Fish vary in their ability to utilize dietary carbohydrate. Certain species have been reported to utilize simple sugars as well as or better than complex carbohydrates, whereas other species are unable to utilize simple sugars effectively as an energy source (NRC 1993). In general, complex carbohydrates such as dextrin are utilized by most fish better than simple sugars such as glucose (Wilson 1994). The exact mechanism for the relatively poor utilization of simple sugars by fish compared with land animals is still unclear. Glucose tolerance tests performed on various fish species have consistently resulted in prolonged hyperglycemia, which resembles diabetes in mammals (Furuichi and Yone 1981, Shiau and Chen 1993, Wilson and Poe 1987).

Trivalent chromium has been shown to improve glucose utilization in humans (Mertz 1993) and land animals (Amoikon et al. 1995, Striffler et al. 1995). Chromium is generally considered to act as a cofactor in the initiation of insulin action (Mertz et al. 1974). Insulin is the primary hormone that controls how the cells absorb, use and store nutrients. Through its association with insulin, chromium is therefore involved in carbohydrate metabolism and works with insulin to move glucose into cells.

Little is known about the nutritional effects of dietary chromium in fish. Most of the research that has been conducted to date has been limited to only one species, the hybrid tilapia. Shiau and Lin (1993) demonstrated that dietary chromium supplementation in the form of chromium chloride significantly improved glucose but not starch utilization by tilapia. Shiau and Chen (1993) later reported that the improvement of glucose utilization by tilapia fed diets supplemented with chromic oxide was markedly higher than with other forms of chromium including chromium chloride. Researchers from the same laboratory group in Taiwan (Shiau and Liang 1995) then went on to report that the level of chromic oxide in the diet alters glucose utilization by tilapia. Fish fed a glucose diet with 5000 mg chromic oxide/kg diet had higher weight gain than fish fed a glucose diet with 20,000 mg chromic oxide/kg diet. They speculated that the poorer growth performance of tilapia fed the higher level of chromic oxide may be due to a toxic effect of dietary chromium. Their most recent paper (Shiau and Shy 1997) reported that maximum growth and glucose utilization in hybrid tilapia were achieved with a glucose diet containing 204.4 mg chromic oxide/kg diet. The effects of dietary trivalent chromium have also been reported in rainbow trout (Tacon and Beveridge 1982) and common carp (Hertz et. al. 1989).

The improvement of glucose utilization by dietary chromic oxide has significant ramifications for digestibility studies in aquaculture nutrition research because chromic oxide is the external marker most commonly used in such studies (Austreng 1978). Among other criteria, an external marker has to be inert, with no toxic or physiologic effects. It should also not be absorbed from the gut. Not only has the level of dietary chromic oxide been reported to influence carbohydrate metabolism in tilapia, but it has also been reported to affect nutrient digestibility estimations (Shiau and Liang 1995). Whole-fish carcass and tissue chromium concentrations have also been reported to increase proportionately in fish fed varying levels of chromic oxide (Shiau and Liang 1995, Shiau and Shy 1997), which seems to indicate that it is being absorbed and retained in the fish body. Some researchers have therefore raised serious concerns about the use of chromic oxide for the determination of nutrient digestibility in fish (Riche et al. 1995, Ringo 1993).

This study was designed to determine if the level of chromic oxide inclusion in the diet would affect glucose utilization in channel catfish. Like most other fish, channel catfish are unable to utilize glucose efficiently as an energy source (Wilson and Poe 1987). We also wanted to determine whether chromic oxide is sufficiently inert to be used as an external marker in digestibility studies of channel catfish on the basis of fish growth performance, whole-body composition and fish carcass chromium concentrations.

MATERIALS AND METHODS

Table 1. Composition of the basal diet [View Table]

At the start of the experiment, 25 catfish fingerlings (mean weight of 4.9 ± 0.1 g) were stocked into each aquarium. The basal dextrin diet (D0) and the eight glucose diets of graded chromic oxide levels (G0, G50, G100, G200, G400, G1000, G5000 and G10,000) were then fed to randomly assigned (by drawing lots) triplicate groups of fish. All fish were fed at a rate of 4% of their body weight per day for 3 wk and then at 3% of their body weight per day for the duration of the experiment. This amount of diet was divided into two equal feedings per day. All diets were consumed by the experimental fish within 1 min. Fish were batch-weighed by aquarium once every week and the daily ration adjusted accordingly. A prophylactic treatment of acriflavin (~3 mg/L, Sigma Chemical) was administered after each weighing to reduce bacteriological infestation caused by handling (Garling and Wilson 1976). The experiment was conducted for 10 wk.

