• Open Access

Blood Arginine Vasopressin, Adrenocorticotropin Hormone, and Cortisol Concentrations at Admission in Septic and Critically Ill Foals and their Association with Survival

Authors


  • Previously presented in abstract form at the American College of Veterinary Internal Medicine (ACVIM) Forum, June 5–9, 2007.

Corresponding author: S.D.A. Hurcombe, Department of Veterinary Clinical Sciences, College of Veterinary Medicine, The Ohio State University, Columbus, OH 43210; e-mail: Samuel.Hurcombe@cvm.osu.edu.

Abstract

Background: Sepsis is an important cause for neonatal foal mortality. The hypothalamic-pituitary-adrenal axis (HPAA) responses to sepsis are well documented in critically ill humans, but limited data exist in foals. The purpose of this study was to evaluate the HPAA response to sepsis in foals, and to associate these endocrine changes with survival.

Hypothesis: Blood concentrations of arginine vasopressin (AVP), adrenocorticotropin hormone (ACTH), and cortisol will be higher in septic foals as compared with sick nonseptic and healthy foals. The magnitude of increase in hormone concentration will be negatively associated with survival.

Animals: Fifty-one septic, 29 sick nonseptic, and 31 healthy foals of ≤7 days of age were included.

Methods: Blood was collected at admission for analysis. Foals with positive blood culture or sepsis score ≥14 were considered septic. Foals admitted with disease other than sepsis and healthy foals were used as controls. AVP, ACTH, and cortisol concentrations were measured using validated immunoassays.

Results: AVP, ACTH, and cortisol concentrations were increased in septic foals. Septic nonsurvivor foals (n = 26/51) had higher plasma ACTH and AVP concentrations than did survivors (n = 25/51). Some septic foals had normal or low cortisol concentrations despite increased ACTH, suggesting relative adrenal insufficiency. AVP, ACTH, and cortisol concentrations were higher in sick nonseptic foals compared with healthy foals.

Conclusions and Clinical Importance: Increased plasma AVP and ACTH concentrations in septic foals were associated with mortality. Several septic foals had increased AVP : ACTH and ACTH : cortisol ratios, which indicates relative adenohypophyseal and adrenal insufficiency.

Neonatal septicemia is often cited as one of the leading causes of morbidity and mortality in foals during the first 7 days of life.1–11 Despite advancement of diagnostic and therapeutic modalities to treat septicemia, the reported survival rate still ranges between 10 and 70%,2,4,6 which indicates the importance of sepsis in foal health.

There are several reports describing prognostic indicators and predictors of survival in critically ill foals3,5,11–14 but few have investigated the value of endocrine variables in association with survival.15

The hypothalamic-pituitary-adrenal axis (HPAA) plays a vital role in maintaining water homeostasis as well as cardiovascular, immunologic, and metabolic functions.16 In health, arginine vasopressin (AVP) is released from the hypothalamus when increases in serum osmolality are detected, and acts to conserve water via action on V2 receptors in the renal collecting duct and induction of aquaporin-2 channels.17 AVP is also essential for cardiovascular homeostasis where its vasopressor abilities on select vascular beds are mediated by V1 receptors to maintain organ perfusion, especially during hypotension and shock states.17 In humans and horses, AVP is also a potent secretagogue of adrenocorticotropin hormone (ACTH) by its action on V3 receptors in the adenohypophysis.18 In early sepsis in humans, increases in the blood concentrations of AVP, ACTH, and cortisol are indicative of HPAA stimulation.19,20 Furthermore, overstimulation in response to severe or prolonged disease can be manifested by relative adrenal exhaustion or insufficiency (RAI) and AVP depletion, both of which are commonly reported in septic humans21–26 and negatively correlated with survival. These findings also provide the rationale for the use of vasopressin, terlipressin, and glucocorticoids in the face of catecholamine-refractory septic shock; however, this approach remains controversial.

