INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

WILD HERBIVORES spend most of their time grazing or ruminating. This behavior ensures a regular flow of digesta and causes small, momentary fluxes of digestive fluids across the alimentary tract. In cattle, sheep, and ponies on a hay diet, the daily parotid secretion is [in liters/100 kg body weight (body wt)] ~9-17 (15), 6-21 (15), and 7-8 (2), respectively. This salivary secretion corresponds to ~1-4 times the plasma volume, and if this fluid shift were to take place within a short period, it would certainly have the potential to affect fluid-balance regulation. Accordingly, a large single meal fed to sheep (5) and ponies (6) has been shown to cause a significant postprandial hypovolemia that did not occur in animals fed more frequently. Single meals can also increase the plasma renin level in sheep (5) and levels of both renin and aldosterone in goats (8). It has been concluded that, in nonexercising horses (7) and ponies (6), feeding is a major stimulus for the renin-angiotensin-aldosterone system (RAAS).

Athletic horses undergo regular exercise training and competition, and the associated sweating results in substantial losses of sodium, which cannot be compensated for by a grass and grain diet. However, herbivores are known to have a well-developed appetite for sodium, and one of the most common ways of supplying sodium to them is by offering a salt block. The mechanisms regulating sodium appetite are not fully understood, but, in some species, aldosterone and angiotensin II may play a role. If feeding frequency affects the RAAS, then sodium appetite may also be influenced by feeding.

Previous studies of the effect of feeding frequency have been made in animals fed concentrated feed at a maintenance level. These studies have also been made under laboratory conditions, with animals kept in metabolism cages or subjected to some other type of restraint. In addition, few studies have been made on the voluntary sodium intake of animals subjected to physiological sodium losses, such as sweating. In the present study, the effects of feeding frequency on fluid shifts, plasma aldosterone concentration, and voluntary sodium intake were investigated in athletic horses without having to encroach on their daily routine to any great extent.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. Six Standardbred geldings (CD, CW, LS, MK, SB, WS) were used (body wt, 407-522 kg; age, 5-10 yr). The horses belonged to the Swedish Academy of Trotting and Thoroughbred Racing in Örebro, Sweden. They were used in the education of apprentices and professional race trainers and performed a simulated 2,140-m trotting race or interval training once a week. On days when no intensive exercise was performed, the horses spent the morning hours in paddocks or had lighter exercise. The study took place between April and June in 1995. During this period, the average ambient temperature per 24 h increased from 10 to 16°C. The study was approved by the Uppsala Local Ethics Committee.

Diets and feeding frequencies. Two feeding frequencies were studied during two 25-day-periods according to a change-over design. The horses were fed either twice a day (2TD; at 1730 and 0530) or six times per day (6TD; at 1730, 2130, 0130, 0530, 0930, and 1330). The horses had been fed the experimental diet for at least 4 wk before the study began. The diet consisted of 7.4 ± 0.3 kg grass hay and 4.1 ± 0.2 kg concentrates per day, corresponding to the Swedish recommendations (15a) for energy and protein supply for hard-working horses (21.6 MJ/100 kg body wt, 124 g digestible protein/100 kg body wt). The daily amounts of hay and concentrates were evenly distributed among the meals. The stable was divided into two compartments, and horses subjected to the same treatment were housed together to avoid psychological disturbances. The horses were housed in individual boxes (10 m²) and had had free access to a 2-kg salt block (99% NaCl) for several months before the study. The salt block was weighed every afternoon throughout the study. Total sodium intake was calculated as the sum of the daily weight loss of the salt block and the amounts of sodium in the hay, concentrates, and mineral supplement. Concentrations of sodium and potassium in the drinking water were negligible. Water was offered ad libitum from graduated buckets, and water intake was measured every fourth hour throughout the study.

Blood sampling and fecal collection. On day 17 in both treatments, blood samples were taken every hour for 24 h to determine a resting profile for each animal. Additional blood samples were taken every 15 min for 1 h after the 1730 and 0530 meals were provided. The blood samples were taken via a catheter introduced into one of the jugular veins (Intranule, 2.0 × 105 mm, SweVet-Piab, Sjöbo, Sweden). An extension tube was attached to the catheter to facilitate blood sampling. The horses could move freely in their boxes during the experiment, and, to avoid the animals' discomfort, no samples were taken when the horses were lying down. At 0330, almost all horses were lying down; therefore, this sample was omitted. All feces produced during 3 days (days 19-21) were collected in both feeding-frequency treatments. Each horse was fitted with a collecting device, consisting of a bag hanging under the tail just below the anus and attached to a girth and a brest. The bag was emptied manually at least every 90 min, and the feces were frozen at 20°C. After the feces from each horse and treatment were mixed, one sample was taken. These samples were dried (65°C, 24 h), milled (1 mm), ashed (600°C, 12 h), and dissolved in 1 M HNO₃ (1 h). After neutralization of the samples (1 M NH₃), the sodium and potassium concentrations were measured by using an ion-selective electrode. Evaluation of this method has shown that the results are similar to those obtained with fresh samples boiled in HNO₃ and perchloric acid. Occasionally, the collecting device was displaced in horse CW; therefore, fecal data were excluded from the feeding-frequency analysis. The horses were hand walked or lunged for 15 min during the collection days.