The remaining 18 fish, which were to be used for plasma glucose determinations, were placed back into their respective aquaria and fed their respective diets for another 2 d . After being deprived of feed for 24 h, the fish were fed 0.75% of their body weight on the day of plasma collection. Three fish were then removed at random from each aquarium immediately after all the feed offered was ingested (within 1 min). Blood was collected from each fish into tubes containing a sodium flouride/heparin mixture by severing the caudal peduncle after the fish were mildly anesthetized with MS222. Blood from three fish was pooled, centrifuged and the plasma collected into microtubes and stored frozen at 90°C until further analysis. This represented plasma glucose levels at the 0 h time interval. At 1, 2, 3, 4 and 6 h after feeding, three fish per aquarium were removed at random for blood collection. Plasma glucose was assayed by the glucose oxidase procedure (Sigma Chemical) the following week.

Crude protein, lipid, moisture and ash analyses of the whole-fish carcass were determined using standard procedures (AOAC 1990). Diet and tissue chromium was determined with the use of the dry ashing procedure of Miller-Ihli and Greene (1992). Homogenized diet or tissue samples were weighed into acid-washed glass ignition test tubes (Pyrex brand) and 100 µL of 0.36% magnesium nitrate was added as a matrix modifier and ashing aid. The samples were then dried at 130°C for 2 h and then ashed overnight in a muffle furnace with a dry ashing temperature of 480°C. When the sample ash was completely white, samples were diluted to 10 mL with 1 mol/L nitric acid (trace element quality) and analyzed with a Perkin-Elmer 2380 Atomic Absorption Spectrophotometer equipped with both flame and graphite (PE HGA-400) furnaces. Flame was used for chromium determinations in the diets, and the graphite furnace atomic absorption spectrometry procedure of Miller-Ihli and Greene (1992) was used for fish tissue samples.

RESULTS

Table 2. Weight gain, feed efficiency ratio (FER), protein efficiency ratio (PER) and hematocrits of channel catfish fingerlings fed purified diets supplemented with various chromic oxide (Cr2O3) levels for 10 wk1 [View Table]

At 1, 2 and 3 h postprandial, plasma glucose concentrations were significantly higher mainly in fish fed the glucose diets than in fish fed the diet containing dextrin at all levels of chromic oxide supplementation (Table 3). Plasma glucose concentrations were not significantly different among the glucose-fed and dextrin-fed fish at the 0, 4 and 6 h time interval irrespective of the level of chromic oxide in the glucose diets. The plasma glucose concentrations of fish fed the dextrin diet remained low and constant from 1 to 6 h after feeding. In fish fed the glucose diets, plasma glucose concentrations peaked at 1-2 h after feeding and remained elevated thereafter. No significant differences were observed in the postprandrial plasma glucose concentrations in fish fed glucose diets with various levels of chromic oxide supplementation. No significant interactions between diet versus time after feeding were detected.

Table 3. Postprandial changes in plasma glucose of channel catfish fingerlings fed purified diets supplemented with various chromic oxide (Cr2O3) levels1 [View Table]

Whole-body crude protein and ash concentrations in channel catfish fed the test diets were not significantly different irrespective of the carbohydrate source or level of chromic oxide supplementation (Table 4). Channel catfish fed the glucose diet without chromic oxide supplementation had significantly lower body lipid concentrations but higher body moisture compared with fish fed the dextrin diet. The body lipid and moisture content of fish fed glucose diets containing various chromic oxide levels were not significantly different from each other for the most part. No obvious trends were observed.

Table 4. Whole-body composition of channel catfish fingerlings fed purified diets supplemented with various chromic oxide (Cr2O3) levels for 10 wk1 [View Table]

Chromium concentrations in the whole-fish carcass were not significantly different among fish fed the dextrin or glucose diets irrespective of the level of dietary chromic oxide supplementation (Table 5). Fish fed glucose diets containing 5000 and 10,000 mg chromic oxide/kg diet had significantly higher chromium concentrations in the intestine compared with fish fed the dextrin diet or glucose diets at all other levels of chromic oxide supplementation except for fish fed the glucose diet containing 1000 mg chromic oxide/kg diet. A mean chromium recovery value of 99.3 ± 0.7% based on analyzed and added dietary chromium was obtained (Table 5) and validated the accuracy of the dry ashing procedure of Miller-Ihli and Greene (1992) as adapted to our laboratory conditions.

