Kidney Function and Damage

Jean-Pierre Braun

Department of Physiopathology and Experimental Toxicology
National School of Veterinary Medicine of Toulouse
Toulouse, France

Herve P. Lefebvre

Department of Physiopathology and Experimental Toxicology
National School of Veterinary Medicine of Toulouse
Toulouse, France

Clinical biochemistry in nephrology is mainly used to diagnose and monitor renal dysfunction or damage. This is of special importance in human medicine, because of the frequent observation of renal failure in elderly people, and in canine and feline medicine, especially for the early detection of chronic renal failure (CRF), which is frequent. In U.S. private practice, however, renal disease was not reported as one of the 29 most common disorders in dogs and was ranked 17th in cats (Lund et al., 1999), and in a survey of Australian practices, the expected frequency of renal failure ranged from one case per week to one case every 2 weeks (Watson et al., 2001). A British survey indicated that 0.2% of dogs were presented with suspected renal disease, which was confirmed in 25% of the cases with no evidence of breed or sex predisposition (MacDougall et al., 1986). The reported prevalence of CRF in a U.S. university hospital was three times higher in cats than in dogs (Polzin et al., 1992). However, the frequency of renal disease increases with age, especially in dogs and cats, as in humans (Cowgill and Spangler, 1981; Gobar et al., 1998; McCall Kaufman, 1984). In U.S. veterinary university hospitals, renal disease was observed in 15% of dogs more than 10 years old, in 33% of cats more than 15 years old (Polzin et al., 1989), in 3% of cats aged 7 to 10 years, and in 30% of cats more than 15 years old (Krawiec and Gelberg, 1989).

Kidney diseases are uncommon in equids and cattle (Fetcher, 1985). The latter show better resistance to a loss of kidney function than monogastric animals because of the filtration function of the rumen epithelium as shown by bilateral nephrectomy (Fetcher, 1986).

The mammalian kidney consists of tens of thousands to millions of nephrons that function as parallel units. The larger the species, the greater the number of nephrons per kidney (Kunkel, 1930; Rytand, 1937–1938; Vimtrup, 1928). This ranges from about 10,000 in mice (Cullen-McEwen et al., 2003), 175,000 in cats (Brown et al., 1993), 300 to 700,000 in dogs (Finco and Duncan, 1972), and 7 million in elephants, as compared to about 1 million in humans. The number of nephrons progressively increases during fetal development and is complete at birth (e.g., in sheep or humans) or during the few days following birth (e.g., in rat) (Gimonet et al., 1998). The number of nephrons in the dog decreases slightly (5%) during the 2 first months of life, whereas the glomerular volume increases by 33% (Horster et al., 1971). However, great interindividual variability is observed within species. In dogs, the size or the weight has little influence on the number of glomeruli, but the size of these latter is larger in larger breeds (Finco and Duncan, 1972; Kunkel, 1930). In sheep, the number of nephrons in twins was about 30% lower than in single lambs (Mitchell et al., 2004) or showed no difference (Bains et al., 1996).

FIGURE 16-1 Schematic representation and international nomenclature of the parts of the nephron. 1: Renal corpuscle including Bowman’s capsule and the glomerulus (glomerular tuft). 2: Proximal convoluted tubule. 3: Proximal straight tubule. 4: Descending thin limb. 5: Ascending thin limb. 6: Distal straight tubule (thick ascending limb). 7: Macula densa located within the final portion of the thick ascending limb. 8: Distal convoluted tubule. 9: Connecting tubule; 9*: Connecting tubule of the juxtamedullary nephron that forms an arcade. 10: Cortical collecting duct. 11: Outer medullary collecting duct. 12: Inner medullary collecting duct. Reproduced with permission of the publisher from Kriz and Bankir (1988).

FIGURE 16-2 Electron micrograph of the glomerular filtration barrier. 1: Footlike processes of the podocytes; arrow shows the slit diaphragm of a filtration slit. 2: glomerular basement membrane. 3: fenestrated endothelium of the glomerular capillaries.

The RBF remains quite stable, because of autoregulation, even with variations in systemic blood pressure. The precise mechanism of autoregulation is unknown but results from vasoconstriction/dilation of the afferent and efferent glomerular arterioles, which maintains an almost constant hydrostatic pressure within the glomerulus. Autoregulation is efficient in healthy dogs between 70 to 180 mmHg and may be altered in disease, especially during CRF (Brown et al., 1995).

FIGURE 16-3 Pressures involved in glomerular filtration within a glomerular capillary. The driving force is the hydrostatic pressure in the capillary (Phc), which is opposed by hydrostatic pressure within Bowman’s space (Phb) and oncotic pressure in the capillary (Po). Values are pressures in dogs and cats.

1. Parts of the tubule

• Urine composition can show large variations in the same healthy or diseased subject.

• A notable overlap of urinary analyte concentrations is observed in healthy and diseased animals.

• The reference intervals for urine analytes are of little relevance when determining spot urine compositions.

a. Small Hydrophilic Molecules

Glucose, amino acids, and low-molecular-weight proteins are mostly reabsorbed in the proximal tubule. Almost 100% of the glucose, amino acids, and proteins are reabsorbed. The former return to the plasma, whereas small proteins are degraded in the tubular cells. Other small hydrophilic molecules are not or only poorly reabsorbed (e.g., creatinine) (see Section II.A.1.a).

Glucose reabsorption is permitted by the SGLT1 (sodium glucose transporter) in the apical membrane, which couples glucose reabsorption with sodium transport down a gradient produced by an Na/K-ATPase in the basolateral membrane, across which glucose diffuses in a concentration gradient by means of GLUT-2 transporters. This glucose reabsorption capacity is limited by the finite number of transporters: maximal tubular reabsorption (Tm) is attained when plasma glucose concentration (P-Glucose) is about 12 mmol/l in dogs, 15 to 18 mmol/l in cats, and 8 to 10 mmol/l in cows.

No urea reabsorption occurs before the medullary part of the collecting duct because of the presence of urea transporters activated by antidiuretic hormone (ADH). This is part of the mechanism creating a high inner medullary osmolality. A small amount of urea is even secreted into the medullary part of the ascending branch of the loop of Henle. Urea reabsorption is increased when urine flow is low (e.g., during dehydration or volume depletion). Urea is central to the concentrating mechanism in the kidney and is ensured by the presence of urea transporters at the thin extremity of the descending loop of Henle and corresponding segments of the vasa recta, which enable urea diffusing from the medullary part of the collecting tubule to enter the nephron and to a lesser degree the blood. See the review on urea transporters in Smith and Rousselet (2001).

b. Electrolytes

Electrolytes are mostly reabsorbed in the proximal tubule. The rate of reabsorption differs considerably according to the internal balance of each ion. Under “normal” conditions it is almost 100% for sodium, chloride, calcium, and phosphates, but much lower for potassium, especially in ruminants owing to their high dietary intake.

The extent of reabsorption can be estimated from the fractional excretion (FE) of solutes (i.e., the fraction of the filtered load of an electrolyte that is finally excreted in urine) (see Section II.C.3):

• Sodium concentration is kept low in tubule cells, as in other cells, by an Na/K-ATPase in the basolateral membrane. See the review in Feraille and Doucet (2001). This active transport of sodium ions from the urine accounts for most sodium reabsorption. It also creates a sodium concentration gradient that allows cotransport of amino acids, glucose and other ions, and so on. See the review in De Weer (1992). Further reabsorption of sodium and chloride occurs in the ascending branch of the loop of Henle via an Na-K-2Cl cotransporter in the luminal membrane. See the review in Russell (2000). Final adjustment in the distal part of the nephron is hormonally controlled by aldosterone and natriuretic peptides.

