Clinical Assessment of Renal Function

Published on 22/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 2104 times

108 Clinical Assessment of Renal Function

Five to 15 percent of patients in intensive care units (ICUs) experience acute deterioration in renal function.^1,² Conversely, renal dysfunction adds substantially to the morbidity and mortality of these patients. Moreover, changes in renal function directly affect drug disposition. Thus, a means to assess renal function is essential for optimal management. The glomerular filtration rate (GFR) is the standard measure of renal function. It reflects overall renal functional capacity and, in renal failure, correlates with structural damage to the kidney. This chapter reviews selected aspects of renal physiology with an emphasis on measurement of renal function, consequences of altered function, and approaches to improving renal function. The focus is on measurement and optimization of glomerular filtration rate (GFR) and renal blood flow (RBF).

Renal Blood Flow

Under physiologic conditions, blood flow to the kidneys is 20% of cardiac output. This high rate of blood flow (1-1.2 L/min) is particularly remarkable in that the kidneys make up only 0.5% of total body weight. The high blood flow rate is due, at least in part, to the unique anatomic arrangement of the renal vasculature, with the interlobar and arcuate vessels offering little resistance to flow. This in turn is because the interlobular arteries originate from the arcuates in a parallel arrangement and because the afferent arterioles also arise in a parallel arrangement from the interlobular vessels. It is this parallel arrangement that accounts for the low resistance, because the total resistance of n equals parallel paths, each with a resistance R, is R/n³. Major resistance vessels in the kidney are the afferent and efferent arterioles that bound the glomerular capillary network. Although total resistance is a function of resistance across each of these vessels, it is a unique feature of the kidney that variations in the individual resistances across the afferent and efferent arterioles, respectively, may lead to alterations in glomerular capillary pressure and, hence, in GFR.³

Despite a wide range of perfusion pressures, RBF and GFR are maintained relatively constant, a process described as autoregulation. The term autoregulation generally refers to the relative constancy of GFR over a range of perfusion pressures but also refers to the regulation of RBF. Emphasis has been placed on the preglomerular vasculature, mainly the afferent arterioles, as the major site at which renal perfusion is regulated. However, studies also suggest that the larger vessels, such as the interlobular vessels, may respond to a variety of vasoactive stimuli and participate in an autoregulatory phenomenon. A variety of hypotheses have been generated to explain the autoregulatory response of the kidney with respect to RBF. There is evidence to suggest mediation by neural, humoral, or intrarenal factors that regulate the renal circulation.⁴

The renin-angiotensin pathway has a significant effect on renal hemodynamics. Renin, elaborated in the juxtaglomerular cells, may be released in response to a decrease in renal perfusion pressure and to altered sodium chloride delivery to the ascending limb and macula densa cells. Increased renin secretion in turn leads to augmented angiotensin II (AII) formation at the local nephron level. AII, in turn, affects renal vascular resistance by an effect on both the afferent and efferent arterioles, with the effect predominating on the latter vessels.

Renal eicosanoids also affect renal hemodynamics. Eicosanoids are biologically active fatty acid products of arachidonic acid and are synthesized in the kidney in response to a variety of stimuli, with local release and effect on the renal vasculature. Stimulation of the cyclooxygenase pathway and prostaglandin synthetases leads to the formation of endoperoxides (PGG₂, PGH₂), prostaglandins (PGD₂, PGE₂, PGF_2α, PGI₂), and thromboxane A₂, (TXA₂). Leukotrienes are synthesized by another major pathway involving the enzyme, lipoxygenase. In the kidney, the major products of arachidonic acid metabolism are PGE₂ and PGI₂ and, to a lesser extent, PGI_2α. These compounds have a predominant effect of relaxing renal vascular smooth muscle and lead to vasodilatation, whereas TXA₂ is a vasoconstrictor prostanoid. It is believed that in disease states, endogenous vasodilator prostaglandins serve a protective function to maintain renal perfusion and GFR in response to vasoconstrictor stimuli, including AII and enhanced sympathetic nervous system activity. In contrast, release is inhibited by nonsteroidal antiinflammatory drugs.