Exercise test. In the morning (0800-1200) of day 25 in both experimental periods, an exercise test (ET) was performed on a field track. In the ET, the horses trotted five 500-m intervals at a speed of 11.1 ± 0.2 m/s (heart rate ~205 beats/min; heart-rate recorder Optipuls, Borlänge, Sweden). After each interval, the horses returned to the start in a slow trot. This return trip took 2-4 min, depending on whether or not blood samples were taken (see below). All ETs were preceded by a warm-up in a slow trot (14 min). The horses were returned to the stable in slow trot (13 min) immediately after the last interval. In the evening before the ET days, a catheter was inserted into a jugular vein and flushed with heparinized saline (0.2% heparin) overnight and with isotonic saline between samplings. Blood samples were taken at rest, after the two last intervals (within 60 s), and 30 min after the last interval. The ETs were performed at the same time of day for each horse on both occasions. The horses were exercised in fixed pairs and driven by the same person on both occasions. The weather and track conditions were almost identical on the two ET days (11-17°C, 62-69% relative humidity, and sunshine). Four of the six horses (CD, CW, MK, SB) completed both ETs, and only these horses are included in the statistical analysis of exercise data for the two feeding frequencies.

Analysis. The blood samples were collected in lithium- heparinized tubes kept on ice until centrifugation and thereafter frozen at 20°C. The packed cell volume (PCV) was determined in duplicate by centrifugation of blood in capillary tubes (12,000 rpm; ALC microhematocrit centrifuge, Milan, Italy). Total plasma protein concentration (TPP) was measured with a refractometer (Cambridge Instruments, Buffalo, NY). Plasma osmolality was analyzed by freezing-point depression (3-W Wide-Range Osmometer, Advanced Osmometer, Roebling, Berlin, Germany). Analyses of fecal and plasma sodium and potassium concentrations were made by using an ion-selective electrode method (system E2A electrolyte analyzer; Beckman Instruments, Brea, CA). Plasma aldosterone concentration (PAC) was determined after extraction of fat and proteins (acetone and petroleum ether extraction) by using a commercially available radioimmunoassay kit for aldosterone (Coat-a-Count; Diagnostic Products, Los Angeles, CA). The samples for the standard curve were extracted in the same way. The quality control was run by using MultiCalc software version 2.0 (Wallac, Turku, Finland). The minimum detectable concentration was 27.4 pmol/l, the within-assay variation was 4.1%, and the between-assays variation was 7.5%.

Statistical analysis. Values are presented as means ± SE. All data were subjected to analysis of variance (GLM procedure in the Statistical Analysis Systems package, SAS Institute, Cary, NC) by using the model where Y_{i j k} is the observation, µ is the mean value, _i is the effect of animal, _j is the effect of treatment, _k is the effect of sample, ()_{j k} is the effect of interaction between treatment and sample, e_{i j k} is the residuals (independent with mean = 0 and variance = ²_e). The P value for significance within and between treatments was <0.05. The individual diurnal data for PAC and plasma sodium concentrations were pooled into three groups of sodium intake, regardless of feeding frequency: low sodium intake (12-16 mg · kg body wt¹ · day¹), medium sodium intake (17-26 mg · kg body wt¹ · day¹), and high sodium intake (37-70 mg · kg body wt¹ · day¹). Intake-level groups were the same during the two periods for each of the horses except for horse CD, which had a high intake on 2TD and a medium intake on 6TD. Individual data were subjected to linear (linear least-squares-regression analysis) and nonlinear regression analysis (Slide Write Plus version 3.0, Advanced Graphics Software, Carlsbad, CA). The P value for significance was <0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Short-term effects of feeding 2TD and 6TD. All horses consumed the concentrate first, immediately after feeding in both the 2TD and 6TD treatments. The hay was consumed next, but some of it was sometimes left uneaten for hours in 2TD. TPP and plasma osmolality increased after feeding in 2TD (Table 1). The PCV, PAC, and plasma sodium concentration were not affected by feeding 2TD or 6TD, but some changes were seen in plasma potassium concentration (Table 1).

