Intact hypothalamic-pituitary-adrenal (HPA) axis function is essential for regulating the physiologic, metabolic, and inflammatory responses to critical illness. Despite the vital role the HPA axis plays in the endocrine response to severe illness, HPA axis dysfunction occurs frequently in human beings with sepsis, septic shock, and systemic inflammatory response syndrome, and is correlated with both increased disease severity and increased mortality. The purpose of this review is to describe recent evidence that suggests a similar prevalence and significance of HPA axis dysfunction in the critically ill neonatal foal.
Definition and Pathophysiology of Relative Adrenal Insufficiency
The body's primary adaptive response to the homeostatic and metabolic disturbances produced in critical illness includes activation of the hypothalamic-pituitary-adrenal (HPA) axis. This results in secretion of corticotropin-releasing hormone by the hypothalamus, and the subsequent release of adrenocorticotropic hormone (ACTH) by the anterior pituitary gland. ACTH, in turn, enters the systemic circulation and stimulates the adrenal cortices to synthesize and secrete cortisol. Cortisol exerts a variety of systemic effects that are vital for adaptation to the physiologic stresses associated with severe illness, including maintenance of adequate blood pressure, provision of nutrients to tissues, and control of an appropriate inflammatory response.1,2 These effects ultimately serve to modulate existing physiologic stressors and restore homeostatic balance.
Despite its vital role in the physiologic response to the stress of disease, dysfunction of the HPA axis is well documented in critically ill humans; this is particularly true in patients with sepsis, septic shock, and systemic inflammatory response syndrome (SIRS).3-7 This may be manifested by absolute adrenal insufficiency due to complete and permanent destruction of the adrenal glands by the primary illness, though this is relatively uncommon in both human and veterinary medicine.7,8 Less severe and transient relative adrenal insufficiency (RAI), however, is much more common during critical illness; this is typically defined as an inadequate cortisol response for the existing degree of illness.8 As a result, the patient is rendered incapable of mounting an appropriate and sufficient cortisol response to cope with the increased physiologic demands of acute severe disease.3,4,7,8 HPA axis dysfunction in illness may occur through complete or partial failure of the axis at one or many of the levels of control of cortisol secretion and action, including: 1) inadequate regulatory hormone (CRH, ACTH) secretion, 2) adrenal resistance to ACTH, 3) impaired or exhausted synthesis of corticosteroids in the adrenal glands ("loss of adrenal reserve"), and/or 4) impaired tissue response to cortisol.6,9
The specific mechanisms contributing to HPA axis dysfunction in critical illness are poorly understood. Direct injury to one or several components of the HPA axis caused by the primary illness can occur and may be manifested as adrenocortical hemorrhage and/or necrosis on histopathologic examination.8 In addition, the HPA axis can be suppressed by specific drugs (e.g., etomidate) and potentially by certain infectious organisms themselves.10 Most often, though, the patient's own immune and inflammatory response is the most likely source of HPA axis suppression during critical illness.9,10 Sepsis-induced inflammatory cytokines can stimulate and/or suppress the HPA axis at several levels.10,11 For example, tumor necrosis factor-α (TNF-α), which is increased in sepsis, impairs both ACTH secretion from the pituitary gland and ACTH activity at the level of the adrenal gland. In addition, increased leukocyte gene expression and/or increased serum concentration of interleukin-6 (IL-6), another pro-inflammatory cytokine, is correlated with both HPA axis dysfunction and a poor prognosis for survival in septic human neonates and in septic foals.1,12-14 Most often, the cause of RAI in critical illness is likely multifactorial, with several of the above factors working in concert in an individual patient.
