Dietary factors affecting urine volume

Data in the literature indicate that feeding high-moisture diets to adult cats leads to a slightly higher total water intake leading to an increased urine volume⁽Reference Gaskell, Rivers and Burger^15, Reference Burger and Smith¹⁶⁾. When high (82 %)-moisture diets are fed, urine output is increased by 57·4 % compared with low (10 %)-moisture diets⁽Reference Gaskell, Rivers and Burger¹⁵⁾. In a retrospective case–control study with 173 cats with CaOx uroliths, it was found that cats fed (canned) diets containing the highest moisture contents (77·4 to 81·2 %) were only about one-third as likely to develop CaOx uroliths compared with cats fed (dry kibble) diets low in moisture (7·0 to 7·9 %)⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. The purpose of the latter study was to identify dietary factors associated with the increase in occurrence of CaOx uroliths. The dietary factors studied were moisture content, protein, carbohydrate, fat and fibre content, but also Ca, P, Mg, Na, K and chloride content and urine-acidifying potential. Of these factors, the moisture content of the diet seemed to have the strongest association with CaOx urolith formation in cats⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. The favourable effect of a high water intake on LUTD in general was also reported by Markwell et al. ⁽Reference Markwell, Buffington and Smith¹⁴⁾, who stated that recurrence rates of signs in cats classified as having idiopathic LUTD, or FIC, may be more than halved when affected animals are maintained on high-, rather than low-, moisture content diets. Gunn-Moore & Shenoy⁽Reference Gunn-Moore and Shenoy¹⁷⁾ observed a significantly lower urine specific gravity in cats with FIC when they were fed more canned cat food as well.

The increased water intake when feeding moist diets might not only be due to its higher moisture content, but also because its higher mineral (ash) and protein content (evoking a higher renal solute load) may stimulate drinking⁽Reference Buffington^18, Reference Funaba, Yamate and Hashida¹⁹⁾. This is in line with the finding that a high intake of animal protein is accompanied by increased water consumption and urine volume in cats⁽Reference Funaba, Hashimoto and Yamanaka²⁰⁾. Taking into account the overall beneficial effect of high-moisture diets by increasing urine volume, and therefore diluting all substances dissolved in the urine, feeding a high-moisture diet may be considered as one of the most effective measures to prevent CaOx formation.

Another way to increased urine volume is to stimulate drinking by increasing dietary Na intake⁽Reference Borghi, Meschi and Schianchi^21, Reference Chandler²²⁾. This effect of a higher Na intake might explain the results of the previously mentioned retrospective case–control study with 173 cats, in which diets high in Na (0·34 to 0·89 g/MJ metabolisable energy (ME)) were correlated with a decreased risk of CaOx urolith formation⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. In humans, an increased salt intake has been associated with elevated urinary Ca excretion⁽Reference Castenmiller, Mensink and van der Heijden^23–Reference Sakhaee, Harvey and Padalino²⁵⁾. Salt-induced calciuria is believed to result from Na–Ca interactions in the kidney, in both the proximal and distal tubule. Reabsorption of Ca parallels the reabsorption of Na in the proximal tubule and loop of Henle. Na may influence renal reabsorption of Ca in the distal tubule by both a direct effect and indirectly through its effects on parathyroid hormone levels⁽Reference Massey and Whiting²⁶⁾. In a study with healthy cats, the total daily amount of Ca excreted in urine was reported to be increased⁽Reference Biourge, Devois and Morice²⁷⁾. Similar results were found in dogs⁽Reference Lulich, Osborne and Sanderson^28, Reference Stevenson, Hynds and Markwell²⁹⁾. However, in two studies with healthy adult cats, dietary Na levels of up to 0·96 g/MJ ME did not increase urine Ca concentrations, but did increase the volume of water drunk, resulting in a larger urine volume and a concomitant decrease in urine specific gravity and CaOx relative supersaturation (RSS)⁽Reference Hawthorne and Markwell^30, Reference Tournier, Aladenise and Vialle³¹⁾. A possible explanation for this discrepancy might be that an increased dietary Na intake enhances urinary total Ca excretion as seen in humans, but by augmenting urine volume, does not lead to substantially elevated urine Ca concentrations⁽Reference Lekcharoensuk, Osborne and Lulich^9, Reference Biourge, Devois and Morice²⁷⁾.

Overall, research indicates that the urine volume in cats can be increased successfully by feeding a high-moisture diet, as well as by increasing dietary Na intake. An increased Na intake might be used as a preventative measure for CaOx formation as there is no evidence that dietary Na induces elevated Ca concentrations in the urine of healthy cats, and there are no adverse long-term effects reported for this nutrient on, for example, blood pressure and kidney functioning as well⁽Reference Xu, Laflamme and Long^32, Reference Luckschander, Iben and Hosgood³³⁾. However, the effectiveness of increased dietary Na intake on reducing CaOx urolith formation in urolith-forming cats has not been tested.

Factors influencing urinary oxalate: exogenous urinary oxalates

Intake of dietary oxalate is a known factor that increases urinary oxalate excretion. The amount of dietary oxalate available for absorption in the gastrointestinal tract is dependent on the amount of free oxalate present in the diet, dietary components which can form complexes with free oxalate (for example, Ca, Mg) and the activity of oxalate-degrading bacteria in the gastrointestinal tract of man and animals.

