Abstract
The postprandial increase in glucose concentration is typically not considered in selecting diets to manage diabetic and pre-diabetic cats. This study describes increases in glucose and insulin concentrations in 24 clinically healthy, neutered adult cats following one meal (59 kcal/kg) of a moderate carbohydrate diet (25% of energy). Median time to return to baseline after feeding for glucose was 12.2 h (1.8−≥24 h) and for insulin was 12.3 h (1.5−≥24 h). Time to return to baseline for glucose was not different between male (10.2 h) and female (17.2 h) cats. There was evidence female cats had a longer return to baseline for insulin (18.9 h versus 9.8 h) and females had higher (0.9 mmol/l difference) peak glucose than males. This demonstrates that the duration of postprandial glycaemia in cats is markedly longer than in dogs and humans, and should be considered when managing diabetic and pre-diabetic cats.
Introduction
Minimising the increase in glucose concentration following a meal and the subsequent demand on beta-cells to secrete insulin is a primary goal for the management of diabetic and pre-diabetic human patients. 1 In human studies, it has been shown to be more important (but also more difficult) to normalise postprandial glycaemia compared with fasting glucose concentrations. 1 In healthy humans and dogs, postprandial glycaemia normally persists for 2–5 h.2–4
In contrast to humans and dogs, cats are strict carnivores and their natural diet typically contains 1–2% of energy from carbohydrates. 5 They have evolved a number of unique physiological mechanisms associated with glucose metabolism which likely affect their postprandial glucose concentration. First, the gluconeogenic pathway provides an almost continuous source of carbon skeletons for glucose or energy production. 6 Secondly, glucokinase concentrations are markedly reduced or absent, and hexokinase activity increased. In contrast, the liver of omnivores contains both glycolytic enzymes, which catalyse the phosphorylation of glucose during glycolysis.7,8 For hepatic clearance of glucose, cats rely almost completely on hexokinase activity. 9 Thirdly, insulin response to increased plasma glucose concentration is lower and peaks later than in dogs.10,11
As a result of these differences, plasma glucose clearance rates are longer in cats compared to dogs after an intravenous or oral glucose challenge, or after a moderate carbohydrate meal.10,11 Although a number of studies have shown the postprandial increase in glucose concentration after eating is prolonged in cats,12,13 this has not been the focus of these studies and it has received almost no attention in the feline literature. Consequently, the implications of a long postprandial increase in glucose concentration is typically not considered in selecting diets to manage diabetic and pre-diabetic cats, ie, cats with impaired glucose tolerance or with mild persistent hyperglycaemia (impaired fasting glucose). In addition, studies reporting the impact of diet on glucose and insulin concentrations in cats are still being performed with inadequate fasting times and postprandial sampling,12,14–16 resulting in limited power to demonstrate differences and, in some cases, inaccurate conclusions. We found no report in the peer-reviewed literature that focussed on the duration or magnitude of postprandial glycaemia in cats following meal ingestion.
Minimising hyperglycaemia is important for reducing the demand on beta-cells for insulin secretion in both diabetic and pre-diabetic cats, and because of the glucose toxic effects on beta-cells of hyperglycaemia. 17 The International Diabetes Federation defines postprandial hyperglycaemia in humans as a plasma glucose concentration of >7.8 mmol/l. 18 Therefore, an understanding of the duration and magnitude of postprandial glucose response in cats is vital for clinicians to effectively evaluate and manage postprandial hyperglycaemia in pre-diabetic and diabetic cats. Recognition of the duration of postprandial glucose response in cats is also important to ensure appropriate study design — especially for those investigating the impact of diet on glucose and insulin secretion — or where fasting samples for glucose and insulin analysis are required. There are currently no reports comparing postprandial glucose and insulin responses following a meal in male and female cats, although male cats are at increased risk of diabetes. 19
Consequently, the objective of this report is to describe the duration and magnitude of increases in plasma glucose and insulin concentrations in 24 clinically healthy cats following consumption of a moderate carbohydrate [25% metabolisable energy (ME) 20 ] meal, and to compare postprandial responses between male and female cats. This study is the first to focus on the duration and magnitude of postprandial glycaemia in the cat.
Materials and methods
Experimental design
In all cats, a glucose tolerance test was performed, followed 48 h later by a meal feeding test during which plasma glucose and insulin concentrations were measured.