Other vasoactive compounds that affect the renal circulation include the plasma and glandular kallikreins and kinins and endothelium-derived vasoactive factors such as nitric oxide and endothelin.⁴ Among the catecholamines, α- and β-adrenergic agonists are known to affect renal vascular tone by causing vasoconstriction and vasodilatation, respectively. In addition, dopamine in low doses leads to renal vasodilatation. Emphasis has more recently been placed on atrial natriuretic peptide and purinergic agents such as adenosine. The effect is likely to be influenced by changes in salt intake and extracellular fluid volume as well as by hydration status. For example, the influence of AII on renal hemodynamics is greater in sodium depletion, which also activates the sympathetic nervous system. In response to mild nonhypotensive hemorrhage, renal hemodynamics are relatively well maintained. However, with further reductions in volume associated with a more severe hemorrhage, renal ischemia mediated by activation of the renin-angiotensin system, renal efferent adrenergic nerves, and circulating catecholamines may occur.⁴

Finally, modification of dietary protein and amino acid intake may affect renal hemodynamics. Dietary protein intake in excess of 1 g/kg/d has been associated with renal vasodilatation, as have infusions of casein hydrolysates and amino acids.^5,⁶ Conversely, chronic consumption of a low-protein diet may be associated with renal vasoconstriction.

Measurement of Renal Blood Flow

Renal blood flow is measured conventionally by the clearance of infused para-aminohippurate (PAH), which is cleared almost totally from the arterial plasma by both filtration and secretion. Thus, its clearance approximates the rate of renal plasma flow (RPF):

where U_PAH and P_PAH refer to urine and plasma PAH concentration, respectively, and V is urine flow rate in milliliters per minute.

RBF can be estimated by correction for the hematocrit (Hct):

Although available, this test is rarely used in clinical practice. In fact, direct quantitation of RPF and RBF is rarely indicated outside research studies; however, sometimes it is necessary to document that the kidneys are being perfused. In this case, one of three additional methods may be utilized: (1) selective arteriography, including CT angiography and MR angiography, (2) Doppler ultrasonography, and (3) external radionuclide scanning.

Because the latter two methods are noninvasive, they are preferred. With respect to the nuclide study, until recently, scanning was usually performed utilizing ¹²⁵I-iodohippurate sodium; however, the poor radiologic characteristics of ¹³¹I limit its use in renal imaging.⁷ More recently, other agents such as ¹²⁷I-orthoiodohippurate and ^99mTc-L, L-ethylenedicysteine may prove to be superior.^7,⁸

Clinical Correlates

Although a significant body of data has been obtained to indicate a complex relationship between neurocirculatory factors and renal hemodynamics, several points can be made from a clinical perspective. Optimization of cardiac output and extracellular fluid (ECF) volume, including the intravascular space, is essential for the maintenance of renal perfusion. Particularly because the effects of vasoactive compounds such as AII and catecholamines are accentuated in the presence of renal hypoperfusion and volume contraction, attention should be directed to an assessment of ECF volume, with correction of any deficits, and to optimizing cardiac function. Frequently, pharmacologic agents have been employed to maintain renal perfusion in situations in which this may be compromised. Specifically, there has been widespread use of so-called low-dose or renal-dose dopamine infusions. This is based on the observation that in low doses (<3 µg/kg/min) dopamine leads to renal vasodilatation.⁹ At higher doses, renal vasoconstriction may occur.

The beneficial effects of dopamine infusion have not been documented in patients who are depleted of sodium chloride and volume, and the use of dopamine has not been shown to be effective beyond a short period of infusion.⁹^–¹¹ That is, infusions of renal-dose dopamine for 24 to 36 hours may be beneficial in the appropriate circumstance, but there is no evidence supporting the long-term use of this agent. Thus, justification for prolongation of its use beyond several days is not supported by available data. Furthermore, reports suggest that adverse outcomes may be associated with the use of dopamine.¹¹ Continuous infusions of fenoldopam mesylate, a potent dopamine A-1 receptor agonist, have been employed in an attempt to preserve renal function in a variety of clinical settings. A meta-analysis of 16 randomized trials in critically ill patients showed that fenoldopam significantly reduced the risk of acute kidney injury, need for renal replacement therapy, and in-hospital death.¹² Beyond anecdotal evidence, there are no compelling data to support the use of other potential vasodilator substances such as prostaglandins. Although high-protein feeding and amino acid infusions may increase RBF by an undefined mechanism, there is no justification for utilizing these therapies solely from a hemodynamic point of view.^5,⁶

