Less is more? Ultra-low carbohydrate diet and working dogs’ performance

Arnon Gal; Williams Cuttance; Nick Cave; Nicolas Lopez-Villalobos; Aaron Herndon; Juila Giles; Richard Burchell

doi:10.1371/journal.pone.0261506

. 2021 Dec 23;16(12):e0261506. doi: 10.1371/journal.pone.0261506

Less is more? Ultra-low carbohydrate diet and working dogs’ performance

Arnon Gal ^1,^*,^#, Williams Cuttance ², Nick Cave ³, Nicolas Lopez-Villalobos ⁴, Aaron Herndon ⁵, Juila Giles ⁶, Richard Burchell ^7,^#

Editor: Balamuralikrishnan Balasubramanian⁸

PMCID: PMC8699952 PMID: 34941910

Abstract

New Zealand farm working dogs are supreme athletes that are crucial to agriculture in the region. The effects that low or high dietary carbohydrate (CHO) content might have on their interstitial glucose (IG) and activity during work are unknown. The goals of the study were to determine if the concentration of IG and delta-g (a measurement of activity) will be lower in dogs fed an ultra-low CHO high fat diet in comparison to dogs fed a high CHO low fat diet, and to determine if low concentrations of IG are followed by reduced physical activity. We hypothesized that feeding working farm dogs an ultra-low CHO diet would reduce their IG concentrations which in turn would reduce physical activity during work. We prospectively recruited 22 farm dogs from four farms. At each farm, dogs were randomized to one of two diets and had a month of dietary acclimation to their allocated diet. The macronutrient proportions as a percentage of metabolizable energy (%ME) for the high CHO low fat diet (Diet 1) were 23% protein, 25% fat, and 52% CHO, and for the ultra-low CHO high fat diet (Diet 2) 37% protein, 63% fat, and 1% CHO. Following the acclimation period, we continuously monitored IG concentrations with flash glucose monitoring devices, and delta-g using triaxial accelerometers for 96 h. Dogs fed Diet 2 had a lower area under the curve (±SE) for IG (AUC _{Diet 2} = 497 ± 4 mmol/L/96h, AUC _{Diet 1} = 590 ± 3 mmol/L/96h; P = 0.002) but a higher area under the curve (±SE) for delta-g (AUC _{Diet 2} = 104,122 ± 6,045 delta-g/96h, AUC _{Diet 1} = 80,904 ± 4,950 delta-g/96h; P< 0.001). Interstitial glucose concentrations increased as the activity level increased (P < 0.001) and were lower for Diet 2 within each activity level (P < 0.001). The overall incidence of low IG readings (< 3.5 mmol/L) was 119/3810 (3.12%), of which 110 (92.4%) readings occurred in the Diet 2 group (P = 0.001). In the Diet 2 group, 99/110 (90%) of the low IG events occurred during the resting period (19:00–06:00). We conclude that feeding Diet 2 (ultra-low CHO high fat diet) to working farm dogs was associated with increased delta-g despite decreased IG concentrations. Interstitial glucose concentrations were positively associated with dogs’ activity levels independent of diet. Lastly, events of low IG occurred at a low incidence and were predominantly seen between 19:00–06:00 in dogs fed the ultra-low CHO high fat diet.

Introduction

There are approximately 200,000 working farm dogs across New Zealand [1]. These dogs are supreme athletes that work in all weather conditions and might run 60–100 km/d on steep terrain, often at 20–30 km/h [1]. Hence, adequate nutrition in the face of high-energy expenditure is paramount to the general health and performance of these dogs. There are currently two common feeding practices for working farm dogs in New Zealand that differ in their dietary macronutrient content: Ultra-low CHO, high fat sources and high CHO, low fat sources of metabolic energy (%ME). However, the effect of dietary macronutrient content on the activity of New Zealand’s working farm dogs is hitherto unknown. Dogs do not require dietary CHO, and exercising sled dogs may have improved performance when fed a high-fat CHO-free diet [2]. However, the concept of “cross-over”, where muscle utilization of CHO for ATP production increases with increased exercise intensity, has been demonstrated in several different mammalian species, including dogs [3]. Since the intensity of activity engaged by working farm dogs in New Zealand is likely to be greater than the sustained moderate-intensity endurance activity of sled dogs, New Zealand’s working farm dogs may be unable to maintain normal body glucose levels when fed a low CHO diet.

One aspect of nutrition is the maintenance of adequate glycemic control. Blood glucose during the fasted and absorptive phases is derived from internal and external sources, respectively. In the fasted state, glycogenolysis of hepatic glycogen stores and gluconeogenesis maintain euglycemia [4, 5]. In contrast, in the absorptive phase, dietary glucose is directly utilized to maintain euglycemia, and excess glucose is converted to storage fuels (i.e., fat and glycogen) as long as the diet has sufficient CHO content [6]. Prolonged physical activity can deplete hepatic and muscle glycogen stores, and maintenance of glucose levels would rely on hepatic gluconeogenesis. However, under conditions of strenuous and prolonged physical exercise, gluconeogenesis might fail to maintain euglycemia. Therefore, feeding diets with high CHO content to dogs that are undergoing strenuous and prolonged physical exercise might support the maintenance of glucose levels for a longer time through maximizing muscle and hepatic glycogen stores and rapidly metabolizable CHO absorption from the gastrointestinal tract.

Healthy human endurance athletes occasionally experience hypoglycemic episodes during exercise [7]. Therefore, hypoglycemia might also occur in working farm dogs and could negatively impact their performance, and general well-being, thus justifying an investigation. This study tested the hypothesis that working farm dogs fed an ultra-low CHO diet would reduce their body glucose levels which in turn would reduce physical activity during work. (Fig 1). The study aims were to determine if IG and delta-g (a measurement of activity) would differ between dietary groups, and if reduced delta-g would follow periods of low levels of IG.

Fig 1 — We hypothesized that an ultra-low CHO high fat diet would reduce IG through decreased hepatic glucose production resulting in decreased delta-g (a measurement of activity). %ME CHO, % of metabolic energy from carbohydrate sources.

Materials and methods

Study design

This prospective, randomized, controlled experimental field study was approved by the Massey University Animal Ethics Committee (Protocol 16/02). We recruited 22 dogs from four farms on the North Island of New Zealand and randomized the dogs into two balanced groups with respect to diet (n = 11 each) using an online randomization tool (www.randomizer.org); there was no attempt to balance the dogs per age, sex, breed, or bodyweight during the randomization and recruitment (Tables 1 and 2). Dogs were recruited after receipt of each owner’s verbal consent to participate in the study. Dogs were only included in the study if they had been assessed by a physical exam by at least one of the authors and by the history provided by the owner and were deemed to be in good health. Acclimation of the dogs to the two diets started one month prior to the commencement of each of the three 96 h study periods, which took place during times of peak seasonal work activity, during which the dogs in each farm worked together as a team. Because of the difficulties in timing the dietary acclimation to match with the unpredictable timing of the periods of peak seasonal work activity (which depends on appropriate weather conditions), the study included three separate 96 h periods between October and November. We recorded the dogs’ bodyweights at recruitment, and at the beginning and at the end of each of the study’s 96 h periods, during which, we continuously monitored IG levels using a flash glucose monitor (FreeStyle Libre, Abbott Diabetes Care, Doncaster, VIC, Australia). Dogs were also fitted with a triaxial accelerometer (TAA; Heyrex^®, Wellington, New Zealand) which was fixed to their neck collar, to continuously quantify their delta-g.

Table 1. Demographics and descriptive statistics of the dog cohort in this study.

Farms	All	A	B	C	D
Number of dogs	22	5	5	8	4
Age (years)^#	3.5 (4)	6^a,b (3.5)	2^a (1)	4.5^b (3.5)	3^a,b (1)
BW (kg) at study entry^†	25.2±2.7	20.7±3.2	25.4±7.8	28.4±7.8	24±4.4
BW (kg) at the beginning of 96 h^†	24.7±6.3	21.8±2.8	24.3±7.7	26.9±7.8	24.2±4.1
BW (kg) at the end of 96 h^†	24.7±6.7	20.9±2.7	25±8.3	27.5±7.9	23.4±4.5
Males, Females	16, 6^*	2, 3	4, 1	8, 0	2, 2
Heading dog, Huntaway, Mixed breed, Border Collie^¶	9, 9, 3, 1	2, 2, 1, 0	3, 1, 0, 1	4, 4, 0, 0	0, 2, 2, 0

Open in a new tab

BW, bodyweight;

^†, results are presented as mean (±SD);

^#, results are presented as median (IQR); different superscript letters represent a significant difference (P < 0.050) between farms;

^¶, significant under-represented breed (P < 0.001);

*, significant difference (P < 0.050) between males to females; All dogs, but one intact female, were neutered.

Table 2. Demographics and descriptive statistics per diets at the beginning of the acclimation period.

	Diet 1	Diet 2	P value
^#Baseline age [median (IQR); year]	3.5 (3.5)	3.5 (3.5)	0.947
*Baseline body weight (mean ± SD; kg)	23.6 ± 4.5	26.5 ± 8.1	0.850
^#Baseline body condition score [median (IQR); 1–9 scale]	4 (2)	4 (1)	0.409
⁺Sex (M, F)	9M, 2F	7M, 4F	0.338
⁺Farms (A, B, C, D)	2A, 2B, 5C, 2D	3A, 3B, 3C, 2D	0.825
⁺Breeds (Bc, Hd, Hw, Mx)	1Bc, 5Hd, 3Hw, 2Mx	0Bc, 4Hd, 6Hw, 1Mx	0.485

Open in a new tab

Bc, Border collie; CHO, carbohydrates; F, female; Hd, Heading dog; Hw, Huntaway dog; IQR, interquartile range; M, male; Mx, mixed breed dog; SD, standard deviation.

Diets

The macronutrient proportions as a percentage of metabolizable energy (%ME) for the first diet (Diet 1) were 23% protein, 25% fat, and 52% CHO (Pedigree Adult real chicken^®, MARS Petcare New Zealand, Auckland, New Zealand), and for the second diet (Diet 2), 37% protein, 63% fat, and 1% CHO (Chicken Feast Freeze Dried^®, K9 Natural, Christchurch, New Zealand). The ingredients and nutritional analysis of the two diets are presented in S1 and S2 Tables. During the acclimation and the 96 h testing periods, the dogs were fed their allocated diet exclusively once daily between 6pm-8pm whilst in their individual kennels. No other source of food was allowed. The dogs had ad libitum access to water. The owners were asked to feed the dogs during the acclimation period to maintain bodyweight similar to other studies in athletic dogs [8, 9]. The empirically determined amount fed by the end of the acclimation period was then maintained throughout the 96 h test periods. We based our rationale for choosing the “feeding to maintain body-weight” strategy for several reasons. Firstly, given the nature of the study design, we were unable to quantify the dietary intake of individual dogs. Secondly, we expected large variations in energy expenditure that would lead to significant differences in bodyweight if fed to a predetermined intake. Thirdly, large variations in maintenance energy requirements between dogs, even of the same bodyweight, have been previously documented [10]. Hence, feeding to maintain bodyweight seemed the most appropriate approach.

Sample collection and handling

Blood collection

Whole blood samples (5 mL) were collected by jugular venipuncture from each of the dogs enrolled in the study between 8am-10am after an overnight fast, and split between serum (BD Vacutainer, Auckland, New Zealand) and sodium fluoride (BD Vacutainer, Auckland, New Zealand) blood collection tubes. The samples were collected following dietary acclimation and at the beginning and at end of the 96 h study periods. Whole blood in the serum tubes was allowed to clot for 10 minutes whereas the whole blood in the sodium fluoride tubes was inverted several times to allow for the anticoagulant to mix with the blood. Then, the samples were stored on ice as long as the veterinarian was in the field and shipped on ice overnight to the principal investigator.

Sample handling and processing

Immediately upon receipt of the blood samples from the field veterinarian, the blood samples were centrifuged at 3000 × g for 15 min at 4°C to facilitate harvesting the serum and plasma for insulin and glucose assays, respectively. A local veterinary diagnostic laboratory (IDEXX New Zealand, Palmerston North, New Zealand) measured plasma glucose concentration (OSR6121, Beckman-Coulter, Sydney, NSW, Australia) on an AU680 Chemistry Analyzer (Beckmancoulter, Sydney, NSW, Australia); whereas serum samples for insulin quantification were stored in -80°C until the end of the study and then shipped on dry ice to the University of Queensland (Gatton, QLD, Australia) for analysis.

Flash glucose monitoring and triaxial accelerometry

The FreeStyle Libre^® (Abbott Diabetes Care, Doncaster, VIC, Australia) was used to measure the IG in the subcutaneous adipose tissue every 15 minutes. To apply the sensor, a 5 cm × 5 cm square on the lower right side of the neck was clipped, and aseptically cleaned. The FreeStyle Libre^® sensor was deployed on the prepared area using the pre-loaded applicator and secured with an elastic bandage. The FreeStyle Libre^® has been previously shown to have a good correlation with plasma glucose in dogs [11–14]. At the same time, the dogs were also fitted with a triaxial accelerometer (Heyrex^®, Wellington, New Zealand) that generates a numerical measurement unit of activity (delta-g) [15–17]. Acceleration in all three axes was summed to produce the delta-g value, which was analyzed as 15 min epochs. The duration of the period of continuous glucose and activity recording was 96 h. During that time, the dogs in each farm worked according to their usual regimen together as a team.

Accelerometers and flash glucose monitors were fitted in all dogs in the morning (between 8am–10am) on the first day of data collection. All data points for IG had a corresponding time-stamped, matched activity datapoint. This permitted the evaluation of the temporal relationship between IG and activity as these could be tracked on a time-matched basis for both IG and activity. To standardize the Time 0 point between dogs, the dogs were normalized based on their delta-g. Hence, the first IG data point that was used for analysis (i.e., Time 0) corresponded to the first nadir of delta-g (i.e., the lowest value for a 15-min epoch) that followed the first recorded peak of delta-g (i.e., the highest value for a 15-min epoch) for each dog. That way, all dogs were aligned based on the elapsed time from Time 0 to the end of the 96 h period of data acquisition.

The measurement of IG in dogs is a relatively new practice, and hence well-established references intervals for canine IG are not available. By way of extrapolation from established blood glucose levels [18], we defined low IG as any IG concentration below 3.5 mmol/L.

Insulin assay, and the homeostatic model assessment of insulin resistance

Serum samples were maintained at -80°C until thawing for the assay. Insulin was measured at a regional veterinary reference laboratory (QML Vetnostics Laboratory, Murarrie, QLD, Australia) that participates in the European Society of Veterinary Endocrinology Quality Assurance Program. The reference laboratory reported quality control results for the assay of less than 1 SD from the mean across all participating laboratories in the European Society of Veterinary Endocrinology Quality Assurance Program. A direct chemiluminescent human assay, which is based on a two-site sandwich immunoassay using a mouse monoclonal capture antibody (ADVIA Centaur Insulin Assay, Siemens-Australia, Bayswater, VIC, Australia) was used. The assay’s lower limit of detection and range are 0.5 μU/L and 1–300 mU/L, respectively. The within run coefficient of variation (CV) at insulin concentrations of 27 mU/L and 65 mU/L are 1.05% and 3.9%, respectively. The between run CV at insulin concentrations of 27 mU/L and 65 mU/L are 4.07% and 6.83%, respectively.

