Endocrine Dysfunction in Critical Illness

Kelsey A. Hart • Ramiro E. Toribio

The endocrine response to illness is vital for ensuring organ, tissue, and cellular health during disease and includes simultane­ous increases and decreases in the activity of multiple endocrine axes.

Assessment of Endocrine Function in the Critically Ill Patient

Endocrine dysfunction during critical illness can be due to a primary underlying and previously unrecognized endocrine disorder (such as undiagnosed PPID in a geriatric horse under­going colic surgery to correct a strangulating small intestinal lesion), or secondary endocrine dysfunction can develop as a result of factors related the underlying severe illness (such as development of insulin resistance in an endotoxemic horse with colitis).

In health, mechanisms for assessment of endocrine function include measurement of basal circulating hormone concentra­tions and dynamic tests, including suppression tests if overactiv­ity of an endocrine access is suspected and stimulation tests if insufficient endocrine activity is suspected. Tests that induce physiologic stress such as the insulin tolerance test can also be used to globally assess endocrine axis function. These same assessment options are available in a clinical setting but are often too dangerous to be used in critically ill patients. Cutoff values for plasma hormone concentrations and dynamic test responses that suggest appropriate function in the unstressed patient are not applicable to the critically ill patient undergoing severe pathophysiologic stress. The critically ill patient is also in a dynamic state, with a constantly changing and evolving endocrine response,⁷ so one-time responses do not always accurately represent endocrine function.

Thus, endocrine assessment in the critically ill patient often requires assessment of function via multiple methodologies. Basal hormone concentrations from multiple levels of an endocrine axis can be assessed simultaneously, and dynamic testing can be assessed in conjunction with basal hormone concentrations to provide a more comprehensive picture of the function of the entire axis. Free, biologically active hormone concentrations can be measured if alterations in binding proteins are anticipated. Methodology for detection of hormone activity at the level of the target tissue would be ideal but is not com­mercially available for most hormones; thus, clinical evaluation of the patient with a high index of suspicion for evidence of endocrine dysfunction can be an important additional “test” to assess tissue hormone activity in the critical care setting. For example, as detailed later, critical illness-related corticosteroid insufficiency (CIRCI) appears to be fairly common in people with septic shock and can be treated with cortisol replacement in the form of low doses of hydrocortisone. Frequently used methodology for CIRCI diagnosis includes documentation of a low basal cortisol concentration, a decreased cortisol response to ACTH stimulation testing with a variety of protocols, or decreased free cortisol concentration.

In sum, because endocrine responses are necessarily, dramati­cally, and variably altered in critical illness, it can be difficult to define what degree of alteration is expected and appropriate and what degree is inappropriate in the critically ill patient. Currently available methodology can be used to assess endocrine function during critical illness in most patients, but expectations for universally applicable guidelines for test interpretation and diagnostic cutoff values are likely unrealistic. Rather, the astute clinician must be aware of common endocrine disturbances seen in critical illness and consider testing and/or endocrine intervention in a timely fashion to prevent development of more severe and potentially fatal complications.

HPA Axis

The HPA axis is the master regulator of various endocrine systems. It modulates cardiovascular, metabolic, and immune activity in response to physiologic needs and stressful events, with the ultimate goal of maintaining organ function. It regulates circulating concentrations of corticosteroid hormones, which are vital for appropriate stress responses in both health and disease. Upon stressful stimulation, the hypothalamus releases corticotropin-releasing hormone (CRH, CRF) and arginine vasopressin (AVP) into the pituitary portal system to stimulate pituitary corticotrophs to secrete adrenocorticotropic hormone (ACTH). In horses, AVP is an important releasing factor for ACTH. ACTH acts on the zona fasciculata of the adrenal cortices to induce the synthesis and secretion of the glucocorticoid hormones. To a lesser extent, ACTH also induces aldosterone secretion by the adrenocortical zona glomerulosa.

