Total and relative deficiency of gut mucosal IgA in German shepherd dogs demonstrated by faecal analysis.

The concentration of immunoglobulins in faecal extracts was investigated as a method of assessing the production of immunoglobulins by the gut mucosa of 137 dogs. There were significant correlations between the concentrations in faecal extracts and the concentrations produced in duodenal organ cultures. Seventy-six German shepherd dogs had significantly lower median immunoglobulin A (IgA) concentrations in their faecal extracts than 63 controls of various breeds. Sixteen of the German shepherd dogs had IgA concentrations below the 95 per cent confidence limit of the control population and six had no demonstrable faecal IgA. The faecal concentrations of immunoglobulin G and albumin were significantly higher in the German shepherd dogs than in the controls, but their immunoglobulin M concentrations were similar.

Localizing the X-linked orange colour phenotype using feline resource families.

Many genes influencing mammalian coat colours are well conserved. While genes responsible for pelage phenotypes in one species provide strong evidence for a candidate gene in a different species, the X-linked orange phenotype of the domestic cat is unique within mammals. The orange locus (O) undergoes X-inactivation, producing females that express both wildtype black (wt) and orange (variant) phenotypes when heterozygous (tortoiseshell). The orange locus has not yet been localized on the X chromosome. Tortoiseshell male cats have been identified but have been shown to be sex chromosome trisomies (XXY). To localize the cat orange locus, 10 feline-derived X-linked microsatellites were analysed in two extended cat pedigrees consisting of 79 and 55 individuals, respectively, segregating for the orange phenotype. Linkage analyses excluded close association of orange in the vicinity of the nine informative X-linked microsatellites. One marker was not polymorphic within either family. Several markers suggested exclusion (Z < -2.0) at distances of 7.5-33 cM. Exclusion analyses suggested a possible location for orange a 14 cM region near Xcen. Recombination distances of markers in the segregating feline pedigrees were reduced as compared with the feline interspecies backcross family. Thus, the presented pedigrees may be useful as reference families for the domestic cat because more accurate recombination rates for domestic cats can be determined.

Prevalence of enteropathic Escherichia coli in dogs with acute and chronic diarrhoea.

Samples of faeces from 57 dogs with acute diarrhoea, 82 dogs with chronic diarrhoea, 34 clinically healthy household dogs and 88 kennelled control dogs were analysed by hybridisation, using DNA probes to detect enteropathogenic Escherichia coli (EPEC) and enterotoxigenic E coli (ETEC), verocytotoxin-producing E coli (VTEC), enterohaemorrhagic E coli (EHEC), enteroinvasive E coli (EIEC) and enteroaggregative E coli (EAggEC). Samples of duodenal juice from 60 of the 82 dogs with chronic diarrhoea were also examined. Significantly more of the dogs with diarrhoea were excreting EPEC (acute 35.1 per cent, chronic 31.7 per cent) and VTEC (acute 24.6 per cent, chronic 28 per cent) than the kennelled dogs (EPEC 17.1 per cent, VTEC 0 per cent) or the household control dogs (EPEC 6 per cent, VTEC 5.9 per cent). Enteropathic E coli was also detected in the duodenal juice of 23 of 60 (38.3 per cent) of the dogs with chronic diarrhoea. The EPEC attaching and effacing A (eaeA) gene and the verocytotoxin 1 (VR1) gene coding for VTEC were often found together. There was good agreement between in vitro studies and hybridisation for the detection of eaeA and VT1. Isolates from the dogs with diarrhoea adhered significantly more to Hep-2 cells, and VT1-positive strains from the dogs with diarrhoea consistently killed more than 50 per cent of Vero cells.

Relationship between faecal character and intestinal transit time in normal dogs and diet-sensitive dogs.

The relationship between stool character and whole gut transit time (WGTT), which is the average time for the passage of material through the lumen of the alimentary tract from ingestion to defecation, was studied in eight control dogs and 12 dogs with non-specific dietary sensitivity. Dogs were fed four diets in a cross-over design, and faecal quality was assessed daily and WGTT determined using plastic pellets. Faecal quality was unaffected by diet in the control dogs. Dogs with dietary sensitivity produced looser faeces compared with the control dogs, and this was significant for two of the diets. There was no significant effect of diet on mean WGTT within or between groups. Minimum WGTT, which was the interval to the first appearance of markers in faeces, was shorter in sensitive dogs compared with controls, and this was significant for two of the four diets. There were significant, inverse relationships between minimum WGTT and both mean faeces score and percentage unacceptable defecations. These data suggest that rapid transit of certain dietary components may impact negatively on stool quality and contribute to loose faeces in dogs with non-specific dietary sensitivity.