Animals
Twenty-four, neutered adult cats (12 male, 12 female) were used in this study (cats used in this study were sourced from municipal council shelters located in Queensland, Australia and were rehomed after the study). All cats were healthy based on physical examination and routine haematological and serum biochemical analyses. Mean bodyweight was 4.9 kg (range 3.6–6.0 kg). All cats had a body condition score of 3 (considered ideal) on a scale of 1 (underweight) − 5. 21 Based on visual assessment, all cats were estimated to be between 2 and 6 years old, although accurate ages were unknown. For 2 weeks prior to sample collection, cats were familiarised with their environment, carers and study procedures to minimise any effects of stress on the study results. The Animal Experimentation Ethics Committee of the University of Queensland, Australia approved the protocol for this study.
Cats were housed individually in 1.0 × 0.6 × 0.6 m holding pens with a litter pan, food and water dishes, bedding and a toy. All cats were fed once daily in their individual holding pens. During the day, compatible groups of 6–8 cats were exercised in a group cat facility for approximately 8 h, which included environmental enrichment aids such as tree bases, boxes, shelving and toys.
Glucose tolerance test
Blood samples (4 ml) were collected prior to (-30 and 0 min) and at 2, 5, 10, 15, 30, 45, 60, 90, 120 and 180 min following glucose administration. Glucose (Astra BP, 1.0 g/kg body weight ) was administered via a jugular catheter as a bolus dose over 60 s. Immediately after glucose injection, 3 ml of saline solution (0.9% NaCl solution) was infused into the catheter.
Meal feeding test
At least 24 h before the test, a jugular catheter [18 gauge × 8 cm polyurethane jugular catheter (Cook Veterinary Products)] was placed into one jugular vein under general anaesthesia for the collection of blood samples. Cats were anaesthetised with propofol (Diprivan 10 mg/ml; Zeneca), given as an initial bolus dose of 6–7 mg/kg followed by additional doses of 5–10 mg as required. Catheters were flushed twice daily with heparinised saline (20 IU of heparin/ml in 0.9% saline solution) to maintain patency until their removal at the end of the meal feeding test.
Four weeks prior to, and during, the meal feeding test cats were fed a commercial, dry extruded feline maintenance diet (Iams Adult Cat Chicken Dry Food: dry matter 36% protein, 23% fat and 30% carbohydrate; approximate energy distribution: protein 29% of dietary ME, fat 46% of dietary ME and carbohydrate 25% of dietary ME calculated using the modified Atwater factors 22 ) (Table 1). Food was withheld for 23.5 h before the meal feeding test.
Ingredient list for the commercial maintenance diet fed in this study
Three-and-a-half days before the meal feeding test, cats were allowed to eat ad libitum for 48 h. Initially, food intake increased substantially then stabilised, and the amount consumed by each cat in the last 12 h was used to determine the amount fed during the meal feeding test. As this study was designed to represent as close to ‘real life’ as possible, the test was designed to allow the variability in food intake between cats to be reflected as between-cat variability. Cats were required to eat ≥90% of their 12-h ad libitum consumption (mean 59 kcal/kg bodyweight, range 27–100 kcal/kg) within 0.5 h of being fed. All cats ingested the required test amount. However, one cat was re-tested 1 month later as a result of catheter malfunction.
Sample collection
Blood samples (either 4 ml or 13 ml) were collected 30 min prior to and immediately (approximately 5 min) before food was offered (time 0), and at 1, 2, 3, 4, 6, 8, 10, 12, 15, 18 and 24 h thereafter.
Blood samples were placed into sterile ethylenediaminetetraacetic acid (EDTA) vacuettes containing the proteinase inhibitor, aprotinin (Trasylol, Kallikrein Inactivator, 10 000 U/ml; Bayer Australia), added to the vacuettes at 0.05 ml per ml of whole blood. After collection, samples were kept on ice for 10–15 min until centrifugation at 1500 g for 8 min. After separation, plasma samples were split and stored in 500 μl vials at −70°C until assayed for glucose and insulin concentration.
Glucose was measured in plasma using an automated glucose analyser (Olympus 400, Olympus; mean intra-assay and interassay variability of 0.6% and 1%, respectively). Insulin was measured using a commercially-available kit (Phadeseph Insulin RIA; Pharmacia and Upjohn Diagnostics AB), validated for the detection of feline insulin with a mean intra-assay and interassay variability of 8.8% and 3.8%, respectively. 23
Statistical analyses
We used descriptive statistics to describe the increases in plasma glucose and insulin concentrations after feeding. Glucose and insulin variables analysed were baseline, peak and 24-h mean concentrations, times to peak, times to return to baseline, and absolute and percentage increase in glucose and insulin concentration.
