The role of the HPA in maintaining homeostasis in humans and animals is well established; however, its role in the presence of systemic disease is less defined.1 In humans, studies2,3 show that SC concentrations increase in the acute phase of systemic disease. Conflicting evidence exists regarding whether SC concentration changes in humans have prognostic value. For instance, findings disagree on whether adverse outcome is associated with SC concentrations that are high2,4–11 versus low or at the lower reference limit,1 and other studies12,13 show no correlation between SC concentration and outcome.
Applicable translation of findings in human medicine to veterinary medicine is hampered by species differences, such as lower SC concentrations in dogs versus humans.14 Furthermore, the potential association between SC concentration and outcome has not been investigated as much in veterinary patients as it has in human patients. However, studies show that high SC concentrations are associated with nonsurvival in dogs with parvovirus infection15 or acute Babesia canis infection.16 A laboratory-based study17 in dogs with experimentally induced Staphylococcus aureus pneumonia similarly shows a correlation between SC concentration at 24 hours after onset of sepsis and nonsurvival to 96 hours.17 Conversely, a small prospective study18 in dogs admitted to an ICU shows no substantial correlation between any markers of HPA function and patient outcome.
The aim of the retrospective study presented here was to investigate whether SC concentration is a useful prognostic indicator for nonsurvival versus survival to hospital discharge in critically ill dogs. We hypothesized that SC concentrations would be higher in dogs that did not survive to hospital discharge, compared with those that were discharged alive, thereby conferring prognostic value for predicting outcome in terms of survival versus nonsurvival to hospital discharge.
Materials and Methods
Animals
In this retrospective study, the patient database of the University of Liverpool Small Animal Teaching Hospital was searched for records of all dogs that were admitted to the ICU for emergency or critical care and that had SC concentration measured with a CLA ≤ 3 days after admission between January 2010 and May 2018. Dogs were excluded if they did not initially require emergency or critical care, were referred for urgent care but not deemed critically ill (eg, dogs that had acute hemorrhagic diarrhea syndrome but were cardiovascularly stable) by the attending veterinarian, had received glucocorticoids or other medications known to affect the HPA (eg, ketoconazole, progestogens, or etomidate) by any route in the previous 6 weeks, had a concurrent condition (eg, hypercortisolism or hypoadrenocorticism) that could affect the HPA, or did not have SC concentration measured with a CLA within the first 3 days of hospitalization. If the suitability of including a dog was unclear, the authors reviewed the dog's medical history and decided by consensus for inclusion or exclusion. The study was approved by the University of Liverpool Veterinary Research Ethics Committee (VREC575).
Data collection
Data collected from the medical records included signalment, SC concentration (μg/dL), presence of sepsis (yes or no), and survival to hospital discharge (yes or no). Dogs were classified as having sepsis if they met the most recent human criteria for sepsis.19 Sepsis was defined as “organ dysfunction caused by a dysregulated host response to infection.”19 Organ dysfunction was defined as an increase in SOFA score20 ≥ 2 during hospitalization. An infection was determined present when the medical record evidenced a confirmed final diagnosis of an infectious process (whether primary [eg, pyelonephritis] or perceived to have substantially contributed to the dog's critical illness even if not the primary complaint [eg, aspiration pneumonia secondary to an esophageal foreign body]) or an attending veterinarian's strong suspicion of infection (eg, severe pneumonia) that could not be confirmed because the dog was too unstable to undergo confirmatory testing at the time. The final diagnosis recorded in the dog's clinical notes was that given by the attending veterinarian. While hospitalized, dogs were managed by interns, residents, or board-certified veterinarians, and for each dog managed by an intern or resident, a board-certified veterinarian supervised the care and confirmed the final diagnosis. On the basis of information in the medical records and with use of the modified Glasgow coma scale,21 a SOFA score was calculated retrospectively for each dog. When medical records lacked sufficient information on individual parameters for scoring, those parameters were awarded a zero score. Dogs were grouped according to whether they died before hospital discharge (nonsurvivors) or were discharged alive (survivors), without attention to the duration of hospitalization. Dogs were then further considered on the basis of whether sepsis was or was not present and whether their CLA-measured SC concentration was or was not above the upper reference limit (5.8 μg/dL).
