exercise, obesity, pregnancy, level of protein intake, hyperglycemia, surfeit or deficit of extracellular fluid, and medications that lower blood pressure, inhibit the renin-angiotensin-aldosterone system, or inhibit sodium-glucose linked transport proteins. Premature birth or intrauterine growth retardation is associated with a decreased number of nephrons and increased risks for AKD and CKD. Loss of nephrons occurs with aging, but the association is variable and the cause is not fully understood (see Chapter 21 ).

True GFR, like many other physiologic properties, cannot be measured directly. Rather, GFR is assessed by measuring the clearance of filtration markers, substances that are eliminated primarily by glomerular filtration and whose plasma concentration varies inversely with the level of GFR (higher plasma concentration at lower GFR and lower plasma concentration at higher GFR). An ideal filtration marker for GFR measurement would have the following characteristics:

• •

  Short equilibration time between plasma and final volume of distribution

• •

  Not bound to plasma proteins

• •

  Freely filtered at the glomerulus (molecular weight <approximately 20,000 daltons)

• •

  Not secreted or reabsorbed at the tubules

• •

  Eliminated wholly by the kidney

• •

  Resistant to degradation

• •

  Easy and inexpensive to measure

None of the currently available filtration markers meets all of these requirements. Inulin, a 5200 polymer of fructose, was identified by the pioneering work of Homer Smith, and the urinary clearance of inulin remains the reference standard for measured GFR (mGFR), but inulin is not available in many countries. Except for creatinine, filtration markers used in clearance procedures for GFR measurement are exogenous substances.

Clearance measurements are not routinely performed in clinical practice; instead, the level of GFR is usually estimated from the plasma concentration of an endogenous filtration marker. Estimated GFR (eGFR) is in widespread use in clinical practice but differs from true GFR because of the influence of physiological processes other than GFR that affect the plasma concentration of the endogenous filtration marker. These processes include generation, renal tubular reabsorption and secretion, and extrarenal elimination of the marker, and collectively, they are termed non-GFR determinants of the filtration marker. The relationships of the non-GFR determinants to GFR and plasma concentration of the filtration marker in the steady state (when GFR and non-GFR determinants are constant) are shown in Figure 23.1 . GFR is inversely related to the plasma concentration of the filtration marker and modified by each of the non-GFR determinants. The inverse relation between GFR and the plasma concentration of the filtration marker allows maintenance of a constant filtered load (the product of GFR times the plasma concentration) and enables continued excretion of the marker despite reduction in GFR, but with the consequence of a higher plasma concentration of the marker ( Fig. 23.2 ).

Determinants of the serum level of endogenous filtration markers.

The plasma level (P) of an endogenous filtration marker is determined by its generation (G) from cells and diet, extrarenal elimination (E) by gut and liver, and urinary excretion (UV) by the kidney. Urinary excretion is the sum of filtered load (glomerular filtration rate [GFR] × P ) minus reabsorption (TR) plus tubular secretion (TS). In the steady state, generation minus extrarenal elimination equals urinary excretion. By substitution and rearrangement, GFR can be expressed as the ratio of the non-GFR determinants ( G, TS, TR, and E ) to the plasma level.

Permission granted from J Am Soc Nephrol for Stevens LA, Levey AS. Measured GFR as a confirmatory test for estimated GFR. J Am Soc Nephrol. 2009;20(11):2305–2313.

Effect of an acute GFR decline on generation, filtration, excretion, balance, plasma concentration of an endogenous filtration marker (P _marker ), and estimated GFR (eGFR) based on the marker.

After an acute GFR decline, generation of the marker is unchanged, but filtration and excretion are reduced, resulting in retention of the marker (a rising positive balance), a rising plasma concentration, and declining eGFR (non–steady state). During this time, eGFR is higher than GFR. While GFR remains reduced, the rise in plasma concentration leads to an increase in filtered load (the product of GFR times the plasma level) until filtration equals generation. At that time, cumulative balance and the plasma level plateau at a new steady state. In the new steady state, eGFR approximates mGFR. GFR is expressed in units of milliliter per minute per 1.73 m ² . Tubular secretion and reabsorption and extrarenal elimination are assumed to be zero.

Permission granted from J Am Soc Nephrol for Stevens LA, Levey AS. Measured GFR as a confirmatory test for estimated GFR. J Am Soc Nephrol. 2009;20(11):2305–2313.