The homeostatic model assessment of insulin resistance (HOMA-IR) is an epidemiological tool used in humans to assess insulin resistance. High HOMA-IR levels indicate an increased resistance to insulin, and lower levels indicate an increased sensitivity to insulin’s action. The HOMA-IR was calculated as previously described [19, 20] according to the following formula: plasma insulin x serum glucose / 22.5.

Dog activity levels

We split the dogs’ delta-g values over the 96 h period into three levels according to the individual dog’s triaxial accelerometer delta-g values as follows: high (≥75% percentile; a total of 981 15-min epochs), moderate (>25% and <75% percentiles; a total of 1803 15-min epochs), and low (≤25% percentile; a total of 915 15-min epochs).

The ‘working period’ and ‘resting period’

For the purpose of comparing the effect of the diets on IG at the end of the working day (when the dogs rested in the kennel) compared with the period of time of the day where the dogs had been working, we divided the day into two periods: the period spanning between 19:00–06:00 (i.e., ‘resting period’; the end of the working day to the beginning of the next working day) and the period from 06:00–19:00 (‘working period’). We assumed that during the ‘resting period’ the dogs spent most of their time sleeping. This assumption was based on 1) the time from sunset to sunrise in Auckland, North Island of New Zealand (https://www.gaisma.com/en/location/auckland.html); and 2) the lower activity quartiles between 19:00–06:00 accounted for 85.6% of the total lower activity quartile over the 96 h periods (S1 Fig).

Power-sample size and statistical analysis

A priori power analysis (G*Power software version 3.1.9.2) indicated that 9 dogs in each group will be sufficient to detect a difference between two dependent IG means of 100 mg/dL and 70 mg/dL with SD of 30 mg/dL, power of 0.8, alpha error probability of 0.05, and a correlation coefficient of 0.6 between measurements on the same dog. To account for potential losses of dogs during the study period (i.e., dislodgement of the sensor), we recruited 11 dogs per group.

The statistical analysis was performed with the statistical package SAS University Edition (SAS Institute Inc., Cary, NC, USA). The variables in the dataset were examined for normal distribution by inspection of Q-Q plots, histogram, and by the Shapiro-Wilk test. Descriptive statistics were calculated by the MEANS procedure. The median (IQR) were used to describe the variables ‘age’ and ‘body condition score’ (Tables 1 and 2). The mean (±SD) were used to describe ‘bodyweight at the beginning of the acclimation period’, ‘bodyweight at the beginning of the 96 h data recording period’, and ‘bodyweight at the end of the 96 h data recording period’ (Table 1). Analysis of variance for the dependent variable IG was performed with the MIXED procedure using a linear mixed model for repeated measures taken on the same dog during the 96 h of data recording [21]. The model included the fixed effects of ‘period’, ‘activity level’, ‘diet’, interaction between ‘activity level’ and ‘diet’, and the random effect of ‘dog’ to account for repeated measures on the same ‘dog’. The time (i.e., the 96 h of data recording) was converted to the class variable ‘period’ defined with 8 levels. The effects of ‘activity level’ and ‘diet’ on the least squares means (LSM) of dependent variable IG were compared by the Fisher’s Least Significant Difference test as implemented in the LSMEANS option.

The NPAR1WAY procedure was used to analyze the dependent variables ‘age’ and ‘body condition score’ (scale of 1–9) with respect to farm (Table 1) and diet (Table 2) allocation, as these variables did not follow a normal distribution. The NPAR1WAY procedure was also used to analyze the dependent variables ‘plasma insulin’ and ‘HOMA-IR’ between the beginning and the end of the 96 h data recording period, as these variables did not follow a normal distribution. Analysis of variance for the dependent variables ‘bodyweight at the beginning of the acclimation period’, ‘bodyweight at the beginning of the 96 h data recording period’, and ‘bodyweight at the end of the 96 h data recording period’ was performed with the MIXED procedure using a linear mixed model. The model included the fixed effects of farm (Table 1) and Diet (Table 2).

Repeated measures of ‘bodyweight’ on the same dog were analyzed using the MIXED procedure with a mixed model that include the fixed effect of ‘diet’, ‘time’ (i.e., the beginning of the acclimation period, and the beginning and the end of the 96 h data recording period), and interaction between ‘diet’ and ‘time’, and the random effect of ‘dog’ to account for repeated measures on the same dog. Repeated measures of ‘serum glucose’ on the same dog were analyzed using the MIXED procedure with a mixed model that include the fixed effect of ‘diet’, ‘time’ (i.e., the beginning and the end of the 96 h data recording period), and interaction between ‘diet’ and ‘time’, and the random effect of ‘dog’ to account for repeated measures on the same dog.

Analysis of the frequency of low IG readings during the ‘working period’ and ‘resting period’, between activity levels, and between diets was performed with Chi-Square Test as it is implemented in the FREQ procedure.

Curves for ‘IG’ and ‘delta-g’ during the period of recorded data were modelled using a mixed model fitting nine-knots splines for each dog using the statistical software ASReml [22, 23]. The model included the fixed effect of ‘diet’ and random effect of ‘dog’. A spline for ‘IG’ and ‘delta-g’ were modelled for each dog and the best predicted curve was estimated for each diet. The differences between the areas under the curves (AUC) for ‘IG’ and ‘delta-g’ were performed by calculating their z statistics according to the following formula: z = |AUC₁-AUC₂| / sqrt (SE₁² + SE₂²). Fitting a random regression model with ASReml, modeling the ‘IG’ and ‘delta-g’ for each dog, and choosing the best predicted AUC for each diet was deemed the most accurate way to estimate the effect of diet on ‘IG’ and ‘delta-g’. The unfitted raw data is presented in S2 Fig.

Results

The effect of the diets on bodyweight

At the beginning of the acclimation period there were no significant differences between the dogs in diet groups 1 and 2 with regards to ‘age’, ‘bodyweight’, ‘body condition score’, ‘sex’, distribution within farm, and breeds (Table 2).

The analysis of variance for bodyweight indicated that Dogs fed Diet 1 maintained their bodyweight from the beginning of the acclimation period to the beginning of the 96 h study period (LSM ± SE 23.6 ± 2.1 kg vs. 23.9 ± 2.1 kg; P = 0.916), whilst dogs fed diet 2 lost approximately 1.2 kg (LSM ± SE 26.5 ± 1.9 kg vs. 25.3 ± 1.9 kg; P = 0.007). The bodyweights of all dogs remained unchanged during the 96 h study period (Diet 1, LSM ± SE 23.9 ± 21. kg vs. 23.9 ± 21. kg, P = 1.000; Diet 2, LSM ± SE 25.3 ± 1.9 kg vs. 25.3 ± 1.9 kg, P = 1.000).

The effect of diets on activity level and interstitial glucose concentration

The predicted areas under the curve for ‘IG’ and ‘delta-g’ per diet are graphically depicted in Fig 2. Dogs fed diet 2 had a significantly lower area under the curve (±SE) for IG when compared with the Diet 1 (AUC _{Diet 2} = 497 ± 4 mmol/L/96 h, AUC _{Diet 1} = 590 ± 3 mmol/L/96 h; P = 0.002). In contrast, dogs fed diet 2 had a significantly higher area under the curve (±SE) for delta-g when compared with the Diet 1 (AUC _{Diet 2} = 104,122 ± 6,045 delta-g/96 h, AUC _{Diet 1} = 80,904 ± 4,950 delta-g/96 h; P < 0.001).

Interstitial glucose concentration as a function of three activity levels

The LSM (±SE) IG was significantly greater with higher activity level (P < 0.001), and in Diet 1 (P < 0.001; Fig 3). In addition, there was a significant interaction between the level of activity and diet (P < 0.001), whereby dogs fed Diet 2 had a greater increase in IG with increased activity than dogs fed Diet 1 (Fig 3).

The effect of diet, and the ‘working period’ and ‘resting period’ on the incidence of low levels of IG

Over the three 96 h periods, there were 3810 IG readings; 2090 (55%) during the ‘working period’, and 1720 (45%) during the ‘resting period’. Overall, the proportion of low IG readings (< 3.5 mmol/L) was 119/3810 (3.12%; S3 Fig). Despite the lower total number of IG readings during the ‘resting period’ relative to the ‘working period’, a significantly (P < 0.001) higher proportion of low IG readings occurred during the ‘resting period’ (101/119; 84.9%) relative to the ‘working period’ (18/119; 15.1%; S4 Fig).

There were 1504/3810 and 2306/3810 IG readings from dogs in Diet 1 and Diet 2 groups, respectively (P < 0.001). We found that 9/119 (7.6%) of the low IG readings occurred in dogs from Diet 1 group (0.6% of the total IG readings for this group), and 110/119 (92.4%) in dogs from Diet 2 group (4.8% of the total IG readings for this group; P = 0.001).

We observed a reciprocal pattern in the timing of the low IG readings between the two diets. Seven of the 9 (77.8%) low IG readings in Diet 1 group occurred during the ‘working period’ and only 2/9 (22.2%) occurred during the ‘resting period’. In contrast, 99/110 (90%) of the low IG readings in the Diet 2 group occurred during the ‘resting period’ and only 11/110 (10%) occurred during the ‘working period’. There was a significant interaction between the diet and the ‘resting period’ / ‘working period’ with regards IG (P < 0.001).

Overall, 10/22 (45.5%) dogs experienced low IG events (median of 6.5; IQR 17.5; range 1–42). Eight of the 10 dogs (80%) were from Diet 2 group and accounted for 110/119 of the low IG readings, whereas 2/10 dogs were from Diet 1 group and accounted for 9/119 of the low IG readings. Two of the 10 dogs had a total of two low IG readings during the ‘working period’ without any accompanying low IG readings during the ‘resting period’. Similarly, 2/10 dogs had seven low IG readings during the ‘resting period’ without any accompanying low IG readings during the ‘working period’. The other 6/10 dogs had low IG readings both during the ‘resting period’ and during the ‘working period’.

The effect of activity level and diet on the incidence of low IG readings

There was a significantly lower proportion of low IG readings (P < 0.001) in dogs during periods of high activity level 5/119 (4.2%) in comparison to dogs during periods of moderate 62/119 (52.1%) and low 52/119 (43.7%) activity levels. The proportion of low IG readings at the high activity level relative to the low and moderate activity levels did not differ between the diets (P = 0.544); however, within each level of activity, the proportion of low IG readings was higher in the group of dogs of Diet 2 relative to the group of dogs of Diet 1 [high activity level Diet 2 vs. Diet 1, 4 (80%) vs. 1 (20%), P = 0.386 moderate activity level Diet 2 vs. Diet 1, 48 (92%) vs. 4 (8%), P < 0.001; low activity level Diet 2 vs. Diet 1, 58 (94%) vs. 4 (6%), P < 0.001].

Glucose, insulin and HOMA-IR

Mean (±SD) serum fasted glucose concentrations at the beginning and the end of the 96 h periods were 4.83 ± 0.43 and 4.66 ± 0.57 mmol/L, respectively. Median (IQR) serum insulin and HOMA-IR at the beginning of the 96 h periods were 4 mU/L (1 mU/L) and 0.80 (0.29), and at the end of the 96 h periods were 4 mU/L (2 mU/L) and 0.84 (0.44). Serum glucose, insulin, and the HOMA-IR did not differ between the beginning and the end of the 96 h periods (P = 0.320; P = 0.489; P = 0.724; respectively). The median (IQR) HOMA-IR was significantly lower (P = 0.047) at the end of the 96 h study period for Diet 2 group 0.78 (0.25) when compared to Diet 1 group 1.04 (0.56); whereas, the median (IQR) HOMA-IR did not differ between diets at the beginning of the 96 h study period (Diet 2 0.67 (0.38) vs. Diet 1 0.90 (0.78), P = 0.061). None of the dogs had evidence of hyperinsulinemic hypoglycemia neither at the beginning nor the end of the 96 h study periods.

Discussion

In this study we found that the adaptation to Diet 2 (high fat and ultra-low CHO with 63% and 1% ME from fat and CHO sources;) resulted in significantly lower IG and higher activity AUCs in New Zealand farm dogs compared to a diet with high CHO and low fat (Diet 1; Fig 2). We did not measure the concentrations of ketone bodies in the blood and urine of these dogs; hence we could not confirm if ketones were increased in these dogs during the 96 h periods. Based on previous studies in dogs fed similar %ME from CHO and fat sources, it is probable that following fat adaptation, ketone production and utilization was increased in these dogs, which could explain their decreased IG concentration and increased activity. In support of that, Beagle dog bitches fed a ultra-low CHO high fat diet (0% and 74% ME from CHO and fat sources) had elevated serum ketones, decreased alanine, and glucose was lower by approximately 0.83–1.1 mmol/L (15–20 mg/dL) during the week before whelping (when negative energy balance develops) compared to matched controls fed a high CHO low fat diet (44% and 30% ME from CHO and fat sources) [24]. Similar findings have been demonstrated in rodents and humans. For example, rats on a high fat diet ran 32% more than control rats that were fed a diet with lower fat and higher CHO [25]. Adaptation to high fat diets in well-conditioned humans increases the rate of fat oxidation [26]. Ultramarathon and iron-man athletes on a high fat diet had 2.3-fold higher peak fat oxidation, and 59% higher mean fat oxidation during submaximal exercise relative to matched control ultramarathon and iron-man athletes on high CHO diets. The peak fat oxidation occurred at a higher percentage of VO²max (70.3 ± 6.3 vs. 54.9 ± 7.8). Nevertheless, these athletes did not have differences in resting and post-exercise levels of muscle glycogen demonstrating the muscle adaptation to oxidative fat metabolism [26]. As we suggested above, feeding dogs a diet of ultra-low %ME CHO and high %ME fat leads to adaptations that offer significant advantages to dogs that perform endurance work, such as New Zealand farm dogs. For example, exhaustive exercise in sled dogs that consumed a diet with 0% of metabolizable energy from a CHO source and high %ME from fat had higher albumin, total calcium, and magnesium than control dogs on diets with higher %ME from a CHO source [27]. The adaptation to high fat diets has been postulated to be beneficial because the higher albumin, total calcium, and magnesium protects against exercise-induced hypovolemia and the adaptation to high rates of muscle fat oxidation protects against muscle injury [27].

The results of this study may not directly translate to other kinds of working dogs. For example, it is unlikely that racing dogs (e.g., Greyhounds) would benefit from ultra-low CHO high fat diet as their race-day performance relies on high levels of muscle glycogen for short anerobic bursts of high-speed running. Similarly, personal assistance dogs (e.g., seeing eye dogs) that are not expected to engage in high endurance activities might not benefit from the advantages of fat adaptation like sled, military/police, search and rescue, and New Zealand farm dogs would. Maximizing the dietary benefits for each of the different fields of activity that dogs engage in our lives would require specific research tailored to the requirements of the individual field.