Baseline concentrations were calculated as the averages of -30 min and -5 min values for each cat. Peak concentration was the highest measured value after time 0 and time to peak was considered to be the time from feeding until that sampling. Mean 24-h glucose and insulin concentrations for each cat over the 24-h period from feeding were calculated by dividing the respective area under the curve (AUC) by 24. AUCs above zero were calculated for glucose and insulin concentrations for each cat using the trapezoidal method.24,25
Each cat was considered to have exceeded its baseline glucose concentration if blood glucose concentration at one or more timepoints was greater than the sum of the cat’s individual baseline concentration and the 90% range of differences. Each cat was considered to have returned to baseline at the estimated timepoint after both the cat’s peak and after the time that it first exceeded baseline at which blood glucose concentration had first declined to equal the sum of the cat’s individual baseline concentration and the 90% range of differences. This timepoint was estimated by linear interpolation between the first timepoint at which blood glucose concentration had declined below this sum and the immediately preceding timepoint. Time to return to baseline was calculated as the interval from time 0 to this estimated timepoint. The 90% range of differences was calculated using previously reported methodology based on the variance of the two baseline samples within cat but with the variance calculated as the residual mean square from analysis of variance after accounting for between-cat variability by fitting cat identity as a categorical variable. 26 The 90% range of differences allows statistical comparison of single values at various timepoints with the baseline concentrations, using the variance of the baseline samples for the cats collectively as an estimate of the unexplained variation at each time-point in each individual cat. Values at timepoints that differ from baseline concentrations by more than the 90% range of differences can be considered to differ significantly (P <0.05) from the cat’s baseline. Because the residuals of the regression of insulin on cat were heteroskedastic, concentrations at -30 min and -5 min were log transformed before estimation of the 90% range of differences. The resulting estimate was then back-transformed. Because the difference between two log-transformed values is equivalent to the ratio of the back-transformed values, the back-transformed 90% range of differences was interpreted on a multiplicative rather than additive scale. Thus, time to return to baseline for insulin was then estimated for each cat as described for blood glucose but using baseline multiplied by the back-transformed 90% range of differences.
Data distributions are summarised as the mean value of each variable ± standard deviation unless otherwise indicated.
In addition to calculating these descriptive statistics for all cats, we also compared the glucose and insulin concentrations at baseline and peak, mean concentrations for 24 h, and times to peak and times to return to baseline between male and female cats. For the glucose tolerance test, we compared glucose and insulin concentrations at baseline, 2 and 3 h, and times for glucose and insulin to return to baseline between male and female cats. Concentrations were compared using linear regression; peak and 24-h mean concentrations, and concentrations at 120 min and 180 min were compared with the baseline concentration for the same compound (ie, for glucose when analysing glucose concentrations, for insulin when analysing insulin concentrations) fitted as a covariate. Because residuals from the model of mean glucose were heteroskedastic, the distribution of the ß-coefficient was estimated using the non-parametric bootstrap with 1000 replications. Normal-based confidence intervals were used, ie, we assumed the underlying sampling distribution of ß-coefficients was normally distributed. Times to peak and times to return to baseline were compared between male and female cats using Kaplan-Meier analyses with log-rank tests. For cats not returning to baseline by 24 h, times to return to baseline were right-censored at that time. These comparisons were performed using Stata (version 11, StataCorp).
Results
Mean glucose concentrations for all cats at each timepoint in the meal feeding test are shown in Figure 1. Mean baseline glucose concentration was 5.0 ± 0.6 mmol/l. Following feeding, mean peak glucose concentration was 7.2 ± 1.2 mmol/l; the mean increase was 2.1 ± 1.0 mmol/l and the mean of each cat’s increase expressed as percentages of the cat’s baseline concentration was 43%. Mean postprandial glucose concentration for 24 h from feeding was 5.8 ± 0.8 mmol/l (Table 2). In 15/24 cats, blood glucose concentrations did not increase to significantly above baseline until more than 2 h after feeding. Median time to peak glucose concentration was 6 h (range 1–12 h) after feeding. The 90% range of differences was 0.63 mmol/l. Median time for glucose concentration to return to baseline was 12.2 h (range 1.8−≥24 h) after feeding. Blood glucose concentrations in 12/24 cats had not returned to baseline by 12 h and for four cats (cats 14, 20, 23 and 25), glucose concentration had not returned to baseline by 24 h.