Septicemia most often is related but not limited to Gram-negative bacteremia27 where endotoxin (lipopolysaccharide) is released into the blood leading to the production of proinflammatory cytokines and systemic inflammation. Endotoxin and proinflammatory cytokines, notably tumor necrosis factor (TNF)-α, have been detected in the blood of septic foals and correlate with the severity of disease.28–30 In humans, TNF-α, interleukin (IL)-1, and IL-6 also elicit prolonged activation of the HPAA at different levels,31,32 which could be upregulated by both Gram-negative and Gram-positive bacteria. There is controversy as to the point at which endotoxin acts on glucocorticoid secretion; however, endotoxin has been shown to increase ACTH, cortisol, corticotropin releasing hormone (CRH), and AVP in humans and horses, and appears to be a physiologic response to decrease systemic inflammation.32,33

Although the role and function of the HPAA have been investigated extensively in the human critical care literature, limited information exists in septic foals. Gold et al15 recently showed that septic foals were likely to have increases in ACTH concentration and ACTH : cortisol ratios, and these derangements were amplified in nonsurvivors, suggesting HPAA dysregulation at the level of the adrenal gland, where cortisol secretion was inappropriately low compared with the ACTH stimulus.

The purpose of this study was to determine the AVP, ACTH, and cortisol concentrations in septic foals and in foals with disease other than septicemia and compare them to healthy controls. In addition, we also sought to determine if relative vasopressin deficiency occurs in the acute stages of sepsis in foals by determining blood hormone ratios. We proposed that endocrine dysregulation of the HPAA occurs in foals and would be associated with severity of disease and survival.

Materials and Methods

Animals

Full-term foals ≤ 7 days old of both sexes and any breed admitted to The Ohio State University Veterinary Teaching Hospital (OSU) and Hagyard Equine Medicine Institute (HEMI) were included. Foals were then classified into one of the 2 groups: septic and sick, nonseptic foals. Foals in the septic group were defined as having a positive blood culture or a sepsis score ≥14.1,14 Sick nonseptic foals were those foals that presented for conditions other than septicemia, including meconium impaction, hypoxic ischemic encephalopathy, various orthopedic conditions, and complete or partial failure of transfer of passive immunity. These foals also had negative blood cultures and sepsis scores ≤10. A 3rd group of foals were used as healthy controls from various breeding farms, and were <72 hours old at the time of examination. They were considered healthy based on physical examination, had a normal CBC, serum biochemistry, and serum immunoglobulin G concentration, and had sepsis scores ≤5. Foals with a history of prior corticosteroid, hypertonic saline solution (HSS), or IV plasma administration within the 24 hours before admission were excluded from the study because these treatments may have altered the HPAA responses in these foals. Foals that had received prior isotonic crystalloid fluids, antimicrobials or PO immunoglobulin supplementation (eg, banked colostrum or plasma) were included for analysis.

Survival was defined as survival to discharge from the hospital. Nonsurvival was defined as death from progressively worsening disease or euthanasia based on a grave medical prognosis (ie, refractory to intensive medical therapy). Any foals that were electively euthanized for reasons related to owner financial limitations or personal decisions to not to proceed with treatment were not included in the study to avoid bias.

This study was approved by OSU executive committee and adheres to the principles for the humane treatment of animals in veterinary clinical investigations as stated by the American College of Veterinary Internal Medicine and National Institute of Health guidelines. Owner consent was obtained before inclusion in the study.

Data Collection

A complete history, including expected foaling date, maternal health during pregnancy, and administered medications, was obtained. Categorical variables assessed included age at presentation, breed, and sex. Physical examination findings were evaluated, including presence of a septic focus that included arthropathy, pneumonia, diarrhea, uveitis, or omphalitis. Variables analyzed and compared among all foals included clinical examination findings (heart rate, respiratory rate, temperature, mucous membrane color, capillary refill time, presence of cold extremities, peripheral pulse quality, mental status, thoracic and abdominal auscultation, ocular examination, and umbilical examination) and calculation of the sepsis score.14 Blood variables assessed included a CBC,a serum biochemistry,b blood l-lactate concentration,c blood fibrinogen concentration, serum immunoglobulin Gb concentration, blood culture results, and blood hormone concentrations (AVP, ACTH, and cortisol).

Calculated serum osmolality was determined in a subset of septic foals (n = 25) to determine the influence of osmotic mechanisms on AVP release. These foals were chosen based on having high AVP concentrations and a sepsis score ≥18.

Sampling

For hospitalized foals, whole blood was obtained by jugular venipuncture within 1 hour from the time of admission before any medical therapy and placed into plain serum clot tubes and chilled aprotinin-EDTA tubes. Aprotinin, a protease inhibitor, was added to preserve sample integrity (500 kU/mL whole blood in an EDTA blood tube). A total of 20 mL of venous blood was obtained for hormone assays, stored in ice water, and centrifuged within 12 hours at 5°C, 2000 rcf for 15 minutes. Serum and plasma were then aliquoted and stored at −80°C until analyzed. For healthy foals, blood samples were collected during a routine newborn foal physical examination at the farm and were not transported. These foals were <72 hours of age.