In general, large amounts of oxalate are present in green leafy vegetables and bran concentrates, moderate amounts in nuts and cereals, and low amounts in dairy products, meat and fish⁽Reference Massey^37, Reference Siener, Honow and Voss³⁸⁾. The contribution of dietary oxalate intake to urinary oxalate excretion in free-roaming cats is unknown, but can be expected to be low, as the natural (carnivorous) diet of the cat is mainly composed of food/prey items that contain low amounts of oxalate^(39, Reference Plantinga, Bosch and Hendriks⁴⁰⁾. In contrast, omnivorous diets (for example, for humans and rats) contain moderate to high levels of oxalate, which may result in a relatively high contribution of exogenous oxalate to total urinary oxalate excretion. In humans, contributions of dietary oxalate to urinary oxalates have been reported, with estimates ranging from 25 to 68 %⁽Reference von Unruh, Voss and Sauerbruch^41, Reference Holmes, Goodman and Assimos⁴²⁾. Unfortunately, no such data are available in cats.

In thirty different dry pet foods for adult small breed dogs, the oxalate content was found to range between 9·6 and 60·0 mg/MJ ME (4–25 mg/100 kcal) with a mean of 26·3 mg/MJ ME (11 mg/100 kcal)⁽Reference Stevenson, Hynds and Markwell⁴³⁾. In a canned diet designed to assist in the management of LUTD in dogs, the oxalate content was 5·9 mg/MJ ME (2·5 mg/100 kcal)⁽Reference Stevenson, Blackburn and Markwell⁴⁴⁾. The average daily intake of oxalate (2·5–15·5 mg/kg body weight (BW) per d) in dogs fed these commercial dry diets can be considered to be relatively high compared with the average oxalate intake (2–3 mg/kg BW per d) by humans⁽Reference Stevenson, Hynds and Markwell⁴³⁾. Stevenson et al. ⁽Reference Stevenson, Hynds and Markwell⁴³⁾ studied the relative effects of dietary Ca and oxalate on the composition of urine by feeding healthy dogs a diet supplemented with 24·0 to 60·0 mg oxalate/MJ (10–25 mg/100 kcal) and varying the dietary Ca contents. These authors found that the oxalate excretion inconsistently increased with a higher oxalate intake only when dietary Ca intake was low (molar Ca:oxalate ratio <  18). With higher dietary Ca:oxalate ratios (> 25) the oxalate excretion remained low and stable, irrespective of dietary oxalate content. The variation between dogs when different oxalate levels were fed, in conjunction with a low dietary Ca content, was very high, with effects ranging from 0 to a 400 % increase in urinary oxalate excretion. These results indicate that the contribution of dietary oxalate to urinary oxalate excretion in dogs is dependent on the dietary Ca content, and is expected to be low with a high (> 40) molar Ca:oxalate ratio in the diet.

Other factors than dietary Ca are also known to influence intestinal oxalate absorption. A similar role for dietary Mg has been found⁽Reference Morozumi, Hossain and Yamakawa⁴⁵⁾. Both minerals can directly interact with oxalate to form an insoluble complex to lower the free oxalate concentration in the gastrointestinal tract⁽Reference Morozumi, Hossain and Yamakawa⁴⁵⁾, resulting in a reduction in the absorption of oxalate⁽Reference Holmes, Goodman and Assimos^42, Reference Liebman and Chai^46, Reference Liebman and Costa⁴⁷⁾. Fat and phosphate can act as scavengers for Ca, thereby indirectly increasing the availability of oxalate for uptake by the intestine⁽Reference Masai, Ito and Kotake^48, Reference Naya, Ito and Masai⁴⁹⁾. In addition, oxalate-degrading bacteria, such as Oxalobacter formigenes present in the intestinal tract of man and rats, are known to reduce the contribution of exogenous urinary oxalates to total urinary oxalates, as this bacterium uses oxalic acid (or its anion oxalate) as its sole energy source by degrading oxalic acid to formate⁽Reference Sidhu, Allison and Chow^50, Reference Troxel, Sidhu and Kaul⁵¹⁾. O. formigenes has also been detected in the stool of cats and dogs⁽Reference Weese and Palmer^52, Reference Weese, Weese and Rousseau⁵³⁾. Lactic acid bacteria have also been shown to degrade oxalates in vitro ⁽Reference Weese, Weese and Yuricek⁵⁴⁾. The impact of oxalate-degrading bacteria in the gastrointestinal tract of humans and animals on urinary oxalate excretion remains to be clarified.

Research in dogs indicates that, although canine foods contain relatively high levels of oxalate, the contribution of exogenous oxalates to the urinary oxalate excretion is expected to be of minor importance mainly due to the high molar Ca:oxalate ratio in commercial canine foods. The influence of factors other than dietary Ca that are known to affect intestinal oxalate absorption in humans is not known in dogs and cats, but is expected to be of minor importance in practice, as the dietary Mg content in the majority of commercial foods is approximately 10-fold lower than the Ca content, and dietary Ca and phosphate are normally balanced. Oxalate and Ca contents of dry commercial feline diets may be expected to be in the same range as canine foods, as similar ingredients are used. This would imply that the potential contribution of exogenous oxalates to total urinary oxalate excretion by domestic cats is expected to be marginal as well, as the Ca:oxalate ratio can be expected to be high. However, the oxalate content of commercial feline foods and its contribution to urinary oxalate excretion remains to be determined in order to ascertain its relative importance.