The non-GFR determinants of filtration markers are difficult to measure, so estimating equations use easily measured demographic (age, sex) and clinical variables (weight) as surrogates. The inclusion of these variables minimizes systematic errors in subgroups defined by these variables and minimizes systematic differences between groups. ^, GFR estimating equations convert the plasma concentration of the marker to the GFR scale and, by design, are more accurate than the plasma concentration alone and facilitate clinical decision making on the basis of the level of GFR. Thus it is recommended that clinical laboratories report eGFR whenever the filtration marker is measured. The principal limitations of GFR estimating equations are that the demographic and clinical surrogates only capture the average relations among the marker and its non-GFR determinants and that the relations among the marker and its non-GFR determinants may vary across populations.

eGFR is more accurate in the steady state than in the non–steady state (when GFR and non-GFR determinants are not constant). In the non–steady state, the rate and direction of change in the concentration of the filtration marker and in level of eGFR reflect the magnitude and direction of the change in GFR, but the change in the concentration of the marker and level of eGFR derived from the marker lags behind the change in GFR. As shown in Figure 23.2 , after a fall in GFR, the decline in eGFR is less than the decline in GFR and eGFR thus exceeds GFR. Conversely, after a rise in GFR, the rise in eGFR is less than the rise in GFR and eGFR is thus less than GFR. As the serum concentration approaches the new steady state, eGFR approaches GFR. The rate of rise in the marker reflects not only the severity of the reduction in GFR but also the non-GFR determinants.

Below we first describe endogenous filtration markers in current use and investigation and then describe frequently used GFR estimating equations. Of note, serum and plasma concentrations of these markers are equivalent; next, we refer to serum concentrations.

The recommended approach for evaluation of GFR begins with an initial test, followed by supportive tests when indicated. eGFRcr is the initial test as creatinine is widely available, inexpensive, and ordered routinely. ^{, , ,} However, in the steady state in most ambulatory cohorts, accuracy is limited; only 75% to 85% of eGFRcr values are within 30% of mGFR (P ₃₀ ). For clinical settings where greater accuracy is required for clinical decision making or for individuals in whom there are concerns that eGFRcr may be even less accurate, supportive tests are recommended. ^{, , , ,} eGFRcr-cys generally provides the most accurate eGFR (P ₃₀ as high as 90%) and is recommended as the primary supportive test. For clinical settings or individuals requiring even more accurate GFR assessment for clinical decision making, measured GFR using an exogenous filtration marker is recommended.

This approach requires recognition of the potential for error in estimated and measured GFR ( Fig. 23.3 ). The most important sources of error in eGFR are non-GFR determinants of either serum creatinine or cystatin C. Discordance between eGFRcr and eGFRcys can be used as an indicator of the presence of non-GFR determinants of creatinine and cystatin C. Cross-sectional studies show that eGFRcr-cys is generally more accurate relative to mGFR than either eGFRcys or eGFRcr irrespective of the sign or magnitude of the difference between eGFRcys and eGFRcr (eGFRdiff). ^, There are exceptions, such as in otherwise healthy populations with large deviations in the non-GFR determinants of serum creatinine, but not serum cystatin C, such as decreased creatinine generation due to reduced muscle mass, decreased creatinine secretion, or extrarenal elimination due to use of specific medications. In these cases, eGFRcys may be the most accurate.

Glomerular filtration rate (GFR) evaluation using initial and supportive tests.

The algorithm describes the approach to the evaluation of GFR. Our approach is to use initial and supportive testing to develop a final assessment of true GFR and to apply it in individual decision making at each point in time. The initial test for evaluation of GFR is often eGFRcr, which will be available in most patients because creatinine is measured routinely as part of the basic metabolic panel. If the eGFRcr is expected to be inaccurate, or if a more accurate assessment of GFR is needed for clinical decision making, such as diagnosis or staging of CKD or drug dosing, then cystatin C should be measured and the discordance between eGFRcr and eGFRcys should be assessed. If eGFRcr and eGFRcys are not discordant (within 15 mL/min/1.73 m ² or <20%–30% of each other), then accuracy of eGFRcr, eGFRcys, and eGFRcr-cys is similar. If eGFRcr and eGFRcys are discordant (not within 15 mL/min/1.73 m ² or > 20%–30% of each other), then eGFRcr-cys is generally more accurate than either eGFRcr or eGFRcys, with some exceptions, such as otherwise healthy populations with increased creatinine generation owing to increased muscle mass, or decreased creatinine secretion or extrarenal elimination because of use of specific medications, when eGFRcys may be more accurate. If an even more accurate assessment of GFR is needed for a clinical decision, then GFR should be measured using plasma or urinary clearance of exogenous filtration markers, if available. This consideration should be applied any time GFR is required for a clinical decision. It is important to determine how accurate an assessment of GFR needs to be for a clinical decision. P30 for eGFR does not generally exceed 90% (90% of eGFR within 30% of mGFR). P15 for mGFR does not generally exceed 90% (90% of mGFR within 15% of true mGFR). At a GFR of 60 mL/min/1.73 m ² , 30% accuracy for eGFR corresponds to 42 to 78 mL/min/1.73 m ² , and 15% accuracy for mGFR corresponds to 51 to 69 mL/min/1.73 m ² . At a GFR of 30 mL/min/1.73 m ² , 30% accuracy for eGFR corresponds to 21 to 39 mL/min/1.73 m ² and 15% accuracy for mGFR corresponds to 26 to 35 mL/min/1.73 m ² . ∗Use eGFRcr or eGFRcr-cys depending on discordance between eGFRcr and eGFRcys.