Development of a technique for the in vivo assessment of flatulence in dogs.

OBJECTIVE: To develop a noninvasive method for the in vivo assessment of flatulence in dogs. ANIMALS: 8 adult dogs. PROCEDURE: Rectal gases were collected via a perforated tube held close to each dog's anus and attached to a monitoring pump fitted with a sensor that recorded hydrogen sulfide concentrations every 20 seconds. Patterns of flatulence were monitored for 14 hours after feeding on 4 days, and within- and between-dog variation was assessed over 4 hours on 4 consecutive days. Rate of hydrogen sulfide production (flatulence index) and frequency and number of emissions were evaluated as potential indicators of flatus characteristics. An odor judge assigned an odor rating to each flatulence episode, and the relationship between that rating and hydrogen sulfide concentration was determined. RESULTS: Flatulence patterns varied within and between dogs. Variation was most pronounced for flatulence index; mean coefficients of variance within dogs over time and between dogs on each day were 75 and 103%, respectively. Flatus with hydrogen sulfide concentrations > 1 parts per million could be detected by the odor judge, and severity of malodor was highly correlated with hydrogen sulfide concentration. Odor ratings were accurately predicted by use of the equation 1.51 X hydrogen sulfide concentration(0.28). CONCLUSIONS AND CLINICAL RELEVANCE: The technique described in this report appears to provide sensitive, reliable, and relevant data and will enable further studies of the factors that influence flatulence in dogs. Use of this technique also has the potential to aid in investigations of colonic physiology and pathology.

Administration of charcoal, Yucca schidigera, and zinc acetate to reduce malodorous flatulence in dogs.

OBJECTIVE: To determine whether feeding activated charcoal, Yucca schidigera, and zinc acetate would ameliorate the frequency and odor characteristics of flatulence in dogs. DESIGN: In vitro screening of active agents followed by a randomized controlled trial. ANIMALS: 8 adult dogs. PROCEDURE: A fecal fermentation system was used to assess the effects of activated charcoal, Yucca schidigera, and zinc acetate alone and in combination on total gas production and production of hydrogen sulfide, the primary determinant of flatus malodor in dogs. All 3 agents were subsequently incorporated into edible treats that were fed 30 minutes after the dogs ate their daily rations, and the number, frequency, and odor characteristics of flatulence were measured for 5 hours, using a device that sampled rectal gases and monitored hydrogen sulfide concentrations. RESULT: Total gas production and number and frequency of flatulence episodes were unaffected by any of the agents. Production of hydrogen sulfide in vitro was significantly reduced by charcoal, Yucca schidigera, and zinc acetate by 71, 38, and 58%, respectively, and was reduced by 86% by the combination of the 3 agents. Consumption of the 3 agents was associated with a significant decrease (86%) in the percentage of flatulence episodes with bad or unbearable odor and a proportional increase in the percentage of episodes of no or only slightly noticeable odor. CONCLUSIONS AND CLINICAL RELEVANCE: Results suggest that activated charcoal, Yucca schidigera, and zinc acetate reduce malodor of flatus in dogs by altering the production or availability of hydrogen sulfide in the large intestine.

Use of an orally administered combined sugar solution to evaluate intestinal absorption and permeability in cats.