Diurnal effects of feeding 2TD and 6TD. The TPP and plasma osmolality were lower in 2TD than in 6TD on several occasions in the late night and morning (Fig. 1). The PCV was not affected by feeding frequency (Fig. 1). Individual variation in TPP and PCV were greater in 2TD than in 6TD (Fig. 1).

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Diurnal changes (means ± SE) in total plasma protein concentration (top), packed cell volume (middle), and plasma osmolality (bottom). Left: , horses fed 2 times/day; right: , horses fed 6 times/day. Arrows, times at which feed was given. Values represented by solid symbols are significantly different from those obtained at 1830. * Significant difference from corresponding value obtained when fed 6 times/day; P < 0.05.

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Diurnal changes (means ± SE) in plasma aldosterone (top), sodium (middle), and potassium concentrations (bottom). Left: , horses fed 2 times/day; right: , horses fed 6 times/day. Arrows, times at which feed was given. Values represented by solid symbols are significantly different from those obtained at 1830. * Significant difference from corresponding value obtained when fed 6 times/day; P < 0.05.

Individual sodium intake and its effects. The daily individual sodium and water intakes are shown in Table 3. Voluntary sodium intake from the salt block was in the range of 0-62 mg · kg body wt¹ · day¹, but, on an individual basis, the daily variation was small. The mean voluntary sodium intake was not affected by feeding frequency. Individual water intake and individual sodium intake showed a positive linear correlation (y = 36.5 + 0.22x; R² = 0.54, P < 0.01).

1

1

1

1

1

1

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Diurnal changes (means ± SE) in plasma aldosterone concentration in relation to daily sodium intake. , Low intake, 12-16 mg · kg body wt1 · day1 (n = 4 horses). , Medium intake, 17-26 mg · kg body wt1 · day1 (n = 5 horses). , High intake, 37-70 mg · kg body wt1 · day1 (n = 3 horses). Solid symbols are significantly different from values at 1830; P < 0.05.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Concentrations of sodium () and potassium () in feces in relation to daily sodium intake (n = 12 horses). Sodium concentration decreased when level of sodium intake decreased [y = 5.00  9.07e(x/15.63); R2 = 0.96, P < 0.001], and potassium concentration increased when sodium intake decreased [y = 7.46 + 19.61e(x/9.62); R2 = 0.78, P < 0.001]. Symbols represent individual horses. DW, dry weight; BW, body wt.

Effects on exercise. All horses completed the ETs without signs of fatigue. There were no differences between 2TD and 6TD in TPP, PCV, osmolality, plasma sodium, potassium, or aldosterone concentration before, during or after the ET (Table 4). The simulated races caused a similar weight loss of 11 ± 3 and 13 ± 4 kg for 2TD and 6TD feeding, respectively (NS).

(x/8.89)

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Packed cell volume after 4th and 5th time intervals. Values are means ± SE in relation to daily sodium intake (n = 10 horses). Packed cell volume fits equation y = 57.43 + 18.41e(x/10.40); R2 = 0.59, P < 0.05. Symbols represent individual horses.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Contrary to earlier reports (6, 7), fluid shifts and PAC showed no major alterations in relation to the number of times per day that the horses were fed (2TD vs. 6TD). Although voluntary sodium intake was not affected by feeding frequencies, it did induce significant alterations in the PAC, PCV, water intake, and fecal excretion of sodium and potassium. There was great individual variation in sodium intake, and four of the horses consumed sodium in amounts less than or equal to the suggested maintenance requirement of 20 mg · kg body wt¹ · day¹ (16). Considering that these were athletic horses, the intakes were surprisingly low.

As expected, the highest PACs were found in the horses with the lowest sodium intake. The capacity to minimize urinary sodium losses is good in horses, both at rest (21) and after exercise (13). However, in herbivores, the sodium balance is not necessarily under primary control of the kidneys, because, under some conditions, it may be affected mainly by fecal excretion (18). In horses (1, 21) and in sheep (18), the sodium excretion in feces may well exceed that in the urine. Aldosterone has been shown to stimulate sodium absorption in the intestinal tract of horses (4), and this mechanism is dependent on the exchange of either potassium or hydrogen ions. In this study, the sodium and potassium contents of the feces were reciprocally related. The most dramatic interactions occurred at sodium intakes <26 mg Na · kg body wt¹ · day¹ when PAC was also elevated. At this level, potassium excretion was about six times higher than sodium excretion. In addition, at sodium intakes <26 mg · kg body wt¹ · day¹, aldosterone was released according to a diurnal rhythm, with elevated plasma levels occurring between 2130 and 0430. To our knowledge, this is the first observation of a diurnal rhythm in PAC in the horse, although a tendency for PAC to be higher in the early morning has been reported earlier (7). It is widely accepted that there is a circadian rhythm in PAC in humans and that body posture affects the RAAS. In humans, it has also been shown that the circadian rhythm is amplified by a restricted sodium intake (14), but the origin of the rhythm seems to be complex. The present study shows that changes in body posture will not have any sustainable effect on PAC in horses, because this level did not change between 0230 and 0430, a period during which almost all horses had been lying down.