• Chloride is the most abundant anion in the extracellular compartment. The plasma concentration and elimination of chloride are usually concomitant with those of sodium, except in the case of acid-base disorders. In metabolic acidosis, bicarbonate ions secreted by the kidney cells are exchanged with chloride by an Na-independent Cl/HCO₃ exchanger, thus increasing urine chloride elimination and the plasma anion gap.

• Potassium is reabsorbed in the proximal tubule (~70%) and in the ascending part of the loop of Henle, the distal tubule, and the medullary collecting duct. Potassium is also secreted by the distal tubule and cortical collecting duct, mainly during hyperkalemia (Berliner and Kennedy, 1948). See the review in Hebert et al. (2005). Part of this process is based on aldosterone secretion, which induces the synthesis of Na/K-ATPases, thereby favoring potassium excretion and sodium reabsorption.

• The inorganic phosphates (P_i) in plasma comprise about 80% HPO₄^2– and 20% . They are reabsorbed in the proximal tubule by a sodium cotransporter, which is inhibited by PTH, thus increasing phosphodiuresis. Some newly identified peptides may also decrease tubular phosphate reabsorption without any alteration of glucose and amino acid reabsorption and thus lead to renal loss of phosphates. See the reviews in Laroche and Boyer (2005)and Ritz et al. (2003). The Tm for phosphate reabsorption is higher in ruminants than in other species (Summerill and Lee, 1985; Symonds and Manston, 1974).

• Magnesium: Non-protein-bound magnesium is filtered by the glomerulus. Only about 25% are reabsorbed in the proximal tubule. Most reabsorption occurs in the ascending branch of the loop of Henle (50% to 60%) (Rosol and Capen, 1996).

c. Water

FIGURE 16-4 Water reabsorption in the kidney tubule (values are osmolalities). The corticomedullary osmotic gradient results from reabsorption of sodium and chloride without concomitant reabsorption of water in the ascending limb of the loop of Henle and from urea reabsorption in the medullary part of the collecting ducts. Owing to water permeability of the descending limb of the loop of Henle plasma isoosmotic urine flowing from the proximal tubule is first concentrated in the descending branch of the loop of Henle, then is diluted by escape of NaCl and water retention in the ascending branch, and finally is concentrated along the collecting ducts. Reproduced with permission of the publisher from Finco (1995b).

e. Endocrine Functions

Two major hormones, erythropoietin (EPO) and 1α, 25-dihydroxycholecalciferol (calcitriol), are synthesized by the kidneys and released into the blood.

EPO is a cytokine that regulates erythrocyte production synthesized in the peritubular cells in response to hypoxia. Minor amounts are also produced in the liver, mainly in the newborn. EPO binds to receptors of bone marrow progenitor cells and acts synergistically with other growth factors to proliferate and mature the erythroid progenitor cells. In advanced chronic renal disease, the synthesis of EPO decreases and is insufficient to meet the demands for new red cell production, resulting in anemia. See the review in Fisher (2003).

Kidney function can be evaluated from the concentrations of plasma or urine analytes, which are mainly dependent on their elimination (e.g., P-Creatinine). These indirect markers can be easily and rapidly measured, but their sensitivity is poor and generally remains unaltered until 75% of renal function has been lost and their concentrations may be modified by extrarenal factors. Direct tests of kidney function are based on the elimination kinetics of markers of glomerular filtration, blood flow, or tubule reabsorption/secretion and are based on the clearance concept. These tests are more difficult and take longer to perform but allow earlier detection of reduced function.

P-Creatinine is the test most often used to diagnose and monitor kidney disease in human and animal clinical pathology. P-Urea is also used frequently but is subject to more numerous extrarenal factors of variation. These molecules are almost totally eliminated by glomerular filtration, so that in the case of kidney failure their plasma concentration increases. However, neither test is sensitive in the early diagnosis of kidney disease because of the large functional reserve of the kidneys. Moreover, variations of P-Urea and P-Creatinine are not proportional to the number of functional nephrons (e.g., a mean increase of 85% in P-Creatinine and of 140% in P-Urea was observed after a two-fold reduction of GFR, with values close to the upper limit of the reference interval) (Lefebvre et al., 1999).

1. Creatinine

a. Creatinine Metabolism

FIGURE 16-5 Schematic representation of creatinine metabolism. The first step is the transamidination of arginine and glycine yielding guanidinoacetic acid (GAA) by the kidney enzyme arginine: glycine amidinotransferase. In the liver, N-methylation of guanidinoacetate into creatine is catalyzed by guanidinoacetate methyltransferase, using methyl groups donated by S-adenosylmethionine. Creatine is distributed to muscle cells where it is reversibly phosphorylated to creatine phosphate by creatine kinase. Creatinine is the product of the spontaneous, irreversible, non-enzymatic, internal dehydration of creatine, and dephosphorylation of creatine phosphate.

As observed in humans and rats, extrarenal intestinal catabolism may be suspected in cases of renal failure. This may involve the bacterial flora, as demonstrated in the rat by accumulation creatine bacterial catabolites including methylguanidine (Jones and Burnett, 1972; Mitch et al., 1980). This hypothesis is supported by the observed increase in cats with ARF (Ohashi et al., 1995). Finally, creatinine is rapidly cleared from plasma (half-life of 3 h in dogs; Watson et al., 2002a) and eliminated in urine, where its total excretion depends on body weight (Gartner et al., 1987).

b. Preanalytical Factors of Variation

• Specimen: Serum and plasma creatinine concentrations are identical, although a slightly higher serum value has been reported in dogs (Thoresen et al., 1992). Differences between jugular and cephalic vein may be overlooked (5 μmol/l in dogs) (Jensen et al., 1994). Canine P-Creatinine is not changed by storage at –20°C or –70°C for up to 8 months (Thoresen et al., 1995) or by three freeze-thaw cycles (Reynolds et al., 2006).

In human urines, the creatinine concentration decreased by about 20% to 30% in 1 week at any temperature between 4°C and –80°C, and was not altered by five freeze-thaw cycles (Schneider et al., 2002), whereas others found that it was stable up to 30 days when stored at 4°C and minimally decreased at 25°C (Spierto et al., 1997).

• Diet and meals: P-Creatinine in dogs and cats was increased by meals containing meat, especially when this was cooked (Epstein et al., 1984; Evans, 1987; Harris and Lowe, 1995; Lowe et al., 1998; Sagawa et al., 1995; Watson and Church, 1980; Watson et al., 1981) or after oral loading with creatine (Lowe et al., 1998), although large interindividual differences were observed. P-Creatinine was higher in dogs fed chicken-based diets than egg- or casein-based diets (Bartges et al., 1995a). No differences in P-Creatinine were observed between cats receiving high- or low-protein diets (Adams et al., 1993). P-Creatinine was moderately higher in young dogs on a low-salt diet (Bagby and Fuchs, 1989) and about 50% higher in goats fed on a low-protein diet (Valtonen et al., 1982).

• Hydration status: P-Creatinine was only moderately increased in dogs deprived of water for 4 days, even when the weight loss was ≥10% (English et al., 1980; Hardy and Osborne, 1979).