Mean ± SEM plasma glucose concentrations (mmol/l) measured over 24 h in 24 healthy cats fed a single standard feline maintenance diet (protein 29%, fat 46% and carbohydrate 25%) consumed over 30 min. Means and SEMs are for all cats at each timepoint. Note: The highest mean for any timepoint (6.7 mmol/l at 6 h) is less than the mean of each cat’s peak (7.2 mmol/l; Table 2) as individual cats peaked at various times from 1–12 h
Means ± SDs for plasma glucose (mmol/l) and plasma insulin (µU/ml) concentrations measured over 24 h in 24 healthy cats fed a single meal of 50% of their daily energy consumed as a standard maintenance diet (protein 29%, fat 46% and carbohydrate 25%) consumed over 30 min. The range is shown in parentheses
Mean insulin concentrations for all cats at each timepoint in the meal feeding test are shown in Figure 2. Insulin concentration increases, expressed as percentages of the cat’s baseline concentration, averaged 152% or 13.2 ± 6.6 µU/ml from a mean baseline of 10.1± 3.6 µU/ml to a mean peak of 23.3 ± 6.8 μU/ml at a median of 4 h (range 1–12 h) after feeding (Table 2). Mean for 24 h from feeding was 15.0 ± 3.8 µU/ml. The 90% range of differences was 1.45. Median time for insulin concentration to return to baseline was 12.3 h (range 1.5−≥24 h) after food was offered. Insulin concentrations in 13/24 cats had not returned to baseline by 12 h and for four cats (cats 9, 12, 14 and 21), insulin had not returned to baseline by 24 h. Of the four cats with insulin concentrations at 24 h that were significantly above their respective baseline concentrations, only one also had glucose concentration at 24 h that was significantly above its baseline.
Mean ± SEM plasma insulin concentrations (µU/ml) measured over 24 h in 24 healthy cats fed a single meal of a standard feline maintenance diet (protein 29%, fat 46% and carbohydrate 25%) consumed over 30 min. Means and SEMs are after eating for all cats at each timepoint. Note: The highest mean for any timepoint (21.1 µU/ml at 6 h) is less than the mean of each cat’s peak (23.32 µU/ml; Table 2) as individual cats peaked at various times from 1–12 h
Baseline glucose concentrations did not differ significantly between male and female cats [estimated difference (male relative to female) -0.2 mmol/l; 95% confidence interval (CI) -0.7–0.3; P = 0.418]. However, males had significantly lower peak concentrations (estimated difference adjusted for baseline concentration –0.9 mmol/l; 95% CI -1.7 to -0.1; P = 0.034). Neither times to peak nor times to return to baseline differed significantly between genders (P = 0.718, median time to peak for both genders 6 h; P = 0.330, median return to baseline 10.2 h for males and 17.2 h for females). One of the 12 males and 3/12 females had not returned to baseline by 24 h. Food intake was not different between male (13.5 gm/kg) and female cats (13.0 gm/kg).
Neither baseline nor peak insulin concentrations differed significantly between male and female cats [baseline: estimated difference (male relative to female) 1.2 µU/ml; 95% CI -1.9–4.3; P = 0.444; peak: estimated difference adjusted for baseline concentration -3.9 µU/ml; 95% CI -9.4–1.6; P = 0.154]. Times to peak did not differ significantly between genders (P = 0.724; medians for females and males: 5 and 3 h, respectively) but there was some evidence that times to return to baseline were shorter in males (P = 0.106; medians for males and females: 9.8 and 18.9 h, respectively). One of the 12 males and 3/12 females had not returned to baseline by 24 h.
From the glucose tolerance test, baseline glucose concentrations did not differ significantly between male and female cats [estimated difference (male relative to female) -0.2 mmol/l; 95% CI -1.0–0.5; P = 0.524]. However, there was some evidence that glucose concentrations were lower in male cats at 2 h [estimated difference (male relative to female adjusted for baseline concentration) -2.3 mmol/l; 95% CI -4.7–0.1; P = 0.064] and 3 h (estimated difference -0.9 mmol/l; 95% CI -1.7–0.0; P = 0.062). The 90% range of differences was 0.43 mmol/l. Times for glucose to return to baseline did not differ significantly between males and females (P = 0.330; medians for males and females: 154 and 155 min, respectively). One of the 12 males and 3/12 females had not returned to baseline by 180 min.