Regardless of the inciting cause, it is most important to note that RAI occurs secondary to a primary, severe illness or stress, and almost always fully resolves if the patient survives the primary illness. However, due to the vital role that the HPA axis plays in the response to the physiologic stress of illness, the occurrence of RAI can greatly impact a patient's survival. Because of the important immunomodulatory and anti-inflammatory effects of cortisol, HPA axis dysfunction during critical illness can ultimately result in immunoendocrine dysregulation, allowing an initially appropriate inflammatory response to become dangerously exaggerated.9 Disruption or dysregulation of the immune/inflammatory response due to HPA axis failure may result in imbalance between pro- and anti-inflammatory cytokines, with two important consequences: impairment of the body's ability to combat infection, and/or overwhelming inflammation that damages host tissues and organs.15 In addition, because cortisol plays a vital role in the maintenance of blood pressure, HPA axis dysfunction can result in significant hypotension. This overwhelming inflammatory response and systemic hypotension is most often clinically manifested as septic shock and/or SIRS. In addition, an overzealous and inappropriate inflammatory response can result in continued and additional HPA axis suppression, as described above. This further perpetuates septic shock and SIRS, ultimately resulting in inadequate tissue perfusion and impairment of the host's ability to clear the infectious organism, often culminating in the death of the patient.3,15
Significance of HPA Axis Dysfunction in Critical Illness
An extensive body of evidence in the human literature clearly links HPA axis dysfunction with an increased incidence of shock, multiple organ dysfunction syndrome (MODS), SIRS, and death in critically ill patients.7,16 For example, the results of a recent comprehensive retrospective cohort study of septic adults indicate that non-survivors have significantly lower cortisol responses to high-dose ACTH stimulation than do survivors.17 Another study by Loisa et al found that in 41 septic patients, those that met criteria for RAI had longer ICU stays and increased severity of multiple organ failure as compared to those with normal adrenal function.7 Annane et al evaluated a larger group of 189 patients with septic shock, and found that patients with RAI had a mortality rate of 82%, versus a mortality rate of 26% in septic patients who did not meet criteria for RAI.18 The incidence of RAI in human pediatrics is less well documented, but a recent study that evaluated 57 children with septic shock determined that the incidence of RAI in these patients was as high as 44%, depending on the diagnostic criteria used for RAI.16 Furthermore, the survival rate in septic children that met criteria for RAI was lower (47%) than the survival rate in patients without RAI (67%).16 The incidence of RAI in critically ill veterinary patients has not been thoroughly investigated, though Burkitt et al recently reported similar findings in dogs with sepsis.19 In this study, dogs with an inadequate cortisol response to a high-dose (250 µg) ACTH stimulation test (delta cortisol < 3 µg/dl) had an increased incidence of systemic hypotension and a decreased survival rate as compared to septic dogs with appropriate cortisol responses to exogenous ACTH (delta cortisol > 3 µg/dl).19
The potential importance of adrenal insufficiency in critically ill patients is further underscored by the results of two comprehensive meta-analyses evaluating the effect of low (physiologic or replacement) dose corticosteroid supplementation in patients with septic shock.15,20 These reviews, as well the results of several smaller, single-center trials, showed a significant beneficial effect of low-dose corticosteroid supplementation in critically ill patients with RAI.3,15,20 In these studies, patients meeting criteria for RAI that were supplemented with low (physiologic) doses of corticosteroids showed more rapid shock reversal, earlier cessation of vasopressor therapy, shorter ICU stays, more rapid ventilator weaning, and improved survival than patients with RAI that did not receive steroid supplementation.3,15,20,21 Thus, the detection of HPA axis dysfunction is clearly an important diagnostic consideration in human critical care, and provides a means by which to identify patients that may benefit from corticosteroid replacement therapy.