Cortisol CLA
Serum cortisol concentration was measured with a cortisol CLA.a The use of a cortisol CLA has been previously validated in dogs.14,22,23 Intra-assay coefficients of variation range from 10% to 55% and are concentration dependent, with increased precision at higher concentrations.22 The standard operating procedure at the facility during the study period was that blood samples intended for cortisol CLA were collected in serum tubes with or without a serum gel separator and then stored at room temperature until tested, typically ≤ 4 hours after collection.
Statistical analysis
An a priori sample size was calculated with statistical software.b To our knowledge, no veterinary study provided sufficient information to perform a sample size calculation for the investigation we intended; therefore, a comparative study6 looking at SC concentrations and nonsurvival in human patients with septic shock was used as guidance. We determined that a sample size of 94 dogs, with 47 dogs in each group (survivors vs nonsurvivors), was required (α = 0.05; β = 0.8; d = 0.6).
Data were retrieved from hospital recordsc and entered into a spreadsheet.d Statistical analysis was performed with available software.e Results were reported as mean ± SD or as median and IQR. Normality was assessed with the Kolmogorov-Smirnov test and visualization of Q-Q plots. To compare results for baseline group characteristics, the independent t test or Mann-Whitney U test was used for continuous variables and the χ2 test was used for categorical variables. Comparisons of SC concentration results for survivors versus nonsurvivors as well as for dogs with versus without sepsis and for other variables of interest were performed with the Mann-Whitney U test. The effect size was calculated with the formula:
where r = effect size (small, r = 0.10 to < 0.30; medium, r = 0.30 to < 0.50; and large, r = > 0.50), Z = score derived from the Mann-Whitney U test, and n = the number of samples.
Univariate analysis to identify a potential relationship between SC concentration and nonsurvival to hospital discharge was performed with binary logistic regression, and ORs were calculated. A ROC curve was created by plotting the sensitivity (true-positive rate) versus 1 - specificity (false-positive rate) to assess the accuracy of SC concentration in predicting nonsurvival to hospital discharge, and the Youden index was used to find the optimum SC concentration cutoff. The predictive value of SC concentration for nonsurvival to hospital discharge was assessed (OR, sensitivity, and specificity) for cutoffs determined from the ROC curve (optimal cutoff determined with the Youden index) or upper reference limit (5.8 μg/dL [160 nmol/L]). Significance was set at P < 0.05.
Results
Animals
Our search of the medical records identified 2,507 dogs that had been admitted to the hospital and had SC concentrations measured between January 2010 and May 2018. Of these, 2,278 were excluded because they did not require emergency care (n = 2,002), had received glucocorticoids or other potentially interfering medication (79), had or were suspected of having hypercortisolism (65), had incomplete medical records (62), had hypoadrenocorticism (23), had intervertebral disk disease (22), or either did not have SC concentration measured ≤ 3 days after hospital admission (17) or had it measured by means other than CLA (8). Of the 229 dogs that met the inclusion criteria, all had initially been triaged and referred by other veterinarians, 184 survived to hospital discharge (survivors), and 45 did not survive to hospital discharge (nonsurvivors). Of the 45 nonsurvivors, 34 were euthanized and 11 died. Thus, the overall mortality rate during the study period was 20% (45/229), with 15% (34/229) from euthanasia and 5% (11/229) from death. The exact reason for euthanasia was not documented in any dog.
Dogs in the nonsurvivor group included 5 Labrador Retrievers; 3 each of Cocker Spaniel, English Springer Spaniel, and mixed-breed dog; 2 each of Boxer, Cavalier King Charles Spaniel, Flat-Coated Retriever, and Pug; and 1 each of 23 other breeds. Dogs in the survivor group included 33 mixed-breed dogs; 17 Labrador Retrievers; 8 each of Cavalier King Charles Spaniel and Cocker Spaniel; 7 each of Border Collie, Boxer, English Springer Spaniel, and German Shepherd Dog; 6 Yorkshire Terriers; 5 each of Border Terrier and Golden Retriever; 4 each of Beagle, Dogue de Bordeaux, Rottweiler, Shih Tzu, Siberian Husky, and Staffordshire Bull Terrier; 3 each of Chihuahua, Jack Russell Terrier, Lhasa Apso, Miniature Schnauzer, Pug, Standard Poodle, Weimaraner, and West Highland White Terrier; 2 each of Akita, Greyhound, Lurcher, Newfoundland, and Pomeranian; and 1 each of 6 other breeds.