The predominant circulating glucocorticoid hormone in animals, including horses, is cortisol, though corticosterone is also present at moderate concentrations in cattle and goats. [Note: because HPA axis dysfunction in critically ill animals is best characterized in horses and the individual glucocor­ticoid hormones have similar effects, cortisol will be used to refer to all glucocorticoid hormones for the purpose of this review.]

Cortisol has diverse systemic effects on a variety of target tissues; during critical illness, the sum of these effects helps to maintain blood pressure, coordinate energy metabolism, and regulate the immune and inflammatory response. However, a number of studies suggest that up to 40% to 60% of people, dogs, and horses with sepsis and septic shock suffer from transient HPA axis dysfunction (CIRCI).^2,4,8,9 In these patients, the cortisol response to illness is inadequate for the existing degree of physiologic stress and ACTH concentrations. Circulating cortisol concentrations may be increased from basal levels found in healthy unstressed individuals but are still not adequate to meet demands to support and regulate cardiovas­cular, metabolic, and inflammatory responses.^4,10 In some patients with CIRCI, cortisol concentrations may be appro­priately increased, but peripheral glucocorticoid resistance results in impaired cortisol activity in target tissues.^4,10

The pathogenesis of CIRCI is poorly understood and is likely multifactorial. Pathophysiologic manifestations of cortisol deficiency can result from dysfunction at one or more levels of the HPA axis, including: (1) inadequate regulatory hormone (CRH, AVP, ACTH) secretion; (2) impaired adrenal sensitivity to ACTH due to receptor antagonism or downregulation; (3) impaired adrenal cortisol synthetic capacity due to loss of substrate or enzymatic failure (“adrenal exhaustion” or “loss of adrenal reserve”); or (4) impaired cortisol activity in the target tissues due to decreased glucocorticoid receptor (GR) expression or activity or intracellular conversion of cortisol to cortisone, which cannot bind the GR.^4,10 This HPA axis dysfunction may be a result of direct injury to one or several components of the HPA axis caused by hemorrhage, ischemia, or necrosis associated with the primary illness.^4,10 In addition, the HPA axis can be suppressed by specific drugs (e.g., etomidate),¹¹ bacterial components such as endotoxin,¹² or the patient's own immune response.^10,13

One key aspect of CIRCI is that the HPA axis dysfunction is transient, in contrast to absolute adrenal insufficiency (Addison's disease), which is characterized by permanent glucocorticoid and mineralocorticoid insufficiency.⁴ Thus, patients with CIRCI that recover from their primary disease regain normal HPA axis function and do not require lifelong steroid supplementation.

Current understanding of HPA axis function in critically ill horses and foals is limited; to date, CIRCI is best character­ized in septic neonatal foals.^17-23 Although basal ACTH and cortisol concentrations are higher in septic foals than healthy age-matched control foals and in nonsurviving foals (as one would expect with the stress of illness), several studies have reported significantly increased ACTH-to-cortisol ratios in 17 19 21 23

nonsurviving septic foals.¹'^,19,21,23 Such high AC!H-to-cortisol ratios, with high ACTH concentrations and low corresponding cortisol concentrations, suggest cortisol synthetic failure at the level of the adrenal gland may occur in the septic full-term foal.

Findings from ACTH stimulation tests in hospitalized foals also provide evidence for adrenocortical dysfunction in this population. Peak and delta cortisol (peak-basal cortisol concentration) responses to ACTH stimulation tests do not consistently differ between healthy and ill foals, but increased disease severity and poorer prognoses were associated with decreased cortisol responses to ACTH stimulation in hospital­ized foals.^20,22 Specifically, nonsurviving foals had significantly lower delta cortisol responses to low-dose ACTH stimulation as compared with survivors,²² and foals that met criteria for CIRCI (as characterized by an inadequate delta cortisol response to a high-dose ACTH stimulation test) had a significantly greater incidence of shock, multiple organ dysfunction syndrome, and nonsurvival than foals with an adequate cortisol response to ACTH.²⁰ When human diagnostic criteria for CIRCI²⁴ based on inadequate delta cortisol responses to high-dose ACTH stimulation were adapted and applied to hospitalized foals 1 week of age or younger, approximately 50% of all hospitalized foals and a subgroup of septic foals met these criteria.²⁰