Factors influencing urinary oxalate: endogenous urinary oxalates

Since hyperoxaluria in humans is recognised as an important risk factor for CaOx urolith formation, many studies (mainly in rodents) have been conducted to unravel the aetiopathogenesis of hyperoxaluria. Studies in human subjects and rodents have revealed that the largest part of the endogenous oxalate formation originates from the conversion of sugars (including glucose and fructose), amino acids (including hydroxyproline, glycine, serine, phenylalanine, tyrosine and tryptophan) and/or glycolate, ending via the metabolism of glyoxylate to oxalate. Another precursor that can give rise to oxalate is ascorbic acid, which can break down non-enzymically in urine to produce oxalate, a reaction which is accelerated at alkaline pH. Originally it was thought that 40 to 50 % of urinary oxalate was derived from the breakdown of ascorbic acid⁽Reference Holmes, Goodman and Assimos^42, Reference Lulich, Osborne and Thumchai^55–Reference Ogawa, Miyazato and Hatano⁵⁷⁾. However, close scrutiny of the experimental procedures indicates that the breakdown occurred due to processing of urine at alkaline pH⁽Reference Holmes, Knight and Assimos⁵⁸⁾. In cats, ascorbate has not been found to be an important dietary precursor of oxalate⁽Reference Yu and Gross⁵⁹⁾, indicating that endogenous oxalate biosynthesis in cats relies predominately on glyoxylate metabolism.

Endogenous biosynthesis of oxalate occurs mainly in the liver⁽Reference Farinelli and Richardson⁶⁰⁾ and is highly dependent on the glyoxylate content in the hepatocytes (Fig. 3). Any glyoxylate that is not reduced to glycolate or transaminated to glycine is oxidised to oxalate, a reaction catalysed by cytosolic l-lactate dehydrogenase⁽Reference Behnam, Williams and Brink⁶¹⁾ (Fig. 3(a), step I). Since oxalate can be considered a metabolic ‘end-waste-product’ it will be quantitatively excreted in the urine, mainly via glomerular filtration. Oxalate excretion also occurs in the proximal tubule after a high oral oxalate load⁽Reference Holmes, Ambrosius and Assimos^62, Reference Knight, Holmes and Assimos⁶³⁾. The most efficient way of reducing urinary endogenous oxalate excretion is to reduce the content of glyoxylate in the hepatocyte (Fig. 3). This can be achieved in two ways: by supplying less dietary precursors for glyoxylate production in the hepatocyte or by metabolic removal of glyoxylate from oxalate synthesis.

Fig. 3 Proposed metabolic pathways of de novo oxalate synthesis via glyoxylate metabolism. Major metabolic conversions that are considered prominent in the feline hepatocytes when precursors from a carnivorous and omnivorous diet are present are shown in panels A and B, respectively. Metabolic conversions expected to play a major role are indicated by thick arrows. Metabolic conversions expected to play a minor role are indicated by dashed arrows. Essential metabolic conversions in situation A are indicated with I, II, III or IV, meaning: I, conversion of glyoxylate into oxalate catalysed by cytosolic l-lactate dehydrogenase (l-LDH)⁽Reference Behnam, Williams and Brink^61, Reference Mdluli, Booth and Brady⁶⁹⁾; II, conversion of glyoxylate into l-glycine catalysed by alanine:glyoxylate aminotransferase 1 (AGT1)⁽Reference Danpure, Guttridge and Fryer^64, Reference Lumb, Purdue and Danpure^74, Reference Danpure^75, Reference Ribaya and Gershoff^82, Reference Takayama, Fujita and Suzuki^83, Reference Zentek and Schulz^92, Reference Nguyen, Kalin and Drouve^100, Reference Noguchi, Okuno and Takada^101, Reference Ogawa, Hossain and Ogawa¹⁰⁵⁾; III, conversion of glyoxylate into glycolate catalysed by glyoxylate reductase/hydroxypyruvate reductase (GR/HPR)⁽Reference Mistry, Danpure and Chalmers^67, Reference Danpure, Jennings and Mistry^68, Reference Mdluli, Booth and Brady^69, Reference McKerrell, Blakemore and Heath^72, Reference Booth, Conners and Rumsby^155, Reference Genolet, Kersten and Braissant¹⁵⁶⁾; IV, conversion of hydroxypyruvate into d-glycerate catalysed by glyoxylate reductase/hydroxypyruvate reductase⁽Reference Mistry, Danpure and Chalmers^67, Reference Danpure, Jennings and Mistry^68, Reference Mdluli, Booth and Brady^69, Reference McKerrell, Blakemore and Heath^72, Reference Booth, Conners and Rumsby^155, Reference Genolet, Kersten and Braissant¹⁵⁶⁾. Essential metabolic conversions in situation B are indicated with Ia–d, II, III and IV, meaning: Ia, conversion of cytosolic d-fructose, d-glucose and d-galactose into d-glycerate⁽Reference Nguyen, Dumoulin and Henriet^86, Reference Nguyen, Dumoulin and Wolf^87, Reference Nguyen, Dumoulin and Henriet⁸⁹⁾; Ib, conversion of d-glycerate into hydroxypyruvate; Ic, conversion of hydroxypyruvate into glycolaldehyde⁽Reference Gambardella and Richardson⁸⁰⁾; Id, conversion of glycolaldehyde into glycolate⁽Reference Gambardella and Richardson⁸⁰⁾; II, conversion of peroxisomal glycolate into oxalate catalysed by glycolate dehydrogenase (GD)⁽Reference Yanagawa, Maeda-Nakai and Yamakawa⁹⁷⁾; III, conversion of peroxisomal glycolate into glyoxylate catalysed by glycolate oxidase (GO)⁽Reference Yanagawa, Maeda-Nakai and Yamakawa⁹⁷⁾; IV, conversion of glyoxylate into l-glycine catalysed by alanine:glyoxylate aminotransferase 1⁽Reference Danpure, Guttridge and Fryer^64, Reference Lumb, Purdue and Danpure^74, Reference Danpure^75, Reference Noguchi, Okuno and Takada^101, Reference Xue, Sakaguchi and Fujie¹⁵⁷⁾.