• Physical exercise: P-Creatinine was decreased (by about 40%) by physical training in sled dogs (Kronfeld et al., 1977; Querengaesser et al., 1994). It was not significantly changed after strenuous physical exercise in untrained dogs (Chanoit et al., 2002). It was increased by about 20 μmol/l after a 400 m sprint in greyhounds (Rose and Bloomberg, 1989; Snow et al., 1988) and by approximately 50% after strenuous sprint efforts in sled dogs (Hammel et al., 1997; Querengaesser et al., 1994), but not after very long races (up to 415 miles) (Hinchcliff et al., 1993; Querengaesser et al., 1994).

• Housing: P-Creatinine was slightly higher (10 to 20 μmol/l) in dogs kept indoors than outdoors (Kuhn and Hardegg, 1988; Rautenbach, 1988), whereas others found no difference (Spangenberg et al., 2006).

• Drugs: P-Creatinine was decreased in dogs receiving glucocorticoids (Braun et al., 1981), whereas U-Creatinine was increased (Iversen et al., 1997). P-Creatinine was unchanged or little affected by a single dose of nonsteroidal anti-inflammatory drugs (NSAID) (Lobetti and Joubert, 2000; Mathews et al., 2001) or by halothane anesthesia (Lobetti and Lambrechts, 2000), and moderately increased by furosemide administration (Adin et al., 2003) and high-dose trimethoprim-sulfadiazine in dogs (Lording and Bellamy, 1978). P-Creatinine could increase (Kitagawa et al., 1997), remain unchanged (Atkins et al., 2002), or decrease (Pouchelon et al., 2004) in dogs treated with ACE inhibitors for heart failure.

c. Analytical Factors of Variation

HPLC is considered to be the reference method (Blijenberg et al., 1994; Hanser et al., 2001), but routine analyses are based on the nonspecific Jaffé reaction (alkaline picrate) and enzymatic procedures (Guder et al., 1986). See the reviews on creatinine measurement in Spencer (1986) and recommendations for improvement in Myers et al., (2006). The enzymatic methods give slightly lower results than HPLC and Jaffé reaction, ~5 and 20 μmol/l, respectively, in dog plasma (Evans, 1987; Palm and Lundblad, 2005). The main interferents in the Jaffé reaction are glucose, ketones, and hemoglobin, but canine plasma was unaffected by hemolysis up to 25 g/l (Jacobs et al., 1991, 1992; O’Neill and Feldman, 1989). Interference in the kinetic Jaffé technique is limited in normal dog plasma as the other chromogens react more slowly than creatinine (Palm and Lundblad, 2005). There is less interference in urine because of the lower proportion of Jaffé chromogens, so that calculated creatinine clearances may differ greatly according to the technique used, as shown in rats (Jung et al., 1987).

d. Reference Intervals and Physiological Factors of Variation

Reference intervals for P-Creatinine have been poorly defined in most species because of animal selection and the chosen analytical method. The reference intervals for dogs were shown to differ considerably from one textbook to another (Lefebvre et al., 1998b), so that any interchanging of reference limits or decision levels would be risky:

• Breed: P-Creatinine is higher in large breeds of dogs (Braun et al., 2002; Feeman et al., 2003; Hilppo, 1986; Medaille et al., 2004), even in puppies (Kühl et al., 2000).

• Gender: No differences related to gender were observed in dogs (Broulet et al., 1986; Passing, 1981), although moderately higher concentrations (~ + 10%) were observed in 1- to 3-year-old and 9- to 11-year-old male beagles (Fukuda et al., 1989).

• Age: A high P-Creatinine concentration was reported in newborn calves, puppies, and foals, probably because of the accumulation of creatinine ingested from the allantoic fluid (9 to 23 mmol/l in bovines) (Edwards et al., 1990; Klee et al., 1985; Kühl et al., 2000; Lupke et al., 1967). This decreased during the following weeks and then showed a moderate increase again. In dogs and cats, P-Creatinine almost doubled from the first weeks to 1 year of age, then remained stable up to 10 years (Kraft et al., 1996; Passing and Brunk, 1981; Strasser et al., 1993, 1997; Swanson et al., 2004; Vajdovich et al., 1997; Wolford et al., 1988), although other authors reported a regular decrease with age in dogs (Fukuda et al., 1989; Lowseth et al., 1990). In piglets, P-Creatinine almost doubled between 2 and 6 months of age (Krogh et al., 1979).

• Biological rhythms: Both a circadian and a seasonal rhythm were observed in dogs, but they were of limited significance (Singer and Kraft, 1988; Sothern et al., 1993; Strasser et al., 2001), whereas no such rhythms were reported in bulls (Boehnke, 1980). In healthy sheep, creatinine was 25% higher in summer than in winter (Nawaz and Shah, 1984).

e. Variations of P-Creatinine in Disease

FIGURE 16-6 Relationship between P-creatinine and GFR in dogs. Observe that (1) on the right-hand part of the graph, GFR can be decreased from about 6 to 2 ml/min/kg without any significant change of P-creatinine; (2) on the left-hand part of the graph, dramatic reductions of P-creatinine, as observed during animal rehydration in clinics, do not imply a significant increase in GFR. Reproduced with permission of the publisher from Finco (1995a).

The sensitivity and specificity of P-Creatinine for the diagnosis of CRF are not very high in dogs and neither are the predictive values (Braun and Lefebvre, 2005; Gleadhill, 1994). The critical difference for P-Creatinine in dogs is 35 μmol/l (Jensen and Aaes, 1993). These criteria have not been reported for other species.

It was suggested that the evolution of renal disease in dogs could be monitored by repeated creatinine measurement and that the time of death could be predicted from the 1/P-Creatinine versus the time curve (Allen et al., 1987); however, this approach gave false estimates in humans (Walser et al., 1988).

b. Preanalytical Factors of Variation

• Specimen: No difference between canine serum and heparin plasma was observed and only minor changes were noted when specimens were stored frozen for up to 8 months (Thoresen et al., 1995). P-Urea is little affected by hemolysis (up to 25 g hemoglobin/l) and icterus in cattle, horses, cats, or dogs; it is decreased by lipemia in dogs (Jacobs et al., 1992; O’Neill and Feldman, 1989). P-Urea is stable in plasma and whole blood stored for up to 3 days at 20°C (Thoresen et al., 1992) and in serum or plasma stored frozen at –20°C and at –70°C up to 8 months (Thoresen et al., 1995).

• Diet and meals: P-Urea is increased in dogs after meals. Peak postprandial increase can be as high as 7 mmol/l about 6 h after the meal and last for more than 18 h; it is greater with high-protein diets or in animals fed large amounts (Anderson and Edney, 1969; Epstein et al., 1984; Evans, 1987; Vogin et al., 1967). In most species, basal P-Urea reflects the balance between nitrogen utilization and excretion and can be greatly influenced by nutrition (Kohn et al., 2005). The fasting concentration of P-Urea was lower in dogs on low-protein diets with normal or reduced renal function (Polzin et al., 1983, 1991; Reynolds et al., 1999), in horses (Doreau and Martin-Rosset, 1985), in sheep (Rabinowitz et al., 1973), in goats (Valtonen et al., 1982), and in cats (Hesta et al., 2005). P-Urea was also increased by prolonged fasting, because of catabolism of body proteins (Rabinowitz et al., 1973).