The incidence of CIRCI in critically ill horses has only been examined in one study to date, but findings suggest it may also occur in this population.²⁵ Twenty-four percent of severely ill horses had inappropriately low basal cortisol concentrations at hospital admission, and 85% of these horses had inadequate delta cortisol responses to ACTH stimulation.²⁵ Marked adrenal hemorrhage at necropsy was also seen in nonsurviving horses 25

in this study.²⁵

In concert, this evidence suggests that CIRCI occurs in critically ill and septic neonatal foals and in critically ill horses with comparable frequency and impact as in people with similar illnesses.

The clinical presentation of CIRCI is typically vague and insidious, as signs related to the primary illness usually pre­dominate. Indeed, it is the primary illness that should prompt the clinician to consider CIRCI as a possible complicating factor because data in people, dogs, horses, and foals suggest it is more likely in certain diseases than others.⁴’⁵’’’ Recent studies in horses suggest that CIRCI should be considered as a possible complication in any critically ill horse or foal, but particularly in septic foals and horses with evidence of SIRS.^19-22,25 Clinicians should also have a high index of suspicion for cortisol insufficiency in premature foals due to the late maturation of the HPA axis and adrenal gland in foals. Human studies suggest that CIRCI also occurs in conjunction with acute respiratory distress syndrome (ARDS), massive trauma, and major surgery,^4,10 so it should also not be overlooked in horses with similar diseases.

Specific manifestations of CIRCI are directly related to inadequate cortisol support for maintenance of blood pres­sure, nutrient metabolism, and regulation of the immune/ inflammatory response, and they can include one or more of the following: (1) persistent hypotension despite appropriate volume resuscitation and vasopressor support; (2) persistent hypoglycemia or hyperlactatemia despite glucose support and adequate perfusion; or (3) persistent signs of SIRS. Diagnostic criteria for SIRS in some groups of ill horses and foals have been investigated and include tachycardia, fever or hypothermia, leukocytosis (neutrophilia), leukopenia (neutropenia), and increased band neutrophils.^26-30 Specific signs of mineralocorti­coid deficiency (such as persistent hyponatremia, hypochloremia, or hyperkalemia) are not a consistent feature of CIRCI but do occasionally occur.

Specific diagnostic criteria for CIRCI in horses or foals are not well defined. Indeed, determination of the ideal diagnostic criteria for CIRCI in any species is inherently difficult as quantifying a universally appropriate cortisol response to variable and constantly changing disease severity in individual patients is next to impossible. Thus, standard diagnostic criteria for CIRCI in people and infants are not widely accepted, but current recommendations include documentation of either (1) an inappropriately low basal cortisol concentration during critical illness or (2) an inadequate delta cortisol response to high-dose (supraphysiologic) ACTH stimulation testing.²⁴ Testing for CIRCI in people is typically done at ICU admission for patients with diseases likely complicated by CIRCI (e.g., sepsis, ARDS) or when clinical signs consistent with CIRCI, such as persistent hypotension after fluid and vasopressor support, are identified. Cutoff values for adequate basal and delta cortisol concentrations in illness are based on responses to environmental stress in healthy individuals.

In general, measurement of basal plasma hormone concentra­tions offers a rapid and safe means of assessing adrenocortical function in the critically ill horse or foal. There are commercial radio- and chemiluminescent immunoassays validated for measurement of plasma or serum equine cortisol concentra- tions.^31,32 In foals, age-related variation in HPA axis function needs to be considered when determining cutoff values for CIRCI diagnosis.^20,22,33 Specific basal cortisol cutoff values for diagnosis in CIRCI in ill foals or horses have not been critically evaluated, but basal cortisol less than 9.9 μg∕dL in foals 4 hours old and younger, less than 7 μg∕dL in foals 12 to 24 hours old, less than 4.4 μg∕dL in foals 36 to 48 hours old, less than 2.5 μg∕dL in foals hours to 7 days old, and less than 9.7 μg∕dL in horses that have been used to diagnose CIRCI in recent reports.^20,25 However, due to the pulsatile nature of cortisol secretion, plasma concentrations can vary widely during a 24-hour period in both healthy and sick horses and foals with time of day, season, emotional state, and moment- 3435