The nonsurvivor group consisted of 28 males (11 castrated and 17 sexually intact) and 17 females (12 spayed and 5 sexually intact); the survivor group consisted of 108 males (66 castrated and 42 sexually intact) and 76 females (65 spayed and 11 sexually intact; Table 1). Although no difference in survival to hospital discharge was detected for male versus female dogs in general or for females that were sexually intact versus spayed, sexually intact males had greater odds (OR = 2.4; 95% CI, 1.2 to 4.6; P = 0.038) for nonsurvival to hospital discharge than did castrated males. Although median age did not differ significantly (P = 0.46) between survivors (6.0 years; IQR, 3.0 to 10.0 years) and nonsurvivors (7.0 years; IQR, 4.5 to 10.0 years), the median age was older in neutered dogs (spayed females and castrated males; 7.0 years; IQR, 4.5 to 10.0 years; P < 0.001) than in sexually intact dogs (4.0 years; IQR, 1.0 to 8.0 years). Median age also differed significantly (P = 0.006 and P < 0.001, respectively) for spayed females (7.0 years; IQR, 4.0 to 10.0 years) versus sexually intact females (4.0 years; IQR, 0.60 to 6.0 years) and for castrated males (8.0 years; IQR, 5.0 to 10.5 years) versus sexually intact males (4.0 years; IQR, 1.0 to 8.0 years).
Results of analysis to identify potential factors associated with survival versus nonsurvival to hospital discharge in critically ill dogs that had SC concentrations measured with a CLA ≤ 3 days after admission between January 2010 and May 2018.
| Variable | Nonsurvivors (n = 45) | Survivors (n = 184) | P value |
|---|---|---|---|
| Age (y)* | 7.0 (4.5–10.0) | 6.0 (3.0–10.0) | 0.46 |
| Weight (kg)* | 20.0 (11.0–29.6) | 19.2 (9.0–31.6) | 0.82 |
| Sex | 0.67 | ||
| Male | 28 | 108 | |
| Female | 17 | 76 | |
| Neuter status | 0.01 | ||
| Sexually intact | 22 | 53 | |
| Castrated or spayed | 23 | 131 | |
| No. (%) of dogs with sepsis | 10 (22) | 25 (14) | 0.15 |
| SC concentration (μg/dL)* | 8.5 (4.8–11.8) | 4.5 (2.5–6.9) | < 0.001 |
| No. (%) of dogs with SC concentration > 5.8 μg/dL | 30 (67) | 66 (36) | < 0.001 |
Reported as the median and IQR.
Median body weight was not substantially different between survivors (19.2 kg [42.2 lb]; IQR, 9.0 to 31.6 kg [19.8 to 69.5 lb]) and nonsurvivors (20.0 [44.0 lb]; IQR, 11.0 to 29.6 kg [24.2 to 65.1 lb]; Table 1). Similarly, the proportion of dogs that met the criteria for sepsis was not meaningfully different between survivors (25/184 [14%]) and nonsurvivors (10/45 [22%]).
SC concentration
Results for SC concentrations were nonnormally distributed, and the overall median SC concentration was 5.0 μg/dL (138 nmol/L; IQR, 2.7 to 8.2 μg/dL [74 to 226 nmol/L]; reference interval, 2.0 to 5.8 μg/dL [55 to 160 nmol/L]). Ninety-six of the 229 (42%) dogs had a SC concentration > 5.8 μg/dL, and the proportion of dogs with high SC concentration was significantly (P < 0.001) greater for nonsurvivors (30/45 [67%]), compared with survivors (66/184 [36%]; Table 1). In addition, the median SC concentration was significantly (P < 0.001) higher for nonsurvivors (8.5 μg/dL [234 nmol/L]; IQR, 4.8 to 11.8 μg/dL [132 to 326 nmol/L]) versus survivors (4.5 μg/dL [123 nmol/L]; IQR, 2.5 to 6.9 μg/dL [69 to 190 nmol/L]; Figure 1). The effect size was medium (r = 0.30). There was a 1.11 increase in the odds of nonsurvival (OR = 1.11; 95% CI, 1.05 to 1.17; P < 0.001) for every 1 μg/dL increase in SC concentration.