Table 5. Chromium concentrations of whole-fish carcass and intestine (wet weight basis) of channel catfish fingerlings fed purified diets supplemented with various chromic oxide (Cr2O3) levels for 10 wk1 [View Table]

DISCUSSION

The prolonged hyperglycemia observed in glucose-fed channel catfish is similar to that observed in other fishes (Furuichi and Yone 1981, Shiau and Chen 1993). In a review of the results from several investigators, Wilson (1994) pointed out that the persistent hyperglycemia observed in fish after glucose tolerance tests may be due to one or several factors including the following: 1) the lack of an inducible glucokinase enzyme in fish because glucose is less potent than certain amino acids as a stimulus for insulin release; 2) the possible inhibition of insulin release by somatostatins, which are released in response to high blood glucose levels; or 3) the relatively low number of insulin receptors in fish compared with land animals.

On the basis of growth performance, plasma glucose concentrations and fish carcass composition, we did not observe any significant effect of dietary chromic oxide on glucose utilization in channel catfish irrespective of the level of inclusion in the glucose diets (from 50 to 10,000 mg chromic oxide/kg diet). On the contrary, significant effects of trivalent chromium on glucose utilization and subsequent growth performance have been reported for the rainbow trout (Tacon and Beveridge 1982), common carp (Hertz et al. 1989) and hybrid tilapia (Shiau and Chen 1993, Shiau and Liang 1995, Shiau and Lin 1993, Shiau and Shy 1997). Even though these studies have provided data concerning the nutritional effects of dietary trivalent chromium in fish, the currrent information on this topic is not conclusive and is often conflicting.

Tacon and Beveridge (1982) reported no significant difference in the growth response of rainbow trout fed 0, 1 or 3 mg chromium/kg diet, but fish fed 6 mg chromium/kg diet displayed a much reduced growth response. On the contrary, chromium supplementation at 2 mg/kg glucose diet significantly increased weight gain and energy deposition in tilapia (Shiau and Lin 1993). In the same study, Shiau and Lin (1993) reported that the same chromium supplementation at 2 mg/kg diet did not improve the growth performance of tilapia when the carbohydrate source in the diet was cornstarch and not glucose. However, Hertz et al. (1989) reported improved glucose tolerance in common carp fed a low protein diet with high levels of wheat starch and supplemented with 2 mg chromium/kg diet. Shiau and Chen (1993) reported that there was no significant difference in weight gain between starch-fed and glucose-fed tilapia when the glucose diet was supplemented with 2 mg chromium (as chromic oxide)/kg diet, which implied optimal utilization of dietary glucose. However, in a later study, Shiau and Shy (1997) concluded that the supplemented chromic oxide level that would maximize glucose utilization in tilapia was 204.2 mg/kg glucose diet based on maximum growth.

It is noteworthy to point out that, apart from the rainbow trout study by Tacon and Beveridge (1982), which was conducted in a flow-through water system, the studies with common carp and hybrid tilapia were conducted either in static water systems in which the water was changed daily (Hertz et al. 1989) or weekly (Shiau and Shy 1997) or in a closed recirculated water system with a common water reservoir to all of the aquaria (Shiau and Chen 1993, Shiau and Liang 1995, Shiau and Lin 1993). The chromium content of the aquarium water throughout the experiment in these studies was not monitored. It is possible that chromium may leach from the feces and feed into the culture water, thus providing an additional source of chromium for the fish under study. Even though we do not know the extent to which gill or skin absorption of chromium may occur in fish, it should always be a concern in trace element studies. As a case in point, Shiau and Shy (1997) reported that no growth differences were observed in tilapia fed glucose diets supplemented with various chromic oxide levels when the experiment was conducted in their closed recirculating water system, but significant growth differences were observed when the same experiment was repeated in a static system. Even with the static system, in which only one third of the water in each aquarium was changed weekly (Shiau and Shy 1997), there is still the possibility of bioaccumulation of chromium by the fish from the water. In this study with channel catfish, we used a flow-through water system in which the aquarium water was completely exchanged every 2 h. The experimental fish were therefore in contact with any chromium leached into the water for only a very short period of time.