Glyoxylate removal can be achieved through the conversion into glycine, which is catalysed by serine:pyruvate aminotransferase/alanine:glyoxylate aminotransferase, also called alanine:glyoxylate aminotransferase 1 (AGT1)⁽Reference Danpure, Guttridge and Fryer⁶⁴⁾ (Fig. 3(a), step II). This enzyme can also convert serine into hydroxypyruvate. An essential cofactor for AGT1 is pyridoxine, or vitamin B₆. A deficiency in pyridoxine can lead to a shift from glyoxylate removal through conversion into glycine, to glyoxylate oxidation to oxalate. This shift has been well documented in cats. In a study with kittens that were fed a pyridoxine-deficient diet, a significantly higher daily urinary oxalate excretion was observed⁽Reference Bai, Sampson and Morris⁶⁵⁾.

Glyoxylate can also be converted into glycolate (Fig. 3(a), step III) via the action of the enzyme glyoxylate reductase/hydroxypyruvate reductase (GR/HPR), a 2-hydroxy-acid dehydrogenase. GR/HPR also catalyses the reduction of hydroxypyruvate to d-glycerate (Fig. 3(a), step IV) and the oxidation of d-glycerate to hydroxypyruvate⁽Reference Williams and Smith⁶⁶⁾. In the human liver, GR/HPR is found predominantly in the cytosol (> 90 %), with a small portion in the mitochondria⁽Reference Mistry, Danpure and Chalmers⁶⁷⁾. In cats, GR/HPR localisation has been reported to predominantly occur in the cytosol⁽Reference Danpure, Jennings and Mistry⁶⁸⁾.

The existence and importance of the conversions of glyoxylate into glycine and glycolate in the hepatocyte has been discovered as a result of extensive research into two severe autosomal-recessive disorders leading to mild to severe hyperoxaluria in humans. These disorders are termed primary hyperoxaluria (PH) and two types can be distinguished. PH type I is an inherited disorder of glyoxylate metabolism arising from a deficiency in AGT1, leading to a disrupted conversion of glyoxylate into glycine (Fig. 3(a), step II). PH type II is an inherited disease caused by mutations in GR/HPR, leading to a reduced conversion of glyoxylate into glycolate (Fig. 3(a), step III) and that of hydroxypyruvate into d-glycerate (Fig. 3(a), step IV)⁽Reference Mdluli, Booth and Brady^69, Reference Cramer, Ferree and Lin⁷⁰⁾. PH type II patients have a limited ability to convert d-glycerate into hydroxypyruvate, resulting in most cases in an elevated concentration of l-glycerate in the urine, as metabolic removal of d-glycerate occurs via conversion into l-glycerate (Fig. 3). The reported cases of PH in cats⁽Reference De Lorenzi, Bernardini and Pumarola^71, Reference McKerrell, Blakemore and Heath⁷²⁾ reflect those of human PH type II⁽Reference Danpure, Jennings and Mistry⁶⁸⁾, while cases of PH in cats analogous to human PH type I are unknown.

Highly notable is the observation that the spatial localisation of AGT1 seems to be species dependent. In carnivores and insectivores, AGT1 is mainly present in mitochondria⁽Reference Danpure, Fryer and Jennings^73, Reference Lumb, Purdue and Danpure⁷⁴⁾, while in the human liver an almost exclusive peroxisomal localisation of this enzyme has been reported⁽Reference Danpure⁷⁵⁾. The mitochondrial localisation of AGT1 is seen in carnivorous and insectivorous species of different genera (mammals, birds, reptiles)⁽Reference Danpure, Fryer and Jennings⁷³⁾, indicating that AGT1 localisation in the mitochondrion might be a necessity when consuming high-protein, low-carbohydrate diets. Interestingly, the Gly170Arg in combination with the Pro11Leu mutation in the gene AGXT, encoding for AGT1, which are commonly found in human patients with PH type I, targets AGT1 to mitochondria instead of to its normal location in peroxisomes⁽Reference Danpure⁷⁵⁾, resembling the mitochondrial localisation of hepatic AGT1 in carnivores⁽Reference Lumb, Purdue and Danpure⁷⁴⁾. Herbivores, such as Old World monkeys (macaques and baboons), rabbits and guinea-pigs, have most, if not all of their AGT1 located in peroxisomes⁽Reference Lumb, Purdue and Danpure⁷⁴⁾. Rodents (rats, mice, hamsters) and marmosets (New World monkey) have AGT1 distributed approximately equally between both organelles. The interspecies difference in intracellular localisation of hepatic AGT1 may very well be the result of dietary selection pressure during evolution⁽Reference Lumb, Purdue and Danpure⁷⁴⁾.