Glomerular Filtration Rate

Of the 500 to 700 mL of plasma delivered per minute to the kidneys (corresponding to a renal blood flow of 1-1.2 L/min), 20% to 25% is filtered. Glomerular filtration is a major function of the kidney and averages approximately 130 mL/min/1.73 m² in normal males and 120 mL/min/1.73 m² in females. Estimation or direct assessment of GFR remains one of the most important measurements of renal function and is widely utilized in clinical practice.

Measurement of Glomerular Filtration Rate

GFR is classically measured as the clearance of inulin (C_In), a fructose polymer with a mean molecular weight of approximately 5 kD. Because this substance is not present endogenously, it must be given by constant infusion after a loading dose. Inulin is available commercially but is expensive, often difficult to obtain, and cumbersome to utilize. As a result, C_In is rarely used in clinical practice except for research protocols. Although inulin is generally measured chemically, ³H-labeled and ¹⁴C-labeled inulin are also available but are expensive.

More recently, other radiolabeled nuclides have been found to be satisfactory substitutes for inulin and have advantages in the measurement of GFR.^7,^8,^13,¹⁴ Particularly ^99mTc-labeled diethylenetriamine pentaacetic acid (DTPA) and ¹²⁵I- or ¹³¹I-labeled iothalamate clearances closely approximate the C_In.^15,¹⁶ ^99mTc-DTPA has been utilized and found to give measurements that correlate closely with C_In in ICU patients.^17,¹⁸ In addition, the clearance of gentamicin has been utilized in a limited fashion to measure GFR.^19,²⁰ At the present time it is not common for GFR to be measured directly. Rather, GFR is estimated by the endogenous creatinine clearance or serum creatinine determination (see later).

The normal values for GFR given previously apply for individuals from the teenage years through approximately age 35. Thereafter, GFR declines in most individuals. Whereas this decline was formerly thought to occur at a relatively constant rate of approximately 10 mL/min per decade,²¹^–²³ more recent data obtained in a longitudinal fashion indicate that this reduction is not so predictable.²⁴ In addition, a circadian rhythm for GFR has been described.^25,²⁶ GFR is maximal in the daytime, whereas a minimal value during the night has been found in normal individuals. Whether this circadian pattern of GFR occurs in critically ill hospitalized patients is not known.

Creatinine Clearance and Serum Creatinine

Creatinine Clearance

The endogenous creatinine clearance (C_Cr) enjoys widespread use as a reasonable gauge of GFR when great precision is not demanded, which it rarely is in clinical practice. The use of creatinine as a marker of GFR has the advantage that creatinine is endogenously produced and is easily measured by inexpensive methods. Creatinine, like inulin, is freely filtered and absorbed minimally if at all by the tubules. However, creatinine is secreted, and the contribution of secretion to total excretion is greater as the GFR decreases and serum creatinine rises. At GFRs below 40 mL/min, C_Cr exceeds C_In by 50% to 100%.^15,²⁷ When GFR is significantly depressed and it is deemed important to get a more precise measurement of GFR, one of the previously mentioned methods to estimate GFR directly might be utilized. Additionally, because C_Cr overestimates GFR and the clearance of urea underestimates GFR, the mean value of simultaneously obtained creatinine and urea clearances has been shown to provide a close estimation of C_In when the latter is below 20 mL/min.²⁸

Because cimetidine competes with creatinine for tubular secretion (see later), administration of cimetidine may increase the accuracy both of creatinine clearance in 24-hour collections (when given for several days beforehand) and of 4-hour, water-loaded clearances.²⁹^–³¹ Taking advantage of this effect results in a more accurate estimate of GFR. Specifically, C_Cr obtained in the presence of cimetidine (400 mg as a priming dose followed by 200 mg every 3 hours) yielded values that closely approximated C_In.^29,³⁰ Volume expansion in humans causes a small rise in GFR, whereas volume depletion, severe heart failure, hypotension, anesthesia, surgery, trauma, sepsis, and even mild intestinal bleeding without frank hypotension may depress GFR substantially.