Diagnosis of HPA Axis Dysfunction and Relative Adrenal Insufficiency
Despite the importance of HPA axis dysfunction in critical illness, definitive diagnosis of RAI can be problematic. Accurate assessment of an adequate cortisol response to severe disease in each individual patient is inherently difficult, due to wide individual variations in systemic cortisol levels during the course of disease and the many host, pathogen, and environmental factors that play a role in the physiologic response to infection. Thus, a consensus on the diagnostic criteria for RAI or HPA axis dysfunction in critical illness has not been fully established in human medicine. Current suggestions for diagnosis of HPA axis dysfunction in septic human patients are based on one or more of the following findings: 1) basal serum cortisol concentrations less than those commonly induced by stress in healthy people (i.e., < 18-25 µg/dl);3,4 2) a blunted cortisol response (delta cortisol < 9 µg/dl) to administration of a high (supraphysiologic) dose of synthetic ACTH (125-250 µg) in the classic high dose ACTH stimulation test;4,22 and/or 3) a blunted response to administration of a more "physiologic" low dose (1 µg) of synthetic ACTH.3,23
In response to experimentally induced stress, healthy people will experience an increase in serum cortisol concentration to 2 to 3 times their resting level, to concentrations > 18 µg/dl. Thus, a basal cortisol level of < 18 to 25 µg/dl is used by some to define RAI in septic patients.8,14 This method, though, can be misleading, as cortisol levels can fluctuate rapidly, vary widely in critically ill patients, and vary considerably by age and species. A more comprehensive method of assessing adrenal function involves assessment of the response of the adrenal gland to stimulation with exogenous ACTH with an ACTH stimulation test.7,24 The classic high-dose ACTH stimulation test is theorized to produce a maximal adrenal response due to administration of a supraphysiologic dose of ACTH, and is most useful for diagnosing absolute adrenal insufficiency, but may produce false positive cortisol responses in some patients. Patients with true relative adrenal insufficiency, who do not respond to physiologic amounts of ACTH, can often produce an appropriate response when administered supraphysiologic doses of ACTH.8,14 Thus, a low dose stimulation test using 1 µg ACTH may more accurately diagnose RAI in critically ill patients.8,14 Siraux et al compared the low dose (1µg) and high dose (250µg) ACTH stimulation tests in 46 human patients with septic shock. Using delta cortisols of < 9 µg/dl as a criteria for RAI, a low dose test identified patients with an inadequate adrenal response that would have been misdiagnosed with the high dose test alone.23 Evaluation of baseline cortisol and response to both a low and high dose of ACTH may provide information regarding the severity and degree of ongoing HPA axis dysfunction; a patient with a low baseline cortisol concentration and an inadequate cortisol response to both low and high doses of ACTH likely has more severe and extensive HPA axis dysfunction than a patient with an appropriate baseline cortisol concentration and appropriate response to the high dose of ACTH, but an inadequate response to the low, physiologic dose of ACTH.
Other methods for evaluating HPA axis function include measurement of serum free cortisol levels (the biologically active fraction of cortisol), the insulin tolerance test, and the metapyrone test.22 Serum free cortisol levels may provide more accurate and clinically relevant information than total cortisol levels, since free cortisol is the biologically active form of cortisol. However, its measurement requires specialized and tedious methodology.22,25 The insulin tolerance test and the metapyrone test both provide a more complete assessment of the entire integrity of the HPA axis than the ACTH stimulation test, which predominantly assesses adrenal function, but both are time and labor intensive and can be associated with potentially significant adverse effects.22 Thus, these methods are not used commonly in clinical settings.