OBJECTIVE: To evaluate intestinal permeability and absorption in healthy cats in association with diet and normal intestinal microflora. ANIMALS: 6 healthy domestic shorthair cats. PROCEDURE: A sugar solution containing D-xylose, 30-methyl-D-glucose, L-rhamnose, lactulose, and 51Cr-EDTA was administered intragastrically to healthy cats, and urinary excretion of ingested sugars was determined 5 hours after administration. After the same cats had received metronidazole for 1 month, the study was repeated. A final study was performed while cats were maintained on a new diet differing in composition and processing. RESULTS: Lactulose-to-rhamnose ratios, reflecting intestinal permeability, were higher in cats, compared with values for humans or dogs, and values obtained before and after metronidazole administration (mean +/- SEM; before, 0.40 +/- 0.08; after, 0.45 +/- 0.09) were not significantly different. Intestinal absorption also was unaltered after antibiotic administration, and the xylose-to-glucose ratio was 0.70 +/- 0.03 before and 0.71 +/- 0.06 after metronidazole administration. Sugar recovery did not differ significantly while cats were maintained on canned or dry food. CONCLUSIONS AND CLINICAL RELEVANCE: Reference ranges were established for the percentage urinary recovery of orally administered D-xylose, 3-0-methyl-D-glucose, L-rhamnose, lactulose, and 51Cr-EDTA obtained after 5 hours in healthy cats. The intestines of cats appear to be more permeable than those of other species, although the normal bacterial microflora does not appear to influence the integrity or function of the feline intestine, because values obtained for the measured variables before or after antibiotic administration were not significantly different. In addition, differences were not detected when the diet was completely altered.

Comparison of the bacterial flora of the duodenum in healthy cats and cats with signs of gastrointestinal tract disease.

OBJECTIVE: To determine whether a colony environment predisposes healthy cats to high bacterial counts, including counts of obligate anaerobes, in the duodenum and whether increased numbers of bacteria could be found in the duodenum of cats with signs of chronic gastrointestinal tract disease. DESIGN: Prospective study. ANIMALS: 20 healthy control cats (10 from a colony environment and 10 pet cats) and 19 cats with a history of chronic gastrointestinal tract disease. PROCEDURE: Undiluted duodenal fluid was quantitatively and qualitatively assessed by bacteriologic culture under aerobic and anaerobic conditions. Serum concentrations of cobalamin and folate were also measured. RESULTS: Significant differences were not detected in the numbers of bacteria found in the duodenum of cats housed in a colony environment, compared with pet cats fed an identical diet prior to sampling. All healthy cats were, therefore, combined into 1 control group. Compared with healthy cats, cats with clinical signs of gastrointestinal tract disease had significantly lower counts of microaerophilic bacteria, whereas total, anaerobic, and aerobic bacterial counts were not significantly different. None of the cats with disease had total bacterial counts higher than expected from the range established in the control cats. Differences were not detected in regard to serum folate or cobalamin concentrations between diseased and healthy cats. CONCLUSIONS AND CLINICAL RELEVANCE: These findings indicated that healthy colony cats and pet cats have high numbers of bacteria in the duodenum, including high numbers of obligate anaerobes. Our findings also suggest that bacterial overgrowth in the small intestine is not a common clinical syndrome in cats with chronic nonobstructive gastrointestinal tract disease.

Effects of oral administration of metronidazole on small intestinal bacteria and nutrients of cats.

OBJECTIVE: To determine effects of oral administration of metronidazole on the number and species of duodenal bacteria and selective nutrients of cats. ANIMALS: 6 healthy domestic shorthair cats. PROCEDURE: Undiluted duodenal fluid was obtained for quantitative and qualitative bacterial culture to determine species and number of bacteria in healthy cats. Blood samples were assayed for taurine, total protein, albumin, cobalamin, and folate concentrations. Cats then were given metronidazole (20 mg/kg of body weight, PO, q 12 h) for 1 month, after which bacterial cultures and serum assays of nutrients were repeated. Nine months after cessation of antibiotic treatment, duodenal bacteria were re-evaluated and serum was assayed for total protein, albumin, cobalamin, and folate concentrations. RESULTS: Oral administration of metronidazole caused a significant decrease in aerobic and anaerobic bacterial counts in the duodenum of healthy cats, accompanied by emergence of Streptococcus spp and Corynebacterium spp. Serum concentrations of cobalamin and albumin increased when duodenal bacterial counts were decreased, although changes in folate or taurine concentrations were not detected. Measured variables did not differ, when comparing results obtained before and 9 months after cessation of metronidazole. CONCLUSIONS AND CLINICAL RELEVANCE: Oral administration of metronidazole decreased the number of aerobic bacteria and altered indigenous flora in the small bowel of cats. Normal duodenal flora appeared to be stable, because species of bacteria were re-established by 9 months after cessation of metronidazole. Bacterial flora appeared to have an impact on nutrients, because albumin and cobalamin increased during antibiotic administration and returned to preadministration concentrations after cessation of the antimicrobial.

Multiplex PCRs for identification of Escherichia coli virulence genes.