Box-and-whisker plots of SC concentration measured within 3 days of hospitalization in 229 critically ill dogs treated between January 2010 and May 2018 and grouped by whether dogs did not survive to hospital discharge (nonsurvivors; n = 45) or were discharged alive (survivors; 184). Each box represents the IQR, the central horizontal line in the box represents the median, whiskers represent the data points farthest from the median that are not outliers (ie, that are within 1.5 times the IQR of the first and third quartiles), and diamonds represent the maximum and minimum values. *Median SC concentration differed significantly (P < 0.001) for survivors versus nonsurvivors.
Citation: Journal of the American Veterinary Medical Association 256, 9; 10.2460/javma.256.9.1034
When SC concentration was evaluated for dogs grouped by those with versus without sepsis, the median SC concentration was significantly (P < 0.001) higher for dogs with sepsis (9.2 μg/dL [253 nmol/L]; IQR, 6.1 to 17.6 μg/dL [168 to 486 nmol/L]) versus those without sepsis (4.5 μg/dL [124 nmol/L]; IQR, 2.6 to 7.0 μg/dL [72 to 193 nmol/L]; Figure 2). The effect size was medium (r = 0.3).
Box-and-whisker plots of SC concentration in the dogs in Figure 1 grouped by those with (n = 35) versus without (194) sepsis. *Median SC concentration differed significantly (P < 0.001) for dogs with versus without sepsis. See Figure 1 for key.
Citation: Journal of the American Veterinary Medical Association 256, 9; 10.2460/javma.256.9.1034
SC concentration as a predictive indicator for nonsurvival to hospital discharge
The area under the ROC curve was 0.72 (95% CI, 0.64 to 0.81; Figure 3), which indicated that SC concentration had good predictive value for nonsurvival to hospital discharge in critically ill dogs. With use of the Youden index, the SC concentration of 7.6 μg/dL (210 nmol/L) was identified as the optimum cutoff, at which the OR, sensitivity, and specificity for nonsurvival to hospital discharge were 5.4 (95% CI, 2.7 to 10.9), 58%, and 80%, respectively. Alternatively, when the upper reference limit (SC concentration, 5.8 μg/dL) was used as the cutoff, the OR, sensitivity, and specificity for nonsurvival were 3.6 (95% CI, 1.8 to 7.1), 67%, and 64%, respectively.
Results of ROC curve analysis to determine the predictive value of SC concentration for nonsurvival to hospital discharge in the critically ill dogs described in Figures 1 and 2. The area under the ROC curve is 0.72 (95% CI, 0.64 to 0.80)
Citation: Journal of the American Veterinary Medical Association 256, 9; 10.2460/javma.256.9.1034
Discussion
Results of the present study indicated that the median SC concentration in critically ill dogs was higher in those that did not survive to hospital discharge, compared with those that were discharged alive. This finding supported our hypothesis and was consistent with studies in dogs with parvoviral diarrhea,15 B canis rossi babesiosis,16 or S aureus pneumonia17 that show associations between high SC concentration and nonsurvival in affected dogs. However, the mortality rate in the present study (20% [45/229]) differed from that in the study16 involving dogs with B canis infection (7% [7/95]), likely because of differences in underlying disease processes in dogs evaluated. However, the mortality rate of the present study did not have a comparative finding in the study17 involving dogs with S aureus pneumonia because of the terminal design of that previous study. Another difference was that the SC concentration was > 5.8 μg/dL for all 7 dogs that died in the Bcanis study, 16 compared with 30 of the 45 (67%) nonsurvivors in the present study.