Hormone Assays

A solid phase, double antibody commercial radioimmunoassayd (RIA) was used to determine AVP concentrations after peptide extractione from plasma. To determine plasma ACTH concentrations, a human-specific immunoradiometric assayf was used. Serum cortisol concentrations were determined by a validated direct RIA.g,34

AVP Assay

A validated assay is not commercially available for the quantification of equine AVP in blood, but an indirect RIA has been validated to measure human AVP. According to amino acid sequence analysis, human and equine AVP are 100% homologous and theoretically the human assay should detect equine AVP. Using previously described methods in horses,35 we used a validated human AVP RIA in the determination of equine AVP by measuring putative blood AVP concentrations in healthy horses (n = 3) administered HSS. Two liters of HSS were administered IV over 10 minutes to induce hyperosmolality (mean calculated osmolality 300 mOsm/L) and resultant AVP release into systemic circulation. Serial blood samples were obtained and analyzed for AVP concentration. Sample handling methods were exactly the same as those used for the foals in this study. We found that median plasma AVP concentrations increased after HSS from 1.8 pmol/L (baseline) to a peak of 59 pmol/L at 10 min and gradually decreased over 4 hours to baseline concentrations. The baseline AVP concentrations in these horses were similar to those reported previously in other studies.36,35

Statistical Analyses

Data were summarized by calculating descriptive statistics and graphically represented when possible using software programs.h,i,j Frequency distributions of categorical variables were evaluated, and means, medians, standard errors of the mean, and ranges were calculated for continuous variables. Continuous variables were further categorized to facilitate analysis and determine crude odds ratios and 95% confidence intervals based on the ranges and interquartile values. The dependent variable was survival, yes or no. All variables were tested using logistic regression (procgenmodi). All variables were screened and any variables with a P value ≤ .25 were tested in a forward stepwise multivariate logistic regression to determine a final model. Variables that resulted in a P value < .05 were retained in the model.

All data were assessed for normality by the Shapiro-Wilk statistic. Appropriate parametric and nonparametric testing was performed depending on the distribution of the data. Comparisons among all groups were assessed using a one-way analysis of variance or Kruskal-Wallis statistic with posthoc Dunn's multiple comparison testing. Comparisons between survivors and nonsurvivors within each group of foals were assessed using Student's t-test or Mann-Whitney U statistic. A comparison of hormone concentration between blood culture-positive and blood culture-negative septic foals also was assessed. Significance was set as P < .05. Values are recorded as median and range, unless otherwise stipulated.

Results

Study Population

A total of 111 neonatal foals fulfilled the criteria for inclusion: 51/111 were classified as septic, 29/111 were classified as sick nonseptic, and 31/111 healthy foals were assessed. The median age for all foals was 24 hours (range: septic, 1–168 hours; sick nonseptic, 1–144 hours; healthy, 24–72 hours).

Breeds represented included Thoroughbred (n = 78) and non-Thoroughbreds breeds (n = 33), including Standardbred (n = 13), Quarterhorse (n = 9), and 1 each of American Paint Horse, Hanoverian, Saddlebred, Dutch Warmblood, Friesian, mixed/grade, Appaloosa, Arabian, and Miniature horse. Colts were overrepresented compared with fillies, 66/111 and 45/111, respectively, which was true for each group of foals (Table 1). For all hospitalized foals (septic and sick nonseptic foals), the mean age at presentation was 34 hours, and all healthy foals were <72 hours of age when examined.

Table 1.   Breed and sex characteristics of foals included in the study (values are expressed as a fraction of the total number of foals and percentage [%]).
VariableTotal NumberSepticSick NonsepticHealthy
Breed
 Thoroughbred78/111 (70%)28/51 (55%)19/29 (66%)31/31 (100%)
 Non-Thoroughbred33/111 (30%)23/51 (45%)10/29 (34%)0/31 (0%)
Sex
 Colt66/111 (59%)31/51 (61%)19/29 (66%)16/31 (52%)
 Filly45/111 (41%)20/51 (39%)10/29 (34%)15/31 (48%)

There was no statistical difference in survival status for age, sex or breed (Thoroughbred versus non-Thoroughbred) in the foals in this study. Similarly, the survival rate in septic and sick nonseptic foals treated at either referral institution was not different (OR 1.4; 95% CI, 0.56–3.3) where the overall survival rate for foals treated at OSU was 69% (25/36), including 62% (16/26) of septic foals and 90% (9/10) of sick nonseptic foals, and HEMI was 59% (26/44), including 36% (9/35) of septic foals and 89% (17/19) of sick nonseptic foals. All healthy foals survived (100%, 31/31).