Various methods are available to measure creatinine. Creatinine is frequently measured using the Jaffé alkaline picric acid reaction. Although this method is widely utilized, this reaction also measures other chromogens, which may lead to a false elevation in the estimated serum creatinine (S_Cr) measurement. Substances such as acetoacetate (in ketoacidosis), pyruvate, ascorbate, 5-flucytosine, certain (but not all) cephalosporin antibiotics, and very high urate artifactually raise S_Cr in normal subjects by 0.5 to 2 mg/dL.³²^–³⁸ These substances are excreted into the urine but contribute trivially compared with overall urine creatinine (U_Cr). Thus, noncreatinine chromogens affect the S_Cr but have little effect on the U_Cr.

In individuals with normal renal function, the contribution of serum noncreatinine chromogens to raising the S_Cr is approximately equal to the contribution of secretion to creatinine excretion, such that the C_Cr closely approximates GFR. As GFR decreases, the contribution of noncreatinine chromogens to the total measured S_Cr becomes less than the secreted moiety, and the C_Cr overestimates GFR to a greater extent. Direct enzymatic creatinine measurements are not affected by noncreatinine chromogens. Very high levels of serum glucose (>1000 mg/dL) and 5-flucytosine may interfere with the enzymatic reaction, whereas high levels of bilirubin (>5 mg/dL) affect the autoanalyzer method ³⁶ and lead to falsely low S_Cr values. It is therefore important to know the method by which a given laboratory measures S_Cr. Competing for the same proximal tubular organic base secretory site as creatinine, certain pharmacologic agents may suppress this process and lead to a rise in S_Cr. Trimethoprim, probenecid, and cimetidine, but not ranitidine, are organic bases that inhibit creatinine secretion competitively and can result in a mild elevation in S_Cr, usually 0.5 mg/dL or less.³⁹^–⁴²

As with all clearance methods, the C_Cr is subject to errors that may amount to as much as 10% to 15% or more. In addition to potential problems in estimating S_Cr and U_Cr, errors in timing of urine collection, incomplete collection, and inaccurate measurement of urine volume are other factors that contribute to errors.⁴³ Although 24-hour U_Cr clearances have been widely utilized, no specified time period is required for the clearance to be obtained. In fact, shorter collection periods of several hours may be more accurate in patients passing adequate amounts of urine (not oliguric), particularly if the patient is not in a steady state (see later). To reduce errors in volume measurement, one can induce a water diuresis in stable subjects before beginning the test,⁴⁴ although this is rarely practical in the ICU setting. Nevertheless, because many ICU patients have indwelling Foley catheters, it should be possible for accurately timed urine collections to be obtained and for C_Cr to be measured with reasonable accuracy.

Serum Creatinine

Because of the practical and technical problems in obtaining estimates of GFR by clearance methods, renal function is most commonly estimated by following the S_Cr in hospitalized patients. Creatinine is formed nonenzymatically from creatine and phosphocreatine in muscle cells and is normally present in the serum at a concentration of 0.8 to 1.4 mg/dL in adults and 0.3 to 0.6 mg/dL in children and pregnant subjects. The measured S_Cr depends on the method of measurement, as discussed previously, GFR, rate of creatinine production, volume of distribution (e.g., S_Cr is lower in anasarca), and extent of its tubular secretion and intestinal degradation.³ Because creatinine production is closely related to muscle mass, S_Cr is generally less in females than in males and decreases as muscle mass is lost with aging or with debilitating illnesses.