Increases in both plasma renin and PAC in response to feeding in horses were shown by Clarke et al. (7); therefore, meal intake was affirmed as a physiological event that causes significant diurnal variation in the RAAS. Increased postprandial renin secretion has also been reported in ruminants (5, 8). In all of these reports, it is suggested that the activation of RAAS is due to the secretion of digestive fluids and the resulting hypovolemia. Therefore, the difference in postprandial PAC between earlier investigations and the present study might be related to the magnitude of the postprandial decrease in plasma volume. In the present study, TPP increased by 3-6% in animals receiving 2TD feeding. In earlier studies on horses (7) and ponies (12) in which the feeding interval was similar to that in our study (12 h) or longer (24 h), TPP increased by as much as 11-12% after feeding. There can be several explanations for the comparatively low increase in TPP during feeding in the present study. First, our horses were athletic and were offered almost twice as much feed per kilogram body weight as the animals in the studies mentioned above. As a result, our horses might have been less excited during feeding and might have fed less voraciously. Blair-West and Brook (5) noticed that renin was secreted in sheep that ate rapidly but not in slow feeders. Nevertheless, in the present study, TPP increased by 10% in the most voracious individual, without any increase in PAC. Second, the ponies and horses in the earlier studies were offered only pelleted feed, which has been shown to be consumed more rapidly than long hay (17). If feed is consumed at a slower rate, the digestive fluids will be reabsorbed, and an equilibrium between fluid secretion and reabsorption will be reached.

There were no differences in daily water intake between 2TD and 6TD protocols. This is in accordance with a study on ponies fed once per day or 6TD (12). However, the amounts of water were more evenly distributed throughout the day when the horses were fed 6TD. This pattern of water intake may have supported the less variable maintenance of the fluid homeostasis, illustrated by the small individual variation in TPP and PCV in 6TD. There was a positive correlation between water intake and sodium intake. Body water content or plasma volume was not measured in the present study, but the increased PCV during exercise may indicate that a reduction of the plasma volume and/or cellular hypervolemia had occurred in horses with a low sodium intake. An increase in PCV, while on low-sodium diets, has been shown in resting rats (10) and humans (22) as well as after exercise in rats (11). That the increase in PCV in the present study was due to a decreased plasma volume and/or to cellular hypervolemia rather than to differences in fluid shifts during exercise is supported by the finding that the individual changes in TPP and osmolality that occurred when animals shifted from rest to exercise were similar, irrespective of the sodium intake. It has previously been reported that a comparatively high red cell volume during exercise could be related to a decrease in the work capacity in Swedish Standardbred horses (20).

In the present study, we did not observe any alterations in the response to exercise when feeding frequency was changed. The horses were not exercised within 2.5 h after feeding, and, according to the results from resting conditions, this should have been long enough to prevent prior feed intake from having any effect on fluid and electrolyte regulation. The exercise-induced increase in TPP probably reflects a decrease in plasma volume due to increased hydrostatic pressure and a true fluid loss due to sweating. The increase in plasma osmolality was mainly due to the accumulation of lactate (data not shown).

The variation in individual voluntary sodium intake is intriguing. Why did so many of the horses regulate their sodium losses through internal strategies instead of increasing their sodium intake? The daily voluntary sodium intake was almost identical during the two periods in four of six horses, suggesting that these horses regulated their sodium intake by some physiological signal. It has been shown in sheep (9) and rats (19) that sodium-deficient animals will not replete their sodium reserves when offered concentrated sodium solutions. These authors suggest that consuming concentrated salt solutions causes taste aversion and/or a temporary inhibition of the sodium appetite. A salt block might, therefore, not be an optimal source of sodium for athletic horses.

The present study has shown that feeding athletic horses large meals had a short-term effect on their hemodynamics but did not influence their diurnal PAC or voluntary sodium intake. PAC, on the other hand, was significantly influenced by voluntary sodium intake.