We also found a temporal relationship between delta-g and IG in which the direction of changes in IG concentrations followed and lagged after similar changes in delta-g (Figs 2 and 3). Mechanistically, the temporal relationship between changes in delta-g and IG could be explained by increased sympathetic tone, increased secretion of glucagon, and inhibition of insulin secretion during a high activity level which would increase IG [28]. Conversely, decreased activity levels lead to decreased sympathetic tone combined with increased insulin release, 5’ adenosine monophosphate (AMP) and 5’ adenosine monophosphate-activated protein kinase (AMPK)-mediated hepatic gluconeogenesis and translocation of insulin-independent glucose transporter type 4 (GLUT4) from the sarcoplasmic reticulum to the striated muscle membrane which would decrease IG [28, 29]. However, further studies are required to provide mechanistic insight to determine whether either or both of the above explanations, or alternative explanations underlie the temporal relationship between delta-g and IG as our study was not designed to investigate it.

During the 96 h of data recording there were neither low peaks of IG that followed high peaks of delta-g, nor low peaks of delta-g followed low peaks of IG. One possible explanation could be that the 96 h study periods of intense monitoring might have been too short to capture these changes. Our alternative explanation is that the dogs included in this study were physically well conditioned, had a low body fat mass, and had very high sensitivity to insulin coupled with high efficiency of energy utilization. In support of this alternative, conditioned sled dogs had higher GLUT4 per g of protein in peripheral blood mononuclear cells, lower HOMA-IR, and lower fasting concentrations of insulin and glucose than sedentary dogs. The HOMA-IR index represents the degree of sensitivity to insulin and has been validated in dogs [19, 20]. The median HOMA-IR of the dogs in this study (0.82–0.85) is similar to that reported in conditioned sled dogs [19] and provides some support to our alternative explanation.

In this study we demonstrated multiple episodes of low IG in healthy dogs. The dogs in this study had their serum insulin and glucose concentrations measured at the beginning and the end of the 96 h study periods, and none of the dogs had results that were suggestive of hyperinsulinemic hypoglycemia. Therefore, healthy dogs could have a low incidence of low IG and the frequency of low IG may be greater in dogs fed ultra-low CHO high fat diet (Diet 2). Whether episodes of low IG occur in all dogs in all activities or is only present following periods of heavy exercise remains to be determined. Also, the clinical significance of low IG is obfuscated by the absence of well-established reference intervals for IG in dogs. One possibility is that our extrapolated reference intervals for IG led to misinterpretation of the lower IG concentrations as being abnormal, and that it is normal for dogs to experience those IG concentrations, especially following exercise. Alternatively, it could be that the dogs in this study truly had physiologically low levels of IG (glucopenia) and that either clinical signs did not develop, or that the farmers did not detect the clinical signs because most occurred during the ‘resting period’ when the dogs were confined in the kennels and physically away from the farmers. As the study was not designed to answer this question, we cannot provide further insights. Nevertheless, the higher incidence of low IG (per our extrapolated reference interval) in dogs fed ultra-low CHO high fat diet (Diet 2) could suggest that this diet either provided insufficient glucose for hepatic output overnight, or that the adaptation to higher dietary fat reduced tissues’ dependency on glucose (e.g., brain), and prevented deleterious effects of glucopenia. At this point we postulate that it is most likely the result of adaptation to higher dietary fat as discussed above.

As stated, most events of low IG occurred during the ‘resting period’ (84.9% vs. 15.1%). It is hard to contrast our finding with those of others because flash glucose monitoring is rarely performed on healthy individuals, and because well-established reference intervals for IG are absent. The risk for the development of hypoglycemia in diabetic humans is highest at night during the sleeping period, as sleep impairs counterregulatory-hormone responses to hypoglycemia [30–33]. In one study, hypoglycemia was induced in healthy human volunteers during early and late sleep [30]. During late sleep, there was a significantly diminished counterregulatory hormonal response (epinephrine, norepinephrine, ACTH, cortisol, and growth hormone) to the induced hypoglycemia [30]. In another study, type-1 human diabetic patients and matched healthy controls were exposed to hyperinsulinemic hypoglycemic clamps during the day, at night when they were asleep, and at night when they were awake [32]. Both, type-1 human diabetic patients and their matched healthy controls had decreased sympathoadrenal responses to hypoglycemia during sleep, but not at night if they were awake. In another study, serum glucagon concentrations were compared between human patients with type-1 diabetes and matched healthy controls [31]. Strikingly, the authors of that study demonstrated that glucagon secretion during the sleeping period was independent of glucose and insulin. Serum glucagon did not differ between type-1 diabetic patients and their matched controls, despite substantial differences in their serum glucose and insulin concentrations [31]. In a similar study, serum glucagon did not differ between human type-1 diabetics and matched healthy controls when hypoglycemia was induced during the sleeping period [33]. The nocturnal hypoglycemia phenomenon has not been previously reported in dogs. In a recent study involving 14 diabetic dogs fed a variety of commercial diets periods of low IG were documented mostly during the day [34]. The sleep patterns between dogs and humans are quite different (dogs often waking multiple times per night, human usually having a single wake/sleep cycle), which could account for inter-species differences in nocturnal hormone patterns. Alternatively, the marked difference between diets in the aforementioned study (commercial diets predominantly low in fat and with moderate to high CHO and fiber vs. ultra-low CHO high fat in our study) combined with marked differences in physical conditioning (diabetic dogs vs. highly athletic farm working dogs), twice vs. once a day feeding, and timing of insulin administration is the essence of the difference in the timing of the low IG (as the majority of low IG reading in the Diet 1 group (high CHO low fat) in our study had also happened during the day).

This study had a few limitations. Firstly, the diets differed in respect to their micronutrient composition. However, both diets were formulated to be within the limits established by the Association of American Feed Control Officials (AAFCO), and within those limits, we are not aware of micronutrient compositions that would differ enough to affect glucose regulation. The diets would also be expected to differ in respects to digestibility and intestinal transit time, in their effects on the microflora and bacteria-derived metabolites, feelings of satiety, and potentially other processes. However, the authors posit that the paramount difference that is responsible for the differences in IG concentrations seen in these dogs is the difference in the macronutrient composition, predominantly in CHO and fat composition. A future study could reduce nutritional confounders by formulating the diets using varying proportions of the same ingredients, such that only the macronutrient proportions differ. A second limitation is that we were unable to separate a possible concurrent postprandial effect from that of resting/sleeping (between 19:00–06:00) on IG, as the dogs in this study were fed in the evening after work. Hence, the higher overall incidence of low IG between 19:00–06:00 might in part be related to a postprandial effect in addition to the inherent decreased autonomic function during the sleeping period as explained above. The length of the postprandial period in dogs has not been well characterized but in one study it was determined to be at least 6 hours [35]. In that study, dietary macronutrients composition (i.e., high CHO low fat vs. low CHO high fat) in dogs fed once daily did not significantly affect the AUC of glucose and insulin during the 6 h postprandial period, possibly implying that this is also the case in our study. However, direct comparison between that study and ours is difficult because the macronutrient compositions of diets between the two studies differed. The last limitation in the study was the unbalanced distribution of IG reading between the Diet 2 group (2306) and the Diet 1 group (1504). We contend that it is unlikely to have affected our interpretation of the results because the distribution of IG readings for each diet was relatively uniform across the 24 h and because the statistical analysis was based on a large number of observations in each group.

Conclusion

Feeding ultra-low CHO high fat diet (Diet 2) to very active dogs was associated with decreased IG and increased activity. Interstitial glucose concentrations were positively associated with the level of activity of dogs independently of the diet. Events of low IG occurred at an incidence of 3.12% per 96 h and were predominantly seen in dogs fed ultra-low CHO high fat diet (Diet 2), occurring mostly between 19:00–06:00 following the day’s work. Further work is necessary to determine if ultra-low CHO high fat diet could significantly affect work performances in New Zealand farm working dogs.

Supporting information

S1 Fig. Distribution of low activity quartiles from three 96-hour periods of 22 New Zealand working dogs.

The period between the two broken red vertical lines denotes the time the dogs were awake and working. Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S2 Fig. Distribution of IG and activity over the elapsed 96-hour periods.

Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet.

(TIF)

Click here for additional data file.^{(16.4MB, tif)}

S3 Fig. Distribution of interstitial glucose readings of 22 New Zealand working dogs over three 96-hour periods.

Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet; IG, interstitial glucose; L-IG, low interstitial glucose (<3.5 mmol/L); N-IG, normal interstitial glucose (3.5 mmol– 5.6 mmol/L). The horizontal dashed blue line separates the L-IG (below) from the N-IG (above). The vertical broken black lines denote the period between 19:00–06:00.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S4 Fig. Distribution of low IG readings (<3.5 mmol/L) of 22 New Zealand working dogs over three 96-hour periods.

Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet. The broken red lines denote the period between 19:00–06:00.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S1 Table. Nutritional analysis of trial diets.

(DOCX)

Click here for additional data file.^{(13.8KB, docx)}

S2 Table. Ingredients of trial diets.

(DOCX)

Click here for additional data file.^{(13.6KB, docx)}

Data Availability

All relevant data are used for the statistical analyses and for the tables and figures are available within the following stable data repository (doi: 10.17632/d76znr6jvk.1).

Funding Statement

AG received funding for this study from the Working Dog Centre, Massey University, New Zealand (http://workingdogs.massey.ac.nz/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1.Dalton C. Farm dogs: Te Ara—the Encyclopedia of New Zealand; 2013. https://teara.govt.nz/en/farm-dogs.
2.Kronfeld DS, Hammel EP, Ramberg CF Jr., Dunlap HL Jr.. Hematological and metabolic responses to training in racing sled dogs fed diets containing medium, low, or zero carbohydrate. Am J Clin Nutr. 1977;30(3):419–30. doi: 10.1093/ajcn/30.3.419 . [DOI] [PubMed] [Google Scholar]
3.Roberts TJ, Weber JM, Hoppeler H, Weibel ER, Taylor CR. Design of the oxygen and substrate pathways. II. Defining the upper limits of carbohydrate and fat oxidation. J Exp Biol. 1996;199(Pt 8):1651–8. . [DOI] [PubMed] [Google Scholar]
4.Felig P, Marliss E, Owen OE, Cahill GF Jr. Blood glucose and gluconeogenesis in fasting man. Arch Intern Med. 1969;123(3):293–8. Epub 1969/03/01. . [PubMed] [Google Scholar]
5.Wasserman DH. Regulation of glucose fluxes during exercise in the postabsorptive state. Annu Rev Physiol. 1995;57:191–218. Epub 1995/01/01. doi: 10.1146/annurev.ph.57.030195.001203 . [DOI] [PubMed] [Google Scholar]
6.Ahmed M, Gannon MC, Nuttall FQ. Postprandial plasma glucose, insulin, glucagon and triglyceride responses to a standard diet in normal subjects. Diabetologia. 1976;12(1):61–7. Epub 1976/03/01. doi: 10.1007/BF01221966 . [DOI] [PubMed] [Google Scholar]
7.Rutherford WJ. Hypoglycemia and endurance exercise: dietary considerations. Nutr Health. 1990;6(4):173–81. Epub 1990/01/01. doi: 10.1177/026010609000600402 . [DOI] [PubMed] [Google Scholar]
8.Schnurr TM, Reynolds AJ, Komac AM, Duffy LK, Dunlap KL. The effect of acute exercise on GLUT4 levels in peripheral blood mononuclear cells of sled dogs. Biochem Biophys Rep. 2015;2:45–9. Epub 2015/09/05. doi: 10.1016/j.bbrep.2015.05.002 . [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Pratt-Phillips SE, Olsen R, Geor R, Zirkle A, Moore A, Harkins C, et al. Effect of reduced protein intake on endurance performance and water turnover during low intensity long duration exercise in Alaskan sled dogs. Comp Exerc physiol. 2018;14(1):19–26. doi: 10.3920/CEP170024 [DOI] [Google Scholar]
10.Bermingham EN, Thomas DG, Cave NJ, Morris PJ, Butterwick RF, German AJ. Energy requirements of adult dogs: a meta-analysis. PLoS One. 2014;9(10):e109681. Epub 2014/10/15. doi: 10.1371/journal.pone.0109681 . [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Corradini S, Pilosio B, Dondi F, Linari G, Testa S, Brugnoli F, et al. Accuracy of a flash glucose monitoring system in diabetic dogs. J Vet Intern Med. 2016;30(4):983–8. Epub 2016/06/20. doi: 10.1111/jvim.14355 . [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Silva DD, Cecci GRM, Biz G, Chiaro FN, Zanutto MS. Evaluation of a flash glucose monitoring system in dogs with diabetic ketoacidosis. Domest Anim Endocrinol. 2021;74:106525. Epub 2020/08/18. doi: 10.1016/j.domaniend.2020.106525 . [DOI] [PubMed] [Google Scholar]
13.Malerba E, Cattani C, Del Baldo F, Carotenuto G, Corradini S, Golinelli S, et al. Accuracy of a flash glucose monitoring system in dogs with diabetic ketoacidosis. J Vet Intern Med. 2020;34(1):83–91. Epub 2019/11/15. doi: 10.1111/jvim.15657 . [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Del Baldo F, Canton C, Testa S, Swales H, Drudi I, Golinelli S, et al. Comparison between a flash glucose monitoring system and a portable blood glucose meter for monitoring dogs with diabetes mellitus. J Vet Intern Med. 2020;34(6):2296–305. Epub 2020/10/31. doi: 10.1111/jvim.15930 . [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Albright JD, Seddighi RM, Ng Z, Sun X, Rezac DJ. Effect of environmental noise and music on dexmedetomidine-induced sedation in dogs. PeerJ. 2017;5:e3659. Epub 2017/08/09. doi: 10.7717/peerj.3659 . [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Lee AH, Detweiler KB, Harper TA, Knap KE, de Godoy MRC, Swanson KS. Physical activity patterns of free living dogs diagnosed with osteoarthritis. J Anim Sci. 2021. Epub 2021/07/04. doi: 10.1093/jas/skab204 . [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Mejia S, Duerr FM, Salman M. Comparison of activity levels derived from two accelerometers in dogs with osteoarthritis: Implications for clinical trials. Vet J. 2019;252:105355. Epub 2019/09/27. doi: 10.1016/j.tvjl.2019.105355 . [DOI] [PubMed] [Google Scholar]
18.Kaneko JJ. Carbohydrate metabolism and its diseases. In: Kaneko JJ, Harvey JW, Bruss ML, editors. Clinical biochemistry of domestic animals. Sixth ed. San Diego, CA, USA: Academic Press; 2008. p. 45–80. [Google Scholar]
19.Schnurr TM, Reynolds AJ, Gustafson SJ, Duffy LK, Dunlap KL. Conditioning causes an increase in glucose transporter-4 levels in mononuclear cells in sled dogs. Int J Biochem Cell Biol. 2014;55:227–31. doi: 10.1016/j.biocel.2014.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Serisier S, Leray V, Poudroux W, Magot T, Ouguerram K, Nguyen P. Effects of green tea on insulin sensitivity, lipid profile and expression of PPARalpha and PPARgamma and their target genes in obese dogs. Br J Nutr. 2008;99(6):1208–16. Epub 2007/12/07. doi: 10.1017/S0007114507862386 . [DOI] [PubMed] [Google Scholar]
21.Littell RC, Henry PR, Ammerman CB. Statistical analysis of repeated measures data using SAS procedures. J Anim Sci. 1998;76(4):1216–31. Epub 1998/05/15. doi: 10.2527/1998.7641216x . [DOI] [PubMed] [Google Scholar]
22.Gilmour AR, Gogel BJ, Cullis BR, Thompson R. ASReml user guide release 3.0 Hemel Hempstead, HP1 1ES, UK: VSN International Ltd; 2009.
23.White IM, Thompson R, Brotherstone S. Genetic and environmental smoothing of lactation curves with cubic splines. J Dairy Sci. 1999;82(3):632–8. Epub 1999/04/09. doi: 10.3168/jds.S0022-0302(99)75277-X . [DOI] [PubMed] [Google Scholar]
24.Romsos DR, Palmer HJ, Muiruri KL, Bennink MR. Influence of a low carbohydrate diet on performance of pregnant and lactating dogs. J Nutr. 1981;111(4):678–89. Epub 1981/04/01. doi: 10.1093/jn/111.4.678 . [DOI] [PubMed] [Google Scholar]
25.Murray AJ, Knight NS, Cole MA, Cochlin LE, Carter E, Tchabanenko K, et al. Novel ketone diet enhances physical and cognitive performance. FASEB J. 2016;30(12):4021–32. Epub 2016/08/17. doi: 10.1096/fj.201600773R . [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Volek JS, Freidenreich DJ, Saenz C, Kunces LJ, Creighton BC, Bartley JM, et al. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism. 2016;65(3):100–10. Epub 2016/02/20. doi: 10.1016/j.metabol.2015.10.028 . [DOI] [PubMed] [Google Scholar]
27.Hammel EP, Kronfeld DS, Ganjam VK, Dunlap HL Jr. Metabolic responses to exhaustive exercise in racing sled dogs fed diets containing medium, low, or zero carbohydrate. Am J Clin Nutr. 1977;30(3):409–18. Epub 1977/03/01. doi: 10.1093/ajcn/30.3.409 . [DOI] [PubMed] [Google Scholar]
28.Pencek RR, Fueger PT, Camacho RC, Wasserman DH. Mobilization of glucose from the liver during exercise and replenishment afterward. Can J Appl Physiol-Rev Can Physiol Appl. 2005;30(3):292–303. doi: 10.1139/h05-122 [DOI] [PubMed] [Google Scholar]
29.Richter EA, Hargreaves M. Exercise, GLUT4, and skeletal muscle glucose uptake. Physiol Rev. 2013;93(3):993–1017. doi: 10.1152/physrev.00038.2012 . [DOI] [PubMed] [Google Scholar]
30.Jauch-Chara K, Hallschmid M, Gais S, Oltmanns KM, Peters A, Born J, et al. Awakening and counterregulatory response to hypoglycemia during early and late sleep. Diabetes. 2007;56(7):1938–42. doi: 10.2337/db07-0044 [DOI] [PubMed] [Google Scholar]
31.Jauch-Chara K, Hallschmid M, Schmid SM, Oltmanns KM, Peters A, Born J, et al. Plasma glucagon decreases during night-time sleep in Type 1 diabetic patients and healthy control subjects. Diabet Med. 2007;24(6):684–7. Epub 2007/03/27. doi: 10.1111/j.1464-5491.2007.02116.x . [DOI] [PubMed] [Google Scholar]
32.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F, et al. Decreased epinephrine responses to hypoglycemia during sleep. N Engl J Med. 1998;338(23):1657–62. doi: 10.1056/NEJM199806043382303 [DOI] [PubMed] [Google Scholar]
33.Schultes B, Jauch-Chara K, Gais S, Hallschmid M, Reiprich E, Kern W, et al. Defective awakening response to nocturnal hypoglycemia in patients with Type 1 Diabetes Mellitus. PLoS Med. 2007;4(2):e69. doi: 10.1371/journal.pmed.0040069 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Shea EK, Hess RS. Assessment of postprandial hyperglycemia and circadian fluctuation of glucose concentrations in diabetic dogs using a flash glucose monitoring system. J Vet Intern Med. 2021;35(2):843–52. Epub 2021/02/02. doi: 10.1111/jvim.16046 . [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Schauf S, Salas-Mani A, Torre C, Jimenez E, Latorre MA, Castrillo C. Effect of feeding a high-carbohydrate or a high-fat diet on subsequent food intake and blood concentration of satiety-related hormones in dogs. J Anim Physiol Anim Nutr (Berl). 2018;102(1):e21–e9. Epub 2017/04/28. doi: 10.1111/jpn.12696 . [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0261506.r001