AVP, ACTH, and Cortisol Concentrations

For each group of foals, AVP, ACTH, and cortisol concentrations were determined (Table 2). Septic foals had significantly higher AVP, ACTH, and cortisol concentrations compared with healthy foals (P < .001). Septic foals also had higher AVP and ACTH concentrations compared with sick nonseptic foals (P < .01). Cortisol concentrations were significantly higher in both septic and sick nonseptic foals compared with healthy controls (P < 0.01), but no difference was observed between these 2 groups. Plasma AVP and ACTH concentrations were significantly higher in septic foals that died (P < .01; Table 3). There was no significant difference in hormone concentration between blood culture-positive septic foals and blood culture-negative septic foals (P > .07).

Table 2.   Blood hormone concentrations and ratios in neonatal foals at admission (values expressed as median and range).
Foal ClassificationAVP (pmol/L)ACTH (pmol/L)Cortisol (nmol/L)AVP:ACTHAVP:CortisolACTH:Cortisol
  1. a Sixteen foals were assessed.
    *P<.05 compared with healthy foals;**P<.01 compared with sick nonseptic.

Septic264035242**7.7*12**
 (n=51)(2.6–97)(3.3–200)(32–1380)(0–72)(0–58)(0–130)
Sick nonseptic5.27.3208633.0**4.9
 (n=29)(1.2–59)(2.8–110)(30–1117)(2.3–310)(0.61–18)(0.69–32)
Healthy4.6a5.25297a6.4a9.0
 (n=31)(2.2–60)(3–140)(10–1100)(27–180)(0–41)(2.4–35)
Table 3.   Blood hormone concentrations in septic foals and their association with survival status (values expressed as median and range).
HormoneSurviving
(n=25/51)
Nonsurviving
(n=26/51)
P Value
  • **

    P < .01

AVP (pmol/L)8.6 (2.6–97)44 (5.3–97).0012**
ACTH (pmol/L)11 (3.3–180)86 (5.8–200).0004**
Cortisol (nmol/L)280 (32–1400)390 (72–1000).142

For each group of foals, the HPAA hormone ratios were determined. Septic foals had a significantly lower AVP : ACTH ratio compared with healthy foals (P < .01) (Table 2). Septic foals also had higher AVP : cortisol and ACTH : cortisol (P < .01) ratios compared with sick nonseptic foals. Sick nonseptic foals had lower AVP : cortisol ratios compared with healthy foals (P<.01).

Survival Results

Descriptive. The sepsis score was highly associated with survival status in these foals. Foals with a sepsis score of ≤11 were 24 times more likely to survive than foals with a score ≥12 (95% CI, 6.5–157). Septic foals in which a septic focus was not evident were 16 times more likely to survive than those with 1 or more foci of infection.

Foals that did not have cold extremities on physical examination were more likely to survive than those whose limbs or ears were cold (OR 12.2; 95% CI, 4.5–37.3).

In addition, the presence of oral mucous membranes that were pink or pale pink with a capillary refill time of ≤2 seconds was associated with survival (Table 4). Foals that did not receive pressor agents (eg, dobutamine, norepinephrine) were more likely to survive than foals in which these medications were used (Table 4).

Table 4.   Univariate analysis for survival among 111 neonatal foals, including specific endocrinologic, clinicopathologic, descriptive, and microbiologic data.
Variable (units)RangeCrude Odds
Ratio
for Survival
95%
Confidence
Interval
  • TB, Thoroughbred; Non-TB, non-Thoroughbred breed; BUN, blood urea nitrogen.

  • *

    P < .05.