PCRs were developed to detect 11 Escherichia coli virulence genes. Primers amplified the respective genes without cross-reaction with other genes. Specificity was maintained in multiplex reactions; excellent amplification of target genes was possible with a minimum of four multiplex reactions. These reactions successfully identified genes in E. coli from the feces of four dogs.

Inheritance of gluten-sensitive enteropathy in Irish Setters.

OBJECTIVE: To establish a model for inheritance of gluten-sensitive enteropathy (GSE) in Irish Setters. ANIMALS: 44 dogs of a 6-generation family of Irish Setters with GSE and 7 healthy Irish Setters. PROCEDURE: Phenotype of each dog was determined after oral administration of gluten in the weaning diet, using morphometric evaluation of jejunal biopsies (all generations) and measurement of small intestinal permeability by use of a lactulose-rhamnose permeation test (generations 1, 2, and 3). Overall probability for each of 4 genetic models of inheritance (autosomal recessive, autosomal dominant, sex-linked recessive, and sex-linked dominant) accounting for segregation of partial villus atrophy within the entire family was calculated. RESULTS: The autosomal recessive model was most tenable and was 56,250 times more likely to account for segregation of partial villus atrophy than the autosomal dominant model, assuming disease prevalence of 0.8%. Both sex-linked models were untenable. These conclusions were robust to the error attached to estimation of disease prevalence. High intestinal permeability without morphometric jejunal abnormalities in 4 of 20 dogs in the 3 youngest generations suggested heterogeneity of lesions associated with GSE. CONCLUSIONS: Genetic transmission of GSE is under the control of a single major autosomal recessive locus.

In situ hybridization as a technique for the immunological investigation of canine intestine: jejunal expression of IFNgamma and IL10 in Irish setters and beagles.

In situ hybridization (ISH) has found numerous applications in biology and medicine. However, its use to demonstrate expression of cytokines within the canine small intestine has not been described. Digoxigenin-labelled riboprobes complementary to mRNA encoding canine IFNgamma and IL10 were used to demonstrate expression of these cytokines within formalin-fixed, paraffin-embedded sections of jejunum obtained from healthy control Irish setter (IS) dogs (n = 4), gluten-sensitive IS in remission (n = 7), and beagles with high enteric bacterial populations (n = 5). Proportional areas of cells within the lamina propria showing one of three mutually exclusive staining intensities were measured, as well as the total stained area. Intensity categories were chosen arbitrarily to represent cells showing weak, moderate or dense staining (grades 1-3 respectively), reflecting increasing expression of mRNA. Control and gluten-sensitive IS showed similar total and grade-by-grade areas of expression of IFNgamma and IL10 in the lamina propria (p>0.05), in contrast to beagles, which showed greater total and grade 1 areas of expression of IFNgamma, and greater total, grade 1 and grade 2 areas of expression of IL10, than both groups of IS (p<0.05). Epithelial expression of both cytokines was demonstrated in beagles and IS, but differences between groups for each cytokine were not apparent (p>0.05). This study has validated the use of in situ hybridization for the detection of IFNgamma and IL10 mRNA within canine intestinal biopsies, andhas shown heightened jejunal expression of both cytokines in beagles with high enteric bacterial populations.

Serum insulin-like growth factor (IGF)-I concentrations are reduced by short-term dietary restriction and restored by refeeding in domestic cats (Felis catus).

Nutritional modulation of insulin-like growth factors (IGF) and their binding proteins (IGFBP) is well established. The effect of nutritional restriction on the serum IGF/IGFBP system of adult cats was investigated to evaluate serum IGF-I as a biochemical marker of nutritional status. Assays for measuring feline serum IGF and IGFBP were validated and normal ranges established in a study population of 46 healthy nonobese adult cats. Serum concentrations of IGF-I and IGF-II correlated significantly with body weight (r = 0.75, P < 0. 0001 and r = 0.34, P < 0.03, respectively). Serum IGFBP profiles were similar to other species, including humans, dogs and guinea pigs. IGFBP-3 was the predominant binding protein reflecting IGF-I concentrations and body size. Serum IGFBP-2 concentrations were high relative to the normal human serum pool (NHS) control. Food withdrawal for 18 h followed by refeeding did not alter circulating IGF or IGFBP concentrations, including IGFBP-1, in nine cats. Short-term dietary restriction of nine adult cats to supply initially 56% (56%M) and then 42.5% (42.5%M) of calculated maintenance energy requirements for 14 d resulted in a significant weight loss (P < 0.01). However, serum IGF-I concentrations fell significantly (-51%, P < 0.01) only with 42.5%M restriction. Serum IGF-II, IGFBP, insulin and albumin concentrations were not altered during the study. We conclude that nutrition does modulate the adult feline IGF/IGFBP system, but to a lesser extent than in other species. Further evaluation is required before serum IGF-I can be used for the assessment of nutritional status in adult cats.