Genomic expression analyses have revealed that the adaptive shift in AGT1 targeting among species can be ascribed to the use of alternative transcription- and translation-initiation sites of the single-copy AGXT gene⁽Reference Oatey, Lumb and Danpure^76, Reference Birdsey, Lewin and Cunningham⁷⁷⁾. The longest transcript of AGXT, as found in cats, encodes a cleavable N-terminal mitochondrial targeting sequence (MTS) and C-terminal peroxisomal targeting sequences for peroxisomal uptake, in which MTS is dominant over peroxisomal targeting sequences in determining the final subcellular destination of AGT1⁽Reference Oatey, Lumb and Danpure^76, Reference Birdsey, Lewin and Cunningham⁷⁷⁾. Species such as man and rabbits expressing AGT1 almost exclusively in peroxisomes have lost the first translation start site during evolution resulting in an MTS-lacking protein. In cats and other species expressing AGT1 almost exclusively in mitochondria, almost all encoded proteins contain the MTS due to the loss of the – more downstream – second transcription-initiation site⁽Reference Lumb, Purdue and Danpure^74, Reference Birdsey, Lewin and Cunningham⁷⁷⁾. Species in which hepatic AGT1 is roughly equally distributed between both organelles utilise both translation-initiation sites⁽Reference Purdue, Lumb and Danpure⁷⁸⁾.

It has been postulated that the dual localisation of AGT1 in the liver is a reflection of its dual physiological function: that is, its role in glyoxylate detoxification in peroxisomes and its role in gluconeogenesis in mitochondria⁽Reference Danpure, Guttridge and Fryer^64, Reference Lumb, Purdue and Danpure⁷⁴⁾. Intraperoxisomal glyoxylate detoxification is essential in herbivores, since their diet is rich in glycolate and carbohydrates. To prevent oxidation of cytosolic glyoxylate to oxalate by l-lactate dehydrogenase (Fig. 3(a), step I), a high activity of intraperoxisomal AGT1 is required. High levels of glycolate and carbohydrates in combination with a scarcity of animal protein (containing l-hydroxyproline, glycine, serine), typically found in diets of herbivores, retard the need for directing hydroxyproline-derived glyoxylate into gluconeogenesis by conversions facilitated by AGT1 (Fig. 3(a), step II). In contrast, carnivores and insectivores are likely to have consumed a very low amount of glycolate and carbohydrates during evolution and these species have not required intensive detoxification of glyoxylate in the peroxisome. In contrast, the high animal protein and low carbohydrate content of the diet of carnivores and insectivores would clearly favour the contribution of gluconeogenesis in the mitochondrion⁽Reference Danpure, Guttridge and Fryer⁶⁴⁾.

As stated earlier, the glyoxylate content in the hepatocyte is also dependent on certain dietary precursors. In human medicine, it has been demonstrated that a high intake of amino acids (i.e. hydroxyproline, serine, glycine, phenylalanine, tyrosine and tryptophan)⁽Reference Purdue, Lumb and Danpure^78–Reference Takayama, Fujita and Suzuki⁸³⁾ or sugars (i.e. glucose, fructose, galactose, xylose)⁽Reference Ribaya-Mercado and Gershoff^84–Reference Nguyen, Dumoulin and Henriet⁸⁹⁾ stimulates oxalate synthesis from the intermediate product glyoxylate. Knight et al. ⁽Reference Knight, Jiang and Assimos^90, Reference Knight, Easter and Neiberg⁹¹⁾ showed that hydroxyproline is a potent contributor to urinary oxalates, with glycine contributing little to the endogenous production of glycolate or oxalate. In cats, the endogenous synthesis of oxalate has not been widely studied. A study by Zentek & Schulz⁽Reference Zentek and Schulz⁹²⁾ describes the effects of various dietary protein sources (collagen tissue, horse meat and soya isolate) on urine composition, including urinary oxalate excretion. Uptake of the (high)-protein diet with collagen as the protein source resulted in an increase in urinary oxalate excretion (2·83 (sd 0·89) mg/kg BW per d) compared with the (high)-protein diets with horse meat and soya protein isolate as the protein source (0·94 (sd 0·29) mg/kg BW per d and 1·17 (sd 0·53) mg/kg BW per d, respectively). Since the hydroxyproline content of collagen tissue is high (about 10–13 %), mitochondrial AGT1, which converts l-hydroxyproline-derived glyoxylate into glycine (Fig. 3(a), step II), might have been saturated in the cats fed the collagen protein, leading to undesirable conversion to oxalate by l-lactate dehydrogenase (Fig. 3(a), step I), resulting in an increased urinary oxalate excretion⁽Reference Takayama, Fujita and Suzuki⁸³⁾. Besides hydroxyproline, the amino acids glycine and serine are also present in relatively high concentrations in collagen tissue⁽Reference Laser Reutersward, Asp and Bjork⁹³⁾.