Endocrinology
 AVP (pmol/L)1.2–2023.6*5.38–131.5
20.1–600.90.16–5.4
60.1–97Referent 
 ACTH (pmol/L)2.8–5015.9*4.01–81.5
50.1–1501.50.25–9.9
150.1–200Referent 
 Cortisol (nmol/L)10–10018.5*4.7–121.5
101–3003.42*1.2–10.5
301–1380Referent 
Hematology
 Total leukocytes (109/L)0.3–4.00.360.09–1.26
4.1–12.02.7*1.39–9.3
12.1–35Referent 
 Segmented neutrophils (109/L)0.0–2.00.80.04–0.4
2.1–7.01.060.32–3.7
7.1–33Referent 
 Band neutrophils (109/L)0.05.3*1.35–21.3
0.01–0.30.70.22–2.34
0.31–2.2Referent 
 Lymphocytes (109/L)0.2–1.00.710.45–2.1
1.1–2.03.70.87–15.3
2.1–5.4Referent 
 Platelets (109/L)5.0–2500.440.18–1.06
251–904Referent 
 Packed cell volume (L/L)15–351.50.3–8.6
36–450.60.18–2.0
46–58Referent 
Biochemistry
 BUN (mg/dL)4–202.30.97–5.6
21–55Referent 
 Creatinine (mg/dL)0.5–1.02.60.6–12.1
1.1–2.014.3*4.12–59.7
2.1–4.03.7*1.16–12.8
4.1–32Referent 
 Total bilirubin (mg/dL)1.0–2.56.0*1.74–24.5
2.6–5.02.8*1.02–7.7
5.1–40Referent 
 Fibrinogen (mg/dL)100–4003.5*1.23–9.9
401–1140Referent 
 l-lactate (mmol/L)1.1–4.07.3*1.97–31.5
4.1–6.01.00.3–3.3
6.1–13Referent 
 Glucose (mg/dL)81–1605.2*1.99–15.1
<80 or >161Referent 
 Total serum protein (g/dL)3.4–5.00.350.11–1.03
5.1–6.9Referent 
 IgG (mg/dL)16–4000.05*0.01–0.16
401–8001.060.13–22
>800Referent 
 Potassium (mEq/L)2.4–4.53.64*1.43–9.7
4.6–7.9Referent 
 Bicarbonate (mEq/L)15–250.50.21–1.18
26–36Referent 
Descriptive data
 Heart rate (per minute)36–800.930.22–3.97
81–1203.00.95–9.1
121–200Referent 
 Respiratory rate (per minute)9–300.480.15–1.44
31–482.00.66–5.9
49–90Referent 
 Temperature (°C)37.6–38.62.00.83–4.95
<37.5 or >38.7Referent 
 Mucous membranesNormal4.3*1.76–11.2
AbnormalReferent 
 Capillary refill (seconds)<24.7*1.67–13.8
24.4* 
3Referent1.16–19.5
 Age (hours)0–120.370.13–1.1
13–242.70.8–9.0
>24Referent 
 Referral institutionHagyard1.40.56–3.3
OSUReferent 
 Sepsis score0–1124.11*6.54–156.69
12 or higherReferent 
 Septic focus presentNo16.03*5.51–59.07
YesReferent 
 SexColt0.670.27–1.6
FillyReferent 
 BreedTB0.910.37–2.38
Non-TBReferent 
 Cold extremitiesNo12.2*4.5–37.3
YesReferent 
Drug administration
 Pressor administrationNo9.7*1.2–200.3
YesReferent 
Microbiology
 Blood cultureNegative8.0*3.2–21.1
PositiveReferent 

A positive blood culture was obtained in 71% (36/51) of septic foals. Of these, Gram-negative bacteria were predominant (21/36; 58%), with E. coli the most prevalent (13/36; 36%) (Table 5).

Table 5.   Blood culture isolates from hospitalized septic neonatal foals at admission (n=36).
IsolateNumber
Escherichia coli13 (36%)
Enterococcus fecalis6 (16.6%)
Actinobacillus equuli4 (11%)
Streptococcus equisimilis4 (11%)
Streptococcus zooepidemicus2 (5.5%)
Salmonella spp.2 (5.5%)
Aeromonas spp.1 (2.7%)
Acinetobacter spp.1 (2.7%)
Clostridium spp.1 (2.7%)
Staphylococcus aureus1 (2.7%)
Staphylococcus spp. coagulase negative1 (2.7%)

Hormone Concentrations and Association with Survival

The survival rate for septic foals in this study was 49% (25/51), where survival was defined as a foal being discharged alive from hospital.

Table 4 shows the crude OR and 95% CI for survival in all foals (n = 111) where historical, descriptive data, clincopathologic data, microbiologic, and endocrinologic data were assessed. In septic foals, plasma AVP concentrations were significantly increased in nonsurvivors compared with survivor foals (median, 44 versus 8.6 pmol/L; P < .01; Table 3). Foals with AVP concentrations in the range of 1.2–20 pmol/L were 23.6 times more likely to survive compared with the referent of >60 pmol/L. As the AVP concentrations increased, the odds for survival decreased, indicating that nonsurvival is associated with high AVP concentrations.