Evidence for Relative Adrenal Insufficiency in Critically Ill Neonatal Foals
Documentation of RAI during critical illness in the equine literature is limited, but may occur with significant frequency in some populations of critically ill horses. Neonatal foals are likely at the greatest risk for HPA axis dysfunction and RAI during illness, for several reasons. First, compelling evidence over the last 30 years has shown that maturation of the HPA axis in the neonatal foal occurs very late in the foal's development, beginning during the last 4 to 5 days prior to parturition and continuing into the first one to two weeks of life.26-28 This HPA axis immaturity is certain to impact the foal's ability to respond to physiologic stress and disease in the neonatal period, and may make the HPA axis more vulnerable to injury and/or dysfunction during periods of critical illness. Furthermore, the leading cause of morbidity and mortality in neonatal foals, both under field conditions and in tertiary referral centers, is bacterial sepsis;29,30 as described above, sepsis is the illness most commonly associated with concurrent HPA axis dysfunction in human critical care.2,6
Although adrenal insufficiency in premature foals and evidence for adrenal hypofunction in healthy term foals in the first few weeks of life has been documented, the prevalence of adrenal insufficiency in septic foals is not well described. A single case report of transient adrenal insufficiency in a neonatal old foal with septicemia and hypovolemic shock was described in 1998 by Couteil and Hoffman.31 In addition to clinical signs correlating with hypoadrenocorticism, this foal showed low basal serum cortisol levels (0.7-1.2 µg/dl) and a poor cortisol response to high dose corticotropin stimulation test (11% increase in serum cortisol after injection) during clinical episodes of collapse. A tapering course of prednisolone supplementation in addition to supportive care and antimicrobial therapy resulted in resolution of the foal's clinical signs and adrenal insufficiency, with a normal adrenocortical response (210% increase in cortisol levels following administration of corticotropin) at 2 months of age. Gold et al recently demonstrated that non-surviving septic foals had markedly elevated ACTH:cortisol ratios (elevated ACTH levels with correspondingly low cortisol levels) as compared to survivors,32 providing additional evidence for HPA axis dysfunction in septic neonatal foals, presumably at the level of the adrenal gland.
During the 2006 and 2007 foal seasons, the potential incidence and significance of HPA axis dysfunction in critically ill neonatal foals was further investigated in our laboratory. HPA axis function was assessed in 71 neonatal foals < 7 days of age at admission to two tertiary referral centers, and correlated with indicators of disease severity and outcome.33 Basal plasma endogenous ACTH and total serum cortisol concentrations were measured using chemiluminescent immunoassays in 71 foals, and a paired low-dose (1 µg) / high-dose (100 µg) ACTH stimulation test was performed in 59 of these foals. The paired low-dose/high ACTH stimulation protocol was based on previous synthetic ACTH dose-response assessment in healthy neonatal foals,34 and evaluated in a healthy, age-matched foals to determine appropriate cortisol responses to exogenous ACTH in foals during the neonatal period.35 HPA axis dysfunction was defined using similar criteria as described in the human literature, as: 1) an inappropriately low basal serum cortisol concentration, and/or 2) an inadequate increase in serum cortisol concentration (delta cortisol) after administration of synthetic ACTH. Specifically, an inappropriately low basal cortisol concentration was defined as a basal cortisol concentration less than the lowest cortisol concentration (mean--1 standard deviation) achieved after administration of a physiologic dose (10 µg) of synthetic ACTH to healthy age-matched foals. An inadequate delta cortisol was defined as a delta cortisol less than the mean delta cortisol achieved in healthy age-matched foals using the same synthetic ACTH stimulation protocol.
Using these two criteria independently, HPA axis dysfunction was diagnosed in 49% and 51% of foals, respectively. Seventy-six percent of foals met criteria for sepsis (sepsis score > 11 and/or positive blood culture); HPA axis dysfunction was diagnosed in 37% and 41% of septic foals, respectively, with the above criteria. As is reported in critically ill human beings, dogs, and adult horses, increased baseline cortisol concentrations were significantly correlated with disease severity and non-survival (death or euthanasia for prognostic reasons during hospitalization) when all foals were considered. In the septic foals, low delta cortisol was also significantly correlated with non-survival. In addition, 90% of the septic foals with an increased basal serum cortisol concentration and a low delta cortisol did not survive, as compared to a non-survival rate of only 9% in foals with a high basal cortisol concentration and an appropriate delta cortisol. This finding suggests that exhaustion of cortisol synthetic capacity ("loss of adrenal reserve") may contribute to HPA dysfunction in septic foals. A significantly higher incidence of MODS and shock was also observed in septic foals with a low delta cortisol than in septic foals with an appropriate delta cortisol. A low-dose (1 µg ACTH) delta cortisol cutoff of at least 3.2 µg/dl and a high-dose (100 µg ACTH) delta cortisol cutoff of at least 5.1 µg/dl were determined to be the most accurate predictors of survival to discharge, while a baseline cortisol concentration cutoff of at least 6.7 µg/dl was most accurate for predicting the occurrence of shock and MODS.