Decision Letter 0

Balamuralikrishnan Balasubramanian

14 Jun 2021

PONE-D-21-02404

Less is more? Ultra-low carbohydrate diet and working dogs’ performance

PLOS ONE

Dear Dr. Gal,

Please submit your revised manuscript by Jul 29 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Balamuralikrishnan Balasubramanian

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Methods section, please provide additional details regarding participant consent from the owners of the animals. In the ethics statement in the Methods and online submission information, please ensure that you have specified (a) whether consent was informed and (b) what type you obtained (for instance, written or verbal). If the need for consent was waived by the ethics committee, please include this information.

3. Thank you for stating the following in the Competing Interests section:

"The authors have declared that no competing interests exist."

We note that one or more of the authors are employed by a commercial company: "VetEnt Te Kuiti, Pet Doctors and North Coast Veterinary Specialist and Referral Centre"

a) Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

b) Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests) . If this adherence statement is not accurate and there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: First, I would like to thank the author for their interesting manuscript. The work described by the authors sheds more light on the effects of macronutrient composition in dogs performing strenuous exercise. I have a few general comments for the authors, followed by some specific remarks.

- The manuscript is quite long, and some paragraphs have a lot of repetition in them. I think the readability of the manuscript may be improved by optimizing conciseness throughout the work.

- Please double check the consistency of your abbreviation use throughout the manuscript. After introducing an abbreviation, you sometimes switch back to the fully written alternative.

- The discussion section may benefit from some editing. I think it is important to focus foremost on the primary goals of this study, i.e. to determine whether IG concentrations and physical activity differ between diets, before focusing on your secondary findings, such as the moments of low IG that you found. You did this well in the abstract, so I would consider using the same order in your discussion.

- Additionally, there is a lot of focus on the cause of your findings and comparisons with previous studies, but it is also important to emphasize the consequences of your findings: what do these findings mean for working dogs? Do we all need to change their diet, or do we need to perform further specific studies? Can we extrapolate these results to all other working dogs?

- The conclusions section may also include some emphasis on why this work was important for the population you researched.

Abstract

- I would suggest changing the order of the abstract, by first stating the goals of the study, before stating your hypothesis.

25- ‘.. their body glucose levels that in turn..’Consider: their body glucose levels which in turn

25- ‘a negative effect’

Could you specify this? Maybe consider: which in turn will reduce physical activity during work

33- I think some editing may be required, consider: ‘Following the acclimation period, we continuously monitored IG concentrations with flash glucose monitoring devices, and activity using triaxial accelerometers for 96 h.

34- Dogs fed the Diet 2

I would rephrase as: dogs fed Diet 2

37- ‘Interstitial glucose’

Consider replacing with IG

39- Please specify your proportions: 119/3810 time points?

42- ‘.. was associated with increased activity, and decreased IG.’

Consider: was associated with increased activity and decreased IG concentrations.

42- ‘Interstitial glucose’ -> IG

43- However, you did not check for specific macronutrient, but only for the two diets that you used. I would consider changing this to: independently of the two diets.

Introduction

53-56 – Do you know why these diets are commonly chosen? Would you be able to give some more background for this?

65-68 – ‘Continuous absorption of glucose…’ But this does not solely depend on dietary CHO content, as the gastrointestinal passage time and interactions with other nutrient have a great effect on this.

72-74 I would suggest changing the order: This study hypothesized that dogs fed an ultra %ME CHO diet will have lower body glucose levels, when compared to dogs fed a high %ME CHO diet’, to emphasize on the ultra-low diet.

Methods

86 – did you use any block randomization to ensure that all farms had an equal part of dogs in either diet 1 or 2?

87: I would omit ‘(see description below)’

I would suggest changing the order in sentence 86-92: ‘Three 96- hour study periods took place during times of peak seasonal work activity, during which the dogs in each farm worked together as a team.’ I would include the dietary acclimation in the ‘diet’-section.

92- ‘body weight’

suggestion: bodyweight

94-96 is there any information on the validation of both methods available?

96-97 – Serum glucose and insulin…

You repeat this information in the part on the insulin and glucose assays, I would recommend omitting this information in line 96-97.

99 – this information is already stated before, I would delete one of the passages. You could, for instance, add the information on the tables earlier to line 86.

‘We recruited 22 dogs from four farms in the ….. randomizer.org) (Table 1 and 2).

This way, you can avoid this information.

101- Were any blood examinations, such as biochemistry or hematology, performed to assess for subclinical disease?

106- You could delete this sentence, as you state this information more extensively below.

116-122: I would state this in the discussion section, and not here. It would have been interesting to see whether the energy intake required to maintain bodyweight is different between both diets, but this would be a subject for further study.

141-173 Suggestion: you could consider combining some paragraphs here, to avoid switching between insulin and glucose, activity levels and back to glucose and the HOMA-IR. You could combine the information of the assays, the IG glucose measurements and the HOMA-IR, and describe the triaxial accelerometry after this.

175- Please provide the full name of HOMA-IR also in the text before using the abbreviation.

192- How did you check for normal distribution? Could you provide any information about whether the assumptions of all tests where met?

Was any power analysis performed for this study?

Results:

222-223 You also state this in your methods section, and could consider omitting it here.

The results section itself may benefit from some editing to improve conciseness. For instance, you could consider removing sentence 282, as you state this in the section heading.

It is important to consider differences between the farms: did you consider including the factor ‘farm’ in your models?

Discussion:

I think it is also beneficial for this report to consider the effects of the other macronutrients of the diets, which also differed widely. The diets also differed widely in dietary fat and protein content, two factors which have also been found to affect glucose tolerance and insulin resistance. It is therefore a possibility that the effects that you found are not due to a lack of carbohydrate content, but maybe an increase in dietary fat content.

Also, differences in housing/environment of the dogs on the different farms may have also affected your results. As an example: maybe the dogs of one of the farms were housed in colder conditions, and had a higher energy requirement (and where therefore more likely to be hypoglycemic?)

Tables and figures:

- Table 1: I think this table may be improved by swapping the ‘Farm’ and ‘variable’ columns. Additionally, I would suggest splitting these tables, as the last two rows do not fit the table headings.

- In table S1, the nutritional analysis of diet 2 is missing. Also, is there any information about the specific nutrient values of both diets (for instance, omega-3 fatty acid concentrations, vitamins and minerals)? I think it is important to provide this information, considering omega-3 fatty acids and omega 6:3 ratios have been found to alter insulin sensitivity, and therefore may affect your results.

Thank you in advance for your response.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: N. R. Blees

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Dec 23;16(12):e0261506. doi: 10.1371/journal.pone.0261506.r002

Author response to Decision Letter 0

10 Aug 2021

The authors thank the Editor and the Reviewer for the time they invested in reviewing the manuscript and providing their valuable comments and suggestions for improving the manuscript and making it more impactful. Below we provide a point-by-point responses to each of the comments made by the Editor and Reviewer.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

AUTHORS’ response: Files were named according to journal’s requirements.

AUTHORS’ response: We added the following sentence in the methods section (L86-87): “Dogs were recruited after receipt of owner’s verbal consent to participate in the study.”

3. Thank you for stating the following in the Competing Interests section:

"The authors have declared that no competing interests exist."

We note that one or more of the authors are employed by a commercial company: "VetEnt Te Kuiti, Pet Doctors and North Coast Veterinary Specialist and Referral Centre"

Please also include the following statement within your amended Funding Statement.

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

AUTHORS’ response: Dr. Burchell was employed at Massey University during the time of the study. He left after the study was completed and has been working in a private referral center. His current employment had no impact on the study whatsoever. Like many private practitioners involved in research his involvement in the paper subsequent to leaving the university sector has been in his own time. Drs. Giles and Cuttance were not employed at Massey University at the time of the study. Their role in the study is articulated in the Authors’ Contributions sections and they did not have any conflict of interests with regards to the aforementioned study.

We added the requested sentence in L441-445 and in the online submission system.

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

AUTHORS’ response: We did not find anywhere in PLOS ONE Editorial system during resubmission a place where we could amend the Competing Interests Statement. In the ‘manuscript data tab’ under ‘Funding information’ there was no area to amend what the editor requested. The roles of each contributor accurately reflect their involvement at different aspects that were listed.

Reviewers' comments:

- The manuscript is quite long, and some paragraphs have a lot of repetition in them. I think the readability of the manuscript may be improved by optimizing conciseness throughout the work.

AUTHORS’ response: We followed the reviewer suggestions and trimmed the parts in which there was unnecessary repetition.

- Please double check the consistency of your abbreviation use throughout the manuscript. After introducing an abbreviation, you sometimes switch back to the fully written alternative.

AUTHORS’ response: We made the suggested change.

AUTHORS’ response: We restructured the discussion similar to the sequence in the abstract, as suggested by the reviewer.

AUTHORS’ response: We followed the reviewer suggestions (see below).

For the question “what do these findings mean for working dogs?”

L321-323: “As we suggested above, feeding dogs a diet of ultra-low %ME CHO and high %ME fat leads to adaptations that offer significant advantages to dogs that perform endurance work, such as New Zealand farm dogs”.

L327-329: “The adaptation to high fat diets has been postulated to be beneficial because the higher albumin, total calcium, and magnesium protects against exercise-induced hypovolemia and the adaptation to high rates of muscle fat oxidation protects against muscle injury”.

Do we all need to change their diet, or do we need to perform further specific studies? Can we extrapolate these results to all other working dogs?

L330-337: “The results of this study may not directly translate to other kinds of working dogs. For example, it is unlikely that racing dogs (e.g., Greyhounds) would benefit from ultra-low CHO high fat diet as their race-day performance relies on high levels of muscle glycogen for short anerobic bursts of high-speed running. Similarly, personal assistance dogs (e.g., seeing eye dogs) that are not expected to engage in high endurance activities might not necessary benefit from the advantages of fat adaptation like sled, military/police, search and rescue, and New Zealand farm dogs would. Maximizing the dietary benefits for each of the different fields of activity that dogs engage in our lives would require specific research tailored to the requirements of the individual field”.

- The conclusions section may also include some emphasis on why this work was important for the population you researched.

AUTHORS’ response: Identifying whether the increased activity that we found is translated to increased work performances in the field should really be the next step before making a recommendation to change the diet for all working dogs.

We added the following sentence (L438-439): “Further work is necessary to determine if ultra-low CHO high fat diet could significantly enhance work performances in New Zealand farm working dogs”.

Abstract

- I would suggest changing the order of the abstract, by first stating the goals of the study, before stating your hypothesis.

AUTHORS’ response: We made the suggested change (L24-28).

25- ‘.. their body glucose levels that in turn..’Consider: their body glucose levels which in turn

AUTHORS’ response: The sentence was changed as follow: “We hypothesized that feeding working farm dogs an ultra-low CHO diet would reduce their IG concentrations which in turn would reduce physical activity during work” (L27-28).

25- ‘a negative effect’

Could you specify this? Maybe consider: which in turn will reduce physical activity during work

AUTHORS’ response: We made the suggested change (L28).

AUTHORS’ response: We made the suggested change (L32-34).

34- Dogs fed the Diet 2

I would rephrase as: dogs fed Diet 2

AUTHORS’ response: We made the suggested change (L34).

37- ‘Interstitial glucose’

Consider replacing with IG

AUTHORS’ response: We made the suggested change (L37).