• Hydration status: Dehydration had little effect on P-Urea in dogs (≤12.5 mmol/l in 4 days, with weight loss up to 16%) (Hardy and Osborne, 1979), but it produced a two-fold increase in calves within 4 days (Bianca et al., 1965).

• Drugs: P-Urea was unchanged by halothane anesthesia in dogs (Lobetti and Lambrechts, 2000) and was increased at high doses of trimethoprim-sulfadiazine (Lording and Bellamy, 1978).

• Physical exercise: P-Urea in sled dogs was higher after 12 weeks of training, probably as a result of increased protein intake (Reynolds et al., 1999). In greyhounds, P-Urea was unchanged by a 235 m sprint and moderately increased 30 min after a 420 m run (Snow et al., 1988).

c. Analytical Factors of Variation

Most techniques are based on the specific action of a bacterial urease. The accuracy of P-Urea measurements is not usually reported.

1 mmol/l = 6 mg/dl. There is no good reason to continue using BUN, as the analytical procedure was abandoned long ago and is merely a source of confusion. If required, the factors are BUN (mg/dl) × 0.356 = Urea (mmol/l), and BUN (mg/dl) × 21.4 = Urea (mg/l).

• Gender: No significant effect of gender was observed in young beagles (Passing and Brunk, 1981) or in adult dogs (Broulet et al., 1986).

• Age: P-Urea decreases in dogs (by about 50%) between birth and 1 to 2 months (Kühl et al., 2000; Wolford et al., 1988), after which irregular but moderate variations have been reported (Broulet et al., 1986; Cowgill and Spangler, 1981; Lowseth et al., 1990; Passing and Brunk, 1981; Strasser et al., 1993, 1997; Vajdovich et al., 1997). P-Urea increased from birth to 6 months in Great Danes, except those fed on low-protein diets (Nap et al., 1991). P-Urea was similar in adult horses and newborn foals, decreasing by about 50% in foals on the first day and remaining low for at least 2 months (Edwards et al., 1990). P-Urea decreased during the first week in calves (Hartmann et al., 1987), although others did not report any change during this time (Klee et al., 1985).

• Individuals: Intraindividual variations in horses were more important than the variations observed before and after foaling (Doreau and Martin-Rosset, 1985).

• Biological rhythms: P-Urea in sheep was 30% higher in summer than in winter (Nawaz and Shah, 1984), but other authors found no seasonal difference in dogs (Strasser et al., 2001).

e. Pathological Factors of Variation

The efficiency of P-Urea in renal failure diagnosis has not been reported. The critical difference in dogs was 2.4 mmol/l around a mean value of 5 mmol/l (Jensen and Aaes, 1993). A threshold of decision of 10 mmol/l has been suggested for cattle with suspected renal disease (Campbell and Watts, 1970).

The relationship between P-Urea and P-Creatinine in dogs was reported to be low (Gabrisch, 1973) or high and linear (Toutain et al., 2000), except for frequent increases in P-Urea concentration, with normal P-Creatinine likely the result of poor preanalytical conditions (Medaille et al., 2004). Because of great interindividual variations, differences in the urea/creatinine ratio cannot be related to specific diseases (Finco and Duncan, 1976).

The variations in P-Urea with disease are similar to those of P-Creatinine, but numerous extrarenal factors may contribute to increased P-Urea, such as gastrointestinal hemorrhage, fasting, or sepsis, which increase protein catabolism; thyrotoxicosis, which lowers P-Urea by increasing GFR; and decreased renal perfusion, which increases renal reabsorption (DiBartola et al., 1996; Prause and Grauer, 1998). Other factors of decreased P-Urea include portosystemic shunts, malnutrition (Davenport et al., 1994), liver insufficiency, and Krebs-Henseleit cycle enzyme defects. See the reviews in Dial (1995) and Sutherland (1989). These extrarenal factors of variation explain why P-Urea is less specific than P-Creatinine for the diagnosis and management of CRF and should not be recommended as a test of renal function. However, as P-Urea greatly depends on protein supply, it is a useful tool for monitoring the effects of dietary protein restriction (Devaux et al., 1996).

P-Urea is less effective in cattle than P-Creatinine to evaluate decreased GFR in diarrheic calves (Brooks et al., 1997), and only 30% of cattle with P-Urea above the upper limit of the reference interval had renal disease (Campbell and Watts, 1970).

f. Effects of Prolonged Increases of P-Urea: Carbamylated Hemoglobin

The carbamylation of proteins is an irreversible nonenzymatic reaction occurring between amino groups of proteins and isocyanate, the active form resulting from cyanate isomerization. This latter is derived from the spontaneous dissociation of urea in solution into ammonium and cyanate ions. Thus, the higher and longer the concentration of P-Urea, the higher the concentration of carbamylated proteins, for instance, of hemoglobin (a process analogous to protein glycation in diabetes mellitus). As proposed in human medicine (Stim et al., 1995; Wynckel et al., 2000), carbamylated hemoglobin may be useful for assessing canine CRF and distinguishing it from ARF (Heiene et al., 2001b), whereas others found a significant overlap with specificity and sensitivity equal to 96% and 84%, respectively, at a 100 μg/g Hb threshold (Vaden et al., 1997a).

3. Cystatin C

Cystatin C is a small constitutive protein (MW ~14,000) synthesized by all nucleated cells and only cleared by glomerular filtration. The plasma concentration is thus increased in cases of renal failure. Cystatin C is considered the most sensitive marker of renal failure in humans. See the reviews in Filler et al. (2005), Laterza et al. (2002), and Price (2000). It can be measured with reagents used in humans in plasma from dogs (Almy et al., 2002; Jensen et al., 2001; Martin et al., 2002) but not cats (Martin et al., 2002). A 50% decrease of P-Cystatin C was reported in dogs following a meal, so sampling should be done after a 12-hour fast (Braun et al., 2002). The upper limit of the reference interval in dogs is 1.3 mg/l (Braun et al., 2002) and similar to the human threshold. P-Cystatin C is well correlated with P-Creatinine and GFR in normal dogs and in dogs with reduced GFR (Almy et al., 2002; Braun et al., 2002). At present, no diagnostic advantage of P-Cystatin C over P-Creatinine has been demonstrated in dogs.

The determination of glomerular filtration rate (GFR)—that is, the volume of ultrafiltrate produced per unit of time (e.g., ml/min) by glomerular filtration—is considered the best way to evaluate kidney function. This is based on the hypothesis of the “intact nephron”: as “the surviving nephrons of the diseased kidney largely retain their essential functional integrity” and “retain a remarkably uniform relationship between glomerular and tubular function” (Bricker et al., 1997).

GFR depends on the size of the animal, so different modes of expression have been proposed. One of the most frequently used is to relate GFR to body weight (ml/min/kg) or better still to lean body mass in humans (Swaminathan et al., 2000). Another approach is to use the body area (ml/min/m²), but as the equations used to calculate this from BW have not been validated in all species, this may introduce further inaccuracy (Price and Frazier, 1998). Another mode of expression consists of relating GFR to the volume of the extracellular compartment, as one kidney function is the regulation of body water content. This indexing is rarely adopted but has been used in humans (Peters et al., 2000) and dogs (Gleadhill, 1994; Gleadhill and Michell, 1996).