to-moment changes in physiologic stressors.^34,35 Thus, correct interpretation of a single random cortisol concentration in an individual animal as appropriate or not can be difficult. Further, diagnosis of CIRCI based on a low basal cortisol concentration was not correlated with increased disease severity of decreased survival in septic foals, as was diagnosis based on ACTH stimulation testing.²⁰

Assessment of cortisol concentrations in conjunction with regulatory hormone (ACTH, CRH) concentrations can provide a more comprehensive picture of HPA axis function. Plasma CRH and ACTH concentrations are easily assessed in most species via commercially available immunoassays, but as the majority of CRH is secreted into the hypothalamic-hypophyseal portal vessels rather than into the systemic circulation, accurate measurement of CRH requires sophisticated sampling meth­odology. In addition, while in healthy individuals ACTH and cortisol concentrations are fairly closely correlated, ACTH- cortisol dissociation is not uncommon during illness in horses and foals.^23,35 Documentation of an increased ACTH-to-cortisol ratio might be helpful for CIRCI diagnosis, especially in foals, but due to wide variation in these parameters in individual animals,^17,19-21 specific cutoff values for appropriate ACTH- to-cortisol ratios are not currently proposed for diagnosis of CIRCI in horses or foals. Even though cortisol and its ratio with ACTH concentrations are used as indicators of HPA axis function in sick foals, progesterone and dehydroepiandrosterone (DHEA) sulfate were found to be good predictors of HPA axis dysfunction and outcome in hospitalized foals.²³

Dynamic tests like the ACTH stimulation test circumvent some of these limitations with diagnosis of CIRCI based on basal hormone concentrations. Typically, in the classic “high- dose” ACTH stimulation test, serum cortisol concentration is measured just before and 30 to 90 minutes after intravenous or intramuscular administration of a supraphysiologic quantity of ACTH (Cosyntropin, 1 to 2 μg∕kg or higher).^20,40,41 This high-dose ACTH stimulation test produces a maximal adrenal response, so an inadequate increase in cortisol concentration in this test is used to diagnose both absolute, irreversible adrenal insufficiency (e.g., Addison's disease) and transient CIRCI.^4,24,36,37

However, the high-dose ACTH stimulation test may be less sensitive for diagnosis of the transient, reversible HPA axis suppression that occurs in CIRCI. In some patients with CIRCI, the adrenal gland fails to produce an appropriate cortisol response to physiologic concentrations of ACTH but may still respond adequately to supraphysiologic ACTH concentra- tions.^38,39 A measurable cortisol response can be produced by administration of lower ACTH doses (0.01 to 0.2 μg∕kg IM) in horses and foals,^{20,22,25,40,41} and such “low-dose” stimulation tests could also be used to diagnose CIRCI in equine patients. Delta cortisol responses vary more with foal age and dose of ACTH administered, but in general an increase from basal cortisol less than onefold to twofold 30 minutes after low-dose ACTH stimulation testing^{20,22,25,40,41} and less than threefold to fourfold 90 to 120 minutes after high-dose ACTH stimulation testing^20,40,41 may be suggestive of CIRCI in horses and foals. However, one study that examined a paired low-dose∕high-dose ACTH stimulation test in hospitalized foals found significant associations between increased disease severity and nonsurvival and inadequate cortisol responses to the high dose but not the low dose of ACTH.²⁰

It is important to note, however, that ACTH stimulation tests only evaluate the adrenal component of the HPA axis. Therefore, patients with cortisol insufficiency due to impaired HPA axis function at the hypothalamic and∕or pituitary levels may produce an appropriate cortisol response to exogenous ACTH that could be falsely interpreted as intact HPA axis function. Insulin tolerance, CRH stimulation, and metyrapone tests are preferred for global assessment of the HPA axis and diagnosis of HPA axis hypofunction and have been described in foals and horses.^42-45 However, risks associated with severe hypoglycemia and the potential for preexisting glucose∕insulin derangements in critically ill horses and foals limit the utility of insulin tolerance testing in patients with potential CIRCI, and the CRH stimulation and metyrapone tests are currently limited by expense and reagent availability, so these tests are not currently recommended for diagnosis of CIRCI in horses and foals.