In contrast, our finding of an association between high SC concentration and nonsurvival to hospital discharge in critically ill dogs was inconsistent with a study18 that shows no substantial difference in SC concentration for dogs that did versus did not survive to hospital discharge. The reason for the discrepancy was not clear but could have been because of a difference in inclusion criteria that resulted in differences between the groups of dogs evaluated. Although the proportion of critically ill dogs with SC concentrations > 5.8 μg/dL in the present study (42% [96/229]) and the previous study18 (37% [19/52]) were similar, the mortality rate was lower in our study (20% [45/229]), compared with the previous study (35% [7/52]).
The area under the ROC curve was 0.72, which suggested a good overall predictive value of SC concentration for nonsurvival to hospital discharge in critically ill dogs of the present study. This finding was similar to the area under the curve (0.62) reported in 2 prospective human studies5,10 involving patients with septic shock. The optimal cutoff calculated from the ROC curve was 7.6 μg/dL (210 nmol/L), at which the OR, sensitivity, and specificity were 5.4, 58%, and 80%, respectively. These findings mirrored those of studies5,6,13 in human patients with septic shock that show high SC concentration is associated (OR, 1.8 to 7.89) with nonsurvival. When the SC concentration cutoff in the present study was lowered to the upper reference limit (5.8 μg/dL), the OR and specificity for nonsurvival also decreased to 3.6 and 64%, consistent with findings in the human studies,5,6,13 and the sensitivity for nonsurvival in our study increased to 67%.
The cause of SC concentration increase during acute, severe illness is likely multifactorial. It is presumed to be in part the result of increased activity of the HPA itself, resulting from catecholamine release,24 proinflammatory cytokines,25 and ACTH-independent adrenal steroidogenesis, among other processes.1,3,24,26 However, non-HPA mediated factors are also believed to be involved, such as those resulting from reduced cortisol binding proteins1,27,28 and cortisol metabolism.3 Thus, it could have been that the difference in SC concentration between survivors and nonsurvivors in our study was truly a pathophysiologic observation, reflecting differences in disease severity and physiologic stress, which our study was not designed to assess.
An unexpected finding in the present study was the greater odds of nonsurvival for sexually intact male dogs, compared with those that had been neutered. This finding could have reflected a greater proportion of certain diseases more commonly seen in young animals (eg, parvoviral enteritis) or a relatively lower tolerance to critical illness in younger dogs because sexually intact male dogs were notably younger than neutered male dogs in the present study. However, a similar association between neuter status and nonsurvival was not detected in female dogs of the present study, suggesting that factors other than age may have been present. Alternatively, this could have been a type I statistical error.
The main limitation of the present study was its retrospective nature. It is not standard practice to measure SC concentration in every critically ill patient at our hospital, and consequently our selection criteria resulted in selection bias toward dogs with suspected HPA dysfunction. This was made apparent during data collection when hypercortisolism and hypoadrenocorticism were identified frequently among differential diagnoses noted in the medical records of dogs with neurologic and gastrointestinal signs, respectively. Therefore, applying the results of the present study to all critically ill dogs must be done cautiously. Relatedly, avoidance of misapplication of results was the principal reason why we did not calculate positive and negative predictive values and believe that a prospective study would better determine the prognostic value of SC concentration in critical illness in dogs.