ACTH concentrations in septic foals were increased among nonsurvivors compared with survivors; this finding paralleled changes observed for AVP.

There was no significant difference in cortisol concentrations between survivor and nonsurvivor septic foals, but high cortisol concentrations were associated with nonsurvival. Foals with cortisol concentrations in the range of 10–100 nmol/L were 18.5 times more likely to survive than those with cortisol concentrations > 300 nmol/L. Foals with cortisol concentrations > 100 nmol/L but <300 nmol/L were 3.4 times more likely to survive compared with foals with cortisol concentrations >300 nmol/L.

Table 3 summarizes the AVP, ACTH, and cortisol concentrations between surviving and nonsurviving septic foals.

Clinicopathologic Findings

The presence of a negative blood culture was highly associated with survival among foals (OR 8.0; 95% CI, 3.2–21.1). A total leukocyte count in the range of 4.0–12.0 × 109/L was associated with survival, where the median values for septic, sick nonseptic, and healthy foals were 4.4, 8.9, and 8.8 × 109/L, respectively. Foals without evidence of a degenerative left shift (band neutrophils) were 5.3 times more likely to survive than those with marked increases in band cell counts (>0.31 × 109/L).

In all foals, plasma fibrinogen concentrations in the 100–400 mg/dL range were associated with survival (OR 3.5; 95% CI, 1.239.9; where >400 mg/dL was the referent) as were blood lactate concentrations of <4.0 mmol/L, where foals with lactate concentrations of 1.1–4.0 mmol/L were 7.3 times more likely to survive than foals with a blood lactate concentration > 4.0 mmol/L.

Foals with serum glucose concentrations < 80  or >161 mg/dL were less likely to survive than those with concentrations in the range 80–160 mg/dL (OR 5.2; 95% CI, 1.99–15.1).

The median calculated serum osmolality in 25/51 septic foals with higher sepsis scores (median, 19) and AVP concentrations (median, 33.1 pmol/L) was 281 mOsm/L, and the median sodium concentration in these same foals was 137 mEq/L.

Variables that were retained in the final logistic regression model included age group, sepsis score, serum creatinine concentration, and plasma AVP concentration (Table 6). Specifically, foal survival was related to age > 24 hours (P= .01), low calculated sepsis score (P < .001), low measured serum creatinine concentration (P= .05), and normal measured AVP concentration (P < .001).

Table 6.   Multivariate logistic regression analysis for survival among 111 neonatal foals (final model).
VariableRangeLikelihood
Ratio
95%
Confidence
Interval
P
Value
  • *

    P<.05.

Sepsis score0–≤1126.24.1–273.9.002*
12 or higherReferent  
Vasopressin (pmol/L)1.2–2014.92.2–144.01*
20.1–600.660.08–5.5.69
60.1–97Referent  
Age (hours)1–120.030.002–0.41.01*
13–240.220.02–1.71.17
24 or higherReferent  
Creatinine (mg/dL)0.5–1.00.030.0004–0.84.05*
1.1–2.00.60.06–5.6.64
2.1–4.00.080.007–0.73.03*
4.1–32Referent  

Discussion

In the current study, we found that increased AVP and ACTH concentrations in septic foals were associated with sepsis score and survival. Plasma AVP concentrations were significantly increased in septic foals and sick nonseptic foals. Specifically, nonsurviving septic foals had significantly higher AVP concentrations than did nonseptic foals. Likewise, plasma ACTH concentrations were increased in septic nonsurviving foals.

Increased AVP release has been demonstrated in the face of health and disease in various species. In healthy, term foals, increases in AVP, ACTH, and cortisol have been demonstrated as a physiologic adaptation to hypovolemia and hypotension experimentally,36 indicating that at birth in term foals, the HPAA is fully functional. This also may be true for septic foals, by virtue of enhanced AVP release. Our results support this, because AVP (and ACTH) concentrations were increased in most septic foals younger than 24 hours.