Finally, some of the physiologic consequences of CIRCI may be associated with cortisol resistance in peripheral tissues rather than circulating cortisol insufficiency; in fact, some patients may have impaired cortisol metabolism and ACTH- cortisol dissociation.^10,46-48 Decreased glucocorticoid receptor binding affinity is described in horses with SIRS and is associated with nonsurvival.⁴⁹ In such cases, all of the previously mentioned tests may suggest intact HPA axis function, with appropriate endogenous cortisol concentrations and normal responses to dynamic testing, despite clinical evidence of cortisol insufficiency such as persistent hypotension, metabolic derangements, or inflammatory dysregulation.

Specific therapeutic protocols for management of CIRCI in horses or foals have not been critically evaluated. It is important to remember that the goal of CIRCI treatment is physiologic cortisol replacement, as a number of studies in septic patients have demonstrated deleterious effects of high-dose (supraphysi­ologic) corticosteroid regimens in sepsis.²⁴ Thus, it may be ideal to avoid more potent and longer-acting corticosteroids such as dexamethasone and prednisolone in favor of hydrocortisone (which is identical to endogenous cortisol), though these drugs are sometimes used in people with CIRCI.

In one case report, a septic foal with adrenocortical insufficiency—as evidenced by a low basal cortisol and impaired cortisol response to ACTH stimulation—demonstrated clinical improvement and eventual normalization of cortisol responses following a course of corticosteroid supplementation in addition to standard therapy for sepsis.¹⁸ Recent evidence also suggests that in healthy 2- to 6-day-old foals, a short tapering course of hydrocortisone (1.3 mg/kg/day divided q4h IV) has potentially beneficial anti-inflammatory effects without significantly impairing innate immune function.⁵⁰ Adjustment of daily cortisol production rates in foals based on a “illness factor,” as has been done in people to determine steroid supplementation regimens in CIRCI, suggests that a hydrocortisone dose of 1 to 4 mg/kg/day divided q4-6h may be appropriate for foals with CIRCI.³⁵ However, this or other corticosteroid regimens have not been critically evaluated in septic foals or horses with CIRCI.

Healthy newborn foals have high pregnane concentrations at birth that decrease gradually over the first 48 hours post foaling.^23,51 In the brain, progesterone, DHEA, and preg­nenolone can be converted into neuroactive steroids (e.g., allopregnanolone) that promote neurogenesis, neuronal plastic­ity, and neuroprotection, modulate CRH and AVP synthesis, and influence the HPA axis. Neurosteroid actions in neurons and glial cell are mediated by the GABA_A receptor. Increased progestogen and neuroactive steroid concentrations have been associated with disease severity, neurologic abnormalities (neonatal maladjustment syndrome [NMS]), prematurity, and mortality in hospitalized foals.^23,52 The mechanisms leading to high neurosteroid concentrations in critically ill foals remain unclear, but it has been speculated to be a consequence of altered intrauterine to extrauterine transition, as well as adre­nocortical enzymatic shifts during illness.^23,52 Progesterone, 17α-hydroxyprogesterone, and allopregnanolone concentrations are often elevated in septic compared with healthy and sick nonseptic foals.^23,52 If progesterone derivatives are involved in the pathogenesis of NMS, reducing their concentrations could be beneficial. Finasteride and dutasteride are 5α-reductase inhibitors used in men for alopecia and prostatic hypertrophy to block the conversion of testosterone to 5α-dihydrotestosterone. These drugs also block the conversion of progesterone to 5α-dihydroprogesterone, a key precursor for allopregnanolone, and have been proposed as potential treatments for specific cases of NMS.