A comparison of endoscopic and surgical collection procedures for the analysis of the bacterial flora in duodenal fluid from cats.

In order to assess an endoscopic collection procedure, populations of bacteria in duodenal fluid from seven adult cats were compared in paired samples obtained by endoscopy and direct needle aspiration during laparotomy. Each sample of duodenal juice was subjected to quantitative and qualitative culture of bacteria under both aerobic and anaerobic conditions. There were no significant differences in total numbers or individual species of bacteria comparing the two collection procedures. These findings indicate that collection of duodenal juice by endoscopy using the procedure described provides a representative sample of small bowel fluid for the assessment of the bacterial flora. Therefore, there appears to be no need for more invasive or complicated sampling techniques when quantitative and qualitative culture of duodenal juice is indicated as part of an investigation of small bowel disease in cats.

Evidence for a compensatory increase in synthesis of brush border enzymes in the jejunum of Irish Setters with gluten-sensitive enteropathy.

OBJECTIVE: To determine whether biosynthesis of aminopeptidase N (ApN), alkaline phosphatase (AP), and total microvillus membrane protein is altered in Irish Setters with gluten-sensitive enteropathy (GSE). ANIMALS: A litter of 6 Irish Setters with GSE and 3 healthy Greyhounds. PROCEDURES: Explants obtained from affected dogs at 4 and 12 months were maintained in vitro and were compared with material from healthy control dogs. Biosynthesis of ApN and AP was monitored by incorporation of [35S]methionine and immunoprecipitation of these enzymes. RESULTS: Jejunal explants from affected Irish Setters had significantly higher rates of biosynthesis of total protein, microvillus membrane protein, AP, and ApN, compared with control tissue. Two forms of ApN with apparent molecular mass of 155 and 135 kd and 4 forms of AP with apparent molecular mass of 210 to 260, 150, 130, and 105 kd were identified in total membrane fractions from control and affected dogs. CONCLUSIONS: Reduced activities of ApN and AP in dogs with GSE are not attributable to decreased synthesis of these proteins and document enhanced synthesis of microvillar membrane proteins, which may be a compensatory response to enterocyte damage. The 150-kd form of AP was most prominent in tissue from the most affected dogs, probably representing an early form of this enzyme. In contrast, the 105-kd form was most intense in tissue from controls and less intense in tissue of affected dogs.

Characterization of intestinal morphologic, biochemical, and ultrastructural features in gluten-sensitive Irish Setters during controlled oral gluten challenge exposure after weaning.

OBJECTIVE: To characterize histologic, biochemical, and ultrastructural changes in the intestine of Irish Setters susceptible to gluten-sensitive enteropathy (GSE) during controlled oral challenge exposure with gluten after weaning. ANIMALS: Six gluten-sensitive and 12 healthy Irish Setters and 3 healthy Greyhounds. PROCEDURE: Jejunal biopsy specimens were taken at 4 and 12 months of age from the 6 gluten-sensitive Irish Setters, which had been reared on a gluten-free diet to which a controlled dose of gluten (0.5 g/kg of body weight/d) was added. Control specimens were obtained at 4 (n = 5) and 12 (7) months of age from the healthy Irish Setters, which had been fed a conventional gluten-containing diet, and at 4 months of age from the healthy Greyhounds fed the controlled dose of gluten. The specimens were subjected to histologic and ultrastructural examinations and assay of brush border enzymes. RESULTS: Gluten-sensitive Irish Setters developed abnormalities characteristic of GSE at 4 months. Abnormalities were comparable to changes not seen previously until 12 months in dogs with GSE fed a conventional gluten-containing diet. In addition, microvilli were stunted and irregular, and a few were vesiculated and reduced in number; the glycocalyx was reduced or absent. By 12 months of age, there was improvement in morphologic and biochemical parameters, indicating partial recovery despite continued exposure to gluten. CONCLUSIONS: Relative early onset of intestinal damage, compared with that previously reported, and subsequent partial recovery suggestive of oral tolerance to gluten may be attributable to oral administration of gluten as a purified extract rather than in dietary cereal, but alternative explanations include differences in environment or genetic susceptibility to gluten.