These findings parallel results in critically ill human patients,23 including children20 and adults19 with early sepsis. Similar increases in AVP have been found in baboons, dogs,37 and rats,38 and in preliminary data in septic foalsk where up to a 10-fold increase in AVP was seen. In children with septic shock, nonsurvivors were reported to have increased AVP concentrations.20 Proposed mechanisms for the increased AVP concentration during sepsis include a physiologic response to stress,36 changes in blood pressure20,36,39 and blood volume in relation to blood pressure,20,36 changes in serum osmolality,17,20,40 and response to circulating endotoxin and proinflammatory mediators, including IL-1β, IL-6, and TNF-α.20,31–33 RAI, a common complication of septic shock in children, also may contribute to maintaining increased AVP concentrations, because cortisol inhibits AVP and ACTH release.41,42 However, increases in AVP are likely to be the result of a complex interaction of several mechanisms.23,40

Although we cannot determine the exact reason for enhanced AVP release in critically ill foals, both osmotic-related and nonosmotic-related mechanisms are possible.17,20,43 Based on normal serum osmolality and sodium concentration in a subset of septic foals, nonosmotic mechanisms appear to be at least partially responsible for enhanced AVP release in most septic foals. These findings are consistent with those reported in the human literature on sepsis.19,40

Studies have shown that bacterial toxins (exotoxins and endotoxin) can induce activation of the mononuclear phagocyte system and production of proinflammatory cytokines in septic foals,28,30 notably TNF-α, IL-1, and IL-6, which are thought to be responsible, in part, for the development of systemic inflammation. Endotoxin and proinflammatory cytokines themselves also are reported as AVP secretagogues in both humans31,40 and horses33 by activating AVP magnocellular neurons and enhancing AVP release. TNF-α is an early mediator in endotoxemia and is correlated with the severity of clinical signs.28–30 Similarly, detectable endotoxin concentrations in plasma of septic foals are correlated with nonsurvival.28 In our study, nonsurviving septic foals may have had higher concentrations of endotoxin or exotoxin and inflammatory cytokines than survivors, and measuring their blood concentrations may have been useful to determine their association with AVP release during sepsis, but such determinations were not the goal of this study.

Systemic hypotension is reported commonly in septic foals,7,8,44–47 and is also associated with increased AVP concentration in the acute stage of sepsis in humans.19 Blood pressure was only sporadically measured in these foals, because of clinician discretion in decision making, difficulties associated with invasive techniques, and reported inaccuracy associated with indirect sphygmomanometry in hypotensive states,45,46,48 and results therefore are not reported. However, given normal calculated serum osmolality in foals with high AVP concentrations, we propose that nonosmotic stimuli for AVP release were related to hypotension or hypoperfusion (baroreceptor-mediated release) and systemic inflammation (stress-mediated release) in these foals.

Septic foals and sick nonseptic foals had proportionally higher plasma ACTH concentrations compared with healthy foals. Median plasma ACTH concentration was significantly higher in septic than in sick nonseptic foals, which, in turn, was significantly higher than in healthy foals. These results are consistent with limited published data in septic foals,15 but differ from some results of sepsis studies in humans, where ACTH and cortisol concentrations often were low. Differences may be explained by species variation, age of subject, duration of illness, and severity of illness.42,49–51 There may also be inherent differences in the data attributable to interpretation of single samples in this study.

In addition to CRH,52,53 AVP is the main secretagogue for pituitary ACTH release,18,33,40,52,54 which may explain why foals with increased AVP also had increases in ACTH, as seen in previous studies in foals.36,k Again, proposed mechanisms for increased ACTH concentrations are likely to be similar to those described for increased AVP release. Other mechanisms may include RAI, where adrenocortical exhaustion and lack of cortisol production provide a positive stimulus for ACTH release in times of extreme stress. This phenonemon has been described in critically ill humans19,42,49,51 and more recently in septic foals.15,55 Currently, the diagnosis of RAI in foals is difficult to establish because it is not well described. In human medicine, RAI more often is diagnosed by a subnormal response in cortisol release after administrations of exogenous ACTH.56 In this study, RAI was especially difficult to identify because only a single cortisol measurement was made at a random time point and the median cortisol concentration in hospitalized foals (septic and sick nonseptic) was increased (352 and 208 nmol/L, respectively). However, in critically ill foals where a marked increase in ACTH and concomitant normal or low cortisol concentration are found, this finding may be supportive of RAI. Our results are consistent with those described by others15 where ACTH and cortisol concentrations in septic foals were significantly increased compared with healthy controls, and nonsurvivors had the highest concentrations of these hormones. Also, the AVP : ACTH and ACTH : cortisol ratios were significantly higher in septic foals than in the other 2 groups. Based on these ratios, we accept the validity of our findings and postulate that nonsurvivors may have more severe disease and that this endocrine response to extreme stress is appropriate in these foals, but insufficiency at the level of the adrenal gland may exist. Further studies are required to document the prevalence of RAI or relative AVP deficiency in foals with sepsis over time, as occurs in pediatric critical care medicine,41 and furthermore whether the utility of therapeutic supplementation with these hormones is warranted in sepsis.