RAAS, Blood Volume, and Blood Pressure Regulation

Hypovolemia and hypotension are common in critically ill people and animals, and regulation of blood pressure and blood volume is quite complex. The RAAS, HPA axis, and arginine vasopressin (AVP, also known as antidiuretic hormone or ADH) work independently and in concert to control these responses during both health and illness. The RAAS is activated when macula densa cells in the renal juxtaglomerular apparatus sense decreases in blood volume or blood sodium concentration or increases in plasma potassium concentration, resulting in renin release. Renin cleaves plasma angiotensinogen to angiotensin I, which is then further processed to the potent vasoconstrictor angiotensin II by pulmonary angiotensin-converting enzyme (ACE). In addition to vasoconstriction, angiotensin II induces aldosterone release from the adrenal cortices. Aldosterone increases blood volume and pressure by activating the renal tubular sodium potassium ATPase and epithelial sodium chan­nels to enhance renal sodium and water reabsorption, as well as potassium excretion. Increases in plasma osmolality and decreases in arterial pressure are also detected by hypothalamic osmoreceptors and arterial stretch receptors, respectively, stimulating hypothalamic AVP synthesis and release from the posterior pituitary. The primary effect of AVP is the insertion of aquaporin channels in the renal medullary collecting ducts to increase water reabsorption, thus increasing blood volume and blood pressure. AVP also can directly induce arteriolar vasoconstriction and activates the HPA axis by inducing pituitary ACTH release. HPA axis activation and resultant increases in plasma cortisol further aid blood pressure regulation via direct cardiac inotropic effects and downregulation of the vasodilator nitric oxide.

In critical illness, RAAS activation is important for blood pressure and electrolyte regulation, but excessive activation can have detrimental effects. Angiotensin II has recently been demonstrated to have proinflammatory effects, and associations between angiotensin II and increased organ failure and mortality have been described.⁵³ Transient hyperreninemic hypoaldo­steronism is observed frequently in people with septic shock and is associated with the development of acute renal failure in these patients.⁵⁴ This hypoaldosteronism could be a reflection of adrenocortical suppression in CIRCI and could play a role in persistent hypotension in affected patients, but as described earlier, specific signs of mineralocorticoid deficiency like electrolyte derangements are uncommon in CIRCI. Information regarding specific RAAS responses in critically ill horses is limited. Hypotensive adult horses demonstrated increased aldosterone concentrations before fluid resuscitation, but aldosterone responses to hypotension appear to be exaggerated in neonatal foals.⁵⁵ RAAS activation in critically ill foals is mainly characterized by an increase in angiotensin II and aldosterone concentrations.⁵⁶ Renin activity and aldosterone concentrations were also increased in horses with experimentally induced laminitis but might represent appropriate responses to concurrent hyponatremia rather than pathologic RAAS activation.⁵⁷ Effects of RAAS on activation on inflammatory status and mortality are not described in horses or foals.

AVP concentrations are initially increased in most acutely critically ill patients, stimulated by the development of hypoten­sion and systemic inflammation (proinflammatory cytokines are potent AVP secretagogues).^3,58 In many patients with septic shock, AVP concentrations tend to fall dramatically despite persistent hypotension, consistent with relative AVP deficiency due to both depletion of stored AVP and inhibition of AVP synthesis.⁵⁸ Tissue AVP activity may also be impaired by inflammatory responses in critical illness.³ However, there are conflicting reports regarding associations between increases or decreases in AVP concentration and mortality in critically ill patients.⁵⁸ Increased AVP concentrations were observed in adult horses with naturally occurring hypovolemia, but not in a small group of hypotensive foals, suggesting AVP responses in foals might be immature.⁵⁵ Larger studies documented higher AVP concentrations in septic foals and an association among high AVP concentrations, hypoperfusion, and decreased survival.⁵⁹ Vasopressin infusions are sometimes used as vasopres­sors in hypotensive people, but variable efficacy and concerns regarding potential side effects associated with decreased skin and splanchnic blood flow limit its clinical application.^21,58 Studies investigating the safety and efficacy of vasopressin infusions in horses and foals are lacking, though vasopressin is used anecdotally as a vasopressor in some critically ill neonatal foals with reported success.