Conclusions and Future Directions
Collectively, the above evidence strongly suggests that HPA axis dysfunction occurs in septic foals with significant prevalence, similar to the approximate 50% prevalence reported in septic adult humans and infants.4,16,36 In addition, as is reported in people,4,16,36 septic foals with evidence of HPA axis dysfunction are more likely to have shock, multiple organ failure, and are less likely to survive to discharge than septic foals with intact HPA axis function. Further investigation of the pathogenesis of HPA axis dysfunction in the septic foal and evaluation for a potential benefit of corticosteroid replacement therapy in septic foals with HPA axis dysfunction is warranted in future studies.
1. Gibson S, et al. Emer Med Clin N Am 2005;23:909.
2. Marik P, Zaloga G. Chest 2002; 122:1784.
3. Aneja R, Carcillo J, et al. Arch Dis Child 2007; 92:165.
4. Annane D, et al. Am J Respir Crit Care Med 2006; 174:1319.
5. Annane D, et al. J Amer Med Assoc 2000; 283:1038.
6. Cooper M, Steward P. New Eng J Med 2003; 328:727.
7. Loisa P, et al. Acta Anaesthesologica Scandanivia 2002; 46:145.
8. Marik P, Zaloga G. Crit Care Med 2003; 31:141.
9. Marik P. Curr Opin Crit Care 2007; 13:363.
10. Turnbull A, Rivier C. Physiolog Rev 1999; 79:1.
11. Haddad J, et al. J Neuroimmunol 2002; 133:1.
12. Gold J, et al. J Vet Int Med 2007; 21:482.
13. Harris M, et al. J Pediatr 2005; 147:462.
14. Kozyra E, et al. Ann Pharm 2005; 39:691.
15. Keh D, Sprung C. Crit Care Med 2005; 32:S527.
16. Pizarro C, et al. Crit Care Med 2005; 33:855.
17. Lipnier-Freidman D, et al. Crit Care Med 2007; 35:1012.
18. Annane D, et al. J Amer Med Assoc 2000; 283:1038.
19. Burkitt J, et al. J Vet Int Med 2007; 21:226.
20. Minneci P, et al. Ann Intern Med 2004; 141:47.
21. Huang C, Hong-Chyuan L. Am J Respir Crit Care 2006; 173:276.
22. Asare K. Pharmacotherapy 2007; 27:1512.
23. Siraux V, et al. Crit Care Med 2005; 33:2479.
24. Rai R, et al. Crit Care Resusc 2004.6:123.
25. Hamrahian A, et al. New Engl J Med 2004; 350:1629.
26. Nathanielsz P, et al. J Reprod Fertil Suppl 1975; 23:625.
27. Rossdale P, Ousey J. Equine Vet J 1984; 16:300.
28. Silver M, et al. Equine Vet J 1984; 16:278.
29. Cohen N. J Am Vet Med Assoc 1994; 204:1644.
30. Morris D, Moore J. J Am Vet Med Assoc 1991; 199:1584.
31. Couteil L, Hoffman A. J Am Vet Med Assoc 1998; .212:1594.
32. Gold J, et al. J Vet Int Med 2007; 21:791.
33. Hart K, et al. Unpublished data 2007.
34. Hart K, et al. J Vet Int Med 2007; 21:314.
35. Hart K, Barton M. Unpublished data 2007.
36. Watterberg K. Seminars in Neonatology 2004; 9:13.