Effect of withholding food versus feeding on creatinine, symmetric dimethylarginine, cholesterol, triglycerides, and other biochemical analytes in 100 healthy dogs

Abstract

Background

Withholding food is often recommended before collection of blood for routine biochemical analysis in dogs despite a paucity of evidence to support this requirement.

Objectives

To compare measurements of selected biochemical analytes collected before and after feeding in clinically healthy dogs.

Animals

One hundred clinically healthy staff‐ and student‐owned dogs weighing ≥15 kg.

Methods

Prospective observational study. Food was withheld from the dogs for 10‐26 hours. Preprandial serum was collected, and then dogs were fed their usual food at an amount equivalent to at least 2/3 resting energy requirement (RER). Selected serum analytes were measured at 2‐, 4‐, 6‐, and 8‐hours postprandially. The proportion of postprandial values that exceeded either the reported allowable total error (TEa), or for symmetric dimethylarginine (SDMA), the reference change value (RCV), was determined. As neither TEa nor RCV is available for lipase, comparison was made to the high end of the reference interval (RI).

Results

The proportion of dogs with at least 1 postprandial measurement that exceeded the TEa or RCV was 92/100 for triglycerides, 66/100 for blood urea nitrogen (BUN), 46/100 for phosphorus, 17/100 for glucose, 9/100 for bilirubin, 5/100 for SDMA, 2/100 for creatinine, and 0/100 for cholesterol and albumin. Postprandial lipase never exceeded the RI in dogs with normal fasted lipase.

Conclusions and Clinical Importance

Withholding food is generally not necessary before performing routine biochemical analysis in clinically healthy dogs. Withholding food might be helpful to limit variability in analytes impacted by feeding, such as triglycerides and phosphorus.

Abbreviations

BUN

blood urea nitrogen

IRIS

International Renal Interest Society

RCV

reference change value

RER

resting energy requirement

RI

reference interval

SDMA

symmetric dimethylarginine

TEa

allowable total error

USG

urine specific gravity

1. INTRODUCTION

Serum biochemical analytes are routinely used for health assessment and disease management in dogs. Withholding food before sample collection is often recommended to reduce preanalytical variation in variables such as hemolysis, lipemia, and other factors involving specimen collection, handling, processing, and storage. Current guidelines from the International Renal Interest Society (IRIS) recommend withholding food overnight before measurement of creatinine and symmetric dimethylarginine (SDMA). ¹ However, studies evaluating the effects of feeding on serum creatinine concentrations relied on small numbers of dogs and produced variable results with creatinine reported to increase, decrease, or remain unchanged after eating commercial dog food. ² , ³ , ⁴ , ⁵ In contrast, blood urea nitrogen (BUN) is consistently higher after ingestion of commercial diets as well as cooked and raw meats. ² , ⁵ The effect of feeding on SDMA has not been reported in dogs.

In dogs, increases in cholesterol and triglyceride concentrations above the reference interval are routinely attributed to feeding. ⁶ , ⁷ Glucose and triglyceride concentrations measured after withholding food and after feeding are greater in obese dogs than appropriately conditioned dogs, while cholesterol concentrations are unaffected by obesity. ⁸ , ⁹ Peak triglyceride concentrations in overweight and obese dogs occur anywhere between 1 and 11 hours after feeding, although the highest values for 6% of dogs are preprandial. ¹⁰ In human medicine, multiple prospective, general population‐based studies indicate that lipid profiles are minimally affected by normal food intake. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Measurements of triglycerides, high density lipoproteins, and low density lipoproteins have similar prognostic value for cardiovascular disease whether from dogs that are fed or have had feed withheld. ¹¹ , ¹⁶ , ¹⁷ Recent guidelines from the American College of Cardiology state that lipid profiles from dogs that have not had feed withheld are adequate for estimating risk of atherosclerotic cardiovascular disease in most adults. ¹⁸

Lipemia is a common source of preanalytical variation, causing interference by light scattering, volume displacement, and hemolysis. ¹⁹ As lipemia is often a postprandial effect, withholding food is frequently recommended to reduce potential interference with analyte measurements.

As owner compliance with protocols to withhold food can be poor, it is important for veterinarians to know the anticipated effects of feeding on blood analytes. In this study, we sought to compare measurements of triglycerides, BUN, phosphorus, glucose, bilirubin, SDMA, creatinine, cholesterol, albumin, and lipase in samples obtained after withholding food overnight versus samples obtained at 2‐, 4‐, 6‐, and 8‐hours after feeding in clinically healthy dogs. Postprandial measurements were compared to preprandial measurements using allowable total error (TEa), a quality requirement based on analytical imprecision and bias along with clinical decision thresholds for each analyte that represents the maximum acceptable analytical error to ensure clinical usefulness. No TEa is reported for SDMA, so change was compared against a recently reported reference change value (RCV), ²⁰ which is the minimum percent change needed to consider 2 consecutive results in an individual to be significantly different.

2. METHODS AND MATERIALS

2.1. Study sample

One hundred two student‐ and staff‐owned dogs were enrolled. Eligible dogs weighed ≥15 kg and were clinically well as reported by the owner. The use of medications and supplements was allowed, provided these had been given at the same dose for >6 weeks before enrollment. Data collected included age, sex, neuter status, breed, and weight. Owners were asked to withhold food after 10:00 pm the night before evaluation, and to answer a questionnaire which included the time of the last meal fed (to determine duration of fasting), current medications, previous medical history, and current diet. The study protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at Texas A&M University (#2019‐0369 CA). Owner consent was obtained for each dog.

2.2. Feeding

Food was withheld for a minimum of 10 hours. Dogs were weighed upon admission, and RER was calculated as 70 (weight in kg^0.75) kcal. Each dog was offered its regular diet at an amount (by weight or volume) that corresponded to its RER. Any remaining food was removed after 10 minutes, weighed, and recorded to determine the calories consumed. Dogs were required to eat a minimum of 2/3 RER to be included. Dogs were provided free access to water.

2.3. Blood sampling, storage, and analysis

Venipuncture was performed before feeding and then at 2‐, 4‐, 6‐, and 8‐hours postprandially. The sample collected before feeding was considered the baseline blood sample. Blood collected by jugular or peripheral venipuncture was placed into serum separator tubes (Vacuette; Greiner Bio‐One, Kremsmunster, Austria). The blood was allowed to clot at room temperature for a minimum of 30 minutes, then was centrifuged at 1730g for 10 minutes at 6°C within 2 hours of collection. Serum was removed, aliquoted, and frozen (−80°C) until submission for batch analysis. All samples from the same dog were analyzed in the same batch. Biochemical analysis was performed at a commercial reference laboratory (IDEXX West Sacramento) using a Beckman Coulter AU680 automated analyzer (Beckman Coulter, Inc, Brea, California). Methodologies for all assays are described in Table 1. Full datasheets with methodological detail are publicly available (www.beckmancoulter.com).

TABLE 1.

Analytical methodology using a Beckman Coulter chemistry analyzer at a single location at IDEXX West Sacramento

Analyte	Method
Triglycerides	Enzymatic colorimetric
BUN	Urease
Phosphorus	Ammonium molybdate
Glucose	Hexokinase
Bilirubin	Diazo coupling
SDMA	Immunoassay
Creatinine	Modified Jaffe
Cholesterol	Cholesterol oxidase, POD colorimetric
Albumin	Bromocresol green
Lipase	Enzymatic colorimetric

Note: Full datasheets with methodological detail are publicly available (www.beckmancoulter.com/support/tech‐docs).

2.4. Urine specific gravity

Free‐catch urine was obtained postprandially on the day of study. Because voluntarily voided urine was used, timing varied among dogs. Urine specific gravity (USG) was measured with a handheld refractometer.

2.5. Statistical analysis

Data were recorded in an electronic spreadsheet (Microsoft Excel, 2019, Microsoft Corp., Redmond, Washington). Postprandial values were compared to baseline data using TEa (Table 2) for those analytes with a published value. These values were determined by the American Society for Veterinary Clinical Pathology and represent quality goals for the maximum tolerable analytical error for each analyte, beyond which clinical decision‐making might be impacted. ²² Changes in SDMA were compared against a recently reported RCV of 43%, which represents the minimum percent change to consider 2 consecutive measurements different (ie, 43% change between 2 consecutive SDMA measurements cannot be attributed to natural variation alone). ²⁰ The number of postprandial results that moved outside of the RI were also determined.

TABLE 2.

Published allowable total error (TEa) for biochemical analytes with concentrations falling below, within, or above the reference interval ²¹

Analyte a	Low analyte values	Within RI	High analyte values
Triglycerides	NCR	25%	25%
BUN	15%	12%	12%
Phosphorus	20%	15%	15%
Glucose	10%	20%	20%
Bilirubin	NCR	30%	30%
Creatinine	20%	20%	20%
Cholesterol	20%	20%	20%
Albumin	15%	15%	15%

Abbreviation: NCR, not clinically relevant.

^aPublished TEa not available for SDMA or lipase.

3. RESULTS

3.1. Study sample

Two dogs were removed from the study because of unfavorable temperament and unwillingness to eat. One hundred dogs completed the study. The median age was 5 years (range 0.75‐16 years). There were 42 castrated male dogs, 49 spayed female dogs, 4 intact male dogs, and 5 intact female dogs. Mixed breed dogs (n = 26) were most commonly represented, followed by 18 Labrador Retrievers, 11 Golden Retrievers, 7 Australian Shepherds, 6 Border Collies, 6 Pit Bulls, 4 German Shepherds, 4 Black Mouth Curs, 3 Rhodesian Ridgebacks, 3 Australian Cattle Dogs, 2 Blue Heelers, and 1 of each of the following: Alaskan Malamute, Boxer, English Springer Spaniel, Great Dane, German Shorthair Pointer, Kelpie, Miniature Pinscher, Siberian Husky, Staffordshire Terrier, and Walker Hound. The median body weight was 26.75 kg (range 15‐41 kg). The median fasting time was 12.5 hours (range 10‐25.75 hours). The median percentage of RER consumed was 100% (70%‐100%).

Fifteen dogs were receiving long‐term prescription medications (oclacitinib [n = 4], meloxicam and gabapentin [n = 1], carprofen and gabapentin [n = 1], meloxicam, gabapentin, phenylpropanolamine, and joint supplements [n = 1], meloxicam, fluoxetine, and joint supplements [n = 1], fluoxetine and probiotics [n = 1], and lokivetmab [n = 1]). Dietary supplements were routinely administered and included fish oils (n = 2), joint supplements alone (n = 1), joint supplements and probiotics (n = 1), and cranberry supplements (n = 1).

All dogs were fed commercial diets (Supplemental Table S1).

3.2. Hemolysis and lipemia

Of 500 samples, 77 samples had 1+ hemolysis and 16 had 2+ hemolysis; none had 3+ or 4+ hemolysis; 139 had 1+ lipemia, 35 had 2+ lipemia, 5 had 3+ lipemia, and 3 had 4+ lipemia.

3.3. Triglycerides

Three dogs had baseline triglyceride concentrations above the RI, with values from 152 to 171 mg/dL (RI 20‐150 mg/dL; Supplemental Table S2). Thirty‐six dogs had 1 or more postprandial triglyceride concentrations that were above the RI; 33/36 of these dogs had normal baseline triglycerides. Ninety‐two dogs had 1 or more postprandial values that exceeded the TEa when compared to the baseline measurement (Table 3). Sequential changes in triglyceride concentration are shown in Figure 1.

TABLE 3.

Number of dogs out of 100 study dogs (with 95% binomial confidence interval) with analyte measurement exceeding TEa or reference change value (for SDMA) in a positive or negative direction or exceeding the reference interval (for lipase) at each postprandial timepoint compared with baseline

	2‐Hour		4‐Hour		6‐Hour		8‐Hour
Analyte	Increase	Decrease	Increase	Decrease	Increase	Decrease	Increase	Decrease
Triglycerides	81 (72‐88)	0 (0‐4)	79 (70‐86)	4 (1‐10)	64 (54‐73)	6 (3‐13)	53 (43‐62)	7 (3‐14)
BUN	19 (12‐28)	0 (0‐4)	61 (51‐70)	0 (0‐4)	61 (51‐70)	2 (0‐7)	51 (41‐61)	3 (1‐9)
Phosphorus	6 (3‐13)	0 (0‐4)	18 (12‐27)	10 (5‐18)	35 (26‐45)	7 (3‐14)	39 (30‐49)	3 (1‐9)
Glucose	10 (5‐18)	1 (0‐6)	8 (4‐15)	1 (0‐6)	6 (3‐13)	1 (0‐6)	7 (3‐14)	1 (0‐6)
Bilirubin	5 (2‐11)	1 (0‐6)	6 (3‐13)	0 (0‐4)	3 (1‐9)	1 (0‐6)	3 (1‐9)	1 (0‐6)
SDMA	2 (0‐7)	0 (0‐4)	1 (0‐6)	0 (0‐4)	3 (1‐9)	0 (0‐4)	2 (0‐7)	0 (0‐4)
Creatinine	1 (0‐6)	0 (0‐4)	1 (0‐6)	0 (0‐4)	1 (0‐6)	0 (0‐4)	1 (0‐6)	0 (0‐4)
Cholesterol	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)
Albumin	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)	0 (0‐4)
Lipase	2 (0‐7)	0 (0‐4)	1 (0‐6)	0 (0‐4)	1 (0‐6)	0 (0‐4)	1 (0‐6)	0 (0‐4)

FIGURE 1.

Box‐and‐whisker plots of serum triglycerides, BUN, phosphorus, and glucose measured in 100 dogs preprandially and at hours 2, 4, 6, and 8 after eating a meal of their normal diet at 70%‐100% RER. Each box‐and‐whisker plot indicates the median and quartiles. The reference interval is depicted by dashed lines

3.4. Blood urea nitrogen

With the exception of 1 low baseline concentration (7 mg/dL), values for all dogs at all timepoints were within the RI (RI 9‐31 mg/dL; Supplemental Table S3). Sixty‐six dogs had 1 or more postprandial elevations in excess of the TEa (Table 3). Sequential changes in BUN concentration are shown in Figure 1.

3.5. Phosphorus

Three dogs had baseline hyperphosphatemia, ranging from 6.3 to 6.9 mg/dL (RI 2.5‐6.1 mg/dL; Supplemental Table S4). These 3 dogs were less than 2 years old and remained hyperphosphatemic at most other time points throughout the day (12/15 timepoints), but values always remained within TEa. Only 1 dog with baseline normophosphatemia became hyperphosphatemic 8‐hours postprandially; however, this measurement remained within the TEa. This dog was also less than 2 years old. Two dogs had low baseline phosphorus concentrations (both 2.4 mg/dL) that were within the reference interval at all other timepoints; 1 dog's postprandial measurements did not exceed the TEa at any timepoint, while the other dog's 6‐hour and 8‐hour measurements exceeded the TEa. In total, 46 dogs had at least 1 postprandial increase in phosphorus that exceeded the TEa, and all of these measurements remained within the reference interval. Of these 46 dogs, 5 were less than 2 years old, and 41 were at least 2 years old. Overall, 12 dogs were less than 2 years old and 88 dogs were at least 2 years old. Twelve dogs had at least 1 postprandial decrease in phosphorus exceeding the TEa, with 1 measurement below the reference interval (2.4 mg/dL; Table 3). Sequential changes in phosphorus concentration are shown in Figure 1.

3.6. Glucose

One dog had a baseline glucose concentration above the RI (117 mg/dL; RI 63‐114 mg/dL; Supplemental Table S5). Six dogs had at least 1 postprandial glucose concentration above the RI, although the change exceeded TEa in only 2 of these 6 dogs. The highest postprandial glucose measurement at any time was 130 mg/dL, which occurred at 2‐hours and exceeded TEa. Overall, 17 dogs had at least 1 postprandial glucose measurement that exceeded the TEa (Table 3). Sequential changes in glucose concentration are shown in Figure 1.

3.7. Bilirubin

Baseline bilirubin concentrations were within the RI (0‐0.3 mg/dL) for all dogs (Supplemental Table S6). One measurement in 1 dog exceeded the RI at 0.4 mg/dL at 8‐hours; this exceeded TEa for the baseline measurement of 0.2 mg/dL. In total, compared to TEa, 9 dogs (17 bilirubin measurements) had at least 1 postprandial increase in bilirubin and 1 dog had a decrease in bilirubin (Table 3). Of these, 12 samples were nonhemolyzed, 3 had 1+ hemolysis, and 2 had 2+ hemolysis.

3.8. Symmetric dimethylarginine

Nine dogs had baseline SDMA measurements above the RI with values ranging from 15 to 20 μg/dL (RI 0‐14 μg/dL; Supplemental Table S7). Ten dogs with an SDMA concentration within the RI at baseline had at least 1 SDMA measurement above the RI postprandially, although none of these postprandial measurements exceeded the RCV (Figure 2). A total of 5 dogs had at least 1 postprandial increase in SDMA concentration that exceeded the RCV. One of these dogs had a preprandial SDMA of 6 μg/dL and exceeded RCV at all postprandial timepoints with postprandial values ranging from 9 to 10 μg/dL. The other 4 dogs had 1 measurement each exceeding RCV: 1 increasing from 7 to 11 μg/dL at 2‐hours, 2 increasing from 9 to 13 μg/dL at 6‐hours, and 1 increasing from 9 to 13 μg/dL at 8‐hours.

FIGURE 2.

Median percent change and 25th to 75th percentile of SDMA measured in 100 dogs at timepoints 2‐, 4‐, 6‐, and 8‐hours after eating their normal diet at 70%‐100% RER, compared to baseline measurement. The percent change in SDMA exceeded the reference change value of 43% in 5 dogs, but none of the measurements in these 5 dogs exceeded the reference limit of 14 μg/dL

Of the 19 dogs with an SDMA above the RI at any timepoint, 9 had a USG ≥1.035, 7 had a USG between 1.030 and 1.034, and 3 had a USG <1.030. (Supplemental Table S12). The latter 3 dogs had SDMA measurements between 15‐19 μg/dL, 16‐19 μg/dL, and 16‐22 μg/dL, respectively. For each of these, the highest SDMA measurement was at 2‐hours, or, for 1 dog, both 2‐ and 4‐hours postprandially. The lowest SDMA measurement was at 8‐hours postprandially for 2 dogs, and preprandially for the other.

3.9. Creatinine

A single dog had a baseline creatinine concentration above the RI (1.6 mg/dL; RI 0.5‐1.5 mg/dL; Supplemental Table S8). This dog had a USG of 1.050 and a baseline SDMA within the RI (9 μg/dL). Two dogs had at least 1 postprandial creatinine measurement above the TEa (Table 3). One of these dogs had a baseline creatinine of 1.5 mg/dL and exceeded the TEa with measurements of 1.9 mg/dL at 2‐, 4‐, and 6‐hours postprandially, with a USG of 1.037. The other dog had a baseline creatinine of 1.1 mg/dL and exceeded the TEa with a measurement of 1.4 mg/dL at 8‐hours postprandially; its USG was 1.022.

3.10. Cholesterol

Eight dogs had baseline cholesterol concentrations above the RI, with values ranging from 352 to 445 mg/dL (RI 131‐345 mg/dL; Supplemental Table S9) and remained hypercholesterolemic at most timepoints (35/40 measurements). In total, 10 dogs were hypercholesterolemic at 1 or more postprandial timepoints. Baseline cholesterol concentrations for the 2 dogs without baseline hypercholesterolemia were 337 and 343 mg/dL. None of the postprandial cholesterol concentrations exceeded the TEa applied to the baseline measurements (Table 3).

3.11. Albumin

Two dogs were hypoalbuminemic preprandially at 2.6 mg/dL (RI 2.7‐3.9 mg/dL; Supplemental Table S10). One of these dogs was persistently below the RI at all other timepoints, while the other dog was within the RI at all postprandial timepoints, with values ranging from 2.7 to 2.8 mg/dL. The only other postprandial hypoalbuminemic measurements occurred in a dog with an albumin concentration of 2.7 mg/dL at all timepoints except 8‐hours postprandially, when the albumin was 2.6 mg/dL and was within TEa. No postprandial measurement exceeded the TEa compared to the baseline measurement (Table 3).

3.12. Lipase

Preprandially, 2 dogs had lipase activities above the RI with values of 812 and 873 U/L (RI 138‐755 U/L; Supplemental Table S11). One of these had lipase activities within the RI at the remaining timepoints, while the other was above the RI at 2‐ and 4‐hours but within the RI at 6‐ and 8‐hours. In dogs whose preprandial lipase was within RI, no postprandial values exceeded the RI.

4. DISCUSSION

Our findings indicate that triglycerides, BUN, phosphorus, glucose, and bilirubin often increased postprandially while SDMA, creatinine, cholesterol and albumin concentrations, and lipase activity were not altered by feeding. However, the magnitude of change observed for most of these analytes would be unlikely to impact clinical decision‐making.

When available, TEa was used to represent the maximum acceptable amount of imprecision and bias in the analytical method in order to ensure clinical usefulness, and changes beyond that were attributed to biological variability, that is, the effect of feeding. The TEa values used in this study were published by the American Society for Veterinary Clinical Pathology and are consistent with Clinical Laboratory Improvement Amendments regulations used in laboratory testing in human medicine. ²² , ²³ Reference change value, or critical difference, describes inherent variation between serial measurements from analytical and biological causes. When 2 consecutive measurements exceed this expected variation, they are determined to represent a true difference, even when the measurements remain within the reference interval. While RCV is ideal for detecting true changes in analytes that have a high index of individuality, use of the RCV assumes that serial results are not correlated, as RCV is based on analytical and biological variation that occurs over extended periods of time. ²⁴ However, there is no clear consensus on a standard to compare values in an animal being monitored daily or multiple times within a day, as biological variation and coefficient of variation are likely smaller over short periods of time than long periods of time. Although a less commonly used approach, we used TEa to identify clinically relevant change because it is a quality standard derived from medically important decision thresholds, representing a limit for error beyond which diagnostic interpretation might be impacted. It can therefore serve as a reasonable guide to determine whether serial values are different in a patient. The TEa values are typically lower than RCV, ²⁵ which is appropriate in this study given the close timing of the sample collections and the batch analysis of the analytes. Because TEa is lower than RCV for most analytes, lack of change in response to feeding compared to TEa would also not show a change compared to RCV. One disadvantage of this approach is that TEa is only published for a subset of analytes and was not available for SDMA or lipase. Changes in SDMA were therefore compared to preprandial values using RCV reported by 1 study, ²⁰ while changes in lipase activity were compared to the RI.

By recent consensus, nonfasted lipid profiles are routinely allowed for initial lipid testing in most human patients evaluated for cardiovascular risk assessment. ¹⁸ Our results show that in this study sample of healthy dogs, triglycerides, but not cholesterol, routinely increased after a meal. In this study sample, dogs with normal cholesterol concentrations in a fasted state rarely developed hypercholesterolemia postprandially. This suggests that, as in humans, withholding food is not necessary when assessing cholesterol concentrations in healthy dogs. However, our study showed that feeding of a meal increased triglyceride concentrations compared to baseline, which might affect clinical interpretation. Traditionally, diagnosis of hypertriglyceridemia is made based on measurement after withholding food for at least 12 hours. ²¹ However, in a report of Miniature Schnauzers with a history of visible lipemia, withholding food for >12 hours reduced the sensitivity for detecting hypertriglyceridemia, suggesting that a triglyceride measurement after withholding food for 12 hours might fail to identify some dogs with delayed very low density lipoprotein clearance or other dyslipidemias. ²⁶ Therefore, triglycerides measured both after withholding food as well as peak concentration after feeding might be useful in identifying dogs for which dietary fat restriction could be beneficial. Similarly, while our study showed that triglycerides increased after a meal, the role of postprandial hypertriglyceridemia interpretation in dogs is still unclear. Further studies are needed to define the optimal conditions for determination of serum triglyceride concentrations in healthy and dyslipidemic dogs; however, to limit preanalytical variability in an individual, we recommend monitoring triglyceride concentrations under standardized conditions of withholding food.

Our results indicate that BUN often increased in response to feeding based on TEa. Our findings are in agreement with previous studies. ² , ⁵ In our study, with the exception of 1 dog with a fasted BUN measurement that was marginally low, all BUN measurements at all timepoints remained within the RI. This indicates that if clinicians' interpretations are based on RIs, postprandial elevations in BUN are unlikely to influence clinical assessment.

To our knowledge, the effect of feeding on SDMA has not previously been evaluated; however, nutritional intervention with a renal test diet improves renal biomarkers, including SDMA, over time. ²⁷ , ²⁸ Our study showed that SDMA was not different postprandially compared to preprandially. While 5 dogs had at least 1 postprandial measurement exceeding the RCV, none had an SDMA above the RI at any timepoint, so it is unlikely that a clinician would interpret the postprandial values of each of these dogs differently than the preprandial values. While IRIS currently recommends withholding food overnight before determination of serum creatinine and SDMA concentrations in order to stage dogs with chronic kidney disease, dogs in our study rarely had increases in postprandial creatinine or SDMA that exceeded the TEa or RCV, respectively.

Phosphorus concentrations are also an important analyte in dogs with kidney disease. Withholding food might be beneficial in standardizing evaluation of serum phosphorus measurements, although dietary differences in inorganic versus organic phosphorus concentration were not assessed in this study. Nearly half of the dogs in this study had an increase in serum phosphorus exceeding the TEa at 1 or more postprandial timepoints, although none of these dogs had values that exceeded the upper boundary of the RI. The proportion of dogs less than 2 years old that had an increase in phosphorus exceeding TEa was similar to that of dogs 2 years and older (5/12, 42%, vs 41/88, 47%). Thus, while hyperphosphatemia was more common in dogs younger than 2 years, juvenile age did not appear to influence the effect of feeding on phosphorus concentrations. When monitoring phosphorus concentrations in an individual, we recommend measurement under consistent conditions of withholding food to limit preanalytical variability.

A possible confounding factor in our study for those analytes largely affected by renal clearance, such as BUN and phosphorus, is that while dogs had free access to water throughout the study period, they might have been hypodipsic in‐hospital. The USG in the majority of dogs was ≥1.035, which could support dehydration; however, timing of sample collection was not standardized because it relied on collection of voided urine.

Postprandial glucose concentrations differed from baseline based on TEa in 17 dogs at 31 postprandial timepoints; however, most values remained within the RI. The highest blood glucose at any point in this study was 130 mg/dL and occurred 2‐hours postprandially in a dog with a fasting glucose of 96 mg/dL. Hyperglycemia of this magnitude is unlikely to affect clinical assessment of dogs. In our dogs, it is possible that greater postprandial fluctuations in glucose occurred before the first postprandial blood collection at 2‐hours and were therefore missed. However, in clinically healthy Beagles, postprandial glucose is not significantly different from prefeeding values at 5, 10, 20, 30 45, 60, 90, and 120 minutes. ²⁹ In dogs undergoing chemotherapy, serum glucose concentrations are unaffected by feeding. ³⁰ Obese dogs have greater postprandial blood glucose, triglyceride, and insulin concentrations compared to lean dogs. ⁸ Body condition scores were not recorded in our study, so we were unable to determine whether obesity affected changes in glucose concentrations postprandially. A potential limitation of our study was that the time between sample clotting, centrifugation, and removal of the serum ranged from 30 minutes to 2 hours. While this parallels common procedure in clinical practice, glycolysis can result in decreases in measured glucose concentrations, so it has been recommended that serum‐clot contact time not exceed 1 hour for canine samples. ³¹

In 9% of the dogs in our study, total bilirubin increased in excess of the TEa after feeding. However, because the TEa of 30% was applied to bilirubin values reported to 1 decimal place, and typical values are low (RI 0‐0.3 mg/dL), almost any change was in excess of TEa. For example, a change from 0.1 to 0.2 mg/dL exceeds TEa, although this change is unlikely to be clinically meaningful.

To parallel dogs that would be screened in a veterinary hospital, we included a heterogenous sample of dogs assessed as being clinically healthy by their owners, including dogs on long‐term medications, provided there had been no change to the medication regimen or dosage in the 6 weeks before enrollment. One limitation of our inclusion criteria is that it is possible some dogs had medical conditions that could affect postprandial responses. Additionally, dogs received their regular diet rather than a standardized diet in order to simulate a typical hospital setting, although variability in the formulation of the diets could have altered some results. There are variable postprandial changes in BUN, creatinine, glucose, and phosphorus in response to feeding different types of foods. ² , ³ , ⁵ , ³² , ³³ , ³⁴ , ³⁵ For our study, owners were instructed to withhold food after 10:00 pm the night before testing, resulting in variable fasting times depending on the dog's feeding schedule. One dog that was fed 1 meal per day in the mornings had food withheld for over 25 hours, while other dogs had food withheld for as little as 10 hours. It is possible that variable duration of fasting affected our results. Withholding food for 8 hours or less affects lipid concentrations when compared to dogs fasted ≥12 hours. ³⁶

Another potential limitation is that only dogs weighing 15 kg or more were included due to the need for repeated blood collection. Small breed dogs, such Miniature Schnauzers, Shetland Sheepdogs, and Beagles with known breed‐related dyslipidemias, were therefore excluded, which might have biased our findings.

Care was taken to prevent preanalytical variability by standardizing sampling techniques and attempting to limit hemolysis. Of 500 samples, 19% had 1‐2+ hemolysis, 35% had 1‐2+ lipemia, and 2% had 3‐4+ lipemia. Based on instructions for the analyzer and a validation study for SDMA (Patch D, Obare E, Xie H, et al. High throughput immunoassay that correlates to gold standard liquid chromatography‐mass spectrometry (LC‐MS) assay for the chronic kidney disease (CKD) marker symmetric dimethylarginine (SDMA) [abstract]. In: 2015 J Vet Intern Med 2015; 29 (4):1216), 1‐2+ hemolysis and lipemia do not influence the measured biochemical values; however, 3‐4+ hemolysis might decrease total bilirubin by 25‐50% and 4+ hemolysis might decrease SDMA by 10‐25%; however, none of our samples had 3‐4+ hemolysis. Instructions indicated that 3‐4+ lipemia might decrease direct bilirubin, but should not impact total bilirubin. Overall, no analytes measured in this study were stated to be changed by the degree of hemolysis or lipemia observed.

In instances where TEa or RCV was exceeded postprandially, changes rarely exceeded the upper boundary of the RI in this sample of clinically stable dogs, and thus, the changes were unlikely to alter test interpretation in a clinical setting. It should be noted that RIs are derived from animals withheld from food, so comparison to values from postprandial animals is likely to show more results outside of the reported RI. Ideally, especially for analytes with high interindividual variation and low intraindividual variation, a subject‐based RI or RCV would be used rather than a population‐based RI. However, individual animals often lack baseline blood work for comparison, so clinicians frequently use the RI in their assessment of biochemical analytes. In this study, comparison of the measurements to RI is meant to highlight how infrequently most values exceeded the RI postprandially, while comparison against the TEa or RCV were used to help determine whether changes observed might be attributed to biological variability (ie, feeding) if they exceed the standard allowable analytical error. While the population‐based reference interval is less sensitive for detecting true change compared to subject‐based metrics, knowledge of how analytes are affected postprandially in the context of the reference interval is important as it might be the only basis for comparison in many animals.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF‐LABEL ANTIMICROBIAL DECLARATION

Authors declare no off‐label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by the Texas A&M University Animal Ethics committee, number 2019‐0369 CA.

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.

Supporting information

Data S1: Supporting Information

Yi KC, Heseltine JC, Jeffery ND, Cook AK, Nabity MB. Effect of withholding food versus feeding on creatinine, symmetric dimethylarginine, cholesterol, triglycerides, and other biochemical analytes in 100 healthy dogs. J Vet Intern Med. 2023;37(2):626‐634. doi: 10.1111/jvim.16630