Aside from illnesses, many other factors can alter SC concentrations in dogs. Potential interpatient differences include, but are not limited to, breed,29 weight,30 sex,31 stage of the estrous cycle,32 activity levels,33 stress,34 recent surgery,35 hypoglycemia,36 and opioid or benzodiazepine administration.37,38 In addition, SC concentration may also be impacted by potential differences in measurement methods (eg, sample timing3,17,27 and handling39) and inter- and intra-assay variation.14,22,23,40 The relatively large number of dogs in the present study may have mitigated differences in group characteristics; however, the influence of sample timing must be highlighted because studies2,11,27 in humans show temporal changes in SC concentrations during the acute phase of illness, with an initial peak concentration followed by a return to baseline in the subsequent 24 to 72 hours. In dogs, one study17 shows a similar pattern, whereas another study15 shows that SC concentrations only correlated with survival 48 and 72 hours after onset of illness but not at 24 hours after onset.15 This temporal pattern may reflect the short half-life of cortisol in dogs41,42 and humans43,44 and has implications for the present study. Although SC concentration was measured ≤ 3 days after admission to the ICU in the present study, time between disease onset and hospitalization likely varied. The higher median SC concentration in the nonsurvivor group may have been the result of those dogs having been deemed less stable by the referring veterinarians, having been referred more quickly, and having had their SC concentration measured earlier in the disease course, compared with dogs in the survivor group. In a nonexperimental study, such as ours, it is difficult to compensate for differences in sample timing relative to onset of critical illness, especially in a referral population. However, the use of a hospital population, rather than an experimental population, allows for more appropriate application of the results of the present study to other hospital populations. An important caveat to highlight is that all dogs in the present study were referred, which may limit the application of results to primary care practices.
In addition, relative differences in cortisol metabolism between dogs that did and did not survive to hospital discharge in the present study could have influenced SC concentration results. Cortisol primarily undergoes hepatic metabolism, with a lesser degree of renal metabolism, and both processes may be reduced in the presence of severe systemic illness.3 Thus, dogs in the present study with concurrent hepatic or renal dysfunction may have had an additive effect on the already reduced cortisol metabolism of systemic illness. However, studies27,45–47 also show a resultant increase in total cortisol concentrations with hepatic dysfunction. The implication for our study was that potential greater organ dysfunction in nonsurvivors, compared with survivors, may have led to lower cortisol metabolism and resulted in higher SC concentration; however, the number of dogs with affected organs assessed with the SOFA score were too small to provide any evidence for this association. Further, the influence of euthanasia on evaluation of survival versus nonsurvival to hospital discharge must be considered, especially in a retrospective study such as ours in which the reason for euthanasia was unclear.
Results indicated that SC concentration could be used as part of an overall assessment for prognosis in critically ill dogs. We believe that the explanation for the difference in SC concentrations between dogs that did and did not survive to hospital discharge in the present study was most likely multifactorial and that further studies to investigate the relationship and explore interventions that may improve survival to hospital discharge in critically ill dogs are warranted.
Acknowledgments
The study was performed at the Institute of Veterinary Science, University of Liverpool, Neston, England.
No third-party funding or support was received in connection with the present study or the writing or publication of the manuscript. The authors declare that there were no conflicts of interest.
Extralabel use of antimicrobials may have occurred with dogs of the present study, but such use was not discussed in the manuscript.
ABBREVIATIONS
| CI | Confidence interval |
| CLA | Chemiluminescent immunoassay |
| HPA | Hypothalamic pituitary axis |
| ICU | Intensive care unit |
| IQR | Interquartile (25th to 75th percentile) range |
| ROC | Receiver operating characteristic |
| SC | Serum cortisol |
| SOFA | Sequential organ failure assessment score |
Footnotes
Immulite 2000 Cortisol, Siemens Healthcare GmbH, Erlangen, Germany.
G*Power, version 3.1.9.2, G*Power Team, Düsseldorf, Germany.
Tristan, version 1.8.3.110, Orion Engineering Services, Aberdeen, Scotland.
Microsoft Excel, version 2016, Microsoft Corp, Redmond, Wash.
SPSS Statistics, version 22.0, IBM Corp, Armonk, NY.
References
1. Annane D, Pastores SM, Rochwerg B, et al. Guidelines for the diagnosis and management of critical illness-related corticosteroid insufficiency (CIRCI) in critically ill patients (Part I): Society of Critical Care Medicine (SCCM) and European Society of Intensive Care Medicine (ESICM) 2017. Crit Care Med 2017;45:2078–2088.
2. Widmer IE, Puder JJ, König C, et al. Cortisol response in relation to the severity of stress and illness. J Clin Endocrinol Metab 2005;90:4579–4586.
3. Boonen E, Vervenne H, Meersseman P, et al. Reduced cortisol metabolism during critical illness. N Engl J Med 2013;368:1477–1488.
4. Rothwell PM, Lawler P. Prediction of outcome in intensive care patients using endocrine parameters. Crit Care Med 1995;23:78–83.