1. Determination of GFR

There is no easy method for determining GFR from a single blood or urine specimen. In human clinical pathology, there are equations to estimate GFR from P-Creatinine, gender, weight, and age. The most frequently used are the Cockroft-Gault’s equations, but these are imprecise and the results depend on the techniques used for P-Creatinine (Grubb and Nordin, 2006; Wuyts et al., 2003). No such equations are available for use in animals. To our knowledge, the only study relating GFR to P-Creatinine showed that the estimation of canine GFR from P-Creatinine was imprecise, and the authors did not recommend using the equation to estimate GFR (Finco et al., 1995).

The measurement of GFR is based on the clearance of markers freely filtered by the glomerulus and having no or minor secretion and reabsorption. See the review in Heiene and Moe (1998). The accepted reference for GFR determination is the urinary clearance of inulin, a low-molecular-weight polysaccharide (Shannon, 1935). The total amount of this marker eliminated in urine over a period of time is equal to the filtered load:

Total amount eliminated during time t = (U-Inulin) × (U-Volume) = GFR × (P-Inulin)

GFR = (U-Volume) × (U-Inulin)/(P-Inulin) × t

in which U-Inulin and P-Inulin are in the same units, volume is in ml, and time in min, thus GFR is in ml/min.

However, such determinations are labor intensive: they necessitate the maintenance of a constant plasma concentration by constant infusion of the marker, the accurate determination of urine volume, and precise and accurate measurement of the marker. Urine collection is often impractical as the total urine is required, which implies careful washing of metabolic cages (Watson et al., 2002a) or the use of indwelling catheters. See the review on urine collection and preservation in cats and dogs in Osborne (1995).

Easier procedures have been proposed, based on the following:

• The use of other markers: Some markers; such as ferrocyanide (Ladd et al., 1956), thiosulphate (Bing and Effersoe, 1948; Dalton, 1968), and sulfanilate (Maddison et al., 1984; Ross and Finco, 1981) have been abandoned. The endogenous or exogenous clearance of creatinine is the most frequently used technique, whatever the species. Iodinated radiocontrast media (e.g., iohexol) provide a good alternative to inulin, but most laboratories are not equipped to measure iodine (Westhoff et al., 1993, 1994). The metabolism of these various markers has not been documented in domestic animals. In some cases extrarenal metabolism can be significant, as shown in cats for diethylenetriaminepentaacetic acid (DTPA), which first concentrates in the heart and lung, then accumulates moderately in the liver before being concentrated in the kidneys (Drost et al., 2000; Uribe et al., 1992). Radiolabeled markers (see the review in Daniel et al. [1999]) allow easier and more accurate measurement of concentrations, but their use is restricted to specially equipped centers: ¹³¹I-Inulin, ⁵¹Cr-EDTA (Biewenga and van den Brom, 1981; van den Brom and Biewenga, 1981), and ^{99 m}Tc-DTPA (Gleadhill et al., 1999). They have also been used in large species such as horses (Matthews et al., 1992; Walsh and Royal, 1992).

• The technique of administration: Single injection instead of continuous infusion and determination of plasma clearance of the marker, with calculations based on the decreased plasma concentration of the marker (Pihl and Nosslin, 1974; Summerville and Treves, 1986).

• The use of nuclear imaging techniques and ^99mTc-labeled tracers such as DTPA. These permit the functions of each kidney to be evaluated separately (Assailly et al., 1977; Drost et al., 2000; Krawiec et al., 1988), as these may be similar in healthy but not always in diseased dogs (Cowgill and Hornhof, 1986; Lourens et al., 1982). Determination of GFR by imaging is not as precise as clearance methods (Barthez et al., 1998; Kampa et al., 2002).

2. Variations of Results of GFR Determination According to Procedure

a. Preanalytical Factors of Variation

• Hydration status: GFR was higher in dogs that were hyperhydrated, so hydration status should be standardized (Kerr, 1958; Kunze et al., 2006; Tabaru et al., 1993). In cats, infusion of Ringer lactate at 1 to 2 ml/min produced on average a 40% increase of GFR (Rasmussen et al., 1985). In sheep, GFR was lower in summer than in winter because of hemoconcentration (Nawaz and Shah, 1984).

• Diet: GFR was higher in dogs fed animal proteins rather than other proteins (Bartges et al., 1995a, 1995b, 1995d; Kerr, 1958). It was higher with high levels of proteins in the diet in normal dogs and cats and after renal mass reduction (Adams et al., 1994; Bartges et al., 1995c; Bovee, 1992; Bovee and Kronfeld, 1981; Robertson et al., 1986). In partially nephrectomized dogs, low- or high-sodium diets had little influence on GFR (Greco et al., 1994). In cats, GFR was moderately reduced by low-sodium diets (Buranakarl et al., 2004), unchanged in cats receiving potassium-depleted food, but lowered when the same food was acidified with ammonium chloride (Dow et al., 1990). In cows, GFR was unchanged by mineral supply (Hartmann et al., 2001). GFR was higher in sheep fed a normal or high rather than a low protein diet (Cirio and Boivin, 1990a, 1990b; Cirio et al., 1990; Eriksson and Valtonen, 1982; Rabinowitz et al., 1973; Valtonen et al., 1982).

• Meals: GFR increases ranging from 10% to 45% were observed after a meal of proteins in normal dogs and in dogs with renal mass reduction (Bourgoignie et al., 1987; Brown, 1992; Ewald, 1967; Jolliffe and Smith, 1932; Moustgaard, 1947; O’Connor and Summerill, 1976; Reinhardt et al., 1975). This postprandial increase of GFR was not observed in dogs with experimental Fanconi’s syndrome (Woods and Young, 1991). It was not reduced by fasting for more than 1 day (Moustgaard, 1947). No difference in GFR was observed when ponies were fed twice a day or received the same amount of food every 2 h (Clarke et al., 1990).

• Exercise: GFR was unchanged by training in horses and dogs (McKeever et al., 1985, 2002) or by submaximal exercise for 20 or 60 min despite hemoconcentration (Hinchcliff et al., 1990; McKeever et al., 1991). During effort under anaerobic conditions in horses, GFR decreased sharply then returned to prerun values within 15 min (Schott et al., 1991).

• Drugs: In dogs and sheep, GFR was increased by glucocorticoids, probably by an increase of plasma flow rate as in rats (Baylis and Brenner, 1978; Gans, 1975). GFR was decreased by furosemide in dogs and cats (Bostrom et al., 2003; Hanna et al., 1988). Controversial effects of NSAIDs were reported in dogs: either no change of GFR with meloxicam, carprofen, and ketoprofen (Crandell et al., 2004; Ko et al., 2000; Narita et al., 2005) or a moderate decrease with carprofen or ketoprofen (Forsyth et al., 2000). GFR was moderately decreased after repeated excretory urograms in dogs (Feeney et al., 1980a). In cats with experimental CRF, the ACE inhibitor benazepril increased GFR up to 30% (Brown et al., 2001), whereas ACE inhibitors could decrease GFR in sodium-restricted dogs (Hall et al., 1979).

• Biopsy: GFR was not affected after repeated renal biopsies in dogs (Drost et al., 2000; Groman et al., 2004).

b. Techniques of GFR Measurement

The procedures used to calculate plasma clearances may also influence the results as they depend on the mathematical model chosen (Heiene and Moe, 1999; Powers et al., 1977; Watson et al., 2002a) and on the proposed limited sampling strategies (Barthez et al., 2001; Blavier et al., 2001; Finco, 2005; Watson et al., 2002a). Others have suggested measuring the concentration of an exogenous marker at a given time after injection (e.g., P-Creatinine after I.V. load in dogs) (Labato and Ross, 1991; Watson et al., 2002a).