Permission granted from Kidney Med for Adingwupu OM, Barbosa ER, Palevsky PM. Kidney Med. 2023;5(12):100727.

This approach also requires that the clinician determine how accurate the assessment of GFR needs to be for clinical decision making. If P ₃₀ of 75% to 85% is acceptable, then eGFRcr may be sufficient, provided there are not large deviations in the non-GFR determinants of serum creatinine. For individuals with expected large deviations in the non-GFR determinants of serum creatinine but not serum cystatin C, eGFRcys may be sufficient. If P ₃₀ of 90% is acceptable, then eGFRcr-cys may be sufficient.

Values for mGFR also may often contain an element of error, which differentiates mGFR from the “true GFR.” Measurement error is related to the specific filtration marker, the clearance method, and analytical errors in the assay. Studies of repeated measurements within an individual show that only 90% of repeat measurements are within 15% of the original value: P ₁₅ of 90%. Thus standardized protocols that accommodate individual patient circumstances, such as reduced level of GFR or volume overload, as well as standardized laboratory measurement procedures, are recommended to ensure the most accurate results.

GFR evaluation in the non–steady state (e.g., during development and recovery from AKI) is less accurate than in the steady state and requires additional considerations. First, as discussed earlier, eGFR lags behind true GFR in the non–steady state due to the time required for concentration of the endogenous filtration marker to reach steady state serum concentration; eGFR is higher than true GFR when GFR is declining, and eGFR is lower than true GFR when GFR is rising. Because of its smaller volume of distribution, serum cystatin C is hypothesized to rise faster than serum creatinine after an acute GFR decline and may be a more sensitive test for the detection of AKI. Second, changes in fluid status may alter the volume of distribution, with fluid retention leading to slower rise in serum concentrations and fluid depletion leading to faster rise in serum concentrations. Third, the effect of non-GFR determinants may be exaggerated in patients with comorbid conditions that are common in AKI, such as decreased creatinine generation due to muscle wasting, and some studies have demonstrated more accurate drug dosing using algorithms based on eGFRcys than eGFRcr. A kinetic eGFRcr equation has been developed and shown to be predictive of kidney recovery in intensive care unit patients and of delayed graft function in kidney transplant recipients, but it significantly underestimates mGFR. More accurate evaluation can be achieved by GFR measurement, although depending on the rapidity of decline in GFR, this test may overestimate GFR. Alternatively, brief-duration urinary Clcr can be performed, assuming that the serum creatinine concentration does not increase rapidly over 2 to 8 hours.

The rationale for needing an alternative approach to GFR evaluation in a specific population or disease is that the currently recommended approach for use in the general population is less accurate in that population than in comparable patients in other populations and that an alternative approach is more accurate in that population. ^, For eGFR equations, variation in accuracy across populations and among subgroups of the population is common and reflects variability in the non-GFR determinants of the serum concentrations of the filtration markers associated with these conditions, which is generally greater for eGFRcr than eGFRcys. For children or pregnancy, specific approaches to GFR evaluation are recommended. In other diseases or conditions, such as oncology, equations that are recommended for the general population are recommended for use. We do not recommend use of disease-specific estimating equations in clinical practice because each patient can have more than one disease, and selecting a particular disease-specific estimating equation would be arbitrary.