Even though there may be species-specific differences in the uptake, retention and excretion of chromium in fish, the high concentrations of chromium reported in the fish carcasses by Shiau and Shy (1997) may also be due in part to the static culture system they used as discussed earlier. For example, the chromium concentration of the fish whole body (without intestine) in their study increased from 56 to 830 nmol/g tissue in fish fed diets containing chromic oxide from 0 to 5000 mg/kg diet. In this study with channel catfish, the chromium concentrations of whole-fish carcasses (with intestine) did not increase with increasing dietary chromic oxide from 0 to 10,000 mg/kg diet, and a relatively low average concentration of ~6.39 nmol chromium/g wet tissue was retained.

Shiau and Shy (1997) also reported a linear increase in body ash content, which they attributed to the increase in fish tissue chromium content. We did not find any body ash increase in channel catfish fed increasing levels of dietary chromic oxide. We did find significantly higher chromium concentrations in the intestine of fish fed the two highest dietary chromic oxide levels, but we believe that this may be due to contamination from unvoided fecal matter in the intestines of a few fish. We observed high variability in the intestinal chromium values within fish from the same dietary treatment, which would support this explanation.

Further investigation is required to determine the reason for the relatively low levels of chromium found in the fish carcasses of channel catfish fed high levels of dietary chromic oxide compared with tilapia. Using radioactively labeled chromium, Doisy et al. (1976) demonstrated that the absorption of inorganic chromium was <1% in humans. Anderson and Kozlovsky (1985) reported that, at dietary intake levels >40 µg/d, chromium absorption appears to remain constant in humans at ~0.4%. Seaborn and Stoeker (1989) reported that glucose inhibited chromium absorption compared with starch in studies with mice. If one assumes that such low levels of chromium absorption also occur in channel catfish, this still does not explain the relatively low levels of chromium found in the fish carcass if the intake of chromium over the 10-wk feeding period is considered. On the basis of experiments with rats, Anderson and Polansky (1995) concluded that body chromium stores appear to be regulated at the excretion levels rather than at the level of absorption. They found that chromium retention was independent of dietary intake, and few if any dietary components or hormones increased chromium absorption above basal levels. In human beings, diets high in simple sugars have been shown to increase urinary chromium excretion (Kozlovsky et al. 1986). It is speculated that channel catfish may have a highly efficient excretion mechanism for the regulation of chromium homeostasis. Tacon and Beveridge (1982) also did not find any significant difference in the whole-carcass and tissue chromium concentrations of rainbow trout fed graded dietary chromium chloride levels from 0 to 31 mg/kg diet.

Despite the fact that chromic oxide may not be completely inert and is suspected of violating other assumptions required of an external indicator (Bowen 1978), the direct influence of such violations in digestibility estimations is still unclear. Shiau and Liang (1995) reported that tilapia fed both starch and glucose diets supplemented with 0.5% chromic oxide gave higher apparent nutrient digestibility estimations than those fed diets supplemented with 2% chromic oxide. On the contrary, Tacon and Rodrigues (1984) obtained nutrient digestibility estimations that were significantly higher in rainbow trout fed diets with 2% chromic oxide compared with values obtained from fish fed 1.0 or 0.5% dietary chromic oxide. Shiau and Shy (1997) reported that a dietary chromic oxide inclusion level of up to 0.5% had no significant effect on nutrient digestibility estimations in hybrid tilapia.

It should be pointed out that dietary chromium has been reported to have no significant effects on growth performance (Shiau and Chen 1993, Shiau and Liang 1995, Shiau and Lin 1993) or tissue chromium retention (Shiau and Liang 1995, Tacon and Beveridge 1982) when experimental fish were fed starch-based diets instead of glucose-based diets. Therefore, other than avoiding glucose as the carbohydrate source in fish digestibility study diets, it may also be advisable to add chromic oxide at the end of the feeding trial just before fecal collection to minimize any possible physiologic effects of dietary chromium. Shiau and Liang (1995) pointed out that chromic oxide supplemented at either 0 or 8 wk of the feeding trial did not affect nutrient digestibility by tilapia. A chromic oxide level no greater than 0.5% of the diet may also be advisable at this time.

In conclusion, on the basis of the results from this study, it would appear that chromic oxide is sufficiently inert to be used as an external marker for digestibility studies in channel catfish. Furthermore, unless more concrete and consistent proof is forthcoming that significant errors in digestibility estimations occur as a direct result of the suspected violations of chromic oxide as an inert marker, we believe that chromic oxide can be safely used when added in small quantities at the appropriate time to starch-based diets for other fish species. It would also appear that the use of chromic oxide should be limited to flow-through culture systems.

ACKNOWLEDGMENT