From the glucose tolerance test, baseline insulin concentrations did not differ significantly between male and female cats at baseline [estimated difference (male relative to female) 0.4 µU/ml; 95% CI -2.3–3.1; P = 0.767], 2 h [estimated difference (male relative to female adjusted for baseline concentration) -3.1 µU/ml; 95% CI -11.1–4.9; P = 0.452] or 3 h [estimated difference -2.0 µU/ml; 95% CI -5.8–1.8; P = 0.303]. The 90% range of differences was 1.39. Times for insulin to return to baseline did not differ significantly between males and females (P = 0.25; medians for males and females: 160 and 161 min, respectively). Two of the 12 males and 5/12 females had not returned to baseline by 180 min.
Discussion
The purpose of this study was to document the postprandial period in cats and to compare postprandial glucose and insulin responses between male and female cats. The long postprandial period we have demonstrated should be considered by veterinarians assessing the impact of diet on postprandial glucose concentrations in diabetic or pre-diabetic cats, and in the design of nutrition studies in cats. These are often reported with inadequate fasting and postprandial sampling times, which can limit the conclusions made or, at worst, lead to invalid conclusions.12,14–16
In our study, glucose and insulin concentrations in healthy cats remained significantly increased above baseline for a median of 12 h following ingestion of a moderate carbohydrate meal (25% ME), and in 21% and 17% of cats, respectively, glucose and insulin concentrations were increased above baseline for 24 h following a meal (59 kcal/kg body weight). These results are consistent with previous reports where diets with 30–33% of ME as carbohydrate increased glucose concentrations for an average of 12 h in lean cats fed a meal of approximately 30 kcal/kg body weight, and reports of return to baseline of 15 h in obese cats.13,27 Our results contrast with findings in dogs using similar methodology, where dogs were fed their daily energy requirement in one meal (63 kcal/kg) with 25% of energy from carbohydrate, and glucose concentrations returned to baseline at a median of 5 h after meal ingestion. 4
A previous study concluded that the postprandial increase in blood glucose concentrations in cats is negligible. 28 This was based on two findings. First, in healthy cats, blood glucose did not increase significantly within 2 h of eating, as occurs in humans and dogs. Secondly, morning blood glucose concentrations in diabetic cats did not differ significantly between cats fed ad libitum overnight and cats fasted for 12 h. 28 However, in our study, in 15/24 cats, blood glucose concentrations did not increase to significantly above baseline until more than 2 h after feeding and, in 12 cats, median time for blood glucose to return to baseline was greater than 12 h after eating. This delayed and prolonged postprandial response, combined with the short postprandial sampling period, likely accounts for the incorrect interpretation of the previous data.
Our results demonstrating that cats take substantially longer than dogs for postprandial glucose concentrations to return to fasting concentrations after eating are consistent with the relative glucose intolerance observed during an intravenous (IV) glucose challenge in cats compared to dogs, 9 and the findings of a study in healthy dogs and cats fed a meal with added glucose (2 g/kg). 10 Glucose concentrations in cats after an IV glucose tolerance test (0.5 g/kg body weight) return to baseline in 90–120 min,10,29,30 compared with 45–60 min in normal, healthy dogs.31,32 Similarly, when healthy cats and dogs were fed a standard meal with 2 g/kg of glucose added, mean values for peak glucose concentrations in cats were higher (10.0 mmol/l vs 6.3 mmol/l), times to peak later (120 min vs 60 min), and mean times to return to baseline three times longer compared to dogs (270 min vs 90 min). 10
The carbohydrate content of the diet tested in our study constituted 25% of energy as carbohydrate and so was in the lower end of the range typical for premium dry feline maintenance diets.20,22 Higher carbohydrate diets, such as those typically sold in supermarkets with 35–50% of energy from carbohydrate, would be expected to result in higher postprandial glucose and insulin concentrations, and potentially longer times to return to baseline.