Zentek & Schultz⁽Reference Zentek and Schulz⁹²⁾ also investigated the effects of different dietary protein levels on daily urinary oxalate excretion. Of each protein source (collagen tissue, horse meat and soya isolate), two diets were fed, one containing a high and the other a low protein level. In the low-protein diets, protein was exchanged with rice (carbohydrate) and animal fat. Remarkably, all low-protein diets showed a consistently higher daily urinary oxalate excretion. The oxalate excretion increased approximately 5-fold (from 2·83 (sd 0·89) to 13·70 (sd 4·32) mg/kg BW per d) when fed the low-protein diet with collagen as the protein source compared with the high-protein diet. Urinary oxalate excretion increased 4-fold when changing from high- to low-protein diets using soya isolate and horse meat as a protein source (soya isolate diet from 1·17 (sd 0·53) to 4·79 (sd 2·73) mg/kg BW per d and the horse meat diet from 0·94 (sd 0·29) to 3·62 (sd 2·44) mg/kg BW per d). However, the observed increase in urinary oxalate excretion might have been the result of a concomitant high carbohydrate intake with the low-protein diets. As stated earlier, in humans sugars (i.e. glucose, fructose, galactose, xylose) are known to be a substrate for the peroxisomal glyoxylate pathway⁽Reference Ribaya-Mercado and Gershoff^84–Reference Nguyen, Dumoulin and Henriet⁸⁹⁾. It is conceivable that when fed a high-carbohydrate diet, cats suffer an overload of hepatic oxalate as a consequence of a deficiency in peroxisomal AGT1 and excrete more oxalate in their urine. Although the urinary oxalate excretion was found to be unaffected in a study feeding healthy cats a dry commercial diet with a low carbohydrate content (19·1 g/MJ ME) v. a high carbohydrate content (34·7 g/MJ ME)⁽Reference Mustillo, Bartges and Kirk⁹⁴⁾, the carbohydrate content of the low-carbohydrate diet in this study can still considered to be high compared with the carbohydrate content of the experimental diets in the study of Zentek & Schultz⁽Reference Zentek and Schulz⁹²⁾ (i.e. collagen tissue diets: 2·7 v. 23·2 g/MJ ME; soya diets 7·1 v. 22·3 g/MJ ME; horse meat diets 5·9 v. 25·3 g/MJ ME).

A proposed metabolic pathway for the synthesis of the undesirable metabolite oxalate in cats from sugars is presented in Fig. 3(b). In this model, the sugars d-fructose, d-glucose and d-galactose are first converted to d-glycerate (Fig. 3(b), step Ia). Since the activity of fructokinase in feline liver is found to be relatively high and that of glucokinase to be low⁽Reference Tanaka, Inoue and Takeguchi⁹⁵⁾, the contribution of d-fructose compared with d-glucose is expected to be higher. d-Glycerate can subsequently be converted into glycolate via the intermediates hydroxypyruvate and glycolaldehyde (Fig. 3(b), steps Ib, Ic and Id). The sugar d-galactose⁽Reference Ribaya-Mercado and Gershoff⁸⁴⁾ can also be converted to glycolate, with only glycolaldehyde as an intermediate. The sugar xylose enters the metabolic pathway to oxalate via glycolaldehyde as well⁽Reference Rofe, Thomas and Edwards⁹⁶⁾. Cytosolic glycolate may partly be directed into the peroxisomes by diffusion where glycolate will be either oxidised directly to oxalate by the enzyme glycolate dehydrogenase⁽Reference Gambardella and Richardson⁸⁰⁾ (Fig. 3(b), step II) or converted to glyoxylate by the enzyme glycolate oxidase⁽Reference Yanagawa, Maeda-Nakai and Yamakawa⁹⁷⁾ (Fig. 3(b), step III). In most animals, the major part of this glyoxylate will be transaminated to glycine by AGT1 (Fig. 3(b), step IV) (i.e. peroxisomal glyoxylate detoxification), but in cats, and other carnivores, most of this glyoxylate will probably be converted back to glycolate due to a deficit of AGT1 activity in the peroxisome. This surplus in glycolate will probably result in an increased conversion of glycolate to oxalate catalysed by glycolate dehydrogenase (Fig. 3(b), step II).

An additional explanation for the increase in oxalate synthesis upon feeding diets high in carbohydrate and low in protein content may be a reduction of gluconeogenesis in the liver, which results in a lower need for glycine and serine as a gluconeogenic precursor. It is conceivable that high concentrations of glycine and/or serine attenuate mitochondrial AGT1 in a negative feedback loop. This might result in an additional shift to oxalate synthesis. Collectively, it is possible that in cats a high carbohydrate intake is a potential risk factor for CaOx urolith formation by increasing endogenous oxalate synthesis.