39- Please specify your proportions: 119/3810 time points?

AUTHORS’ response: To explain what the proportion relates to we added the word ‘readings’ (L39): “The overall incidence of low IG readings (< 3.5 mmol/L) was 119/3810 (3.12%)…”

42- ‘.. was associated with increased activity, and decreased IG.’

Consider: was associated with increased activity and decreased IG concentrations.

AUTHORS’ response: We changed the sentence as follow: “We conclude that feeding Diet 2 (ultra-low CHO high fat diet) to working farm dogs was associated with increased activity despite decreased IG concentrations” (L41-42).

42- ‘Interstitial glucose’ -> IG

AUTHORS’ response: We made the suggested change (L42).

43- However, you did not check for specific macronutrient, but only for the two diets that you used. I would consider changing this to: independently of the two diets.

AUTHORS’ response: We made the suggested change (L43).

Introduction

53-56 – Do you know why these diets are commonly chosen? Would you be able to give some more background for this?

AUTHORS’ response: The answer is a combination of practical convenience, and success of a single dominant brand. Dogs were historically fed meat and offal from livestock killed on farm (“home-kill”). When home-kill was fed alone, it constituted a very low to zero carbohydrate diet.

For decades, the most popular commercial food was Tux® (Nestle-Purina), originally about 45% of ME as CHO, then more recently 35% ME (Tux Energy®).

During that time, farmers have varied the amount of home-kill and Tux fed, which ranged from 100% home-kill, through 50:50, to 100% Tux (Singh et al).

Singh, I., Tucker, L. A., Gendall, P., Rutherfurd-Markwick, K. J., Cline, J., & Thomas, D. G. (2011). Age, breed, sex distribution and nutrition of a population of working farm dogs in New Zealand: results of a cross-sectional study of members of the New Zealand Sheep Dog Trial Association. New Zealand Veterinary Journal, 59(3), 133-138. 10.1080/00480169.2011.567967

AUTHORS’ response: The reviewer is correct. What we meant to say was that higher dietary glucose would translate to an overall increased absorption of glucose from the GI tract compared with diets low in CHO.

We changed the sentence (L65-68) to “Therefore, feeding diets with high CHO content to dogs that are undergoing strenuous and prolonged physical exercise might support the maintenance of glucose levels for a longer time through maximizing muscle and hepatic glycogen stores and glucose absorption from the gastrointestinal tract”.

AUTHORS’ response: The sentence was revised as follow (72-74): “This study tested the hypothesis that dogs fed an ultra-low CHO high fat diet will have lower body glucose levels than dogs fed a high CHO low fat diet, evidenced by lower IG levels and reduced physical activity (Fig. 1)”

Methods

86 – did you use any block randomization to ensure that all farms had an equal part of dogs in either diet 1 or 2?

AUTHORS’ response: Yes. The exact allocation of number of dog/diet/farm is evident in Table 2 under Farms. We did our best to balance the dogs between diets within farms however when you’re dealing with a small uneven number of dogs/farm it is somewhat challenging…

87: I would omit ‘(see description below)’

AUTHORS’ response: We omitted ‘(see description below)’ and changed the sentence as follow: “Acclimation of the dogs to the two diets started one month prior to the commencement of each of the three 96 h study periods, which took place during times of peak seasonal work activity, during which the dogs in each farm worked together as a team.” (L89-92)

If we had completely removed the acclimation part of the sentence, then the sentence below that explains why the study spun between Oct-Nov becomes out-of-context. We hope the reviewer finds it suitable.

92- ‘body weight’

suggestion: bodyweight

AUTHORS’ response: We made the suggested change (L95).

94-96 is there any information on the validation of both methods available?

AUTHORS’ response: Yes.

For the Flash Glucose Monitoring system:

• Corradini S, Pilosio B, Dondi F, Linari G, Testa S, Brugnoli F, Gianella P, Pietra M, Fracassi F. Accuracy of a Flash Glucose Monitoring System in Diabetic Dogs. J Vet Intern Med 2016;30:983-988.

• Del Baldo F, Canton C, Testa S, Swales H, Drudi I, Golinelli S, Fracassi F. Comparison between a flash glucose monitoring system and a portable blood glucose meter for monitoring dogs with diabetes mellitus. J Vet Intern Med 2020;34:2296-2305.

• Malerba E, Cattani C, Del Baldo F, Carotenuto G, Corradini S, Golinelli S, Drudi I, Fracassi F. Accuracy of a flash glucose monitoring system in dogs with diabetic ketoacidosis. J Vet Intern Med 2020;34:83-91.

• Silva DD, Cecci GRM, Biz G, Chiaro FN, Zanutto MS. Evaluation of a flash glucose monitoring system in dogs with diabetic ketoacidosis. Domest Anim Endocrinol 2021;74:106525.

For the Heyrex accelerometer:

• Albright JD, Seddighi RM, Ng Z, Sun X, Rezac DJ. Effect of environmental noise and music on dexmedetomidine-induced sedation in dogs. PeerJ 2017;5:e3659.

• Lee AH, Detweiler KB, Harper TA, Knap KE, de Godoy MRC, Swanson KS. Physical Activity Patterns of Free Living Dogs Diagnosed with Osteoarthritis. J Anim Sci 2021.

• Mejia S, Duerr FM, Salman M. Comparison of activity levels derived from two accelerometers in dogs with osteoarthritis: Implications for clinical trials. Vet J 2019;252:105355.

96-97 – Serum glucose and insulin…

You repeat this information in the part on the insulin and glucose assays, I would recommend omitting this information in line 96-97.

AUTHORS’ response: We made the suggested change.

99 – this information is already stated before, I would delete one of the passages. You could, for instance, add the information on the tables earlier to line 86.

‘We recruited 22 dogs from four farms in the ….. randomizer.org) (Table 1 and 2).

This way, you can avoid this information.

AUTHORS’ response: We made the suggested change.

“We recruited 22 dogs from four farms in the North Island of New Zealand and randomized the dogs into two balanced groups with respect to diet (n=11 each) using an online randomization tool (www.randomizer.org) (Tables 1 and 2). Dogs were recruited after receipt of each owner’s verbal consent to participate in the study. Dogs were only included in the study if they had been assessed by a physical exam by at least one of the authors and by the history provided by the owner and were deemed to be in good health” (L84-89).

101- Were any blood examinations, such as biochemistry or hematology, performed to assess for subclinical disease?

AUTHORS’ response: The short answer is no. The longer explanation is that dogs that are involved in high level of work such as these dogs cannot perform their work if they have underlying conditions and the owners quickly see that there is alteration in their performances. Also, it was not in our budget to prescreen with intensive bloodwork and additional imaging which would have been required in order to have confidence that the dogs do not have an underlying condition. We qualify our observations and wrote that the dogs were ‘deemed’ in good health and ideal body condition following physical exam and owner’s history.

106- You could delete this sentence, as you state this information more extensively below.

AUTHORS’ response: We made the suggested change.

AUTHORS’ response: We prefer to keep these sentences in the Diets paragraph within the Methods section (unless the reviewer would strongly object) for the reason that the reader immediately understands the rationale for why we chose to do what we did.

AUTHORS’ response: We combined the insulin and HOMA-IR paragraphs together and kept the FGM and triaxial accelerometry in one paragraph. We think that it keeps the flow better and the structure more coherent.

175- Please provide the full name of HOMA-IR also in the text before using the abbreviation.

AUTHORS’ response: We made the suggested change (L171)

192- How did you check for normal distribution? Could you provide any information about whether the assumptions of all tests where met?

Was any power analysis performed for this study?

AUTHORS’ response: We added the following sentence (L196-197) indicating how we assessed for normal distribution: “The data were examined for normal distribution by inspection of Q-Q plots, histogram, and by the Shapiro-Wilk test.”

Presence or absence of normal distribution was a criterion that determined whether ANOVA (normally distributed data) or nonparametric (nonnormally distributed data) tests had been used (indicated in L207-210).

We also added the following paragraph (L189-193): “A priori power analysis (G*Power software version 3.1.9.2) indicated that 9 dogs in each group will be sufficient to detect a difference between two dependent IG means of 100 mg/dL and 70 mg/dL in an individual dog with SD of 30 mg/dL, power of 0.8, alpha error probability of 0.05, and a correlation coefficient of 0.6 between measurements. To account for potential losses of dogs during the study period (i.e., dislodgement of the sensor), we recruited 11 dogs per group.”

The modelling of IG and activity using ASReml do not require the traditional assumptions to perform analysis of variance using the least-squares method (normality, homogenous variance, independent residual errors and additive effects), because the solutions of the model were obtained using restricted maximum likelihood procedures, which do not require the homogeneity of variance assumption; and the correlations among random residuals were accounted for in the modelling of observations for each dog using splines.

Results:

222-223 You also state this in your methods section, and could consider omitting it here.

AUTHORS’ response: We made the suggested change.

The results section itself may benefit from some editing to improve conciseness.

AUTHORS’ response: We trimmed a few sentences in different paragraphs of the results that were repetitive or redundant.

For instance, you could consider removing sentence 282, as you state this in the section heading.

AUTHORS’ response: We made the suggested change.

It is important to consider differences between the farms: did you consider including the factor ‘farm’ in your models?

AUTHORS’ response: We tested for a ‘farm’ effect when we analyzed the data and since ‘farm’ did not significantly affect the model it was omitted.

Discussion:

AUTHORS’ response: The reviewer is correct. Throughout the revised manuscript we referred to Diet 2 as “ultra-low CHO high fat” and Diet 1 as “high CHO low fat” as they are linked together and decreasing either CHO or fat in the diet leads to reciprocal increase in the other component. While the diets also differed in %ME protein we do not think that this difference [23% (Diet 1) vs 37% (Diet 2)] had the same effect as the marked differences in CHO and fat. In the discussion we also ascribe the dietary effect that we found in our study to the process of fat adaptation.

AUTHORS’ response: The four farms in the study were from the Waikato region in New Zealand. Housing conditions of farm dogs in New Zealand are quite standard (outdoors kennel). Therefore, we think that it is very unlikely that the dogs in the different farms were subjected to very different terrains or ambient temperatures. Also, as we knew that we could not completely control for an identical workload between farms (although we think the workloads were quite similar given we aimed the 96-hour periods to occur at peak seasonal activity) we made sure to balance the diet within the farms. Hence, despite the inherent minimal differences between farms in ambient temperature, terrain, and workload we think that the major effects on IG and activity are foremost due to the diet.

Tables and figures:

AUTHORS’ response: We made the suggested change.

- In table S1, the nutritional analysis of diet 2 is missing.

AUTHORS’ response: We added the nutritional analysis of diet 2 that was accidently deleted.

Also, is there any information about the specific nutrient values of both diets (for instance, omega-3 fatty acid concentrations, vitamins and minerals)? I think it is important to provide this information, considering omega-3 fatty acids and omega 6:3 ratios have been found to alter insulin sensitivity, and therefore may affect your results.

AUTHORS’ response: A full nutrient profile for commercial diets is $2,500 and beyond the budget of this trial, thus unfortunately, we don’t have the full nutritional analysis of each diet. That is proprietary information, and we did not have access to that, beyond what information was included in the tables for Tux (Diet 1). A little more was available for the K9 Natural product (Diet 2), but not for both diets.

Within the scope of AAFCO requirements, we are not aware that insulin sensitivity will be affected by micronutrient concentrations in a manner that exceeds the effect from the massive difference in CHO content.

The effect of n-3 PUFA has been shown in diabetic humans. The effects of n-3 PUFA in healthy dogs are less clear. One study showed that enrichment of the diet with n-3 PUFA at 0.2% (as fed basis) had no effect on insulin sensitivity (Irvine et al 2002), and another showed no relationship between serum n-3 PUFA concentrations in dogs and serum insulin (when the dogs were eating various unreported diets), whilst obesity did have a significant effect.

The ultra-low CHO diet (Diet 1) contains a fish oil (Hoki oil), and the n-3 PUFA content is 0.28% (as fed basis), and thus is almost exactly the amount tested by Irvine et al (2002). The dry diet contains a small amount of the n-3 alpha linolenic acid, but no significant amount of long chain n-3 PUFA, and thus would be similar to the control diet in the study by Irvine et al.

We think that there is no reason to suspect that the different PUFA contents of the diets have a significant effect on insulin sensitivity. Compared with the effect of the difference in macronutrients (i.e. CHO, fat, and protein), any effect from micronutrient differences is argued to be small.

Streeter RM, Struble AM, Mann S, Nydam DV, Bauer JE, Castelhano MG, Todhunter RJ, Cummings BP, Wakshlag JJ. The associations between serum adiponectin, leptin, C-reactive protein, insulin, and serum long-chain omega-3 fatty acids in Labrador Retrievers. Vet Med (Auckl). 2015 Apr 8;6:103-110.

Irvine AJ, Butterwick R, Watson T, Millward DJ, Morgan LM. Dietary supplementation with (n-3) polyunsaturated fatty acids does not affect insulin sensitivity in healthy Labrador retriever dogs. J Nutr. 2002 Jun;132(6 Suppl 2):1709S-10S

Thank you in advance for your response.

AUTHORS’ response: Thank you for providing the feedback required to improve and better our manuscript.

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(36.7KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0261506.r003

Decision Letter 1

Balamuralikrishnan Balasubramanian

19 Oct 2021

PONE-D-21-02404R1Less is more? Ultra-low carbohydrate diet and working dogs’ performancePLOS ONE

Dear Dr. Gal,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please submit your revised manuscript by Dec 03 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Balamuralikrishnan Balasubramanian

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments (if provided):

Amend the Reviewer 2 Comments.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: (No Response)

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Reviewer #1: All comments are thoroughly addressed, thank you to the authors for their revised version of the manuscript.

Reviewer #2: Overall, a very interesting manuscript. Additional work investigating the effect of nutrition on active, working dogs is much needed. My biggest concern is about statistical analysis of the data, which is outlined in the methods section below.

Abstract:

L29 – Sentence reads awkwardly. The dogs were randomized to a diet then fed for a month, but the sentence almost reads as if the randomization took a month.

L31 – Remove additional parenthesis around “Diet 1”

Introduction:

General comment – Introduction would benefit from an additional paragraph on why a low CHO diet is selected. The second paragraph of the introduction describes a need for CHO during intense exercise, but then why are low CHO diets used at all? In the discussion, the focus shifts to how low CHO, high fat diets may be advantageous, but this should be introduced in introduction.

L55 – Remove extra space in “(%ME )”

L56 – Suggest expanding on why this is an important area of study (it is more than it being unknown)

L72-76 – Suggest switching the hypothesis and study aims sentences to first describe study aims then hypothesis/expected outcome

Materials and Methods:

L85 – Please give additional information regarding how groups were randomized. Was the randomization balanced across farms, sex, and breed? It does not appear so based on Table 2. It may be simply due to the dogs available at each farm, but please state that here if that is the case.

L86 – Italicize “n”

L88 - Within what time period were the dogs assessed? Within weeks? A year?

L90 – Remove extra space between “which” and “took”

L94 – Suggest removing first person here and throughout - “Bodyweight was recorded at recruitment, then at the beginning and end of each of the 96 h periods.”

L97 – When was the accelerometer fitted? At recruitment? Was there a baseline activity collection period?