The prolonged postprandial period in cats, compared to humans and dogs, may reflect specific feline adaptive mechanisms to meet the requirements of a carnivorous diet. The prolonged postprandial period in cats is owing, at least partly, to a minimally suppressible gluconeogenic pathway, lack of hepatic glucokinase activity, reduced disaccarharide activity in the small intestine, reduced and delayed insulin secretion, and delayed gastric emptying. 33
Cats have a highly developed gluconeogenic pathway relative to omnivores, which provides a continuous source of endogenous glucose; it exhibits minimal suppression in response to high plasma glucose concentrations, which might reduce the capacity of cats to quickly metabolise ingested carbohydrates. 7 The liver of most omnivorous animals contains two glycolytic enzymes — hexokinase and glucokinase — which are responsible for the phosphorylation of glucose to glucose-6-phosphate.7,8 In carnivorous species, such as the cat, the liver displays normal hexokinase activity. However, glucokinase activity in cats is almost absent. Glucokinase is primarily responsible for the conversion of glucose to glucose-6-phosphate, which is stored in the liver. If glucose enters the bloodstream at a rate similar to its rate of cellular uptake, the body is more likely to maintain glucose homeostasis and the likelihood of glycosuria is reduced. 34 As cats rely almost completely on hexokinase to facilitate hepatic glucose utilisation, this is a potential factor contributing to the prolonged postprandial glycaemia observed in this study.
The inability of cats to metabolise a diet high in carbohydrates also results, in part, from reduced activity of the digestive enzymes, such as amylase, in the small intestinal mucosa. 35 This results in the inefficient breakdown of disaccharides and starches to monosaccharides for absorption. In contrast to dogs, cats appear to be unable to induce changes in digestive enzymes to suit their diet. Although the reported digestibility of carbohydrates is similar in dogs and cats, the measure typically used is apparent digestibility, which includes the carbohydrate digested by mammalian enzymes, and fermentation of undigested carbohydrate in the colon by bacteria, and is therefore not an accurate comparison of small intestinal digestion between the species. 36 Cats also have significantly lower intracellular glucose transport activity compared to other species, which lowers their hepatocyte and erythrocyte uptake of glucose for metabolism and storage. 37
A feature of the difference between cats and dogs is that cats respond to a glucose load with a lower and later insulin peak. For example, in a meal-modified oral glucose tolerance test, peak insulin secretion in cats was half that of dogs and occurred three times later. 10 This suggests that the prolonged clearance of a glucose load in cats compared to dogs appears, at least partly, to be mediated by lower and delayed insulin secretion. Another factor implicated in the prolonged postprandial glycaemia of cats is delayed gastric emptying, with mean times for gastric emptying of 24 h and 14 h reported after cats ate 100% and 50% of their daily energy requirement in one meal. 38
The carbohydrate source has also been shown to affect postprandial glucose and insulin concentrations in cats and dogs.13,39 The carbohydrate source in our study was corn meal and corn grits; these have been shown to be beneficial in improving glucose intolerance and insulin sensitivity in overweight and obese cats compared with rice-based diets, but result in higher postprandial glucose concentrations compared with novel carbohydrate sources, such as lentil and cassava flour.12,13
Feeding management and behaviour may also play a role in postprandial glucose and insulin concentrations. The feeding behaviour in our study mimics that commonly seen in cats where energy intake is restricted, for example obese cats on a weight reduction programme. In this scenario, food is typically consumed soon after it is offered. The feeding behaviour is also observed in research cats if fed once daily and energy intake is restricted to maintain ideal weight; many cats consume much of their food soon after it is offered. Domesticated cats are often fed twice daily, which our study mimicked by feeding 50% of daily ad libitum energy intake as one meal. However, this varies between households and both ad libitum and once-daily feeding management are also used.40–42 Although the cats in our study were fed 50% of their daily ad libitum intake, the amount consumed (59 kcal/kg body weight) was similar to what many inactive pet cats require for the maintenance of ideal body weight. Therefore, it is clear that further studies are required to determine the range of postprandial glucose and insulin responses that occur with higher and lower carbohydrate loads, when either the amount fed and/or the carbohydrate content are varied. Of note is the considerable individual variation in postprandial glucose concentrations. Therefore, clinicians need to assess glucose concentrations for at least 12 h after eating when evaluating postprandial hyperglycaemia in diabetic and pre-diabetic patients, and be aware that there might be a carry-over effect lasting more than 24 h in some cats eating a similar carbohydrate load to the cats in our study.