5. Annane D, Sébille V, Troché G, et al. A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin. JAMA 2000;283:1038–1045.
6. Bollaert PE, Fieux F, Charpentier C, et al. Baseline cortisol levels, cortisol response to corticotropin, and prognosis in late septic shock. Shock 2003;19:13–15.
7. Sam S, Corbridge TC, Mokhlesi B, et al. Cortisol levels and mortality in severe sepsis. Clin Endocrinol (Oxf) 2004;60:29–35.
8. Ho JT, Al-Musalhi H, Chapman MJ, et al. Septic shock and sepsis: a comparison of total and free plasma cortisol levels. J Clin Endocrinol Metab 2006;91:105–114.
9. Lipiner-Friedman D, Sprung CL, Laterre PF, et al. Adrenal function in sepsis: the retrospective Corticus cohort study. Crit Care Med 2007;35:1012–1018.
10. Riché FC, Boutron CM, Valleur P, et al. Adrenal response in patients with septic shock of abdominal origin: relationship to survival. Intensive Care Med 2007;33:1761–1766.
11. Boonen E, Meersseman P, Vervenne H, et al. Reduced nocturnal ACTH-driven cortisol secretion during critical illness. Am J Physiol Endocrinol Metab 2014;306:E883–E892.
12. Schein RM, Sprung C, Marcial E, et al. Plasma cortisol levels in patients with septic shock. Crit Care Med 1990;18:259–263.
13. Rady MY, Johnson DJ, Patel B, et al. Cortisol levels and corticosteroid administration fail to predict mortality in critical illness: the confounding effects of organ dysfunction and sex. Arch Surg 2005;140:661–668.
14. Reimers TJ, Salerno VJ, Lamb SV. Validation and application of solid-phase chemiluminescent immunoassays for diagnosis of endocrine diseases in animals. Comp Haematol Int 1996;6:170–175.
15. Schoeman JP, Goddard A, Herrtage ME. Serum cortisol and thyroxine concentrations as predictors of death in critically ill puppies. J Am Vet Med Assoc 2007;231:1534–1539.
16. Schoeman JP, Rees P, Herrtage ME. Endocrine predictors of mortality in canine babesiosis caused by Babesia canis rossi. Vet Parasitol 2007;148:75–82.
17. Cortés-Puch I, Hicks CW, Sun J, et al. Hypothalamic-pituitaryadrenal axis in lethal canine Staphylococcus aureus pneumonia. Am J Physiol Endocrinol Metab 2014;307:E994–E1008.
18. Prittie JE, Barton LJ, Peterson ME, et al. Pituitary ACTH and adrenocortical secretion in critically ill dogs. J Am Vet Med Assoc 2002;220:615–619.
19. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA 2016;315:801–810.
20. Vincent JL, Moreno R, Takala J, et al. The SOFA (sepsisrelated organ failure assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22:707–710.
21. Platt SR, Radaelli S, McDonnell J. The prognostic value of the modified Glasgow coma scale in head trauma in dogs. J Vet Intern Med 2001;15:581–584.
22. Singh AK, Jiang Y, White T, et al. Validation of nonradioactive chemiluminescent immunoassay methods for the analysis of thyroxine and cortisol in blood samples obtained from dogs, cats, and horses. J Vet Diagn Invest 1997;9:261–268.
23. Russell NJ, Foster S, Clark P, et al. Comparison of radioimmunoassay and chemiluminescent assay methods to estimate canine blood cortisol concentrations. Aust Vet J 2007;85:487–494.
24. Bornstein SR, Chrousos GP. Clinical review 104: adrenocorticotropin (ACTH)- and non-ACTH-mediated regulation of the adrenal cortex: neural and immune inputs. J Clin Endocrinol Metab 1999;84:1729–1736.
25. Dunn AJ. Cytokine activation of the HPA axis. Ann N Y Acad Sci 2000;917:608–617.
26. Polito A, Sonneville R, Guidoux C, et al. Changes in CRH and ACTH synthesis during experimental and human septic shock. PLoS One 2011;6:e25905.