The relationship between S_Cr and C_Cr (and hence GFR) can be described by a rectangular hyperbola ⁴³; however, this relationship applies in the steady state and assumes a constant rate of creatinine production (Figure 108-1). Thus, a doubling of the S_Cr, reflects a 50% decrease in C_Cr, a fourfold increase in S_Cr, a 75% drop in GFR, and so on. Because creatinine production may not remain constant, S_Cr may underestimate the decrease in GFR in critically ill patients who have a decrease in muscle mass secondary to an ongoing catabolic state. Moreover, it should be appreciated that S_Cr is an insensitive marker of change early in the course of renal disease. Thus, a 33% fall in GFR may raise the S_Cr from 0.8 to 1.2 mg/dL, a value still within the normal range. If the prior value is not known, this fall in GFR may go unrecognized.

Figure 108-1 Relationship between creatinine clearance and serum creatinine. In steady state, serum creatinine should increase twofold for each 50% reduction in creatinine clearance. Inset represents enlarged view of changes in serum creatinine as creatinine clearance decreases from 120 to 60 mL/min. If serum creatinine is 0.8 mg/dL when creatinine clearance is 120 mL/min, creatinine clearance can decrease by 33% such that increased serum creatinine is still within normal range.

S_Cr provides a close estimate of GFR only in the steady state. With an abrupt decrease in GFR, as may occur in acute renal failure, creatinine production would be expected to continue unchanged, but because of the decrease in GFR, creatinine excretion will be impaired. As a result, the S_Cr increases until a new steady state is obtained, at which time the amount of creatinine produced equals the amount filtered (GFR − S_Cr) and excreted (U_Cr − V). Depending on the extent of damage and decrease in GFR, it may take several days for a new steady state to be achieved (Figure 108-2). Therefore, following an insult leading to an abrupt decrease in GFR, the S_Cr rises progressively over the next several days. This should not be interpreted as a new insult each day, but rather that a steady state has not yet been obtained. While the S_Cr is changing, its absolute value cannot be used as an accurate measure of the decrease in GFR. If an accurate measurement of GFR is needed during this time, a short C_Cr can be obtained.

Figure 108-2 Expected changes in serum creatinine resulting from acute fall in glomerular filtration rate (GFR) and attainment of a new steady state. Between days 0 and 1, patient is excreting all creatinine produced, and serum creatinine is stable at 1 mg/dL. A 50% reduction in GFR on day 1 results in abrupt fall in filtered (and, therefore, excreted) creatinine. Release of creatinine from muscle remains constant; as a result, creatinine is retained and its serum concentration is increased. As creatinine concentration rises progressively, filtered (and excreted) creatinine also increases until excreted creatinine returns to control levels and matches creatinine production. New steady state (days 3 to 4) is achieved by doubling of serum creatinine concentrations, which maintains filtered creatinine load at control levels in the face of halving of GFR. Larger decrease in GFR would lead to greater increase in steady state (e.g., 90% reduction in GFR would lead to 10-fold rise in serum creatinine) and would take longer to achieve.

(From Kassirer JP. Clinical evaluation of kidney function-glomerular function. N Engl J Med 1971;285:385. Reprinted with permission from The New England Journal of Medicine.)

A variety of equations have been developed to estimate C_Cr based on the S_Cr without collection of urine.^45,⁴⁶ Table 108-1 is a compilation of the more commonly used equations.⁴⁷ These equations generally take into consideration muscle mass (estimated as body weight), sex (males having a higher GFR than females), and age. Aging, hepatic diseases, excessive muscle wasting, severe muscular atrophy or dystrophy, hyperthyroidism, paralysis, and chronic glucocorticoid therapy have been associated with reduced creatinine generation.¹⁷ In addition, particularly at low levels of GFR, correction for nonrenal creatinine metabolism is also recommended.^48,⁴⁹ One of the most commonly utilized equations is that developed by Cockcroft and Gault⁵⁰:

TABLE 108-1 Common Equations for Estimating Glomerular Filtration Rate or Creatinine Clearance

Cockcoft-Gault (C_Cr · BSA/1.73 m²)
For men: C_Cr = [(140 − age) · weight (kg)]/S_Cr · 72
For women: C_Cr = ([(140 − age) · weight (kg)]/S_Cr · 72) · 0.85

