Plasma Arginine Vasopressin and Serum Copeptin Concentrations at Rest and After Osmotic‐ and Non‐Osmotic‐Based Stimulation Tests in Dogs

Authors: Mathieu Victor Paulin, Thomas Schermerhorn, Dorsa Mehrabanpour, Suraj Unniappan, Elisabeth Snead

Abstract

Background

Serum copeptin (sCoP) is used as a surrogate for plasma arginine vasopressin (pAVP) measurement in humans.

Objective

To measure pAVP and sCoP at rest and after osmotic‐ and non‐osmotic stimulation testing in dogs.

Animals

Eight young castrated/spayed healthy research Beagles, eight young intact dogs, and eight old neutered healthy client‐owned dogs.

Methods

In this prospective longitudinal study, pAVP and sCoP were measured under iso‐(baseline), hypo‐(water load followed by intravenous administration of desmopressin [WLT]), and hyper‐(water deprivation test [WDT]) osmolar conditions assessed by measured plasma osmolality (pOsm~(m)~), and after administration of arginine (AST), IV, and Bovril (BST), PO. The fraction of change (F) in a variable y (e.g., pAVP) between baseline (T 0) and a timepoint X (T ~ X ~) during testing was defined as Fy = [y (T ~ X ~) – y (T 0)]/y (T 0).

Results

Baseline sCoP had wide inter‐individual variations. Mean [range] FpAVP and FsCoP at the end of WDT were +110% [+80; +142] and +18% [+0.4; +38] compared to baseline, respectively. Mean [range] FpAVP after water load and FsCoP after water load followed by desmopressin administration were −22% [−48; −0.5] and −29% [−39; −14] compared to baseline, respectively. Both FpAVP and FsCoP were strongly correlated to FpOsm~(m)~ (r = +0.76, p = 0.004; r = +0.78, p = 0.002; respectively). When sCoP was measured at T 4h instead of T 2h during WLT, to reflect its longer half‐life reported in humans, the correlation between FpAVPP800 and FsCoP became excellent (r = +0.90, p < 0.001). No stimulation of sCoP secretion occurred during AST or BST.

Conclusion and Clinical Importance

Serum CoP could be used as a surrogate for pAVP measurement in healthy dogs.

Introduction

Arginine vasopressin (AVP), also referred to as anti‐diuretic hormone (ADH), is a key peptide hormone in maintaining water homeostasis and vascular tone [1, 2, 3, 4]. However, plasma AVP (pAVP) measurement has never been implemented as a routine diagnostic test [5, 6]. This is in part due to AVP's short half‐life, its lability in isolated plasma (even when stored at −20°C) [7, 8], the need for protease inhibitors during collection, and poor assay sensitivity requiring extensive incubation and extraction steps [5, 8]. Furthermore, no AVP assay is commercially available in North America for veterinary clinical use. For these reasons, pAVP is seldom measured; instead, indirect evidence of AVP secretion and action through changes in urine and plasma osmolality after water deprivation or hypertonic saline administration has been relied on to differentiate between polyuria‐polydipsia (PUPD) disorders [1, 5, 9, 10].

In human medicine, serum copeptin (sCoP) measurement has mostly replaced pAVP measurement over the last decade [5, 6]. Copeptin, also referred to as “AVP‐associated glycopeptide”, is one of the three main end products of cleavage of the prepro‐AVP hormone, along with AVP and neurophysin II [6, 8, 11]. Copeptin is composed of 39 amino acids, comprises the C‐terminal part of the prepro‐AVP hormone, and is stoichiometrically secreted with AVP from the neurohypophysis in a 1 ratio [5, 6, 8]. Unlike AVP, CoP is extremely stable in plasma or serum ex vivo, and the use of protease inhibitors in the collection tube is not required [3, 5, 12]. Baseline CoP measurement, coupled with CoP‐based stimulation tests, such as hypertonic saline and arginine stimulation tests (AST), offers excellent diagnostic accuracy in humans for diagnosing diabetes insipidus and primary polydipsia, especially in complete forms of central or nephrogenic diabetes insipidus [6, 13, 14, 15, 16, 17, 18]. Oral administration of Bovril, a food supplement containing arginine, is as an alternative to intravenous administration of arginine for growth hormone stimulation testing; however, its use in AVP/copeptin stimulation testing remains undescribed [19].

Salivary and serum CoP measurements are reported in dogs in a single paper [20] and two abstracts [21, 22], all using research enzyme‐linked immunosorbent assay (ELISA) kits [20, 21, 22]. To the authors' knowledge, no studies have looked at sCoP concentrations under variable pOsm conditions. The aims of this study were (1) to investigate sCoP as a potential surrogate of pAVP measurement in healthy dogs, by measuring both peptides under hypo‐, iso‐, and hyper‐osmolar conditions and (2) to report sCoP measurement coupled with arginine stimulation testing after intravenous administration of L‐arginine hydrochloride and oral administration of the food supplement Bovril.

Materials and Methods

The experiment and animal care protocol were reviewed and approved by the Animal Care and Use Committee and the Institutional Review Board of the Western College of Veterinary Medicine, University of Saskatchewan, Canada (#20220024), and informed consent was obtained from all owners for inclusion of their dog in this study. This prospective longitudinal study was divided into four phases, as outlined in Figure 1.

Protocol Shared Between All Four Phases of the Study

Before enrollment, all dogs enrolled were deemed healthy based on the absence of abnormalities detected in the medical history, on physical examination findings, non‐invasive indirect blood pressure (NIBP) measurement, complete blood count (CBC), serum biochemistry, and urinalysis. CBC and serum biochemistry were performed by Prairie Diagnostic Services (PDS) provincial laboratory (Saskatoon, SK, Canada), using the ADVIA 2120i (Siemens Healthineers, Oakville, Ontario) and Cobas C311 (Roche Diagnostics, Laval, Quebec) analyzers, respectively. Manual evaluation of blood smears was conducted for all CBCs by a clinical pathologist or clinical pathology technician. Urinalysis was performed using a refractometer, urine dipstick (Chemstrip 9 urine test strips, Roche Diagnostics, Laval, Quebec) and urine cytology. NIBP was measured with the same high definition oscillometry blood pressure device (S + B medVET GmBH, Babenhausen, Germany) in all dogs and in accordance with the ACVIM guidelines [23, 24]. A single investigator recorded NIBP after a 15‐min period of acclimation in a quiet room and without physical or chemical restraints, using an appropriate size cuff and the same limb for each dog.

Dogs were cohoused in stable compatible groups when not subjected to testing, and individually housed in a cage during testing. Dogs were allowed ≥ 5 days without any interventions between the different osmotic and non‐osmotic tests. For all blood and urine collection, dogs were food‐deprived for 12–18 h before sampling, but water was available ad libitum. Blood was collected via cephalic, saphenous, or jugular venipuncture, while urine was collected by cystocentesis or urethral catheter placement. Plasma and urine osmolality (pOsm~(m)~ and uOsm~(m)~) were measured with a freezing point depression osmometer (FreezePoint, ELITechGroup Inc. Puteaux, France).

For pAVP measurement, blood samples were collected into chilled EDTA tubes both with (pAVPP800; BD Biosciences P800, NJ, USA, containing K2EDTA anticoagulant and a proprietary cocktail of protease, esterase and DPP‐IV inhibitors) and without (pAVPEDTA; BD Biosciences K2EDTA, NJ, USA) a cocktail of protease inhibitors at the timepoints outlined in Figure 1. All samples were kept on ice and centrifuged at 2000×g within 10 min of collection, with the harvested plasma stored at −80°C within 2 min of separation until analyses were performed. pAVP concentrations were measured with a multi‐species AVP ELISA (Arg8‐vasopressin ELISA, Arbor Assays, MI, USA) that included an extraction step, as reported and validated in a previous study in dogs [25]. According to the manufacturer's information, the assay has a sensitivity of 3.7 pg/mL with a detection range of 4 to 1000 pg/mL. All standards, controls, and samples were run in duplicate. The assay protocol was followed per the manufacturer's instructions and included an extraction step. For sCoP measurement, blood samples were collected into serum tubes (BD Vacutainer Plus, NJ, USA) with a clot activator and silicone‐coated interior. All samples were centrifuged at 2000×g 20 min after collection, with the harvested serum stored at −80°C until analyses were performed. Serum CoP concentrations were measured with a canine ELISA (Canine Copeptin competitive ELISA Kit, catalog number #MBS736743, MyBioSource.com, San Diego, USA), as previously validated [22] and reported [22, 26]. According to the manufacturer's product information, the assay has a sensitivity of 1.0 pg/mL with a detection range of 0 to 1000 pg/mL. All standards, controls, and samples were run in duplicate. The assay protocol was followed per the manufacturer's instructions. For both pAVP and sCoP, measurements were performed by the same investigators, the optical density was generated from each well using a plate reader capable of reading at 450 nm, and concentrations were calculated using the online platform MyAssays Ltd. 2021 and AssayFit.com, respectively. The mean intra‐assay coefficient of variation (CV) was determined by using all sCoP and pAVP measurements run in duplicate on a single ELISA plate. The mean inter‐assay CV was determined by using sCoP and pAVP measurements from an arbitrary dog repeated on each respective plate. Throughout the manuscript, the fraction of change (F) in a variable y (pAVPP800, sCoP, pOsm~(m), or uOsm(m)~) between baseline (T 0) and a timepoint X (T ~ X ~) during testing (referred to as the interval [T 0 − T ~ X ~]) was defined as Fy = [y (T ~ X ~) – y (T 0)]/y (T 0).

Phases of the Study

Phase 1 consisted of measuring baseline pAVP and sCoP in healthy dogs (Figure 1). For this purpose, three groups of healthy dogs were the young‐neutered group (four male and four female research Beagles, ≤ 4 years old), the young‐intact group (four male and four female client‐ or staff‐owned dogs, ≤ 4 years old), and the old‐neutered group (four male and four female client‐ or staff‐owned dogs, ≥ 8 years old). Serum and plasma samples were collected for measurement of baseline pAVPP800, pAVPEDTA and sCoP as described above.

Phase 2 consisted of measuring pAVP and sCoP under different pOsm~(m)~ (Figure 1) and included a water deprivation test (WDT) and a water load test (WLT), separated by ≥ 5 days (corresponding to the time needed by the investigators to complete the first test in all dogs). The WDT included eight young neutered research Beagles, while the WLT included the same individuals plus a young intact staff‐owned Border Terrier. All dogs were subjected to a 1‐week acclimation period before WDT. For the WDT, dogs were individually housed in cages with no access to food or water until the endpoint (T end) was T end was defined by either a ≥ 5% body weight loss or a 48‐h period of food and water withholding, whichever came first. A physical examination was performed every 6 h to monitor for the following adverse clinical pale mucosal membranes, capillary refill time > 2 s, weak femoral pulses, cold extremities, tachycardia, altered mentation; if any of these signs arose, medical intervention was started for rehydration, correction of hypovolemia, and the dog was excluded from the study. Blood (collected as previously described) and urine (via cystocentesis) samples were obtained at baseline (T 0) and T end for the measurement of packed cell volume (PCV), plasma total protein concentration ([TP]p), blood urea nitrogen (BUN), serum creatinine concentration ([Creatinine]s), serum electrolytes (sodium [Na^+^]s, potassium [K^+^]s, and chloride [Cl^−^]s) concentrations, pOsm~(m), uOsm(m), urine specific gravity (USG), pAVPP800~, pAVPEDTA, and sCoP. For the WLT, dogs were administered 44 mL/kg of warm tap water over 30 min (T 0 to T 30min) via a nasogastric tube (placed after topical proparacaine hydrochloride analgesia, and tube placement radiographically confirmed). Two hours later (T 2h), 0.3 μg/kg of desmopressin acetate (DDAVP 4 μg/mL, Ferring inc., North York, ON, Canada) diluted in 100 mL of 0.9% NaCl was administered intravenously over 20 min [27] as a continuous rate infusion (CRI) using a 5‐μm filter needle, in a cephalic vein which was then not sampled further. Physical examination and NIBP measurements were performed every 30 min for 4 h (until T 4h), and the following adverse clinical signs were tachypnea, systemic hypertension, nasal discharge, altered mentation, and peripheral edema; if any of these signs arose, medical intervention was started for correction of hypervolemia, and the dog was excluded from the study. Blood (collected as previously described) and urine (via indwelling urinary catheterization) samples were obtained at T 0, T 2h and T 4h for the measurement of PCV, TP, BUN, electrolytes ([Na^+^], [K^+^], [Cl^−^]), pOsm~(m), uOsm(m), USG, urine volume, pAVPP800~, AVPEDTA, and sCoP.

Phase 3 consisted of measuring pAVP and sCoP during AST, after intravenous administration of L‐arginine (AST) and oral administration of Bovril (BST) done 5 days apart (Figure 1). The AST included eight young neutered research Beagles plus a young intact Border Terrier, while the BST only included the eight research Beagles only. For the AST, 500 mg/kg of L‐arginine diluted 3 with 0.9% NaCl (L‐arginine hydrochloride 21%, Braun, Melsungen, Germany) was administered intravenously over 30 min (T 0 to T 30min) as previously reported in humans [14, 28]. For the BST, 14 g/m^2^ of Bovril paste (Unilever, London, UK) mixed with 40 mL of warm water was administered orally by spontaneous feeding or via nasogastric tube when the dog refused to eat the paste. For both tests, physical examination and NIBP measurement were performed every 30 min for 150 min (until T end = T 150min), and the following adverse clinical signs were hypersalivation, vomiting and diarrhea; if any of these signs was not self‐limiting, medical intervention was started and the dog was excluded from the study. Blood samples (collected as previously described) were obtained at T 0, T 60min, and T end, for the measurement of PCV, TP, pAVPP800, pAVPEDTA, and sCoP.

Phase 4 consisted of comparing pAVPP800 with pAVPEDTA, and in determining the feasibility of measuring sCoP with the human B·R·A·H·M·S KRYPTOR platform analyzer (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The former was conducted using all paired plasma samples collected throughout Phases 1, 2, and 3, as outlined in Figure 1. For the latter, 30 arbitrarily chosen serum samples left over from Phases 1, 2, and 3 were measured on the Human B·R·A·H·M·S KRYPTOR platform as a complimentary service trial provided by the Clinical Diagnostics Division at Thermo Fisher Scientific (Mississauga, ON, Canada).

Statistics

Statistical analyses and graphs were performed using a commercial software package (GraphPad PRISM10 software version 10.4.1, La Jolla, CA, USA). Data distributions were assessed by the Shapiro–Wilk test. The Cook‐Weisberg test and Levene's test were used to confirm or reject homoscedasticity. Sphericity was not assumed throughout the manuscript. For non‐matched measures with Gaussian distribution and equal standard deviations (SDs), an unpaired t‐test or an ordinary one‐way ANOVA test along with Tukey's multiple comparison test were performed. For paired measures with Gaussian distribution, a paired t‐test was used. For paired measures without Gaussian distribution, a Wilcoxon matched pairs signed rank test was used. For matched repeated measures with Gaussian distribution, a repeated‐measures one‐way ANOVA with the Geisser–Greenhouse correction was performed along with the Holm‐Šídák multiple comparison test and individual variances computed for each comparison. For matched repeated measures without Gaussian distribution, a Friedman test was performed along with the Dunn's multiple comparison test. Correlation was assessed using a Pearson correlation coefficient or a Spearman correlation coefficient for measures with and without Gaussian distribution, respectively. A p‐value of < 0.05 was used to determine statistical significance.

Results

Marked hemolysis of plasma or serum interfered with optical density readings and led to the exclusion of all samples of two young‐neutered dogs in the WDT and WLT, one young‐neutered dog in the AST and BVT, and one young‐intact and two old‐neutered dogs for sCoP measurement at rest. Aggression and overweight body condition were the most common causes for difficult sampling and subsequent iatrogenic hemolysis. The signalments of each group after exclusion of dogs with marked plasma or serum hemolysis are detailed in Table 1.

In our study, the mean intra‐assay and inter‐assay CV for the multispecies AVP ELISA (Arg8‐vasopressin ELISA, Arbor Assays, MI, USA) when used to measure AVP in canine plasma were 9.5% and 11.9% (n = 8 plates); respectively. The mean intra‐assay and inter‐assay CV for the canine CoP ELISA (MyBioSource.com, CA, USA) when used to measure CoP in canine serum were 3.5% and 6.0% (n = 8 plates); respectively.

During Phase 1 (Figure 1), mean baseline pAVPP800 was significantly higher in old neutered dogs (42.4 pg/mL, 24.4–56.2) than in young neutered dogs (24.7 pg/mL, 20.0–29.9; p = 0.006), while no significant difference was reported between old neutered and young intact dogs (30.8 pg/mL, range 22.0–46.3; p = 0.073) or between young intact and young neutered dogs (p = 0.418; Figure 2A). This difference was not appreciated when considering sCoP, as mean baseline sCoP was not statistically different between all three young intact (308 pg/mL, 223–455), young neutered (351 pg/mL, 183–755), and old neutered (365 pg/mL, 170–557; p = 0.766; Figure 2A). After including all dogs from these 3 groups, median pAVP was similar in male (29.0 pg/mL, 20.0–56.2) and female dogs (28.8 pg/mL, 23.9–54.8; p = 0.897); similarly, median sCoP was similar in male (327 pg/mL, 170–557) and female dogs (307 pg/mL, 223–755; p = 0.912). Dogs included in the osmotic and non‐osmotic stimulation tests had multiple baseline sCoP values available; therefore, for the latter analyses, baseline sCoP was determined by averaging all baseline sCoP values. The serial baseline sCoP values per dog measured before osmotic and non‐osmotic stimulation tests are represented in Figure 2B.

For the WDT (Phase 2, Figure 1), all six dogs reached the endpoint (T end): four dogs had a body weight loss ≥ 5%, while the two remaining dogs first reached a 48‐h period of food and water withholding. No adverse clinical signs were recorded during the WDT. All recorded variables are detailed in Table 2. After water deprivation (T end), and when compared to baseline (T 0), body weight (p = 0.0001) and [K^+^]s (p = 0.011) decreased significantly, while [creatinine]s, [Na^+^]s (p = 0.002), pOsm~(m)~ (p = 0.009; Figure 3A), and uOsm~(m)~ (p = 0.010; Figure 3B) all significantly increased.

For the WLT (Phase 2, Figure 1), a mean volume of 457 mL (range: 312–525) of warm tap water was orally administered over 20 min via nasogastric tube. Only one dog developed adverse clinical signs during WLT, which was characterized by shivering at T 2h30 (30 min after starting DDAVP IV CRI). All recorded variables are detailed in Table 3. After oral administration of tap water (T 2h) and compared to baseline, [Na^+^]s (p = 0.002) and [Cl^−^]s (p = 0.001) concentrations, pOsm~(m)~ (p = 0.010; Figure 4A), uOsm~(m)~ (p = 0.006; Figure 4B) and USG (p = 0.004; Figure 4C) all significantly decreased, while mean arterial pressure (MAP; p = 0.033) and urine volume (p = 0.001; Figure 4D) significantly increased. After IV administration of DDAVP (T 4h), median pOsm~(m)~ (p = 0.023) remained significantly lower than baseline, mean urine volume decreased (p = 0.0002), while both mean uOsm~(m)~ (p = 0.001) and USG (p = 0.004) significantly increased. At the end of the WLT (T 4h), PCV (p = 0.003), [TP]p (p = 0.002), and BUN (p = 0.023) had significantly decreased compared to baseline (T 0).

Mean FpAVPP800 and FsCoP at the end of WDT were +110% (range: +80 to +142) and +18% (range: +0.4 to +38) compared to baseline (T0), respectively. Conversely, the mean decrease in pAVPP800 after oral administration of tap water (WLT at T 2h) was −22% (range: −48 to −0.5), and the mean decrease in sCoP after oral administration of tap water and IV administration of DDAVP (WLT at T 4h) was −29% (range: −39 to −14) compared to baseline (T 0). Mean FpAVPP800 and FsCoP were both significantly different (p < 0.0001) between WDT and WLT (Figure 5).

During osmotic stimulations, the correlation between pAVPP800 and pOsm~(m)~ was positive and strong (Spearman r ~ s ~ = + 0.71, p < 0.0001; Figure 6A), and the correlation between pAVPP800 and uOsm~(m)~ was positive and moderate (Pearson r = + 0.56, p = 0.0027; Figure 6B). Regarding sCoP, the correlation between sCoP and pOsm~(m)~ was positive and fair (r = + 0.40, p = 0.045; Figure 6A), while there was no correlation between sCoP and uOsm~(m)~ (p = 0.153; Figure 6B). FpAVPP800 was positively and strongly correlated to FpOsm~(m)~ (r ~ s ~ = + 0.76, p = 0.004) and FuOsm~(m)~ (r ~ s ~ = + 0.76, p = 0.004; Figure 7A,B). Similarly, FsCoP was positively and strongly correlated to FpOsm~(m)~ (r = + 0.78, p = 0.002) and FuOsm~(m)~ (r = + 0.84, p = 0.0009; Figure 7C,D).

When considering all available paired pAVPP800 and sCoP samples obtained from Phase 1 (baseline) and 2 (osmotic stimulation testing; n = 39), the correlation between pAVPP800 and sCoP was positive but only moderate (r = + 0.40, p = 0.011; Figure 8A). However, the correlation between FpAVPP800 and FsCoP was positive and strong (r = + 0.80, p = 0.001) between the intervals [T 0–T end] in the WDT and [T 0–T 2h] in the WLT. When the interval [T 0–T 2h] in the WLT was extended to [T 0–T 4h] for sCoP measurement, the correlation between FpAVPP800 and FsCoP was positive and excellent (r = + 0.90, p < 0.001), and even near perfect after exclusion of one outlier (r = + 0.96, p < 0.001; Figure 8B).

During the AST (Phase 3, Figure 1), 4/7 (57%) dogs developed adverse clinical signs, with nausea being the only sign (e.g., hypersalivation, licking lips, dry heaving) and only occurring at T 30min. The signs of nausea were self‐limiting and resolved within 5 min without medical intervention. All recorded variables are detailed in Table 4. When compared to baseline (T 0), both PCV (p = 0.004) and [TP]p (p = 0.038) significantly decreased after IV administration of arginine (T end = T 150min), while SBP and heart rate stayed unchanged. For the BVT (Phase 3, Figure 1), 5 dogs spontaneously ate the Bovril paste, while a nasogastric tube placement was necessary in the remaining 3 dogs. No clinical adverse effects were reported in any of the dogs. All recorded variables are detailed in Table 4. When compared to baseline (T 0), both PCV (p = 0.0007) and [TP]p (p = 0.001) significantly decreased after oral administration of Bovril (T end = T 150min) while SBP and heart rate remained unchanged.

During AST, median sCoP was 229 pg/mL (149–773) at T 0, 196 pg/mL (162–879) at T 30, 197 pg/mL (158–696) at T 60, 186 pg/mL (159–691) at T 90, 209 pg/mL (148–772) at T 120, and 191 pg/mL (161–935) at T 150. When compared to baseline, FsCoP for each interval [T 0–T 30]: −6% (range: −21 to +14; ns), [T 0–T 60]: −15% (range: −31 to +10; p = 0.020), [T 0–T 90] = −14% (range: −30 to +13; p = 0.020), [T 0–T 120]: −9% (range: −20 to 0) (ns), [T 0–T150]: −6% (range: −19 to +21; ns; Table 4, Figure 9A). For the BST, there was no significant difference between mean sCoP at any timepoints (Table 4, Figure 9B).

During Phase 4 (Figure 1), all available paired pAVPP800 and pAVPEDTA samples were considered (n = 39), and median pAVPP800 was significantly higher (26.3 pg/mL, 11.9–56.2) than pAVPEDTA (16.5 pg/mL, 2.3–41.7; p < 0.0001; Figure 10A). There was a positive but only fair correlation between pAVPP800 and pAVPEDTA (Pearson r = + 0.44, p = 0.020; Figure 10B). Furthermore, sCoP was measured in 30 arbitrarily chosen samples obtained from the aforementioned osmotic stimulation tests (Phase 2) using the human B·R·A·H·M·S KRYPTOR platform assay. The mean sCoP was 0.62 pmol/L (range: 0.27–0.92), and most results (17/30, 57%) were below the limit of detection of the assay (< 0.70 pmol/L).

Discussion

This study reported a wide basal inter‐individual variation of sCoP and provided grounds for the measurement of sCoP as a surrogate for pAVP under hypo‐, iso‐, and hyper‐osmolar conditions. No stimulation of sCoP secretion occurred during AST or BST.

The first noteworthy finding of this study was the wide inter‐individual variations of sCoP when serial baseline sCoP concentrations were measured. This finding led the authors to primarily consider fractions of change in sCoP (FsCoP) during osmotic and non‐osmotic stimulation tests as opposed to means or medians of sCoP values within a group at a given timepoint. During osmotic stimulation tests, all dogs had a decrease or increase of pOsm~(m)~ of ≥ 1% after WLT and WDT, respectively. As the synthesis of AVP is stimulated by an absolute change in pOsm greater than 1% perceived by hypothalamic osmoreceptors, we concluded that proper osmotic stimulation occurred [5, 8, 10]. Furthermore, the clinical and paraclinical changes observed during WLT and WDT were as water load led to a decrease in pOsm~(m), uOsm(m), and USG with a concurrent increase in urine volume, while water deprivation led to an increase in pOsm(m)~ and uOsm~(m)~ with a concurrent decrease in body weight. The use of IV DDAVP after oral administration of tap water maintained a significantly lower pOsm~(m)~ at T 4h compared to baseline (T 0), by increasing free water reabsorption [1]. This was illustrated by a decrease in urine volume and a concurrent increase in uOsm~(m)~ and USG. Mean FpAVPP800 and FsCoP at the end of WDT were +110% and +18% compared to baseline (T 0), respectively. Conversely, the mean decrease in pAVPP800 after oral administration of tap water (WLT at T 2h) was −22%, and the mean decrease in sCoP after oral administration of tap water and IV administration of DDAVP (WLT at T 4h) was −29% compared to baseline (T 0). The similar changes in pAVP and sCoP observed under different osmotic conditions constitute the first step for the validation of sCoP as a surrogate for pAVP measurement. In a human prospective study enrolling 24 healthy volunteers, sCoP significantly increased from 4.6 ± 1.7 pmol/L to 9.2 ± 5.2 pmol/L during WDT, and decreased from 6.2 ± 2.4 pmol/L to 2.4 ± 2.1 pmol/L after hypotonic saline IV CRI [29]. Furthermore, our study found a positive and strong correlation between pAVPP800 and pOsm~(m)~ (r = + 0.71) which was favorably compared to the moderate correlation between pOsm and AVP (r = 0.49) reported in one human study during a WLT and hypertonic saline test [30]. However, the latter study showed a stronger correlation for pOsm~(m)~ with CoP (r = 0.77) [30]. While the correlation between pOsm~(m)~ and sCoP seen in our study (r = +0.40) was lower, the strength of the correlation was hampered by the aforementioned wide inter‐individual variations of baseline sCoP. Thus, when fractions of change in sCoP, pAVPP800 and pOsm~(m)~ were considered during osmotic stimulation tests, both sCoP and pAVP were strongly correlated to FpOsm~(m)~ (r = +0.78 and r = +0.76), with sCoP performing slightly better than pAVP.

When measuring sCoP with the canine CoP ELISA from MyBioSource.com (CA, USA), the mean intra‐assay and inter‐assay CVs found in our study were 3.5% and 6.0%, respectively, which is comparable to a recent study conducted by the authors using the same assay (3.6% and 4.4%, respectively) [26], and better than in a previous abstract (intra‐assay CV at 14%, inter‐assay CV at 12%) [22]. While the three groups each initially included eight dogs, one to two dogs (depending on the phase of the study) were ultimately excluded from the analysis due to marked hemolysis of their serum which interfered with optical density readings of sCoP. The interference of marked hemolysis with optical density readings of sCoP was already observed in a previous study conducted by the authors [26]. In humans, plasma or serum CoP is routinely measured using the original sandwich immunoluminometric assay (ILMA) [12], or its automated immunofluorescent successor on the B·R·A·H·M·S KRYPTOR platform [6], and most CoP diagnostic cut‐off values have been developed and validated using these assays only [3, 31]. For these reasons, 30 arbitrarily chosen samples obtained from our CoP‐based osmotic stimulations were processed with the human B·R·A·H·M·S KRYPTOR assay. Unfortunately, the mean sCoP was below the limit of detection of the assay, which made this assay unsuitable for sCoP measurement in dogs in this study. Regarding pAVP measurement, mean pAVPEDTA was significantly lower than mean pAVPP800, which further confirms the results of a previous study but in a bigger population (39 vs. 7 paired samples) [26] and suggests an accelerated degradation of pAVP without the use of protease inhibitors in collection tubes [5, 8].

As sCoP is stoichiometrically secreted with AVP from the neurohypophysis in a 1 ratio [6, 8, 11], a correlation between pAVP and sCoP was expected. When considering all available paired pAVPP800 and sCoP samples obtained from Phase 1 and 2, the correlation between pAVPP800 and sCoP was only moderate (r = +0.40). However, after limiting the effect of the wide inter‐individual variations of baseline sCoP by considering FpAVPP800 and FsCoP, the correlation between FpAVPP800 and FsCoP was strong (r = +0.80), which is similar to what has been reported in humans (r = 0.80) [30]. When FpAVPP800 and FsCoP were reported in the WLT between the intervals [T 0–T 4h] and [T 0–T 2h], respectively, the correlation between FpAVPP800 and FsCoP was excellent (r = +0.90), or even near perfect after the exclusion of one outlier (r = +0.96). Due to concerns that IV DDAVP, a synthetic analogue of AVP [1], might interfere with pAVP measurement, pAVP was not measured at T 4h during WLT; however, the better correlation between FpAVPP800 and FsCoP obtained by considering an endpoint time twice as long for sCoP (T 4h) than for pAVP (T 2h) might corroborate the hypothesis that CoP has a longer half‐life than AVP. Indeed, human plasma CoP half‐life was reported to be approximately 2 times longer than pAVP half‐life [32]. This longer half‐life of sCoP might provide some advantages over pAVP due to the pulsatile secretion of the latter [33].

This study also aimed to investigate the effects of AST on CoP secretion. While arginine is commonly used as a stimulus for growth hormone secretion in children with stunted growth [34, 35, 36] and adults with traumatic brain injury [37], arginine in humans is a stimulus for AVP secretion from the neurohypophysis possibly via the L‐arginine–nitric oxide pathway [38]. Median arginine‐stimulated plasma CoP concentrations significantly increased in healthy adults and patients with primary polydipsia, and a cut‐off plasma CoP concentration of 3.8 pmol/L measured 60 min after the start of IV administration of arginine had a diagnostic accuracy of 93% to distinguish central diabetes insipidus from primary polydipsia [14]. While AST was safe and relatively well tolerated in our study, with the most common adverse effect being nausea in humans [14], sCoP conversely decreased at 60 min (mean decrease of −15%) and 90 min (mean decrease of −14%) after the start of IV administration of arginine. The protocol used in our study was comparable to the one used in the latter human study in terms of concentration of L‐arginine, route of administration, and dilution. The cause of these discrepant results remains unknown but might include a decrease in pOsm due to the 3 dilution of L‐arginine with 0.9% NaCl IV solution, which represented a mean total administered volume of 8.1 mL/kg, and a weaker stimulus of AVP secretion via the non‐osmotic pathway [1]. Similarly, no significant difference between mean sCoP or FsCoP was reported at any timepoints after oral administration of Bovril, a food supplement which contains arginine [19]. To the authors' knowledge, copeptin measurement has yet to be reported after Bovril administration in humans, but the authors were interested in exploring it as an alternative to IV administration of arginine, as it is a cheaper and more convenient option (it is a beef flavored food product). Furthermore, arginine for IV administration was challenging to acquire in Canada.

Limitations of the study first include the small sample size, further altered by the exclusion of dogs with marked plasma or serum hemolysis as outlined above, along with the heterogeneity of the groups including different breeds, client‐owned, staff‐owned, and purpose‐bred research dogs. In addition, the control of osmotic and non‐osmotic regulation loops of AVP secretion was suboptimal. Although water was available ad libitum before testing, water intake was not controlled or measured, and large variations of pOsm~(m), uOsm(m)~ and USG were present at the start of osmotic and non‐osmotic testing. The non‐osmotic regulation loops of AVP secretion (e.g., stress, pain, exercise, nausea) [10, 39, 40] were minimized as much as possible by including a 1‐week acclimation period before the experiments, reducing pain by collecting blood from the intravenous catheters when the blood flow was adequate, and by having a consistent feeding and exercise schedule in between tests. While the placement of a nasogastric tube might have caused stress, an increase in pAVP secretion would have underestimated and not overestimated our results. The effect of inconsistent sampling sites (cephalic, saphenous or jugular venipuncture) and inconsistent sampling time throughout the day on pAVP and sCoP is unknown. Regarding the latter, the authors hypothesized that the use of fraction of change (F) for a given variable as opposed to baseline measurements would have limited the effect of inconsistent sampling time, similar to other hormonal stimulation tests. It is also worthwhile to mention that in humans, CoP measurement can be reliably interpreted independently of the time point of sampling [41]. The differences in intestinal absorption between spontaneous feeding and administration via nasogastric tube of Bovril are also unknown.

By measuring pAVP and sCoP in healthy dogs under different osmotic conditions, this study supports sCoP as a surrogate for pAVP measurement. It highlights the wide baseline inter‐individual variability of sCoP, a strong FpAVP–FsCoP correlation, better performance of FsCoP to correlate with FpOsm and FuOsm than FpAVP, and encourages further assay development for clinical use in PUPD disorders.

Disclosure

The authors declare no off‐label use of antimicrobials.

Ethics Statement

The study protocol (#20220024) was reviewed and approved by the Animal Care and Use Committee and the Institutional Review Board of the University of Saskatchewan, Canada. The authors declare human ethics approval was not needed.

Conflicts of Interest

Some serum copeptin samples were generously performed by the Clinical Diagnostics Division at Thermo Fisher Scientific (Mississauga, ON, Canada) on their B·R·A·H·M·S KRYPTO platform analyzer (Thermo Fisher Scientific, Waltham, Massachusetts, USA).