L126 – Serum tubes allowed to clot before what? Assuming this means allowed to clot before being put on ice since the samples were not processed on site, but please clarify.

L136 – Remove comma after “whereas”

L149-150 – Why no baseline measurement during month of acclimation?

L159-160 – Please include any relevant citations for established blood glucose levels

L168 – Please provide inter- and intra-assay variation and sensitivity for this assay

L171 – Suggest brief explanation of what HOMA-IR is. What is normal? What is expected in this population of dogs?

L174-177 – I do not see the purpose of this arbitrary distinction into activity levels, particularly with so few animals who were all highly active working dogs. This becomes a problem later on in the statistical analysis section when activity (a response variable) is used as a fixed effect for IG analysis.

L176 – Is the =< supposed to be ≤ or L198-199 – Were data transformations attempted?

L201-203 – several questions:

- Why is time not included as a fixed effect?

- Above, you mention activity and inactivity periods, but these are also not reflected in this analysis

- What about farm, sex, age, breed, etc.? If these were evaluated as fixed effects then dropped, that should still be documented here.

- Activity is a response variable, not a fixed effect. If you wanted to include “activity level” as a fixed effect, you would need baseline activity data by which to block the dogs by activity level prior to diet allocation & acclimation

L207-208 – Again, were any transformations attempted prior to analysis as a nonparametric variable?

L208 – Remove comma after “whereas”

L209 – Why GLM here but MIXED above? What was included in the GLM model?

L213 – Why not include time & diet x time interaction here?

L214 – Is “activity level” here referring to the “activity levels” described above (L174-177)? Why examine activity level as a dependent variable here but a fixed effect in the linear models above?

Results:

L221 – How was BW statistically analyzed? This is not described in methods. The description of BW effects is a bit clunky. Were diet and time fixed effects significant? The interaction? You jump straight to describing the individual diet x time interactions without presenting results of main effects.

L225 – Suggest authors use consistency with respect to P-value reporting – 2 decimals used here versus three in next line. Recommend confirming if journal has standard for reporting P-values.

L226 – With the standard errors reported here, these weights should not have a statistical difference. Something is not correct with this analysis.

L229 – What about time effect? Diet x time interaction?

L242 – Your hypothesis is that low CHO diets lead to low IG which in turn reduces activity. Now you are trying to attribute activity to IG levels! Again, activity is a response variable, not a fixed effect. Dogs’ activity is likely related to IG, but that is not what this study was designed to assess. This section should be removed.

L256 – Period of the day is not included in your statistical analysis (L201-203) as a fixed effect.

L256 – Is the incidence of low IG more relevant than time spent at low IG? How much total time was spent at low IG? The potential diet x time interaction is important.

L266 – Remove extra space between “Diet 2” and “group”

L269-270 – Would it be expected for the majority of low IG readings to occur in the activity period? Would be nice to mention whether or not this was expected.

L275 – Remove comma after “whereas”

L281 – Again, activity is not a fixed effect, and your hypothesis has the reverse – low IG reducing activity. This section should be removed.

L299 – Consider re-wording as there is no such thing as “inappropriate” serum insulin.

Discussion:

L304 – These results were presented above, the reference to the figure is not needed here.

L307 – Suggest removing “most”

L314 – Reword “people” as “humans” throughout

L321-322 – Where is it suggested that low fat/high CHO diet offers advantages? Your hypothesis is that lower CHO/high fat diet will reduce activity.

L323-326 – These two sentences are very repetitive. Suggest condensing into one sentence.

L334 – Reword to “necessarily”

L339-340 – Here you could discuss how activity could have an impact on IG and recommend additional research since your study was not designed to examine this effect.

L340 – Remove reference to figures from discussion

L347 – Suggest further discussion at the end of this paragraph as to the potential significance of this observed relationship.

L349-350 – Awkward wording, suggest re-working. Since the hypothesis is that low IG leads to reduced activity, suggest focusing on that. May discuss potential for higher activity to reduce IG as avenue for future research, but not really relevant here.

L351 – Expand on the potential limitation of the 96 h collection period. What time period would have been better?

L358 – Suggest “of this alternative”

L360-361 – Suggest removing “The incidence of low IG episodes was 3.12% per 96 h and involved 10/22 (45.5%) of the dogs” from the discussion since it has already been reported in results section.

L363 – Similar to comment in L299 above, please remove “inappropriate” and re-word.

L364 – Suggest removing “we conclude that”

L365 – Is the fact that healthy, active dogs can have low IG really a surprising conclusion/finding of this study? Suggest writing different sentence to link discussion between serum insulin/glucose and IG

L373 – What effect might feeding time/frequency have on IG?

L377 – How does the evening feeding time play into this? Suggest further discussion

L409 – Any suggestions for future investigation regarding IG monitoring? Potential future work could be worked in throughout discussion.

L418 – Suggest addition of how this limitation could be overcome, such as formulating the diets to be similar in every respect except for source of energy (CHO vs. fat)

Conclusion:

L435 – Spell out abbreviation when it begins a sentence.

Tables and Figures:

Fig. 1 – in caption, should read “an ultra-low CHO…” not “a”

Table 2 – Diet randomization is clearly not balanced across farms, particularly farm 3. If there is a reason for this, please include in methods section.

Table 3 – Recommend removal since activity should not be used as a fixed effect.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: N. R. Blees

Reviewer #2: No

PLoS One. 2021 Dec 23;16(12):e0261506. doi: 10.1371/journal.pone.0261506.r004

Author response to Decision Letter 1

1 Dec 2021

PONE-D-21-02404R: Less is more? Ultra-low carbohydrate diet and working dogs' performance

The authors thank the Editor and both Reviewers for the time they invested in reviewing the manuscript and providing their valuable comments and suggestions for improving the manuscript and making it more impactful. Below we provide a point-by-point responses to each of the comments made by the Editor and Reviewer 2.

REVIEWER 1

Authors’ response: Thank you!

REVIEWER 2

Authors' response: After we went through the rebuttal process, we think that the main problem that Reviewer 2 had with our statistical analysis stemmed from a problem with terminology on our end for which we apologize. Specifically, we think that the use of ‘activity’ and ‘activity level’ interchangeably confused Reviewer 2 with regards to what was used as a response variable and what was used as a fixed effect in our analyses. As we explain below, we therefore used ‘delta-g’ as a response variable (derived from the accelerometer recordings), and we stratified ‘delta-g’ to three levels (L/M/H) of activity to assess the effect of ‘activity level’, ‘diet’, and ‘time’ (across the 96-hour of data recording and split between ‘working period’ and ‘resting period’) and interactions between diet x time on IG.

Abstract:

L29 – Sentence reads awkwardly. The dogs were randomized to a diet then fed for a month, but the sentence almost reads as if the randomization took a month.

Authors' response: We changed the sentence as follows (L29): "At each farm, dogs were randomized to one of two diets and had a month of dietary acclimation to their allocated diet. "

L31 – Remove additional parenthesis around "Diet 1"

Authors' response: The parenthesis was removed.

Introduction:

Authors' response: There are two primary feeding practices for working dogs in New Zealand. The reasons why these are the two common practices are multifactorial and involves farmers' personal beliefs and financial concerns (commercially available diets are generally more expensive than home-kill meat-based homemade diets) (L53-56).

We added the following in L56: “Dogs do not require dietary CHO, and exercising sled dogs may have improved performance when fed a high-fat CHO-free diet [2]. However, the concept of “cross-over”, where muscle utilization of CHO for ATP production increases with increased exercise intensity, has been demonstrated in several different mammalian species, including dogs [3]. Since the intensity of activity engaged by working farm dogs in New Zealand is likely to be greater than the sustained moderate-intensity endurance activity of sled dogs, New Zealand’s working farm dogs may be unable to maintain normal body glucose levels when fed a low CHO diet.”

Reviewer 1 was already concerned about the overall length of the manuscript and we are reluctant to add additional information beyond this.

L55 – Remove extra space in "(%ME )"

Authors' response: we removed the extra space.

L56 – Suggest expanding on why this is an important area of study (it is more than it being unknown)

Authors' response: We added the following in L56: “Dogs do not require dietary CHO, and exercising sled dogs may have improved performance when fed a high-fat CHO-free diet [2]. However, the concept of “cross-over”, where muscle utilization of CHO for ATP production increases with increased exercise intensity, has been demonstrated in several different mammalian species, including dogs [3]. Since the intensity of activity engaged by working farm dogs in New Zealand is likely to be greater than the sustained moderate-intensity endurance activity of sled dogs, New Zealand’s working farm dogs may be unable to maintain normal body glucose levels when fed a low CHO diet.”

L72-76 – Suggest switching the hypothesis and study aims sentences to first describe study aims then hypothesis/expected outcome

Authors' response: We would prefer not to do so unless the Reviewer/Editor is adamant that we change the order. Our rationale is that the aims are derived from the hypothesis and not vice versa.

Materials and Methods:

Authors' response: The randomization was based on balancing the diets across farms and dog availability and not on sex, age, breed, or bodyweight because it was challenging to find farms and farmers that were willing to participate. Hence, we were unable to be too selective.

The sentence was changed as follows (L91): "We recruited 22 dogs from four farms on the North Island of New Zealand and randomized the dogs into two balanced groups with respect to diet (n=11 each) using an online randomization tool (www.randomizer.org); there was no attempt to balance the dogs per age, sex, breed, or bodyweight during randomization and recruitment (Tables 1 and 2)."

L86 – Italicize "n"

Authors' response: we italicized it.

L88 - Within what time period were the dogs assessed? Within weeks? A year?

Authors' response: One of the authors (WC) has been providing regular veterinary care to the farmers in this region. It is important to mention that WC gauged the interest of the farmers to participate in the study and enroll their dogs with respect to the high level of work those dogs were expected to do in the month to come. WC assessed the dogs on the farms of the farmers that indicated they were keen to participate and determined if the dogs were deemed appropriate to be enrolled in the study. Once WC decided to enroll the dogs, they were enrolled within days (i.e., the time it took us to ship the diets to the farmers in those farms). Then the acclimation period started and a month after that, the 96-hour of data recording commenced, which took place at the peak seasonal workload.

L90 – Remove extra space between "which" and "took"

Authors' response: we removed the extra space.

L94 – Suggest removing first person here and throughout - "Bodyweight was recorded at recruitment, then at the beginning and end of each of the 96 h periods."

Authors' response: We prefer to use an 'active voice' rather than a 'passive voice' as much as possible. This is a stylistic preference, and the Journal does not have a specific guideline regarding it. Unless the Reviewer is adamant about that, we prefer to leave it like that. The Editor may also have a preference to which we would defer. We think that there has been some shift in the style of scientific writing for the purposes of clarity and because English is a universal communication language, we want our writing to be easy to understand to a wide range of first and second language speakers.

L97 – When was the accelerometer fitted? At recruitment? Was there a baseline activity collection period?

Authors' response: We fitted BOTH the accelerometer and Freestyle Libre at the same time; it was at the end of the one-month dietary acclimation period and at the beginning of the 96-hour data recording period. The 96-hour data recording period was at the PEAK OF SEASON ACTIVITY (see our previous comment for L88). So, there was no baseline recording period. We fitted the dogs with the devices, collected the blood (baseline samples), and off they went to work.

L126 – Serum tubes allowed to clot before what? Assuming this means allowed to clot before being put on ice since the samples were not processed on site, but please clarify.

Authors’ response: We added the word ‘Then’ (L136) and we hope that it is now clear: “Whole blood in the serum tubes was allowed to clot for 10 minutes whereas the whole blood in the sodium fluoride tubes was inverted several times to allow for the anticoagulant to mix with the blood. Then, the samples were stored on ice as long as the veterinarian was in the field and shipped on ice overnight to the principal investigator."

L136 – Remove comma after "whereas"

Authors' response: We removed the comma.

L149-150 – Why no baseline measurement during month of acclimation?

Authors' response: There are two reasons for that. We simply did not have the resources and were working on a tight budget (we paid per the time the accelerometer recorded; also, each FSL sensor is quite expensive and has a maximum lifetime of two weeks, albeit it works for a much shorter time in most dogs). Also, it is one thing to ask the busy farmers to flash the readers over the sensors at least every 6 hours for 96 hours and a different thing to do so for a more extended period. Compliance, recruitment, and logistics were significant hurdles in this study. The second reason is that there was a big difference in the farm workload between the beginning of the acclimation period and the beginning of the 96-hour data recording period. We agree that it would have been interesting to have this data and make comparisons. But we do not think that not having it changed our ability to answer our study question.

L159-160 – Please include any relevant citations for established blood glucose levels

Authors' response: We inserted a citation from a veterinary clinical biochemistry book (L168) that has a population-based reference interval for canine blood glucose.

L168 – Please provide inter- and intra-assay variation and sensitivity for this assay

Authors' response: We added the requested information (L178): “The assay’s lower limit of detection and range are 0.5 µU/L and 1-300 mU/L, respectively. The within run coefficient of variation (CV) at insulin concentrations of 27 mU/L and 65 mU/L are 1.05% and 3.9%, respectively. The between run CV at insulin concentrations of 27 mU/L and 65 mU/L are 4.07% and 6.83%, respectively.”

L171 – Suggest brief explanation of what HOMA-IR is. What is normal? What is expected in this population of dogs?

Authors' response: We added the following (L182): “The homeostatic model assessment of insulin resistance (HOMA-IR) is an epidemiological tool used in people to assess insulin resistance. High HOMA-IR levels indicate an increased resistance to insulin, and lower levels indicate an increased sensitivity to insulin’s action. The HOMA-IR was calculated as previously described [17, 18] according to the following formula: plasma insulin x serum glucose / 22.5.”

Authors' response: stratification to three levels of activity allowed us to see the diet x activity level interaction on IG levels. As the Reviewer can see, there is a different metabolic ‘behavior’ at low/moderate levels of activity between the high and low CHO diets. This opens an intriguing question of why this is happening and what are the underlying physiologic mechanisms involved. Furthermore, as the Reviewer can see, the level of IG for the low/moderate activity levels in the high CHO diet is not different than that of the high level of activity in the low CHO diet. Reenforcing the argument above, it possibly suggests different utilization of energy resources for metabolism (i.e., fat vs. CHO) between the diets (and possibly between different levels of activity). Lastly, stratification of activity to three levels also allowed us to determine in what level of activity the dogs are more likely to have low glucose readings (an interesting question on its own).

L176 – Is the =< supposed to be ≤ or

Authors' response: We changed the signs to be consistent with the Reviewers.

L198-199 – Were data transformations attempted?

Authors' response: Variables that did not follow normal distributions were analyzed with non-parametric tests to avoid the transformation and provide clear biological explanation on a nominal scale. We think that it is a valid statistical approach.

L201-203 – several questions:

- Why is time not included as a fixed effect?

Authors' response: Thank you for this important comment. Following your comment, we included ‘period’ as a fixed effect after converting the 96 hours into an 8-level categorical variable, ran the model again, and the only thing that changed is that there was no difference between the low and moderate levels of activity within diet 1 (high CHO low fat). Fig 3 was revised to reflect this change. This is now stated in L213-220.

- Above, you mention activity and inactivity periods, but these are also not reflected in this analysis

Authors' response: To prevent confusion, we changed ‘activity period’ and ‘inactivity period’ to ‘working period’ and ‘resting period’, respectively. The description for the statistical analysis of the frequencies of low vs. normal/high IG readings during the ‘working’ and ‘resting’ periods was explained in L241.