The prevention of oxalate synthesis in cats by expressing AGT1 almost exclusively in the mitochondrion can be considered as one of the many enzymic adaptations as a result of their obligatory carnivorous lifestyle throughout evolution. By expressing AGT1 almost exclusively in the mitochondrion, the formed glyoxylate (mainly originating from hydroxyproline) is efficiently transaminated to glycine (Fig. 3(a), step II), which can directly be used for intramitochondrial gluconeogenesis. This enzymic adaptation appears to be part of a long list of adaptations in this species, such as cysteine dioxygenase and cysteinesulfinate decarboxylase (taurine synthesis), pyrroline-5-decarboxylase and ornithine aminotransferase (citrulline synthesis), dioxygenase (retinol synthesis), glucokinase (glucose metabolism), 7-hydroxycholesterol-Δ7 reductase (25-hydroxycholesterol synthesis) and picolinic carboxylase (nicotinic acid synthesis)^(39, Reference Morris⁹⁸⁾.

Dietary factors influencing endogenous urinary oxalate synthesis

In cats, urinary oxalate excretion has been found to be inversely correlated with protein intake and dependent on the protein source⁽Reference Zentek and Schulz⁹²⁾, although the inverse correlation with protein intake might be the result of a higher carbohydrate intake. In contrast, in human subjects some studies reported that urinary oxalate excretion increased with increased dietary protein intake and decreased with protein restriction⁽Reference Curhan, Willett and Rimm^99, Reference Nguyen, Kalin and Drouve¹⁰⁰⁾. This might be explained by the distinct subcellular AGT1 distribution in humans and cats.

AGT1 catalyses the hepatic transamination of glyoxylate, being an important precursor of oxalate, to glycine⁽Reference Noguchi, Okuno and Takada¹⁰¹⁾. Pyridoxine, or vitamin B₆, is a known cofactor for AGT1⁽Reference Takada, Mori and Noguchi¹⁰²⁾ and, as such, pyridoxine deprivation could indirectly elevate endogenous oxalate biosynthesis and, in turn, its urinary excretion⁽Reference Teerajetgul, Hossain and Yamakawa^103–Reference Ogawa, Hossain and Ogawa¹⁰⁵⁾. Indeed, the daily urinary oxalate excretion was higher in kittens fed a pyridoxine-deficient diet compared with those receiving adequate amounts of pyridoxine⁽Reference Bai, Sampson and Morris^65, Reference Bai, Sampson and Morris¹⁰⁶⁾. Moreover, pyridoxine can act as a cofactor for various enzymes in the tricarboxylic acid cycle as well, leading to a decreased synthesis of citrate. Thus, under the condition of pyridoxine deficiency, citrate metabolism may be impaired, leading to a lower urinary citrate concentration and increased risk of precipitation with CaOx⁽Reference Teerajetgul, Hossain and Yamakawa¹⁰³⁾. To the authors' knowledge, the effect of pyridoxine on urinary citrate excretion in cats has not been studied.

Although an excess consumption of ascorbate (i.e. vitamin C) has been associated with increased endogenous oxalate synthesis (by the non-enzymic oxidation of ascorbate to oxalate)⁽Reference Chai, Liebman and Kynast-Gales^36, Reference Auer, Auer and Rodgers^107–Reference Baxmann, De O G Mendonça and Heilberg¹¹⁰⁾ in humans, in healthy cats dietary supplementation with ascorbate up to 193 mg/kg did not affect urinary oxalate concentrations⁽Reference Yu and Gross⁵⁹⁾. However, only the effect of a moderate and not a high amount of ascorbate on the urinary oxalate excretion has been tested. As cats, in contrast to man, can synthesise ascorbate in sufficient amounts de novo from glucose, ascorbate is not an essential nutrient⁽¹¹¹⁾. Although there is no evidence that dietary ascorbate can increase urinary oxalate levels in cats, it is generally recommended to avoid excessive amounts of dietary ascorbate. This recommendation is purely based on the results obtained in human subjects.

Dietary factors influencing urinary calcium excretion

Conflicting results exist in the literature on the role of dietary Ca in inducing and resolving hypercalciuria. For decades, the prevailing consensus was that dietary Ca restriction would reduce urinary Ca excretion, and as such CaOx formation⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. However, recent literature in humans and dogs indicates that there might be an advantage of increased dietary Ca intake, which is thought to be related to interactions between Ca and oxalate in the intestinal lumen⁽Reference Stevenson, Blackburn and Markwell^44, Reference Penniston and Nakada¹¹⁷⁾. When sufficient dietary Ca is available, complexation with oxalate occurs in the intestinal lumen, which in turn results in a reduction of intestinal absorption and renal excretion of both exogenous Ca and oxalate. Furthermore, a retrospective case–control study in cats revealed that consuming diets with a relatively low amount of Ca (0·23 to 0·49 g/MJ ME) had a significantly greater risk for CaOx urolith formation than cats that consumed higher dietary levels of Ca⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. This association may be explained by a higher oxalate absorption from the gastrointestinal tract.

Low dietary intake of P may be related to an increased urinary Ca excretion due to a lower binding of Ca by phosphate ions in the gastrointestinal tract resulting in an increased Ca absorption⁽Reference Pastoor, Van't Klooster and Mathot¹¹⁸⁾. An increase in urinary Ca excretion due to a low dietary P content could explain the increased risk for developing CaOx uroliths in cats fed diets containing a low P content (0·20 to 0·42 g/MJ ME)⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. On the other hand, the same study indicates that diets with a high P content (0·76 to 1·13 g/MJ ME) correlate with an increased risk of CaOx urolith formation compared with diets with moderate P content (0·66 to 0·76 g/MJ ME). Excessive amounts of P present in the gastrointestinal tract could compete with oxalate, preventing intestinal Ca complexation with oxalate, which in turn could increase the availability of free oxalate for intestinal absorption and eventually renal excretion⁽Reference Lekcharoensuk, Osborne and Lulich^9, Reference Curhan, Willett and Speizer^119, Reference Bushinsky, Bashir and Riordon¹²⁰⁾. The precise influence of dietary P on hypercalcaemia and CaOx formation in cats remains unclear and should be further studied.