The finding of higher postprandial concentrations in female cats after eating or after an IV glucose challenge was unexpected because male cats have been reported to be less insulin sensitive than female cats and are at greater risk of developing diabetes.43,44 However, pre-menopausal women are more likely than men to have postprandial hyperglycaemia rather than fasting hyperglycaemia, although there is no sex predisposition to diabetes in humans.45,46 Despite this similarity, meaningful cross-species comparisons are difficult because the cats in our study were neutered. The reason for higher postprandial glucose concentrations in women remains to be elucidated and has been attributed to multiple factors, such as mitochondrial homeostasis, the redox state, specific genes and sexual hormones.45–47 There was also a trend for female cats to have a longer return to baseline than male cats, which was not evident in the glucose tolerance test, although glucose concentrations were not measured after 3 h, so this may have limited the sensitivity to detect a difference. Although no reports were found in the literature of comparison between male and female cats for glucose concentrations at 2 and 3 h in a glucose tolerance test or time to return to baseline, it was reported in one study that there was no difference between lean or obese male and female cats for t1/2, Kglucose and area under the glucose curve in a glucose tolerance test. 43 Further research is warranted in cats to determine if this is a consistent finding in other cat populations.
There is no consensus in the veterinary literature about methods for estimating the time for glucose and insulin concentrations to increase after eating and time for these to return to baseline after eating (or after insulin administration). We considered two methods for defining return to baseline. For each of glucose and insulin, we used the first timepoint after peak concentration where concentrations were equal to or less than the sum of the cat’s baseline (average of concentrations at -30 min and -5 min), plus: (i) 10% of the cat’s baseline and (ii) the 90% range of differences. 26 Return to baseline was defined only if the cat had first exceeded baseline based on the same method. These estimates gave somewhat different results, with the 10% method giving the longest times for glucose (median 15 h; results not shown). The advantage of the 90% range of differences method is that it is based on the actual variability from the study population rather than the arbitrary value of 10%, and uses the t-distribution to estimate unexplained variability of the values of glucose and insulin at each timepoint; when sample sizes are small, this is more appropriate than the normal distribution as is used in the critical difference method. A disadvantage of the 90% range of differences method is that it assumes the same variability for all timepoints for all cats as the pooled variance at baseline for all cats.
Cats are prone to stress hyperglycaemia which may occur in experimental and clinical settings. To minimise the effect of stress responses during the current study, cats were familiarised with their environment, carers and study procedures for 2 weeks prior to sample collection. Insulin concentrations significantly increased after eating in our study, but have been shown to decrease with stress hyperglycaemia in cats. 48 This suggests our results were unlikely to be affected by responses to stress stimuli. In addition, the prolonged postprandial period has been previously reported in lean and obese cats,11,13,27 and has been observed previously in our laboratory in other dietary studies. 49
A fasted control group was not included in this study because our aim was to describe the duration and magnitude of postprandial increases in plasma glucose and insulin concentrations in clinically healthy cats fed a meal after fasting for almost 24 h, rather than to compare effects of feeding to those of fasting. We considered that fasting for almost 24 h prior to feeding was necessary, as there was prior evidence in the literature that the postprandial period is prolonged in cats.12,13,26,32 To compare effects of feeding to those of fasting would have required fasting one group of cats for almost 48 h (almost 24 h before feeding and 24 h post-feeding), raising important welfare and ethical concerns.
Conclusions
In conclusion, results from this study show that the postprandial period of cats fed a single meal of 59 kcal/kg with 25% of energy from carbohydrate is 12 h and is frequently longer than 15–24 h. The duration of the postprandial increase in glucose concentrations is substantially greater than in dogs and people when diets of similar composition are considered. This prolonged postprandial period should be considered when treatment protocols are developed for optimising glycaemic control in diabetic and pre-diabetic cats. It is also an important consideration for researchers requiring fasting blood samples for measurement of blood glucose and insulin concentrations in cats. Withholding food for 24 h will not always result in glucose concentrations at true fasting values. Females had higher peak glucose concentrations; this result requires further investigation in a larger population.
Footnotes
Acknowledgements
The authors wish to thank Linda Fleeman, Rebekah Scotney, Elizabeth Jolly, Dawn Herd, Elizabeth Wilcox, Maree Maher and Roger Dyke. The staff of the Small Animal Clinic and Hospital at the University of Queensland are also thanked for their assistance with animal care and data collection.
Funding
The authors are grateful for the funding provided by The Iams Company, Lewisburg, OH, USA for this project.
Conflict of interest
The authors report no real or perceived vested interests that relate to this article (including relationships with the granting body or other entities whose products or services are related to topics covered in this article) that could be construed as a conflict of interest.