Increases in AVP and ACTH indicate an appropriate HPAA response to critical illness. Despite increases in these hormones, however, affected foals were likely to have systemic perfusion impairment. This observation may indicate an inappropriate target organ response, such as adrenocortical unresponsiveness or exhaustion, or inappropriate vascular endothelium responsiveness, where physiologic increases in AVP concentration were insufficient to mediate vasoconstriction through unknown mechanisms. One could postulate potential V1 receptor refractoriness or exhaustion as possible causes.

We found that hypercortisolemia was present in critically ill foals, regardless of the cause of disease. Stress associated with disease, transportation to the hospital, malnutrition, or unknown mechanisms may account for these increases. In the face of disease, an increase in cortisol concentration offered little insight on the prognosis for survival in both septic and sick nonseptic human patients who were hypercortisolemic, in contrast to some other reports in humans.57,58 A normal or modestly increased cortisol concentration was strongly associated with survival in these foals in the context of health (Table 4), which is consistent with early studies on adrenocortical activity in healthy foals in the immediate postnatal period.59 However, the inherent problem of single random cortisol determinations is the assessment of appropriate adrenocortical function in relation to disease status. As previously discussed, hypocortisolemia or normocortisolemia in the face of septic shock actually may be associated with greater severity of disease, adrenal dysfunction, or RAI and is negatively associated with survival in humans.19

We found that several nonendocrinologic variables were strongly associated with prognosis in these foals. Blood l-lactate concentration was significantly associated with survival, which is consistent with Corley et al,13 who found that mean l-lactate concentrations were significantly lower in survivors.

Other findings associated with survival were negative blood culture, sepsis score ≤ 11, absence of band neutrophils, serum immunoglobulin G concentration > 800 mg/dL, serum glucose concentration 80–160 mg/dL, plasma fibrinogen concentration < 400 mg/dL, and absence of use of vasopressor agents in therapy during hospitalization. Similar findings have been reported in other studies,1–3,12,28 and their clinical relevance relates to survival likelihood in critically ill foals.

This study provides some evidence that the HPAA is stimulated and functional in critically ill foals, especially septic foals. At admission to 2 independent referral hospitals, the AVP, ACTH, and cortisol concentrations were increased in critically ill foals. Moreover, we found that the magnitude of increase in hormone concentration was associated with disease state and survival outcome, where nonsurviving septic foals had the highest increases in AVP, ACTH, and cortisol concentrations and were more likely to die. Further studies are needed to conclusively define the HPAA endocrine profile of septic neonates in relation to duration of disease, severity of sepsis, and response to conventional therapy. Also, additional research is required to better define RAI and relative vasopressin insufficiency in septic foals, as occurs in human sepsis patients, and to justify therapeutic use of exogenous corticosteroid and AVP in sepsis.

Footnotes

aCell-Dyn 3500R analyzer, Abbott Laboratories, Abbott Park, IL

bBoehringer Mannheim/Hitachi 911 system, Boehringer Mannheim Corp, Indianapolis, IN

cAccutrend Lactate analyzer, Roche, Mannheim, Germany

dArginine vasopressin double antibody radioimmunoassay, DSL, Webster, TX

eSep-Pak C18 columns, Waters Corporation, Milford, MA

fAdrenocorticotropic hormone immunoradiometric assay, DiaSorin, Stillwater, MN

gCoat-A-Count Cortisol direct radioimmunoassay, Diagnostic Products Corporation, Los Angeles, CA

hPrism, version 4.0a GraphPad Software Inc, San Diego, CA

iSAS version 9.1, SAS Institute Inc, Cary, NC

jExcel, Microsoft Corporation, Mountain View, CA

kGold JR, Divers TJ, Barton MH, et al. ACTH, cortisol and vasopressin levels of septic (survivors and non-survivors) in comparison to normal foals. J Vet Intern Med 2006;20:720 (abstract)

Acknowledgments

We thank all of the technical staff and veterinarians at Hagyard Equine Medical Institute and Galbreath Equine Center, The Ohio State University for their dedication to and assistance with this project. Special thanks are extended to Dr Holly Aldinger for healthy foal sample collection and to Kelly Rourke for her assistance with laboratory techniques.

Funding provided by the American College of Veterinary Internal Medicine (ACVIM) Mary Rose Paradis grant for multicenter research and Equine Research Grants, College of Veterinary Medicine, The Ohio State University.

Ancillary