The measurement of creatinine concentrations by Jaffé technique (discussed earlier) led to erroneously high values for creatinine clearance in early studies. Because of the larger proportion of interfering substances in plasma than in urine with the Jaffé technique, P-Creatinine was overestimated in comparison to U-Creatinine (e.g., in dog urine) (Shannon et al., 1932). This discrepancy between creatinine and inulin clearances does not exist when the measurements are based on the more accurate enzymatic techniques (Finco et al., 1993).

The results of GFR determination may vary according to the technique so their transferability is limited (Driehuys et al., 1998; Finco, 2005; Gagnon et al., 1971; Izzat and Rosborough, 1989; Oester et al., 1968; Rogers et al., 1991), and caution is required when using decision limits from other laboratories or from the literature. Correcting factors have been proposed in some cases, for instance, for iohexol plasma clearance and exogenous creatinine urinary clearance in dogs (Finco et al., 2001; Gleadhill and Michell, 1996). However, such results remain controversial. The estimation of GFR based on iohexol clearance, for example, was considered reliable in dogs and cats (Brown et al., 1996), whereas others showed fdifferences of up to 0.95 ml/min/kg with creatinine clearance (Miyamoto, 2001b). The urine clearance of endogenous creatinine and the urine and plasma clearance of exogenous creatinine gave similar results (Finco et al., 1981, 1991; Lee et al., 1983; Watson et al., 2002a), but plasma DTPA clearance was 1.15 time higher than plasma iohexol clearance in dogs (Moe and Heiene, 1995).

FIGURE 16-7 Examples of mean GFR values in the dog, cat, horse, cattle, sheep, and swine. Each point is the result of an original study.

Dog: Asheim et al., 1961; Bostrom et al., 2003; Brown and Finco, 1992; Ellis et al., 1973a; Ewald, 1967; Fettman et al., 1985; Finco, 1971, 2005; Finco and Cooper, 2000; Finco et al., 1993; Forsyth et al., 2000; Gagnon et al., 1971, 1982; Greco et al., 1994; Haller et al., 1998; Hartenbower et al., 1974; Henegar et al., 2001; Houck, 1948; Kampa et al., 2003; Kaneko et al., 1978; Kunze et al., 2006; Labato and Ross, 1991; Lee et al., 1983; Lefebvre et al., 1998, 1999; Lora-Michiels et al., 2001; Newell et al., 1997; Pihl and Nosslin, 1974; Russel et al., 1987; Tabaru et al., 1993; Watson et al., 2002; Westhoff et al., 1993.

Cat: Adams et al., 1993, 1994; Bolliger et al., 2005; Brown et al., 1996; Buranakarl et al., 2004; Deguchi and Akuzawa, 1997; Drost et al., 2000; Fettman et al., 1985; Finco et al., 1997; Haller et al., 2003; Miyamoto, 1998; Miyamoto, 2001a, b; Osbaldiston and Fuhrman, 1970; Reichle et al., 2002; Rogers et al., 1991; Ross and Finco, 1981; Russo et al., 1986; Uribe et al., 1992.

Horse: Clarke et al., 1990; Finco and Groves, 1985; Gelsa, 1979; Hinchcliff et al., 1990; Knudsen, 1959; Kohn and Strasser, 1986; Lane and Merritt, 1983; Matthews et al., 1992; Morris et al., 1984; Rawlings and Bisgard, 1975; Schott et al., 1991; Traver et al., 1977; Walsh and Royal, 1992; Zatzman et al., 1982.

Cattle: Boehnke, 1980; Deetz et al., 1982; Fleming et al., 1991; Mercer et al., 1978.

Sheep: Bickhardt and Dungelhoef, 1994; Cirio and Boivin, 1990a, 1990b; English et al., 1977; Filippich et al., 1985; Gans, 1975; Garry et al., 1990b; Nawaz and Shah, 1984; Nesje et al., 1997; Rabinowitz et al., 1973.

Goat: Brown et al., 1990a, 1990b; Valtonen et al., 1982.

Swine: Chaiyabutr et al., 1987; Mercer et al., 1979; Wendt et al., 1990.

c. Reference Values and Physiological Factors of Variation

d. Variations with Disease

Decreased GFR is the gold standard of renal failure, whatever its cause or subsequent evolution (discussed later). A critical approach should always be applied in the clinical evaluation of renal function (Finco and Barsanti, 1989), especially as the results obtained differ according to the method used (discussed earlier).

Alterations of GFR may be secondary to many extrarenal diseases. GFR is decreased in hypothyroidism (Adams et al., 1997) and hypoxia (Lobetti et al., 1996). It may be decreased or not in experimental or spontaneous diabetes mellitus (Kaneko et al., 1978, 1979). GFR evaluation is also useful for monitoring the toxicity of drugs eliminated by renal filtration (e.g., carboplatin) (Bailey et al., 2004; Shapiro et al., 1988) and as an indicator of the effects of renal dysfunction on the pharmacokinetics of drugs (e.g., oxytetracyclin in dogs) (Duffee et al., 1990).

1. Concentrating ability

a. Urine Osmolality versus Urine-Specific Gravity

One of the most important functions of the kidney in mammals is to reabsorb more than 99% of the filtered water, so that the final urine is more concentrated than the glomerular ultrafiltrate.

Urine concentration is best evaluated by osmolality (Bovee, 1969)—that is, the concentration of particles in a solution, independently of their chemical characteristics (1 sodium ion has the same osmotic effect as 1 molecule of urea or as 1 molecule of albumin). The main determinant of specific gravity in dog urine is urea (Meyer et al., 1997). The laws of osmolality are only valid for dilute solutions in which the different particles are completely independent and this is not verified in urine or plasma. See the review in Sweeney and Beuchat (1993).

The urine concentration is determined in routine tests from the specific gravity (U-SG)—that is, the ratio of the weight of 1 l of solution to 1 l of water, which depends on the relative proportions and the molecular weight of all the compounds in solution (for instance, 300 mOsm/kg solutions of NaCl, urea, and glucose would have SGs of approximately 1.009, 1.014, and 1.054, respectively; the latter would correspond to ~1 mOsm/kg solution of albumin).

Urine osmolality and specific gravity were highly correlated in dogs (Bovee, 1969; Dossin et al., 2003; Harvey, 1973; Hendriks et al., 1978; Meyer et al., 1997), sheep (English and Hogan, 1979), and cats (Lees and Osborne, 1979). The correlation was weaker in calves (Thornton and English, 1976). Maximum concentrations could be as high as 1400 mOsm/l in rabbits, 2400 mOsm/l in dogs, and 3200 mOsm/l in cat and sheep (Anderson, 1982).

b. Preanalytical Factors of Variation

• Specimen: In cats, U-SG was not changed by freezing (Lees and Osborne, 1979). Spot urines could give very different U-SG results in the same animal depending on the time of sampling.

• Diet: U-SG was lower in cats supplied with moist food than with dry food (Palmore et al., 1978). In cows, U-osmolality was little changed when the mineral supply was reduced by 50% or increased to 200%, but the urine volume was greatly increased during the high mineral supply period (Hartmann et al., 2001).

• Physical exercise: In greyhounds, U-osmolality was decreased (~8%) by training because of increased diuresis resulting from plasma volume expansion (McKeever et al., 1985). U-SG was transiently decreased for 1 to 2 h after an effort under anaerobic conditions in horses (Schott et al., 1991). The 25% to 75% interval for U-SG in postrace thoroughbred horses was 1.021 to 1.033 (Cohen et al., 2002).

• Environment: Water intake, urine volume, and osmolality differed significantly in sheep depending on whether the environment was cool or hot, dry or humid (Guerrini et al., 1980).

• Drugs: U-SG was increased in dogs after the administration of radiographic contrast media (Feeney et al., 1980b). In dogs, urine volume was increased and osmolality decreased after medetomidine (Burton et al., 1998) and after glucocorticoids administration, which interferes with the action of vasopressin on the kidney (Joles et al., 1980; Sirek and Best, 1952; Waters et al., 1997). In horses, U-Osmolality was increased after oral sodium bicarbonate loading (Rivas et al., 1997) and was low after treatment with furosemide (median 1.018) (Cohen et al., 2002).

• Anesthesia: U-SG was not changed by halothane anesthesia in dogs (Lobetti and Lambrechts, 2000) but was decreased by isoflurane anesthesia in horses because of increased diuresis, maybe from the fluid support (Watson et al., 2002b). In horses, U-Volume was increased, and U-SG and U-Osmolality were decreased by xylazine and detomidine (Gasthuys et al., 1986, 1987; Nunez et al., 2004; Steffey and Pascoe, 2002; Thurmon et al., 1984; Trim and Hanson, 1986).

c. Analytical Factors of Variation

Measuring osmolality requires expensive equipment, thus is not easy to do in most veterinary clinics. Freeze-point osmometers are usually unsuitable for the direct measurement of highly concentrated urines, especially in cats (Lees and Osborne, 1979).

d. Reference Values and Physiological Factors of Variation

The range of variations of U-SG or U-Osmolality in healthy animals is large in all species, so that reference intervals are devoid of any relevance for spot urines.

• Breed: U-SG was higher in miniature schnauzers than in labradors and might be a factor in oxalate stone formation (Stevenson and Markwell, 2001).

• Gender: Sex had no effect on canine U-SG (van Vonderen et al., 1997).

• Age: U-SG was lower in aged dogs than in young adults (van Vonderen et al., 1997). U-SG was low at birth in puppies, then increased over the first 2 months to values higher than those of mature dogs (Faulks and Lane, 2003; Laroute et al., 2005). U-SG/osmolality in newborn foals was similar to or moderately lower than that of adults, then it decreased to a state of hyposthenuria during the first day of life and remained so for at least 2 months (Edwards et al., 1990).

• Inter- and intraindividual variability: U-SG high in canine urine, CVs were in the range of 30% to 40% (van Vonderen et al., 1997).

• Biological rhythms: U-SG in dogs was slightly lower in evening than in morning samples (van Vonderen et al., 1997).

e. Pathological Factors of Variation

As variations in healthy animals can be large, hypo- or isosthenuria must either be confirmed on repeated samples or observed in moderately dehydrated or azotemic animals to be interpreted as an indicator of kidney dysfunction.

A concentrating ability below the commonly accepted limits of 1.030 and 1.035, in dogs and cats, respectively, is considered inadequate. See the review as this pertains to dogs in Watson (1998).

2. Tests of water deprivation

When plasma osmolality increases, osmoreceptors in the hypothalamus stimulate the release of ADH in blood, thus increasing water reabsorption and equilibrating plasma osmolality. Diabetes insipidus is an uncommon condition characterized by polyuria and polydipsia without glucosuria. It can be congenital or acquired (Harb et al., 1996) and results from decreased secretion of ADH by the hypothalamus (central diabetes insipidus) or from insensitivity of kidney cells to the effects of ADH. See the reviews in Cohen and Post (2002) and Neiger and Hagemoser (1985).

Tests of water deprivation are based on the fact that sudden or progressive withholding of water produces progressive dehydration. The resulting increase in extracellular fluid osmolality triggers the release of ADH, thus an increase of U-osmolality and U-SG in healthy subjects. The concentrating ability after exogenous ADH administration can be measured to test whether the absence of urine concentration results from a central defect of ADH secretion or from peripheral resistance to ADH. See the review in Finco (1995a). Protocols for test combination have been proposed (Mulnix et al., 1976), as well as for ADH measurement, but they have not gained wide acceptance (Biewenga et al., 1987). In dogs as in humans, urinary excretion of aquaporin-2 occurs in parallel to ADH action and can be used as a marker of collecting duct responsiveness (van Vonderen et al., 2004).

Water deprivation tests may be hazardous, so should only be used to investigate polydipsia-polyuria once significant kidney damage has been ruled out (MacDougall, 1981). They are contraindicated in dehydrated, azotemic, or hypercalcemic dogs or cats (Barsanti et al., 2000).

Water restriction in 10-month-old beagles produced maximum U-SG = 1.070 after up to 23 h of water deprivation (Balazs et al., 1971). In adult dogs, maximum U-osmolality was 2738 mOsm/kg with corresponding U-SG of 1.076 after 72 h and a weight loss of 16% (Hardy and Osborne, 1979). In horses, water deprivation for 72 h led to an average loss of weight of 8%, and an increase in U-SG (mean ~1.050), the hematocrit, and P-Proteins (Genetzky et al., 1987; Rumbaugh et al., 1982).

3. Urine excretion of Ions

a. Fractional Excretion

Many electrolytes are intensely reabsorbed after filtration, mainly in the proximal tubule; their excretion is thus increased when tubule dysfunction occurs. Urine electrolyte concentrations also depend on the alimentary supply as the homeostatic mechanisms aimed to stabilize plasma concentration modulate tubule reabsorption. They can thus greatly differ as a function of the diet in all species and on the proximity of meals in monogastric animals.

The most meaningful information would be obtained from daily urine excretion, which is often impossible to obtain because of the difficulties associated with urine collection. Expressing the urinary elimination of a solute (X) as the ratio of the filtered load that is found in urine has been proposed, whence the name fractional excretion (FE):

FE _× = amount in urine/amount filtered
= (U-X × U-Volume)/(P-X × GFR)

in which U-X and P-X are the urine and plasma concentration of X respectively.

Such spot measurements are often well correlated with daily elimination. See the reviews in Coffman (1980) and King (1994). However, spot determinations in cats are highly variable compared with 72-h values, and should thus be interpreted with caution (Finco et al., 1997). FE_Na and FE_K are poor indicators of the daily excretion of these ions in animals with renal failure (Adams et al., 1991). FEs and standard clearances are highly correlated in horses for sodium, potassium, and phosphates (Traver et al., 1977) (e.g., spot measurements of FE_Pi are within 0.1% of the calculated 24 h value) (Lane and Merritt, 1983). In sheep FE_Na, FE_K, FE_Cl, and FE_Pi are highly correlated with the respective daily urine excretion of these ions (Garry et al., 1990c).

b. Preanalytical Factors of Variation of Ion Excretion

• Diet: Daily elimination and FE_Na, FE_K, FE_Ca, and FE_Mg are higher in fed than in nonfed healthy dogs, whereas the excretion of phosphate is unchanged (Lulich et al., 1991). Diuresis and urine mineral composition differ according to food composition and intake in dogs (Zentek et al., 1994) and cats (Sauer et al., 1985a, 1985b) (e.g., cats fed a low K diet had lower FE_K and higher FE_Cl) (Dow et al. 1990); cats supplemented with magnesium showed a four-fold increase of FE_Mg (Norris et al., 1999a), and FE_Ca was increased in phosphate-depleted dogs (Goldfarb et al., 1977). FE_Pi is lower in dogs fed low-phosphate diets (Polzin et al., 1991) and low-protein diets (Polzin and Osborne, 1988). In cattle (see a review in Lunn and McGuirk [1990]), the composition of the diet greatly influences electrolyte balance, especially for calcium and magnesium (Gray et al., 1988): FE_Mg was about three times higher in cows receiving oral magnesium hydroxide than in controls (Kasari et al., 1990) and is mainly used to test magnesium status (Sutherland et al., 1986). In cows receiving a high mineral diet, FE_Ca, FE_Mg, and FE_Pi were increased (Hartmann et al., 2001). In sheep artificially loaded with NaCl by oral or intraruminal route, GFR, FE_Na, and FE_K increased, whereas P-Sodium remained stable (Meintjes and Engelbrecht, 1993). In cattle, sodium bicarbonate loading increased FE_Na and FE_HCO3, decreased FE_Mg and FECa, but had no effect on FE_K, FE_Cl, and FE_Pi (Roby et al., 1987). In mares, FE_Ca and FE_Pi are dependent on the diet (e.g., FE_Pi ranged from almost 0% to 20% when mares fed on pasture alone or with a high-P mineral supplement) (Caple et al., 1982a, 1982b). FE_Ca, FE_Mg, and FE_Pi were increased in cows receiving a high-mineral diet (Hartmann et al., 2001).

• Meals: In cats, urine excretion of phosphates and magnesium is moderately higher in the postprandial period (Finco et al., 1986). Postprandial variations of FE_Na and FE_Pi were observed in cats fed a high-phosphate high-sodium diet (Finco et al., 1989). FEs were altered after large meals in ponies (Clarke et al., 1990)

• Physical exercise: In horses, FE_K, FE_Na, and FE_Cl were decreased by training (McKeever et al., 2002). FE_Na was increased, whereas FE_Cl was decreased, and FE_K remained unchanged during submaximal exercise (McKeever et al., 1991) or transiently decreased immediately after effort under anaerobic conditions (Schott et al., 1991).

• Drugs and fluid therapy: In dogs, FE_K is increased in volume expansion by NaCl infusion (Massry et al., 1969). FE_K and FE_Cl but not FE_Na are increased after medetomidine administration in dogs (Burton et al., 1998). In horses, I.V. infusions of glucose or saline solutions increased FE_Na, FE_Cl, and FE_Pi, and FE_K was also increased after glucose infusion (Roussel et al., 1993). In ponies, administration of xylazine increased FE_K, FE_Na, and FE_Cl (Trim and Hanson, 1986).

c. Analytical Factors of Variation

Ion concentrations in urine are usually measured with the same techniques as in plasma with ad hoc dilutions when necessary, but techniques have not been validated in animal species. An interfering substance in the urine of sheep, cattle, horses, and cats, but not dogs, causes falsely low results for potassium but not sodium with ion-selective electrodes (Brooks et al., 1988).

d. Reference Intervals and Physiological Factors of Variation

FIGURE 16-8 Mean values of FE of sodium, potassium, chloride, and inorganic phosphates. Each point is the result of an original study in dogs (D), cats (C), horse (H), bovine (B), and sheep (S). The y-axis has a logarithmic scale.

FE_Na *Dog: Burton et al., 1998; Clarke et al., 1990; Deguchi and Akuzawa, 1997; Finco et al., 1997; Fleming et al., 1991; Gagnon et al., 1982. *Cat: Gans, 1975; Garry et al., 1990a, 1990b; Gelsa, 1979. *Horse: Izzat and Rosborough, 1989; Kohn and Strasser, 1986; Lobetti and Joubert, 2000; Lobetti and Lambrechts, 2000; Lulich et al., 1991; McKeever et al., 1991; Morris et al., 1984; Neiger and Hagemoser, 1985. *Sheep: Roussel et al., 1993. *Cattle: Adams et al., 1991; Bickhardt and Dungelhoef, 1994; Buranakarl et al., 2004.

FE_K *Cattle: Adams et al., 1992; Beech et al., 1993; Bickhardt and Dungelhoef, 1994. *Dog: Buranakarl et al., 2004; Burton et al., 1998; Clarke et al., 1990. *Cat: Clarke et al., 1990; Deguchi and Akuzawa, 1997; Dow et al., 1990; Edwards, 1989; Finco et al., 1997; Fleming et al., 1991. *Horse: Gans, 1975; Garry et al., 1990a, 1990b; Gelsa, 1979; Lulich et al., 1991; McKeever et al., 1991; Morris et al., 1984; Neiger and Hagemoser, 1985. *Sheep: Roussel et al., 1993; Russo et al., 1986; Toribio et al., 2005; Traver et al., 1977; Ulutas et al., 2003.

FE_Cl *Cattle: Buranakarl et al., 2004; Burton et al., 1998; Dow et al., 1990. *Dog: Edwards, 1989. *Cat: Fleming et al., 1991; Garry et al., 1990a, 1990b. *Horse: Gelsa, 1979; Kohn and Strasser, 1986; McKeever et al., 1991; Morris et al., 1984; Neiger and Hagemoser, 1985; Roussel et al., 1993; Russo et al., 1986. *Sheep: Toribio et al., 2005; Ulutas et al., 2003.

FE_Pi. *Cattle: Bickhardt and Dungelhoef, 1994; Caple et al., 1982. *Dog: Caple et al., 1982; Edwards, 1989. *Cat: Finco et al., 1997. *Horse: Fleming et al., 1991; Garry et al., 1990a, 1990b; Gelsa, 1979; Kohn and Strasser, 1986; Lane and Merritt, 1983; Lulich et al., 1991; Neiger and Hagemoser, 1985; Roussel et al., 1993; Russo et al., 1986. *Sheep: Toribio et al., 2005; Traver et al., 1977; Ulutas et al., 2003.

e. Pathological Factors of Variation

The main difficulty when interpreting FEs in the diagnosis of renal disease is that they are greatly influenced by all extrarenal factors involved in the regulation of plasma electrolyte balance, mainly by dietary supply (Finco and Barsanti, 1989; Finco et al., 1992a). As a result, the interpretation of increased FEs in terms of tubular dysfunction is often hypothetical. Only repeated measurements under well-controlled conditions (e.g., experimental settings) may offer some relevance.

In cats with severe CRF, the FEs of ions were normal in most animals, and their measurement did not seem to improve diagnosis (Filippich, 1992). In dogs, it was shown that FE_Pi was a less accurate indicator of CRF than P-Creatinine (Gleadhill, 1994).

FEs are also changed in nonrenal diseases. In diabetic cats, FE_Mg was about 20 times higher than in controls and could be responsible for the frequent hypomagnesemia observed in diabetes mellitus (Norris et al., 1999b). In the cat, a case of increased FE_Pi with normal P-Phosphates is described, probably because of deficient reabsorption and resulting in rickets-like symptoms (Henik et al., 1999). In horses, FE_K is moderately but insignificantly lower during rhabdomyolysis (Beech et al., 1993)

1. Proteinuria