- What about farm, sex, age, breed, etc.? If these were evaluated as fixed effects then dropped, that should still be documented here.

Authors' response: We explored each one as a single effect and dropped from the final model because they were not significant.

- Activity is a response variable, not a fixed effect. If you wanted to include "activity level" as a fixed effect, you would need baseline activity data by which to block the dogs by activity level prior to diet allocation & acclimation

Authors' response: We did not analyze activity per se, but we used delta-g as a measure of activity. We think that it created a confusion (for which we apologize) and we corrected the manuscript to indicate that the dependent variables were IG and delta-g (not activity as you correctly indicated). The study was designed to assess the effect of diet on the level of IG and delta-g as a measure of activity. The classification by L/M/H activity levels enabled us to find biologic/metabolic changes in IG that are related to the interaction between diet x activity level.

In order to establish a ‘baseline’, we would have had to fit the dogs with accelerometers and FSL sensors at the beginning of the acclimation period, which as we explained above in response to one of your previous comments (L149-150) was not feasible financially and logistically. We think that the procedure to classify dogs as having low, moderate, and high activity levels was the best way we could classify the dogs. We think that our approach is valid for answering the research question.

L207-208 – Again, were any transformations attempted prior to analysis as a nonparametric variable?

Authors' response: Variables that did not follow normal distributions were analyzed with nonparametric tests to avoid the transformation and provide clear biological explanation on a nominal scale. We think that it is a valid statistical approach.

L208 – Remove comma after "whereas"

Authors' response: We removed the comma.

L209 – Why GLM here but MIXED above? What was included in the GLM model?

Authors' response: Thank you again for your comment. To be consistent, we changed the analysis from GLM to the MIXED procedure and described the MIXED model more precisely (L226).

L213 – Why not include time & diet x time interaction here?

Authors' response: Thank you for the suggestion. We modeled the curves of activity and IG through the time for each dog. This description of the model was modified in the text. To further explain, the modeling of the curves for level of glucose and the delta-g were done at the level of the dog. Later, we wanted to obtain the average curve of level of glucose and delta-g for each of the diets, which was achieved by averaging individual dog’s splines coefficients to achieve the mean and standard error of the coefficients for each diet. The interaction between diet and time can be evaluated but it is difficult as in this case time is a continuous variable. From Fig 2 it is seen that the level of glucose is different across time for each of the diets. If diet and time interaction would have been tested, they would have been significant. However, it was hard to achieve it because the use of splines. If the modeling of the dependent variables IG and delta-g would have been done with a polynomial the interaction could have been achieved but we chose not to use a polynomial because a polynomial will not describe adequately the variation for these variables.

L214 – Is "activity level" here referring to the "activity levels" described above (L174-177)? Why examine activity level as a dependent variable here but a fixed effect in the linear models above?

Authors' response: We apologize for the confusion. We changed the text to IG and delta-g which refers to the dependent variables that were modeled.

Results:

Authors' response: We added the description of analysis of BW in the methods L232 and the results are presented in the result section L259.

L225 – Suggest authors use consistency with respect to P-value reporting – 2 decimals used here versus three in next line. Recommend confirming if journal has standard for reporting P-values.

Authors' response: We changed all P values to include 3 decimal places.

L226 – With the standard errors reported here, these weights should not have a statistical difference. Something is not correct with this analysis.

Authors' response: Thanks for this observation which led us to review and revise this part of the statistical analysis in the statistical methods. The LSM were similar but the standard error changed which make the results significantly different.

L229 – What about time effect? Diet x time interaction?

Authors' response: Thanks for this important point! The modeling of the curves for level of glucose and the delta-g were done at the level of the dog. Later, we wanted to obtain the average curve of level of glucose and delta-g for each of the diets, which was achieved by averaging individual dog’s splines coefficients to achieve the mean and standard error of the coefficients for each diet. The interaction between diet and time can be evaluated but it is difficult as in this case time is a continuous variable. From Fig 2 it is seen that the level of glucose is different across time for each of the diets. If diet and time interaction would have been tested, they would have been significant. However, it was hard to achieve it because the use of splines. If the modeling of the dependent variables IG and delta-g would have been done with a polynomial the interaction could have been achieved but we chose not to use a polynomial because a polynomial will not describe adequately the variation for these variables.

L242 – Your hypothesis is that low CHO diets lead to low IG which in turn reduces activity. Now you are trying to attribute activity to IG levels! Again, activity is a response variable, not a fixed effect. Dogs' activity is likely related to IG, but that is not what this study was designed to assess. This section should be removed. “activity is a response variable, not a fixed effect”

Authors' response: We did not analyze activity per se, but we used delta-g as a measure of activity. We think that it created a confusion and we corrected the manuscript to indicate that the dependent variables were IG and delta-g (not activity as you correctly indicated). The study was designed to assess the effect of diet on the level of IG and delta-g as a measure of activity. The classification by L/M/H levels of activity enabled us to find biologic/metabolic changes in IG that are related to the interaction between diet x activity level.

L256 – Period of the day is not included in your statistical analysis (L201-203) as a fixed effect.

Authors' response: Good observation. We revised the description of the statistical analysis of the frequencies of low vs. normal/high IG readings during the ‘working’ and ‘resting’ periods, which is now explained in L241. We looked at frequencies of low vs. norm/high IG and used the Chi Square test to determine if there were differences.

L256 – Is the incidence of low IG more relevant than time spent at low IG? How much total time was spent at low IG? The potential diet x time interaction is important.

Authors’ response: The incidence of low glucose is important as well as the overall time of low IG. The incidence is important because until recently when veterinarians started to use the FSL (FreeStyle Libre) we did not think that healthy animals have low IG readings. We are saying it carefully because we do not have reference intervals for IG (as we discussed it in length in the discussion). The point is that now we know (from this study and growing clinical experience) that healthy animals could have low IG readings without clinical signs of neuroglycopenia. How many times a dog has low IG is important because its sympathoadrenal responses should prevent it from happening too often. If the incidence starts to increase above a certain level (which is yet to be defined) then there might be a physiologic or clinical problem. However, the ‘normal’ incidence and frequency of low IG in dogs is unknown and this study would be the first to show it. At the same note, how long a dog has low levels of IG is also important. If there are prolonged periods of low IG, the probability of development of clinical signs related to neuroglycopenia increases as well as the chance that an underlying physiologic or clinical problem exist.

In L312 we described that median (range) of low IG events is 6.5 (IQR 17.5; range 1-42). A median of 6.5 equals to 6.5 x 15 min = 97.5 min per the 96-hour of data recording. You can see that one dog spent 10.5 hours (not continuously) per the 96 hours of data recording with low IG level. This dog as much as we know has been completely normal and did not have any clinical problems or changes in its quality of work (in NZ working dogs are “decommissioned” if they have a decrease in work quality).

We analyzed the interaction of time x diet when IG is low by doing a Chi-Square test and found it to be significant (p < 0.001). The result is now indicated in the text in L308.

L266 – Remove extra space between "Diet 2" and "group"

Authors’ response: Extra space removed.

L269-270 – Would it be expected for the majority of low IG readings to occur in the activity period? Would be nice to mention whether or not this was expected.

Authors’ response: Yes, we expected that most low IG would be during the working period. We prefer to discuss our expectations and interpretation of the results in the discussion rather than in the result section.

L275 – Remove comma after "whereas"

Authors’ response: Extra space removed.

L281 – Again, activity is not a fixed effect, and your hypothesis has the reverse – low IG reducing activity. This section should be removed.

Authors' response: We did not analyze activity per se, but we used delta-g as a measure of activity. We think that it created a confusion and we corrected the manuscript to indicate that the dependent variables were IG and delta-g (not activity as you correctly indicated). The study was designed to assess the effect of diet on the level of IG and delta-g as a measure of activity. The classification by L/M/H activity levels enabled us to find biologic/metabolic changes in IG that are related to the interaction between diet x activity level.

L299 – Consider re-wording as there is no such thing as "inappropriate" serum insulin.

Authors’ response: The sentence was changed to “None of the dogs had evidence of hyperinsulinemic hypoglycemia…” (L338)

Discussion:

L304 – These results were presented above, the reference to the figure is not needed here.

Authors’ response: We started the discussion with the statement of the most important finding in the study and developed the paragraph over this statement. It was specifically requested by Reviewer 1 during the previous revision and we made extensive changes in the structure of the discussion to accommodate Reviewer 1’s request. We think that it is legitimate to refer to the figures in the discussion. If the editor would require removing the reference to the figures in the discussion, we will remove it.

L307 – Suggest removing "most"

Authors’ response: We removed ‘most’

L314 – Reword "people" as "humans" throughout

Authors’ response: The word ‘humans’ substituted the word ‘people’ throughout the manuscript.

L321-322 – Where is it suggested that low fat/high CHO diet offers advantages? Your hypothesis is that lower CHO/high fat diet will reduce activity.

Authors’ response: The word ‘above’ in L360 refers to the beginning of this paragraph where we discuss the advantages of adaptation to a high fat low CHO diet.

L323-326 – These two sentences are very repetitive. Suggest condensing into one sentence.

Authors’ response: We condensed the two sentences to one (L362) “For example, exhaustive exercise in sled dogs that consumed a diet with 0% of metabolizable energy from a CHO source and high %ME from fat had higher albumin, total calcium, and magnesium than control dogs on diets with higher %ME from a CHO source [27].”

L334 – Reword to "necessarily"

Authors’ response: The word ‘necessary’ was removed.

L339-340 – Here you could discuss how activity could have an impact on IG and recommend additional research since your study was not designed to examine this effect.

Authors’ response: At the end of the paragraph we added the following L385: “However, further studies are required to provide a mechanistic insight to determine whether either or both of the above explanations, or alternative explanations underly the temporal relationship between delta-g and IG as our study was not designed to investigate it.”

L340 – Remove reference to figures from discussion

Authors’ response: We think that it is legitimate to refer to the figures in the discussion. We will refer to the Editor’s decision about this point.

L347 – Suggest further discussion at the end of this paragraph as to the potential significance of this observed relationship.

Authors’ response: We do not think that in isolation, the temporal relationship that we described in this paragraph, has a significant implication as both 1) increased glucose production by the liver and reduced glucose utilization by insulin-dependent tissues at high activity levels, and 2) increased muscle and liver glucose uptake immediately after low activity levels have been described in dogs. We think that the pronounced fluctuation in IG and a lower AUC glucose in Diet 2 vs. Diet 1, combined with the fact that dogs in Diet 2 had higher AUC delta-g is really the interesting finding here that has a biological significance and we discussed it in the first paragraph.

Authors’ response: The sentence was rephrased as follow (L390): “During the 96 h of data recording there were neither low peaks of IG that followed high peaks of delta-g, nor low peaks of delta-g followed low peaks of IG.” What we are saying is that the temporal relationship seen in Fig 2 was that low IG always followed low delta-g and that we think that it was because of the large flux of glucose into the liver and muscle as dogs started to rest after high activity level. We discussed it in the previous paragraph and this paragraph basically indicate that they did it in high efficiency because they were very sensitive to insulin.

L351 – Expand on the potential limitation of the 96 h collection period. What time period would have been better?

Authors’ response: This is a hard question to answer and we do not have an answer. It maybe never if the explanation that we gave is correct (i.e., low IG followed low delta-g because large flux of glucose into the liver and muscle occurred as dogs started to rest immediately after periods of high activity level) or a slightly longer or much longer than 96 h if low peaks of IG would follow high peaks of delta-g, or low peaks of delta-g would followed low peaks of IG.

L358 – Suggest "of this alternative"

Authors’ response: We made the requested suggestion (L395)

L360-361 – Suggest removing "The incidence of low IG episodes was 3.12% per 96 h and involved 10/22 (45.5%) of the dogs" from the discussion since it has already been reported in results section.

Authors’ response: We deleted the sentence.

L363 – Similar to comment in L299 above, please remove "inappropriate" and re-word.

Authors’ response: L401 now reads: “The dogs in this study had their serum insulin and glucose concentrations measured at the beginning and the end of the 96 h study periods, and none of the dogs had results that were suggestive of hyperinsulinemic hypoglycemia.”

L364 – Suggest removing "we conclude that"

Authors’ response: The sentence now reads as follows (L404): “Therefore, healthy dogs could have a low incidence of low IG and the frequency of low IG may be greater in dogs fed ultra-low CHO high fat diet (Diet 2).”

Authors’ response: Yes, it was a surprise. Having low levels of glucose of this magnitude has been thought of as an abnormal or pathological condition. Seeing that normal healthy dogs could have it on occasion changes the perspective about finding low IG in dogs. As we eluded, low IG ≠ low blood glucose (BG) especially when you do not have reference intervals. Apparently, there could be many things that can induce low IG and this study identified diet and activity as two factors that were related to it.

L373 – What effect might feeding time/frequency have on IG?

Authors’ response: Interesting question! If dogs are not adapted to alternative sources of energy (i.e., fat) then we speculate that low glucose levels would be more likely to happen as the time elapsing between the incidence of low glucose and feeding increases. This could aggravate if the dog is under any chronic stress (physical, pathological, mental). On the other hand, low levels of glucose could also happen postprandially if the incretin-mediated insulinemic effect is not met with a matching opposing effect mediated by glucagon ± cortisol.

L377 – How does the evening feeding time play into this? Suggest further discussion

Authors’ response: We do not know the answer to this question. We explain just above how low levels of glucose might happen postprandially. But there was a strong diet related effect, so it is not simply a timing issue. We think that we discussed it adequately in this paragraph.

L409 – Any suggestions for future investigation regarding IG monitoring? Potential future work could be worked in throughout discussion.

Authors’ response: To know, one would need to perform invasive studies in dogs that are way beyond the scope of this study. These would include catheterization of the portal and hepatic veins to measure pattern of pancreatic hormone secretion and hepatic glucose output in dogs that will be on high CHO low Fat vs. low CHO high fat combined with EEG monitoring of sleep patterns, IG monitoring, BG monitoring, and VO2 monitoring in combination with periods of strenuous activity (on a treadmill or alike) and rest. This is just not something you can achieve with a simple field study such as this.

L418 – Suggest addition of how this limitation could be overcome, such as formulating the diets to be similar in every respect except for source of energy (CHO vs. fat)

Authors’ response: We added the following in L457: “A future study could reduce nutritional confounders by formulating the diets using varying proportions of the same ingredients, such that only the macronutrient proportions differ.”

Conclusion:

L435 – Spell out abbreviation when it begins a sentence.

Authors’ response: (L476) we spelled out interstitial glucose.

Tables and Figures:

Fig. 1 – in caption, should read "an ultra-low CHO…" not "a"

Authors’ response: (L85) we revised to "an ultra-low CHO…"

Table 2 – Diet randomization is clearly not balanced across farms, particularly farm 3. If there is a reason for this, please include in methods section.

Authors’ response: We did our best to balance the diets across the four farms. In farm 3 there was a problem and it was not as balanced as in the other farms. Obviously now we cannot do anything about it but looking at the variables in Tables 1 and 2 one can tell that the randomization process was successful because there were no significant differences in the demographics across farms and diets.

Table 3 – Recommend removal since activity should not be used as a fixed effect.