27. Tsai M-H, Peng Y-S, Chen Y-C, et al. Adrenal insufficiency in patients with cirrhosis, severe sepsis and septic shock. Hepatology 2006;43:673–681.
28. Chan WL, Carrell RW, Zhou A, et al. How changes in affinity of corticosteroid-binding globulin modulate free cortisol concentration. J Clin Endocrinol Metab 2013;98:3315–3322.
29. Höglund K, Lequarré AS, Ljungvall I, et al. Effect of breed on plasma endothelin-1 concentration, plasma renin activity, and serum cortisol concentration in healthy dogs. J Vet Intern Med 2016;30:566–573.
30. Reimers TJ, Lawler DF, Sutaria PM, et al. Effects of age, sex, and body size on serum concentrations of thyroid and adrenocortical hormones in dogs. Am J Vet Res 1990;51:454–457.
31. Frank LA, Rohrbach BW, Bailey EM, et al. Steroid hormone concentration profiles in healthy intact and neutered dogs before and after cosyntropin administration. Domest Anim Endocrinol 2003;24:43–57.
32. Gallelli MF, Monachesi N, Miceli DD, et al. Plasma ACTH, α-MSH and cortisol variations in the dog during the oestrous cycle in different photoperiods. Vet Med (Praha) 2015;60:567–577.
33. Kolezvská J, Brunclík V, Svoboda M. Circadian rhythm of cortisol secretion in dogs of different daily activities. Acta Vet Brno 2003;72:599–605.
34. Beerda B, Schilder MBH, Van Hooff JARAM, et al. Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs. Appl Anim Behav Sci 1998;58:365–381.
35. Church DB, Nicholson AI, Ilkiw JE, et al. Effect of non-adrenal illness, anaesthesia and surgery on plasma cortisol concentrations in dogs. Res Vet Sci 1994;56:129–131.
36. Beerda B, Schilder MB, Janssen NS, et al. The use of saliva cortisol, urinary cortisol, and catecholamine measurements for a noninvasive assessment of stress responses in dogs. Horm Behav 1996;30:272–279.
37. Thomson I, Fraser R, Kenyon CJ. Regulation of adrenocortical steroidogenesis by benzodiazepines. J Steroid Biochem Mol Biol 1995;53:75–79.
38. Vuong C, Van Uum SH, O'Dell LE, et al. The effects of opioids and opioid analogs on animal and human endocrine systems. Endocr Rev 2010;31:98–132.
39. Behrend EN, Kemppainen RJ, Young DW. Effect of storage conditions on cortisol, total thyroxine, and free thyroxine concentrations in serum and plasma of dogs. J Am Vet Med Assoc 1998;212:1564–1568.
40. European Society of Veterinary Endocrinology Quality Committee. ESVE veterinary endocrinology external quality assessment scheme. ESVE report release number 010. Nottingham, England: ESVE, 2017.
41. Kuipers F, Ely RS, Hughes ER, et al. Metabolism of steroids. II. Half-Life of various steroids in dogs. Proc Soc Exp Biol Med 1957;95:187–189.
42. Thomasson B, Steenburg RW. Plasma clearance of cortisol and 11-deoxycortisol in dogs. Am J Physiol 1965;208:84–89.
43. Hindmarsh PC, Charmandari E. Variation in absorption and half-life of hydrocortisone influence plasma cortisol concentrations. Clin Endocrinol (Oxf) 2015;82:557–561.
44. Keenan DM, Roelfsema F, Veldhuis JD. Endogenous ACTH concentration-dependent drive of pulsatile cortisol secretion in the human. Am J Physiol Endocrinol Metab 2004;287:E652–E661.
45. Peterson RE. Adrenocortical steroid metabolism and adrenal cortical function in liver disease. J Clin Invest 1960;39:320–331.
46. Zumoff B, Bradlow HL, Gallagher TF, et al. Cortisol metabolism in cirrhosis. J Clin Invest 1967;46:1735–1743.
47. Tan T, Chang L, Woodward A, et al. Characterising adrenal function using directly measured plasma free cortisol in stable severe liver disease. J Hepatol 2010;53:841–848.