Intestinal permeability of Irish setter puppies challenged with a controlled oral dose of gluten.

A combined test of intestinal permeability using lactulose (L) and rhamnose (R), and absorptive function using xylose (X) and 3-O-methylglucose (G), was carried out at four, six, eight and 16 weeks of age in 22 healthy control and six gluten-sensitive Irish setter (IS) dogs fed a diet containing a controlled dose of gluten from weaning. Comparisons were made with two groups of 12 healthy control dogs of breeds other than IS, one fed the same diet as the setters and the other fed a gluten-free diet. Gluten-sensitive IS showed a rise in permeability (mean [SEM] urinary L/R) from 0.23 (0.07) at four weeks to 0.39 (0.05) at eight weeks, remaining at 0.36 (0.04) at 16 weeks. These results were significantly higher in gluten-sensitive than control IS at six, eight and 16 weeks, compatible with jejunal biopsy lesions characteristic of gluten-sensitive enteropathy demonstrated in affected dogs at 16 weeks. Urinary L/R ratios of control dogs of breeds other than IS peaked at six weeks 0.27 (0.02), and were significantly higher than those of control IS at six and eight weeks, demonstrating differences in permeability between Irish setter dogs and other breeds at this age. There were no significant differences in urinary X/G ratios at six, eight and 16 weeks of age between any of the groups of dogs challenged with gluten. Urinary L/R and X/G ratios were similar in the control dogs of breeds other than IS fed gluten-containing and gluten-free diets. These findings indicate that intestinal permeability testing of puppies during controlled oral gluten challenge provides a practical screening test for gluten sensitivity in Irish setter dogs at an early age.

Insulin-like growth factor axis of gluten-sensitive dogs during a gluten challenge.

Eight gluten-sensitive Irish setters underwent a gluten challenge to investigate changes in the insulin-like growth factor (IGF) axis. In the first study, they were challenged with an acute intraduodenal administration of tryptic-peptic gluten digest and then maintained on dietary gluten for three months. In the second study, the challenge came solely from dietary gluten fed for three months. After the acute intraduodenal administration of gluten, serum IGF-I levels decreased significantly by 21.8 per cent, (P = 0.01) on day 3 after challenge and then returned to normal. There was also a decrease (52.5 per cent, P < 0.03) in the levels of serum IGF-binding protein-3 (IGFBP-3) until day 14 after challenge but they had returned to normal by day 28. In two dogs IGFBP-3 levels decreased through specific serum protease activity. There were no changes in serum IGF-I or IGFBP levels during the second study after the dietary gluten challenge alone, or in four non-gluten-sensitive beagles studied as controls during the acute intraduodenal/dietary gluten challenge.

Nutritional modulation of canine insulin-like growth factors and their binding proteins.

The response of canine insulin-like growth factors (IGFs) and IGF-binding proteins (IGFBPs) to moderate nutritional restriction followed by refeeding has not previously been studied in detail. The purpose of these studies was to examine the effects of nutritional restriction on the IGF system of adult dogs. Normal serum IGF values were established after validation of heterologous RIAs for measuring canine IGFs-I and -II. Canine serum IGFBP profiles were examined by Western ligand blotting (WLB), using radiolabelled recombinant human (rh) IGF-I as the ligand, and were found to be similar to those of other species. IGF-I and IGFBP-3 concentrations correlated with body weight, thus reflecting breed size as previously shown, whereas IGF-II concentrations did not. IGFBP-2 serum concentrations and band intensity on WLB were increased compared with normal human serum IGFBP-2. Overnight fasting had no effect on IGF or IGFBP concentrations, including IGFBP-1, nor did refeeding. Prolonged restriction to 56% and then 42.5% of maintenance energy requirements for 2 weeks decreased IGF-I concentrations by 20.4% and 32.7% respectively. Feeding of the same diet ad libitum for 2 weeks normalised IGF-I concentrations. There were no changes in IGF-II or insulin levels. Serum IGFBP-2 concentrations increased with 56% restriction of maintenance energy (P = 0.03). We conclude that serum IGF-I is potentially a useful marker of short-term change in nutritional status in the adult dog.