Authors’ response: We assume the Reviewer meant Fig 3 as the manuscript does not have Table 3. We explained above that we did not analyze activity per se, but we used delta-g as a measure of activity. We think that it created a confusion and we corrected the manuscript to indicate that the dependent variables were IG and delta-g (not activity as you correctly indicated). The study was designed to assess the effect of diet on the level of IG and delta-g as a measure of activity. The classification by L/M/H levels of activity enabled us to find biologic/metabolic changes in IG that are related to the interaction between diet x activity level. Fig 3 is important because it shows the interaction between activity level and diet with respect to IG. As we explained before, the stratification to three activity levels allowed us to see the diet x activity level interaction on IG levels. There was a different metabolic ‘behavior’ at low/moderate levels of activity between the high and low CHO diets. This opens an intriguing question of why this is happening and what are the underlying physiologic mechanisms involved. Furthermore, as the Reviewer can see, the level of IG for the low/moderate activity levels in the high CHO diet is not different than that of the high level of activity in the low CHO diet. Reenforcing the argument above, it possibly suggests different utilization of energy resources for metabolism (i.e., fat vs. CHO) between the diets (and possibly between different levels of activity). Lastly, stratification of activity to three levels also allowed us to determine at what activity level the dogs are more likely to have low IG readings (an interesting question on its own). We request to keep the figure from these reasons.

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(44.8KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0261506.r005

Decision Letter 2

Balamuralikrishnan Balasubramanian

6 Dec 2021

Less is more? Ultra-low carbohydrate diet and working dogs’ performance

PONE-D-21-02404R2

Dear Dr. Arnon Gal,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Balamuralikrishnan Balasubramanian

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0261506.r006

Acceptance letter

Balamuralikrishnan Balasubramanian

14 Dec 2021

PONE-D-21-02404R2

Less is more? Ultra-low carbohydrate diet and working dogs’ performance

Dear Dr. Gal:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Balamuralikrishnan Balasubramanian

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. Distribution of low activity quartiles from three 96-hour periods of 22 New Zealand working dogs.

The period between the two broken red vertical lines denotes the time the dogs were awake and working. Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S2 Fig. Distribution of IG and activity over the elapsed 96-hour periods.

Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet.

(TIF)

Click here for additional data file.^{(16.4MB, tif)}

S3 Fig. Distribution of interstitial glucose readings of 22 New Zealand working dogs over three 96-hour periods.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S4 Fig. Distribution of low IG readings (<3.5 mmol/L) of 22 New Zealand working dogs over three 96-hour periods.

Diet 1, high %ME CHO low %ME fat diet; Diet 2, ultra-low %ME CHO high %ME fat diet. The broken red lines denote the period between 19:00–06:00.

(TIF)

Click here for additional data file.^{(11.2MB, tif)}

S1 Table. Nutritional analysis of trial diets.

(DOCX)

Click here for additional data file.^{(13.8KB, docx)}

S2 Table. Ingredients of trial diets.

(DOCX)

Click here for additional data file.^{(13.6KB, docx)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(36.7KB, docx)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(44.8KB, docx)}

Data Availability Statement

All relevant data are used for the statistical analyses and for the tables and figures are available within the following stable data repository (doi: 10.17632/d76znr6jvk.1).

[pone.0261506.ref001] 1.Dalton C. Farm dogs: Te Ara—the Encyclopedia of New Zealand; 2013. https://teara.govt.nz/en/farm-dogs.

[pone.0261506.ref002] 2.Kronfeld DS, Hammel EP, Ramberg CF Jr., Dunlap HL Jr.. Hematological and metabolic responses to training in racing sled dogs fed diets containing medium, low, or zero carbohydrate. Am J Clin Nutr. 1977;30(3):419–30. doi: 10.1093/ajcn/30.3.419 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref003] 3.Roberts TJ, Weber JM, Hoppeler H, Weibel ER, Taylor CR. Design of the oxygen and substrate pathways. II. Defining the upper limits of carbohydrate and fat oxidation. J Exp Biol. 1996;199(Pt 8):1651–8. . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref004] 4.Felig P, Marliss E, Owen OE, Cahill GF Jr. Blood glucose and gluconeogenesis in fasting man. Arch Intern Med. 1969;123(3):293–8. Epub 1969/03/01. . [PubMed] [Google Scholar]

[pone.0261506.ref005] 5.Wasserman DH. Regulation of glucose fluxes during exercise in the postabsorptive state. Annu Rev Physiol. 1995;57:191–218. Epub 1995/01/01. doi: 10.1146/annurev.ph.57.030195.001203 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref006] 6.Ahmed M, Gannon MC, Nuttall FQ. Postprandial plasma glucose, insulin, glucagon and triglyceride responses to a standard diet in normal subjects. Diabetologia. 1976;12(1):61–7. Epub 1976/03/01. doi: 10.1007/BF01221966 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref007] 7.Rutherford WJ. Hypoglycemia and endurance exercise: dietary considerations. Nutr Health. 1990;6(4):173–81. Epub 1990/01/01. doi: 10.1177/026010609000600402 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref008] 8.Schnurr TM, Reynolds AJ, Komac AM, Duffy LK, Dunlap KL. The effect of acute exercise on GLUT4 levels in peripheral blood mononuclear cells of sled dogs. Biochem Biophys Rep. 2015;2:45–9. Epub 2015/09/05. doi: 10.1016/j.bbrep.2015.05.002 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref009] 9.Pratt-Phillips SE, Olsen R, Geor R, Zirkle A, Moore A, Harkins C, et al. Effect of reduced protein intake on endurance performance and water turnover during low intensity long duration exercise in Alaskan sled dogs. Comp Exerc physiol. 2018;14(1):19–26. doi: 10.3920/CEP170024 [DOI] [Google Scholar]

[pone.0261506.ref010] 10.Bermingham EN, Thomas DG, Cave NJ, Morris PJ, Butterwick RF, German AJ. Energy requirements of adult dogs: a meta-analysis. PLoS One. 2014;9(10):e109681. Epub 2014/10/15. doi: 10.1371/journal.pone.0109681 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref011] 11.Corradini S, Pilosio B, Dondi F, Linari G, Testa S, Brugnoli F, et al. Accuracy of a flash glucose monitoring system in diabetic dogs. J Vet Intern Med. 2016;30(4):983–8. Epub 2016/06/20. doi: 10.1111/jvim.14355 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref012] 12.Silva DD, Cecci GRM, Biz G, Chiaro FN, Zanutto MS. Evaluation of a flash glucose monitoring system in dogs with diabetic ketoacidosis. Domest Anim Endocrinol. 2021;74:106525. Epub 2020/08/18. doi: 10.1016/j.domaniend.2020.106525 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref013] 13.Malerba E, Cattani C, Del Baldo F, Carotenuto G, Corradini S, Golinelli S, et al. Accuracy of a flash glucose monitoring system in dogs with diabetic ketoacidosis. J Vet Intern Med. 2020;34(1):83–91. Epub 2019/11/15. doi: 10.1111/jvim.15657 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref014] 14.Del Baldo F, Canton C, Testa S, Swales H, Drudi I, Golinelli S, et al. Comparison between a flash glucose monitoring system and a portable blood glucose meter for monitoring dogs with diabetes mellitus. J Vet Intern Med. 2020;34(6):2296–305. Epub 2020/10/31. doi: 10.1111/jvim.15930 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref015] 15.Albright JD, Seddighi RM, Ng Z, Sun X, Rezac DJ. Effect of environmental noise and music on dexmedetomidine-induced sedation in dogs. PeerJ. 2017;5:e3659. Epub 2017/08/09. doi: 10.7717/peerj.3659 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref016] 16.Lee AH, Detweiler KB, Harper TA, Knap KE, de Godoy MRC, Swanson KS. Physical activity patterns of free living dogs diagnosed with osteoarthritis. J Anim Sci. 2021. Epub 2021/07/04. doi: 10.1093/jas/skab204 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref017] 17.Mejia S, Duerr FM, Salman M. Comparison of activity levels derived from two accelerometers in dogs with osteoarthritis: Implications for clinical trials. Vet J. 2019;252:105355. Epub 2019/09/27. doi: 10.1016/j.tvjl.2019.105355 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref018] 18.Kaneko JJ. Carbohydrate metabolism and its diseases. In: Kaneko JJ, Harvey JW, Bruss ML, editors. Clinical biochemistry of domestic animals. Sixth ed. San Diego, CA, USA: Academic Press; 2008. p. 45–80. [Google Scholar]

[pone.0261506.ref019] 19.Schnurr TM, Reynolds AJ, Gustafson SJ, Duffy LK, Dunlap KL. Conditioning causes an increase in glucose transporter-4 levels in mononuclear cells in sled dogs. Int J Biochem Cell Biol. 2014;55:227–31. doi: 10.1016/j.biocel.2014.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref020] 20.Serisier S, Leray V, Poudroux W, Magot T, Ouguerram K, Nguyen P. Effects of green tea on insulin sensitivity, lipid profile and expression of PPARalpha and PPARgamma and their target genes in obese dogs. Br J Nutr. 2008;99(6):1208–16. Epub 2007/12/07. doi: 10.1017/S0007114507862386 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref021] 21.Littell RC, Henry PR, Ammerman CB. Statistical analysis of repeated measures data using SAS procedures. J Anim Sci. 1998;76(4):1216–31. Epub 1998/05/15. doi: 10.2527/1998.7641216x . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref022] 22.Gilmour AR, Gogel BJ, Cullis BR, Thompson R. ASReml user guide release 3.0 Hemel Hempstead, HP1 1ES, UK: VSN International Ltd; 2009.

[pone.0261506.ref023] 23.White IM, Thompson R, Brotherstone S. Genetic and environmental smoothing of lactation curves with cubic splines. J Dairy Sci. 1999;82(3):632–8. Epub 1999/04/09. doi: 10.3168/jds.S0022-0302(99)75277-X . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref024] 24.Romsos DR, Palmer HJ, Muiruri KL, Bennink MR. Influence of a low carbohydrate diet on performance of pregnant and lactating dogs. J Nutr. 1981;111(4):678–89. Epub 1981/04/01. doi: 10.1093/jn/111.4.678 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref025] 25.Murray AJ, Knight NS, Cole MA, Cochlin LE, Carter E, Tchabanenko K, et al. Novel ketone diet enhances physical and cognitive performance. FASEB J. 2016;30(12):4021–32. Epub 2016/08/17. doi: 10.1096/fj.201600773R . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref026] 26.Volek JS, Freidenreich DJ, Saenz C, Kunces LJ, Creighton BC, Bartley JM, et al. Metabolic characteristics of keto-adapted ultra-endurance runners. Metabolism. 2016;65(3):100–10. Epub 2016/02/20. doi: 10.1016/j.metabol.2015.10.028 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref027] 27.Hammel EP, Kronfeld DS, Ganjam VK, Dunlap HL Jr. Metabolic responses to exhaustive exercise in racing sled dogs fed diets containing medium, low, or zero carbohydrate. Am J Clin Nutr. 1977;30(3):409–18. Epub 1977/03/01. doi: 10.1093/ajcn/30.3.409 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref028] 28.Pencek RR, Fueger PT, Camacho RC, Wasserman DH. Mobilization of glucose from the liver during exercise and replenishment afterward. Can J Appl Physiol-Rev Can Physiol Appl. 2005;30(3):292–303. doi: 10.1139/h05-122 [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref029] 29.Richter EA, Hargreaves M. Exercise, GLUT4, and skeletal muscle glucose uptake. Physiol Rev. 2013;93(3):993–1017. doi: 10.1152/physrev.00038.2012 . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref030] 30.Jauch-Chara K, Hallschmid M, Gais S, Oltmanns KM, Peters A, Born J, et al. Awakening and counterregulatory response to hypoglycemia during early and late sleep. Diabetes. 2007;56(7):1938–42. doi: 10.2337/db07-0044 [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref031] 31.Jauch-Chara K, Hallschmid M, Schmid SM, Oltmanns KM, Peters A, Born J, et al. Plasma glucagon decreases during night-time sleep in Type 1 diabetic patients and healthy control subjects. Diabet Med. 2007;24(6):684–7. Epub 2007/03/27. doi: 10.1111/j.1464-5491.2007.02116.x . [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref032] 32.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F, et al. Decreased epinephrine responses to hypoglycemia during sleep. N Engl J Med. 1998;338(23):1657–62. doi: 10.1056/NEJM199806043382303 [DOI] [PubMed] [Google Scholar]

[pone.0261506.ref033] 33.Schultes B, Jauch-Chara K, Gais S, Hallschmid M, Reiprich E, Kern W, et al. Defective awakening response to nocturnal hypoglycemia in patients with Type 1 Diabetes Mellitus. PLoS Med. 2007;4(2):e69. doi: 10.1371/journal.pmed.0040069 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref034] 34.Shea EK, Hess RS. Assessment of postprandial hyperglycemia and circadian fluctuation of glucose concentrations in diabetic dogs using a flash glucose monitoring system. J Vet Intern Med. 2021;35(2):843–52. Epub 2021/02/02. doi: 10.1111/jvim.16046 . [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0261506.ref035] 35.Schauf S, Salas-Mani A, Torre C, Jimenez E, Latorre MA, Castrillo C. Effect of feeding a high-carbohydrate or a high-fat diet on subsequent food intake and blood concentration of satiety-related hormones in dogs. J Anim Physiol Anim Nutr (Berl). 2018;102(1):e21–e9. Epub 2017/04/28. doi: 10.1111/jpn.12696 . [DOI] [PubMed] [Google Scholar]

PERMALINK

Less is more? Ultra-low carbohydrate diet and working dogs’ performance

Arnon Gal

Williams Cuttance

Nick Cave

Nicolas Lopez-Villalobos

Aaron Herndon

Juila Giles

Richard Burchell

Roles

Abstract

Introduction

Fig 1. Hypothesis model.

Materials and methods

Study design

Table 1. Demographics and descriptive statistics of the dog cohort in this study.

Table 2. Demographics and descriptive statistics per diets at the beginning of the acclimation period.

Diets

Sample collection and handling

Blood collection

Sample handling and processing

Flash glucose monitoring and triaxial accelerometry

Insulin assay, and the homeostatic model assessment of insulin resistance

Dog activity levels

The ‘working period’ and ‘resting period’

Power-sample size and statistical analysis

Results

The effect of the diets on bodyweight

The effect of diets on activity level and interstitial glucose concentration

Fig 2. The predicted IG concentration and activity level for dogs stratified by two levels of diet.

Interstitial glucose concentration as a function of three activity levels

Fig 3. Least squares means (±SE) of IG concentration derived from a flash glucose monitor as a function of three activity levels derived from triaxial accelerometer readings and stratified by two levels of diets.

The effect of diet, and the ‘working period’ and ‘resting period’ on the incidence of low levels of IG

The effect of activity level and diet on the incidence of low IG readings

Glucose, insulin and HOMA-IR

Discussion

Conclusion

Supporting information

Data Availability

Funding Statement

References

Decision Letter 0

Balamuralikrishnan Balasubramanian

Roles

Author response to Decision Letter 0

Decision Letter 1

Balamuralikrishnan Balasubramanian

Roles

Author response to Decision Letter 1

Decision Letter 2

Balamuralikrishnan Balasubramanian

Roles

Acceptance letter

Balamuralikrishnan Balasubramanian

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases