<em>Journal of Veterinary Internal Medicine</em> | ACVIM Journal | Wiley Online Library (2025)

1 INTRODUCTION

Synthetic glucocorticoids are commonly used in both human and veterinary medicine for the treatment of inflammatory and immune-mediated disorders.^{1, 2} Polyuria and polydipsia (PUPD) are common adverse effects of glucocorticoid treatment in dogs, and appear to be dose-related.³ Arginine vasopressin (AVP), or anti-diuretic hormone (ADH), is a key peptide hormone involved in maintaining fluid balance and vascular tone, and plays a central role in many disorders associated with PUPD.⁴ In dogs, proposed mechanisms of PUPD secondary to glucocorticoid excess include AVP deficiency and AVP resistance.⁵ In a study in dogs, a single dose of dexamethasone given IM (approximately 0.1 mg/kg) caused significant blunting of the AVP response to hypertonic saline.⁶ A recent study also reported that use of desmopressin acetate (DDAVP) at the dose of 5 μg SC q12h decreased PUPD in dogs during short-term prednisolone administration (1 mg/kg/day PO for 12 days).⁷ Endogenous glucocorticoid excess also was shown to cause marked impairment of osmoregulation and AVP secretion in dogs with pituitary-dependent hypercortisolism and hyperfunctioning adrenocortical tumors.⁵ To our knowledge, basal plasma AVP (pAVP) concentrations have not been reported in dogs treated long-term with glucocorticoids.

The use of protease inhibitors during collection is essential when measuring AVP, and samples must be kept on ice and centrifuged promptly, with the resulting plasma frozen immediately at −80°C.^{8, 9} Previous publications reporting pAVP measurement in dogs have used various protease inhibitors, including aprotinin, D-phenylalanine-argininechloromethketone, benzamidine, leupeptin hemisulfate, and DL-thiorphan.^{10, 11} To our knowledge, the use of BD Biosciences P800 (New Jersey, US) tubes has not been reported in dogs. These tubes contain spray-dried dipotassium ethylene diamine tetraacetic acid (K₂-EDTA) anticoagulant and proprietary protease inhibitor additives, and they have been designed for the preservation of hormones prone to enzymatic cleavage and resulting loss of function, including glucagon-like peptide-1, glucagon, non-acyl ghrelin, gastric inhibitory polypeptide, and oxyntomodulin, according to the manufacturer's description.¹²

Measurement of plasma or serum copeptin (sCoP) has replaced the measurement of pAVP in humans for diagnosing PUPD disorders, especially for distinguishing between complete forms of central or nephrogenic diabetes insipidus and primary polydipsia.⁹ Copeptin, also known as AVP-associated glycopeptide,^{8, 13} is a 39-amino acid glycopeptide that consists of the C-terminal part of the 164 amino acid AVP prohormone.^{8, 9, 13} Copeptin is secreted from the neurohypophysis in a 1 : 1 ratio with AVP.^{9, 13, 14} Unlike AVP, human CoP is extremely stable in plasma or serum ex vivo, and no extraction step or other preanalytical procedures such as addition of protease inhibitors are needed.⁸ To our knowledge, sCoP measurement in dogs has only been reported in 2 abstracts: the first validated a copeptin enzyme-linked immunosorbent assay (ELISA) designed for dogs (MyBioSource, San Diego, US) in diabetic and non-diabetic dogs,¹⁵ and the second validated a copeptin ELISA designed for humans (Phoenix Pharmaceutical Inc, Burlingame, US) in healthy dogs participating in a voluntary blood donor program.¹⁶ Saliva CoP also has been measured in dogs with separation anxiety-related behavioral problems before and after a short separation test, using a copeptin ELISA designed for dogs (BlueGene Biotech, Shanghai, China).¹⁷

We aimed to evaluate the effect of a long-term immunosuppressive dosage of orally administered prednisolone on pAVP and sCoP concentrations in healthy adult research dogs. We hypothesized that both pAVP and sCoP concentrations would decrease during PO prednisolone administration. A secondary aim was to compare pAVP measurements when collected into chilled EDTA tubes with or without the addition of a combination of protease, esterase, and dipeptidyl peptidase 4 inhibitors (BD Biosciences P800). We hypothesized that the BD Biosciences P800 tubes would provide a suitable and convenient option for pAVP storage and measurement when compared to chilled EDTA tubes without the addition of protease inhibitors.

2 MATERIALS AND METHODS

3 RESULTS

All 8 dogs received the intended 35-day course of PO prednisolone at the mean ± SD immunosuppressive dosage of 2.6 ± 0.13 mg/kg/day (range, 2.35-2.75 mg/kg/day) which represented a mean ± SD total dose of 25 ± 5.3 mg/day (range, 20-35 mg/day), and completed the prospective longitudinal study as described (Figure1). However, 1 dog was excluded from the analyses because serum and plasma samples had marked hemolysis, which interfered with optical density readings; the remaining 7 dogs were included in the analyses. All dogs were 40 months old (3.3 years old), 4 were male and 3 were female, and the mean ± SD weight was 9.5 ± 1.8 kg (range, 7.9-14.6 kg). No adverse effects of prednisolone administration were noted throughout the experiment, including after abrupt discontinuation of the prednisolone.

The difference (Δ) between post- and pre-ACTHST cortisol concentration (Δ_cortisol) was significantly lower after the 35-day PO course of prednisolone (D49, mean ± SD Δ_cortisol = 1.8 ± 0.9 μg/dL, range, 0.7-3.4) compared with baseline (D13, mean ± SD Δ_cortisol = 12.8 ± 3.5 μg/dL, range, 9.8-19.9; P = .001). Two weeks after discontinuation of PO prednisolone, mean Δ_cortisol was still significantly lower (D63, mean ± SD Δ_cortisol = 8.9 ± 3.2 μg/dL, range, 5.1-14.0) than baseline (P = .04). However, no significant difference was found between mean Δ_cortisol at 4 weeks after discontinuation of PO prednisolone (D77, mean ± SD Δ_cortisol = 9.7 ± 2.4 μg/dL, range, 7.4-14.6) compared to baseline, and all dogs had post-ACTH cortisol concentration >5.5 μg/dL, suggesting that the pituitary-adrenal axis had recovered after 4 weeks (Figure2).

The mean intra-assay coefficient of variation for the multispecies AVP ELISA (Arg8-vasopressin ELISA, Arbor Assays, Michigan, US) when used to measure pAVP concentration in canine plasma was 7.0%. A single ELISA plate was used for plasma AVP measurement and therefore inter-assay coefficient of variation was not assessed. For the canine CoP ELISA (MyBioSource.com, California, US), the mean intra-assay coefficient of variation was 3.6%, and the inter-assay coefficient of variation was 4.4%. Mean ± SD pAVP_P800 concentration was significantly lower after the 35-day course of PO prednisolone (D49, 25.8 ± 8.1 pg/mL, range, 15.6-38.3) compared with baseline (D13, 34.1 ± 5.4 pg/mL, range, 26.6-43.4; P = .02; Figure3A). A single arbitrarily chosen dog had pAVP measured at 2 additional time points: 2 weeks (D63) and 4 weeks (D77) after prednisolone discontinuation. Plasma AVP_P800 at D77 (22.7 pg/mL) was persistently lower than at D13 (33.3 pg/mL), despite having been off prednisolone for 4 weeks (Figure3B). Median sCoP concentration was also significantly lower after the 35-day course of PO prednisolone (D49, 166 pg/mL, range, 131-223) compared with baseline (D13, 243 pg/mL, range, 157-336; P = .02; Figure3C,D). The decrease in pAVP_P800 and sCoP between D13 and D49 occurred despite unchanged mean ± SD systolic blood pressure measurements (D13, 141 ± 13 mm Hg, range, 122-158; D49, 141 ± 11 mm Hg, range, 126-162 mm Hg; P = 1), and a further decrease in mean ± SD systolic blood pressure occurred 4 weeks after prednisolone discontinuation (D77, 129 ± 14 mm Hg, range, 111-151; P = .004; Figure4).

Mean ± SD calculated pOsm (pOsm_calculated) significantly increased, albeit minimally, after the 35-day course of PO prednisolone (D49, 306.1 ± 1.8 mOsm/kg, range, 303.5-308.8) compared with baseline (D13, 303.3 ± 1.8 mOsm/kg, range, 300.9-305.8; P = .04), before returning to baseline by D77 (304.8 ± 2.8 mOsm/kg, range, 301.6-309.4; not significant; Figure5A). Conversely, mean ± SD USG significantly decreased after the 35-day course of PO prednisolone (D49, 1.018 ± 0.010; range, 1.006-1.040) compared with baseline (1.041 ± 0.012, range, 1.020-1.051; P = .01) before returning to baseline by D77 (1.042 ± 0.010, range, 1.021-1.056; not significant; Figure5B).

The correlation between pAVP_P800 and sCoP was positive (Pearsoncoefficient, r = .77; P = .001; Figure6A). No correlation was found between pAVP_P800 and calculated pOsm (P = .64) or between sCoP and calculated pOsm (P = .36; Figure6B). However, a positive correlation was found between pAVP_P800 and USG (Pearson coefficient, r = .61; P = .02) although a correlation was not found between sCoP and USG (P = .25; Figure6C).

For paired plasma samples at baseline (D13), mean ± SD pAVP_EDTA was significantly lower (5.0 ± 2.5 pg/mL, range, <4.0-9.0 pg/mL) than mean pAVP_P800 (34.1 ± 5.4 pg/mL, range, 26.6-43.4 pg/mL; P < .001), including 3 of 7 EDTA samples without protease inhibitors measuring below the limit of detection of the assay (4 pg/mL; Figure7). No correlation was found between pAVP_P800 and pAVP_EDTA (P = .86).

4 DISCUSSION

We showed that pAVP and sCoP both decreased in response to a long-term immunosuppressive PO course of prednisolone, and reported the use of BD Biosciences P800 tubes containing a proprietary combination of protease, esterase, and dipeptidyl peptidase 4 inhibitors, for the measurement of pAVP.

Routine measurement of pAVP has not been used as a diagnostic test for patient care in either human or veterinary medicine because of concerns regarding the methodologic reliability of laboratory assays.^{8, 13} This situation likely explains the paucity of data regarding the pathophysiology of water regulation in dogs. The in vivo half-life of AVP is very short and has been reported to be <30 minutes in humans and <6 minutes in dogs.^23-25 This short half-life allows for rapid changes in pAVP in response to osmotic and non-osmotic stimuli but significantly hampers its detection even when samples are processed quickly and stored at −20°C.²⁴ Ultimately, AVP is inactivated by endopepitases that are mainly located in plasma, the kidneys, and the liver.^{4, 13} Intact AVP and its metabolites then undergo renal clearance.^{4, 13}

Plasma osmolality significantly increased by approximately 1% by the end of the 35-day course of PO prednisolone. Although this increase is minimal, a subsequent increase in AVP secretion would be expected in healthy dogs in response to increases in pOsm of as little as 1%, especially in the absence of an increase in systemic blood pressure.^{4, 26, 27} Conversely, pAVP significantly decreased by the end of the 35-day PO course of prednisolone, with no correlation found between pAVP and pOsm. Therefore, our results suggested that exogenous glucocorticoid treatment at the mean immunosuppressive dosage of 2.6 mg/kg/day (range, 2.35-2.75 mg/kg/day) induced marked impairment in the normal osmoregulatory control of AVP secretion. This observation correlates with what was reported previously in dogs after a single IM injection of dexamethasone (approximately 0.1 mg/kg IM),⁶ and in humans after a 5-day PO course of prednisolone (anti-inflammatory dose of 30 mg/day).²⁸ In the former study including 5 mongrel dogs, AVP response to hypertonic saline stimulation testing was significantly blunted when performed 1 day after IM administration of dexamethasone.⁶ In the second study, 7 healthy men were subjected to water deprivation testing before and after PO prednisolone administration, and AVP secretion was attenuated despite a documented increase in pOsm with prednisolone treatment.²⁸ Similarly, impairment of the osmoregulation of AVP secretion has been shown with endogenous glucocorticoid excess in dogs with hypercortisolism.⁵ In this study, 9 dogs with pituitary-dependent hypercortisolism and 6 dogs with adrenal-dependent hypercortisolism underwent a hypertonic saline stimulation test and the sensitivity of AVP secretion in response to increasing plasma hyperosmolality was found to be blunted in 9 dogs, with complete absence of a response to hypertonicity in 4 of them.⁵

A direct effect of glucocorticoids on renal responsiveness to AVP also has been speculated in dogs but remains unproven.⁵ Vasopressin regulates the water permeability of the renal distal tubules and collecting ducts by increasing the insertion of aquaporin-2 (AQP2) channels into the apical cell membrane.^{4, 29} In our study, the increase in pAVP was correlated with an increase in USG, which suggests at least partially preserved renal responsiveness to AVP. However, the role of other mechanisms involved in water regulation, such as plasma atrial natriuretic peptide (which tends to increase glomerular filtration rate and inhibit sodium and chloride reabsorption at various levels of the nephron) cannot be excluded.^{30, 31} In humans, plasma atrial natriuretic peptide was increased with exogenous administration of methylprednisolone and spontaneous hypercortisolism,^{30, 31} but similar findings were not obtained in dogs with spontaneous hypercortisolism.³²

Chronic exogenous glucocorticoid administration leads to suppression of the hypothalamic-pituitary-adrenal (HPA) axis through negative feedback, resulting in decreased production of corticotropin-releasing hormone (CRH), ACTH and endogenous cortisol.³³ Atrophy of the adrenal glands also has been reported in dogs receiving chronic glucocorticoids.³⁴ In humans, AVP acts synergistically with CRH to stimulate the release of ACTH from the adenohypophysis.^{28, 35, 36} This pathway is stimulated by non-specific somatic stressors and constitutes a backup system for the release of ACTH.^35-37 In our study, an arbitrarily chosen dog had pAVP_P800 measured at 4 timepoints, and pAVP_P800 measured at 4 weeks after prednisolone discontinuation (D77, 22.7 pg/mL) was persistently lower than at baseline (D13, 33.3 pg/mL), despite apparent recovery of the pituitary-adrenal axis as defined by a non-significant difference of Δ_cortisol between D77 and baseline. When all 7 dogs were considered, the median sCoP was similarly decreased (albeit not significantly) 4 weeks after prednisolone discontinuation (D77, 203 pg/mL) compared with baseline (D13, 243 pg/mL), also suggesting partial hypothalamic recovery. Interestingly, mean USG significantly decreased after the 35-day course of PO prednisolone but returned to baseline by 4 weeks after prednisolone discontinuation. This capacity to restore urinary concentrating ability despite decreased pAVP and sCoP concentrations may be explained by other compensatory mechanisms such as the renin-angiotensin-aldosterone system.

In humans, plasma or serum CoP are routinely measured using the original sandwich immunoluminometric assay (LIA),³⁸ or its automated immunofluorescent successor on the KRYPTOR platform.¹³ Various ELISAs are also available for research purposes but are not recommended for clinical studies in humans.¹³ Both LIA and KRYPTOR assays have good agreement but in 1 study, CoP measured by an ELISA assay correlated poorly with both the LIA and KRYPTOR assays.³⁹ Unfortunately, the only options currently available for veterinary use are ELISA assays. For the canine ELISA used in our study (Canine Copeptin competitive ELISA Kit, MyBioSource.com, San Diego, US), the mean intra-assay coefficient of variation was 3.6%, and the inter-assay coefficient of variation was 4.4%. In an abstract reporting sCoP concentrations in diabetic and non-diabetic dogs using the same ELISA, the inter-assay CV was 12% and the intra-assay CV was 14%.¹⁵ In humans, a close positive correlation between pAVP and plasma CoP concentrations (using a sandwich immunoluminometric assay) was found (Spearman's rank correlation coefficient, r_s = .80).²⁷ Similarly in our study, the correlation between pAVP and sCoP concentrations was positive (Pearson coefficient, r = .77), suggesting that sCoP may be a sensitive surrogate for pAVP measurement. In humans, both serum and plasma can be used for CoP sandwich immunoluminometric assays, and ex vivo CoP stability, defined by ≤20% loss of recovery, was shown in both plasma and serum for ≥7 days when samples were stored at room temperature and ≥14 days when stored at 4°C.³⁸ The decision to measure canine CoP in serum was made arbitrary and solely based on the routine and convenient use of serum in small animal medicine.

No reference interval for canine sCoP concentration has been established. We did not report reference intervals because doing so did not comply with the American Society for Veterinary Clinical Pathology (ASVCP) standards for the determination of reference intervals in veterinary species.⁴⁰ Because only 7 values of baseline sCoP were obtained in healthy dogs (off prednisolone) at a single timepoint, the reference change value (RCV), which defines the boundaries within which an analyte might be expected to vary in a healthy individual owing to analytical and biological variation alone, could not be calculated.^{40, 41} The only published sCoP concentrations are available in abstract form, using a copeptin ELISA designed for humans (Phoenix Pharmaceutical Inc, Burlingame, US) in 9 healthy dogs participating in a voluntary blood donor program.¹⁶ The mean sCoP concentration was 1670 pg/mL (range, 850-4090 pg/mL), which is significantly higher than the mean sCoP concentration obtained in our study at baseline (D13, 234 pg/mL, range, 157-336 pg/mL). Another abstract reported that CoP concentrations in dogs with naturally-occurring diabetes mellitus were higher when compared with non-diabetic dogs, but no values or reference intervals were provided.¹⁵ The reason behind the discrepancy between the range of sCoP measurements obtained in our study vs the aforementioned abstract is unknown, but may be explained by the different assays used. Reference intervals in healthy humans have been established in several studies, using the CoP sandwich immunoassay LIA: 17.5 (4.17-57.5) pg/mL,³⁸ 15.8 (1.83-184.6) pg/mL,⁴² and 20.8 (14.6-34.6) pg/mL.⁴³ The molecular mass of canine CoP is unknown, which prevents inter-species comparisons. Although CoP is released into the capillaries of the neurohypophysis in an equimolar amount to AVP,^{8, 13, 44} the mean pAVP concentration at baseline (D13) was 34.1 pg/mL, which was about 7 times lower than the mean sCoP concentration (234 pg/mL). A similar discrepancy is seen in humans where the reference interval for pAVP (0-5.9 pg/mL) is lower than for plasma CoP. This observation likely can be explained by the difference in in vivo half-lives of AVP (24 minutes) and CoP (approximately 48 minutes),⁴⁵ the lability of AVP in isolated plasma even when stored at −20°C,^{9, 46} and cumbersome current AVP assays that feature low sensitivity,⁹ extensive incubation and extraction steps.^{8, 9} Furthermore, AVP secretion has been shown to occur in a pulsatile manner in healthy dogs, although this is controversial in humans under basal conditions.⁴⁷

The mean intra-assay coefficient of variation for the multispecies AVP ELISA (Arg8-vasopressin ELISA, Arbor Assays, Michigan, US) was 7%, which is lower than previously reported in dogs.¹⁰ In paired samples, the mean pAVP concentration in EDTA tubes was significantly lower (5.0 pg/mL) than the mean pAVP concentration in BD Biosciences P800 tubes (34.1 pg/mL), and 43% of EDTA samples measured below the limit of detection of the assay. These results emphasize the importance of protease inhibitors to preserve the fragile AVP peptide, even when the harvested plasma is immediately stored at −80°C until analysis.^{8, 9} Unlike AVP, human CoP is extremely stable in plasma or serum ex vivo, and the use of protease inhibitors in collection tubes is not required in human medicine.⁸ It is unknown if canine CoP has similar stability.

Limitations of our study include small sample size and suboptimal control of osmotic and non-osmotic regulation loops of AVP secretion. Although water was available ad libitum throughout the study, water intake was not controlled or measured. The non-osmotic regulation loops of AVP secretion (eg, stress, pain, exercise, nausea)^35-37 were minimized as much as possible by including a 2-week acclimation period before the experiments, minimizing pain by collecting blood from IV catheters when possible, and by having a consistent feeding and exercise schedule. Reference intervals for pAVP and sCoP also were not determined in our study because it did not comply with the American Society for Veterinary Clinical Pathology (ASVCP) standards for the determination of reference intervals in veterinary species.

5 CONCLUSION

Chronic exogenous glucocorticoid excess induced marked impairment of the osmoregulation of AVP secretion, suggesting a state of partial central diabetes insipidus in dogs with iatrogenic hypercortisolism. Serum CoP, a glycopeptide consisting of the C-terminal part of the AVP prohormone and co-secreted from the neurohypophysis in a 1 : 1 ratio with AVP, was positively correlated with pAVP and may be a promising surrogate for pAVP measurement in dogs. Because most research into water metabolism is limited by the lability of AVP and serious issues regarding the methodologic reliability of laboratory assays, sCoP measurement may facilitate understanding of the pathophysiology of PUPD disorders in dogs. Alternatively, the use of BD Biosciences P800 tubes for plasma collection is an accessible and convenient commercial option, albeit costly, for pAVP measurements in dogs.

CONFLICT OF INTEREST DECLARATION

For the adrenocorticotropic hormone-stimulation tests, The synthetic adrenocorticotropic hormone (Tetracosactide CosACTHen) for the adrenocorticotropic hormone stimulation testingwas generously provided by Dechra Pharmaceuticals, Canada. Dechra Pharmaceuticals did not participate in study design, data collection and interpretation, or writing of the manuscript.

OFF-LABEL ANTIMICROBIAL DECLARATION

Authors declare no off-label use of antimicrobials.

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

The study protocol (#20220019) was approved by the Animal Care and Use Committee and Institutional Review Board of the Western College of Veterinary Medicine, University of Saskatchewan, Canada.

HUMAN ETHICS APPROVAL DECLARATION

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