Results of a retrospective case–control study revealed that cats fed diets with relatively high K contents (0·52 to 0·77 g/MJ ME) were less than half as likely (OR 0·45) to develop CaOx uroliths compared with cats fed diets with low K contents (0·23 to 0·38 g/MJ ME)⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. In cats, no studies have been conducted that may explain this relationship. However, in human subjects, studies have observed reduced Ca excretion after K supplementation⁽Reference Lemann, Pleuss and Gray^121, Reference Lemann, Pleuss and Gray¹²²⁾. The authors of these studies suggested that the decrease in urinary Ca during K administration may be related to the natriuretic effects of K, resulting in extracellular fluid-volume contraction or to K-induced phosphate retention and/or suppression of 1,25-dihydroxyvitamin D₃ synthesis⁽Reference Lemann, Pleuss and Gray¹²¹⁾. Another explanation for the relationship of a higher K content with a decrease in CaOx formation is the alkalising effect of K salts, leading to a higher urine pH and therefore a lower potential for Ca and oxalate to precipitate. Whether these suggested mechanisms are effective in cats should be investigated further.

Dietary factors influencing urine pH

Dietary protein of animal origin may decrease the base excess of the diet and increase the acid intake. As described above, a high acid intake can result in an increase in urinary Ca excretion. It can also reduce urinary citrate excretion since protons are able to associate with citrate in the intestine and kidney⁽Reference Dussol, Iovanna and Rotily¹³⁶⁾. This mechanism might explain why humans do have a higher risk for CaOx formation following the consumption of high amounts of animal protein⁽Reference Curhan, Willett and Rimm^99, Reference Trinchieri, Mandressi and Luongo¹³⁷⁾. In cats, however, an opposite effect has been found. In a retrospective case–control study it was demonstrated that cats fed diets containing high amounts of protein (25·1 to 33·0 g/MJ ME) were less than half as likely (OR 0·44) to develop CaOx urolithiasis compared with cats receiving low amounts of protein (12·4 to 18·9 g/MJ ME)⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. A possible explanation might be that in cats the consumption of animal protein is often accompanied by stimulated water consumption, eventually resulting in an increase in urine volume, and an increased P excretion without altering Ca excretion⁽Reference Funaba, Hashimoto and Yamanaka²⁰⁾. The observed lower oxalate excretion in healthy cats fed a high-protein diet compared with a low-protein diet⁽Reference Zentek and Schulz⁹²⁾ might explain this association as well (Fig. 3).

The precise mechanism in which urinary Mg may influence CaOx formation is still largely unknown. Originally, Mg was thought to act as a mild inhibitor of CaOx crystallisation⁽Reference Johansson, Backman and Danielson¹³⁸⁾. Another plausible explanation is that Mg, given as alkali salts, increase the urine pH, which in turn stimulates urinary citrate excretion by the kidney, which reduces the risk of CaOx crystal formation⁽Reference Rattan, Sidhu and Vaidyanathan^139, Reference Bonny, Rubin and Huang¹⁴⁰⁾.

A retrospective case–control study in cats revealed that diets with a low Mg content (22 to 43 mg/MJ ME) were associated with CaOx urolith formation⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. Several studies (in other species) have reported an association between low dietary Mg and CaOx urolith formation as well⁽Reference Rattan, Sidhu and Vaidyanathan^139, Reference Schwartz, Bruce and Leslie^141–Reference Ogawa, Yamaguchi and Morozumi¹⁴³⁾. In addition, in human medicine an increased dietary intake of Mg salts (containing citrate) is recommended in patients suffering from CaOx urolithiasis⁽Reference Rattan, Sidhu and Vaidyanathan^139, Reference Ettinger, Pak and Citron^144, Reference Tiselius¹⁴⁵⁾. An additional mechanism of action is the formation of complexes between Mg and oxalate, thereby reducing the supersaturation with CaOx. However, in a study with cats fed a diet with a high Mg concentration, despite a significant increase in urinary Mg, no significant effect was observed regarding urinary CaOx crystal growth inhibition, agglomeration inhibition or solubility, compared with the base diet⁽Reference Reed, Markwell and Jones¹⁴⁶⁾. Interestingly, in the earlier mentioned retrospective case–control study, diets high in Mg content (86 to 336 mg/MJ ME) were associated with a higher risk for CaOx urolith formation compared with diets with a moderate Mg content (43–86 mg/MJ ME)⁽Reference Lekcharoensuk, Osborne and Lulich⁹⁾. An explanation might be that in cats the contribution of dietary oxalate to urinary oxalate excretion is less prominent and therefore complexation of oxalate in the gastrointestinal tract by Mg less effective. Moreover, in cats there is no evidence that hypocitrauria is a risk factor for CaOx urolith formation.