MDRD (1)
GFR = 170 · [S_C_r]^–0.999 · [age]^–0.176 · [0.762 if patient is female] · [1.18 if patient is black] · [BUN]^–0170 · [Alb]^0.318

MDRD (2)
GFR = 186 · [S_C_r.]^–1.154 · [age]^–0.203 · [0.742 if patient is female] · [1.212 if patient is black]

Jellife (1) (C_Cr · BSA/1.73 m²)
For men: (98 − [0.8 · (age − 20)])/S_Cr
For women: (98 – [0.8 · (age − 20)])S_Cr · 0.90

Jellife (2)
For men: (100/S_Cr) − 12
For women: (80/S_Cr) − 7

Mawer
For men: weight · [29.3 − (0.203 · age)] · [1 − (0.03 · S_Cr)]
For women: weight · [25.3 − (0.175 · age)] · [1 − (0.03 · S_Cr)]

Bjornsson
For men: [27 − (0.173 · age)] · weight · 0/S_Cr
For women: [25 − (0.175 · age)] · weight · 0.07/S_C_r

Gates
For men: (89.4 · S_C_r^–1.2) + (55 − age) · (0.447 · S_C_r^–1.1)
For women: (89.4 · S_Cr^–1.2) + (55 − age) · (0.447 · S_C_r^–1.1)

Salazar-Corcoran
For men: [137 − age] · [(0.285 · weight) + (12.1 · height²)]/(51 · S_Cr)
For women: [146 − age] · [(0.287 · weight) + (9.74 · height²)]/(60 · S_Cr)

where age is expressed in years. The preceding expression is used for men. The formula for women is the preceding formula multiplied by 0.85.

The reliability of this equation as a measure of GFR has been assessed in patients with diabetes, pregnant women with renal disease,⁵¹ obese individuals,⁵² elderly individuals,^53,⁵⁴ and black Americans with hypertensive renal disease.⁵⁵ It has also been assessed in critically ill patients.⁵⁶ These studies have indicated that the accuracy of GFR estimates using the Cockcroft-Gault equation is similar to or greater than 24-hour C_Cr, and the precision is better. This equation seems to be most accurate for estimating GFR when the latter is in the range of 10 to 100 mL/min.^52,^55,⁵⁶ The advantage of this formula is that it is simple and underscores the essential determinants of C_Cr.

The MDRD (Modification of Diet in Renal Disease) study equation has gained widespread acceptance by most clinical laboratories, which now routinely report estimated GFR values for blacks and nonblacks when a serum creatinine is ordered.⁵⁷^–⁵⁹ Its major limitations are imprecision and underestimation of measured GFR at high GFR values (GFR > 60 mL/min/1.73 m²). The MDRD equation is generally more precise than the Cockroft-Gault equation.⁶⁰

Limitations at higher GFR values prompted a recent modification by the Chronic Kidney Disease Epidemiology Collaboration Research Group.⁶¹ This equation offers improved precision, especially with higher GFR values up to 90 mL/min/1.73 m².

CKD-EPI equation for estimated GFR (natural scale):

Blacks:

Female (S_Cr ≤0.7) GFR=166·(S_Cr/0.7)^–0.329 · (0993)^Age

Female (S_Cr >0.7) GFR = 166·(S_Cr/0.7)^–1.209 · (0993)^Age

Male (S_Cr ≤0.9) GFR = 163·(S_Cr/0.9)^–0.411 · (0993)^Age

Male (S_Cr >0.9) GFR = 163·(S_Cr/0.9)^–1.209 · (0993)^Age

White or other:

Female (S_Cr ≤0.7) GFR = 144·(S_Cr/0.7)^–0.329 · (0993)^Age

Female (S_Cr >0.7) GFR = 144·(S_Cr/0.7)^–1.209 · (0993)^Age

Male (S_Cr ≤0.9) GFR = 141·(S_Cr/0.9)^–0.411 · (0993)^Age

Male (S_Cr >0.9) GFR = 141·(S_Cr/0.9)^–1.209 · (0993)^Age

Serum Urea Nitrogen

Less accurate as a marker of GFR than the S_Cr, serum urea nitrogen (SUN) (or blood urea nitrogen [BUN]) is still used extensively in clinical practice to estimate renal function. Although this was the earliest available indicator of renal function, several other factors should be appreciated regarding the use of this substance. Urea, like creatinine, is freely filtered and is retained in the blood as GFR falls. However, in contrast to creatinine, urea may be reabsorbed to a significant extent, its excretion tending to be increased with increasing urine flow rates, whereas its excretion is reduced when tubular fluid reabsorption is enhanced. Of greater importance, urea production is more variable than creatinine. Produced in the liver, urea increases with high protein intake, amino acid infusions, and hypercatabolic states. In addition, endogenous sources of protein such as absorbed hemoglobin from gastrointestinal bleeding may contribute to increased urea synthesis. Even at a constant GFR, SUN may rise in subjects on high protein intake and fall with protein restriction or on refeeding of previously starved, nonhypercatabolic subjects.

Several pharmacologic agents also may affect urea nitrogen formation. Tetracyclines may lead to an increase in SUN by an anti-anabolic effect without any detectable change in GFR, whereas glucocorticoids and severe illnesses or trauma do the same by inducing endogenous protein hypercatabolism. Because of the widespread use of hyperalimentation in ICU patients, an impairment in renal function is often associated with a marked disproportion in the elevation of SUN compared with S_Cr. For this reason, the issue is raised as to whether SUN elevation itself poses an important threat to the patient if the GFR is in a range that should not lead to enhanced morbidity by itself. In those circumstances, it is useful to measure the rate of urea appearance (or generation) to estimate whether other factors such as gastrointestinal bleeding, excessive amino acid infusions, and protein administration are contributing to the increase in SUN above that expected by the decrease in GFR.^47,⁴⁸ Urea nitrogen (UN) appearance can be determined from urine urea nitrogen (UUN), SUN, and body weight as follows:

where UUN • V is 24-hour UN excretion, and Δ body pool UN = 0.6 − nonedematous weight (kg) • Δ SUN/day.

If the weight is changing ^47,⁴⁸:

Nitrogen balance (BN) is equal to:

where IN is urea nitrogen intake, and NUN is nonurea nitrogen excretions.⁴⁸

NUN, which includes fecal nitrogen, urinary creatinine, uric acid, and unmeasured nitrogen, averages 0.031 g nitrogen/kg/d.⁴⁸ The data obtained from the just-described measurements may be quite useful in evaluating the cause of disproportionate elevations in SUN. If the patient is in a steady state (with a stable weight and SUN), BN = 0, and IN can be estimated from UN + NUN.⁴⁸ Because catabolism, except for severe trauma and burns, is usually 2 to 4 g nitrogen per day, additional conclusions can be drawn if the patient is not in the steady state. For example, if it is known that IN is less than UN + NUN, gastrointestinal bleeding with or without excess catabolism would be suggested. Similarly, one can evaluate if the increase in SUN is a reflection of excessive exogenous protein and amino acid administration (usually >1.5 g/kg/d; g UN 0.16 = g protein or amino acids). If IN is above UN, such as in severe liver disease, the clinician might more carefully evaluate changes in weight and SUN as well as clearances, because the latter may be more severely depressed than initially suspected.

Sodium Balance and Extracellular Fluid Volume

Sodium is the primary cation of the ECF, present in a concentration of 140 to 142 mmol/L. The volume of the ECF is approximately 20% of total body weight and represents a third of total body water. Regulation of ECF volume is governed by factors regulating sodium balance and sodium excretion. The reader is referred to an excellent review on this topic.⁶² For the purposes of this discussion, several factors are emphasized. Under physiologic conditions and in the steady state, sodium balance is maintained because the amount of sodium excreted equals that which enters the body by oral and intravenous routes. Sodium excretion and the fraction of filtered sodium that is excreted (FE_Na) can be readily determined. Absolute sodium excretion is measured as the product of the urine sodium concentration and the urine volume:

FE_Na can be determined as follows:

For practical reasons, the C_Cr (= U_Cr • V/S_Cr) is used to estimate GFR, such that:

Because the V term in the numerator and denominator cancels out:

Thus, FE_Na can be calculated from the sodium and creatinine determined in a random urine sample and serum (or plasma) simultaneously. The resulting calculation is expressed as a percentage by multiplying by 100. This test is of value in the setting of acute renal failure to aid in distinguishing a prerenal from a renal parenchymal etiology.⁶³ It is not usually helpful in aiding in the diagnosis of urinary tract obstruction or in the presence of underlying chronic renal insufficiency. The reason for the difficulty in interpretation in chronic renal insufficiency can be illustrated by the following considerations. At a GFR of 130 mL/min and a dietary sodium intake of 3 g of sodium (130 mmol), an individual in sodium balance will excrete 0.5% of the filtered load (FE_Na = 0.5%). For sodium balance to be maintained at lower levels of GFR with the same sodium intake, FE_Na must be increased progressively. Successive decreases in GFR by 2 from 130 would result in an FE_Na of 1%, 2%, 4%, and 8%, respectively. Thus, interpretation of the FE_Na in a patient with acute renal failure superimposed on chronic renal insufficiency is problematic unless the prior steady-state FE_Na is known. This is rarely the case.

The fractional excretion of chloride (FE_Cl) has been suggested to be more accurate than that of sodium in helping to distinguish prerenal from parenchymal causes of acute renal failure.⁶⁴ This is particularly so in the situation in which acute renal failure occurs with simultaneous metabolic alkalosis. If the urine contains substantial amounts of bicarbonate urinary pH (U_pH > 7), sodium excretion increases to maintain electroneutrality. Under these circumstances, the FE_Na may give misleading information, but the FE_Cl can be used to obtain the same information.

Although urinary sodium excretion can be used to help make determinations with respect to ECF volume under certain circumstances, this may be fraught with potential errors. No laboratory test is available to provide this information. Rather, the astute clinician must rely on bedside evaluation complemented, where appropriate, with measurements of central venous pressure and pulmonary capillary wedge pressure to assist in making determinations with respect to ECF volume status. For example, a low FE_Na (<1%) in the setting of acute renal failure usually indicates a decrease in renal perfusion but does not provide information on the status of the patient’s ECF volume. Because a low FE_Na can be seen with either ECF volume contraction or severe congestive heart failure, these conditions must be distinguished at the bedside. Moreover, sometimes a low FE_Na exists even in the presence of parenchymal renal disease, such as acute glomerulonephritis, severe burns, and radiocontrast nephropathy. Finally, administration of potent diuretic agents can alter the FE_Na and may result in misleading interpretations. For this reason, urine samples should be obtained before diuretics are administered. However, it may not be possible to obtain urinary sodium or chloride values while a patient is not receiving diuretics. In this setting the fractional excretion of urea nitrogen has been employed to distinguish prerenal from renal causes of acute kidney injury. In a well-hydrated individual, the Fe_UN is 50% to 65%,⁶⁵ whereas in oliguric prerenal azotemia, the Fe_UN is below 35%. The use of Fe_UN in the setting of acute kidney injury has not attained widespread acceptance owing to variable results on comparative trials.^65,⁶⁶

A few additional points are worthy of note with respect to diuretic use. There is now ample evidence that in a patient in positive sodium balance, diuretic therapy should not be utilized without simultaneously restricting sodium intake, including intravenous saline, if negative sodium balance and reduction in edema fluid are desired.⁶⁷ In general, this requires restriction of dietary sodium intake, usually to less than 2 g of sodium per day (0.88 mmol) if the patient is in an edema-forming state. Although a diuresis can be effected even with liberal sodium intake, this requires higher doses of diuretics and more frequent administration of these agents. The coexistence of hyponatremia should not deter clinicians from restricting sodium intake, but rather should cause them to address solute-free water intake as well. Of course, under certain circumstances, obligatory intakes make it difficult to achieve optimal restriction to assist diuresis. That is, with various pharmacologic drips, blood products, and feeding regimens necessary in acutely ill patients in the ICU, this may become a difficult problem. Under those circumstances, increasing doses of diuretics, including continuous infusions of loop diuretics, may be required.