Approximately 50% of total urine protein is derived from plasma proteins via glomerular filtration and incomplete tubular reabsorption; the remainder is derived from tubular secretion and production by the lower urinary tract and, in men, by the genital tract. Plasma proteins have a wide spectrum of molecular weight and charge and are traditionally classified as albumin (molecular weight [MW] 68,000 daltons, most anionic) and globulins, including γ globulins (immunoglobulins, MW approximately 150,000 daltons, most cationic) and α ₁ , α ₂ , and ß globulins (MW 15,000–90,000 daltons, intermediate charge). The glomerular filtration barrier acts as a size-, shape-, and charge-dependent permselective molecular sieve, restricting the passage of larger and more anionic macromolecules (MW >20,000 daltons) across the glomerular basement membrane. (See Chapter 4 for additional details.) Consequently, only about 0.01% or 1 g/day of protein passes into the glomerular filtrate, which is composed primarily of low-MW globulins, including α ₂ -microglobulin, ß-2 microglobulin, apoproteins, enzymes, peptide hormones, and immunoglobulin light chains. ^, Filtered serum proteins are broken down by proximal tubular luminal endopeptidases and resorbed primarily via apical endocytic receptors, megalin, and cubilin. ^, Uromodulin (also known as Tamm-Horsfall glycoprotein) is the predominant urine protein not of plasma origin. It is a large glycoprotein (MW 23 × 10 ⁶ Da) formed on the epithelial surfaces of the thick ascending limb of the loop of Henle and early distal convoluted tubule and secreted into the urine. Tubular secretion of IgA and, in men, proteins produced in seminal, prostatic, and urethral secretions contribute a small percent to total urine protein.

Urinary protein loss is often referred to as excretion. Normal mean values for urine protein and albumin excretion rates (PER and AER, respectively) are less than 50 and 10 mg/daily, but because of the wide variability, clinical laboratories generally consider the upper limits of normal to be approximately 150 mg and 30 mg/day ( Table 23.4 ). In addition to AKD and CKD, a number of factors influence AER, likely through alterations in glomerular capillary Starling forces and K _F . AER is higher during upright than supine posture and can increase after exercise, during fever, and with weight gain. ^{, ,} Activation of the renin-angiotensin-aldosterone system and inhibition of tubuloglomerular feedback can increase AER, and medications that interfere with these systems, such as ACE inhibitors, ARBs, MRA, and SGLT2 inhibitors, can lower AER.

Postural proteinuria has been noted in young people without other evidence of kidney disease. It is defined as increased proteinuria with upright posture and normal values with supine posture. It usually does not exceed 1 g total protein in 24 hours. Characterization of the type of protein has not been reported. Kidney histologic examination generally yields normal or nonspecific findings, ^, and patients with postural proteinuria have not been observed to develop kidney disease. Hypothesized mechanisms are that this is a normal variant or that it reflects a subtle glomerular abnormality, an exaggerated hemodynamic response, or left renal vein entrapment. As discussed next, the finding of proteinuria in a young person should prompt a testing of an early morning specimen to exclude postural proteinuria.

For types of urine specimens and techniques for obtaining them, see the following section on urine dipstick.

The recommended approach for evaluation of proteinuria in adults begins with an initial test, followed by confirmatory tests when indicated ( Table 23.5 ). Spot urine ACR is the initial test in adults, but there is a hierarchy of tests if ACR is not available as an initial test, with subsequent confirmation by ACR. AER in a timed urine sample is the generally accepted confirmatory test for urine ACR but is performed only when urine ACR is not accurate enough for clinical decision making. Equations are available to convert measures of proteinuria to ACR or AER, but accuracy is limited at the lower range (ACR <500 mg/day).

As discussed earlier, measurement of urine albumin and not urine protein is recommended in adults. ^, However, this approach may overlook “tubular” and “overflow” proteinuria, in which nonalbumin proteins predominate. If these conditions are suspected, they may be detected by measurement of total urine protein in addition to urine albumin; increased level of total urine protein with albumin-protein ratio of <0.4 suggests presence of an increased level of a nonalbumin protein. Confirmation of tubular proteinuria can be obtained by immunoassays directed at specific low-MW proteins in urine. Confirmation of immunoglobulin light chains is best obtained by measuring serum assays for immunoglobulin light chains.

This approach requires recognition of the potential for errors in measurement of urine albumin and total urine protein ( Table 23.6 ). Because of substantial day-to-day biological variability in urine albumin and urine protein, a positive result should be followed with a second measurement, ideally in an early morning sample, to confirm the result. ^,

Urine specimens for dipstick and microscopy can be obtained by spontaneous voiding, urethral catheterization, and percutaneous bladder puncture. Technique is important in collecting a sample to avoid contamination. For spontaneously voided urine, a midstream sample should be collected after cleaning of the external genitalia. If a patient has an indwelling catheter, a fresh specimen should be submitted for analysis; samples that have been stagnant in the catheter tubing or bag may have undergone degradation. Suprapubic needle aspiration of the bladder is used when urine cannot easily be obtained by other means, most commonly in infants. Whatever the collection method, it is recommended that a sample be analyzed within 2 to 4 hours of the time of collection to avoid precipitation of solutes and cell lysis.