Adrenomedullin also plays a role in hemodynamics. Adre- nomedullin is a vasodilator peptide of the calcitonin gene family produced by chromaffin cells of the adrenal medulla, cardiac myocytes, and cells in other tissues. It has antiapoptotic and antiinflammatory properties, attenuates ischemia and reperfusion injury, modulates vascular tone, enhances microcirculatory function, promotes angiogenesis, reduces intestinal mucosa permeability, counteracts the RAAS, inhibits aldosterone secretion, increases natriuresis, and supports cardiovascular function during SIRS and sepsis.^60,61 Adrenomedullin concentra­tions are increased in people with sepsis and endotoxemia, hypertension, heart failure, and myocardial infarction.^61-63 Its production by cardiac myocytes is stimulated by TNF-α, IL-1β, angiotensin II, endothelin-1, and hypoxia. Adrenomedullin concentrations were valuable to individualize human patients who require hemodynamic support.⁶⁴ Adrenomedullin analogs have been proposed as therapeutic agents and there are ongoing clinical studies using monoclonal antibodies to the N-terminus of adrenomedullin to prolong its half-life in septic human patients.⁶⁰ A recent study found high plasma adrenomedullin concentrations in critically ill compared with healthy foals; however, concentrations were not different between septic and nonseptic foals and were not associated with survival.⁶⁵

Thyroid Axis

In critical illness, activity of the hypothalamic-pituitary-thyroid axis is suppressed, resulting in decreased plasma triiodothyronine (T3) and thyroxine (T4) and increased reverse T3 (rT3) concentrations, a syndrome called nonthyroidal illness syndrome (NTIS) or euthyroid sick syndrome. This syndrome is well described in humans, dogs, and cats suffering from systemic illnesses, and recent reports suggest that NTIS is common in sick horses and foals.^66-69 The syndrome is described in detail earlier in the Thyroid Gland section.

Parathyroid Gland and Calcium Regulation

Calcium dysregulation is common in critically ill people and animals and is presumed to be associated with parathyroid gland dysfunction, vitamin D deficiency, or both.^70-72 Ionized hypocalcemia has been documented in septic foals and horses with intestinal disease, as well as in horses with experimental endotoxemia.^71,73 High parathyroid hormone (PTH) concentra­tions were found in most horses and foals with hypocalcemia and are consistent with an appropriate endocrine response to low calcium, but some hypocalcemic horses had low PTH concentrations that could be indicative of parathyroid gland dysfunction in this population.^71,73,74 In both people and foals, increased PTH concentrations were associated with decreased survival, but factors other than hypocalcemia likely contributed to this finding as no difference in calcium concentration was found in nonsurviving foals.⁷¹ Proinflammatory cytokines can impact PTH secretion and activity and might play a role in these calcium disturbances in critically ill horses and foals.⁷¹ Calcitonin and PTH-related protein (PTHrP) concentrations were not different between healthy and sick foals,⁷¹ and renal calcium excretion was not different in horses with enterocolitis,⁷⁴ suggesting these factors are not responsible for hypocalcemia in critically ill horses and foals. Procalcitonin (PCT), the precursor of calcitonin, is used in human critical care medicine as a highly specific marker of sepsis. Increased PCT concentra­tion was documented in horses with clinical evidence of SIRS, and endotoxin administration to horses induces a rapid PCT response.^75,76

Information on vitamin D metabolites in critically ill horses and foals is lacking. It was recently shown that hypovitaminosis D is common in critically ill foals.^77,78 Hospitalized foals had lower serum concentrations of 2 5-hydroxyvitamin D and

1,25- dihydroxyvitamin D (active vitamin D, calcitriol) that were associated with hypocalcemia, disease severity, and mortal- ity.^77,78 The mechanisms by which critically ill foals develop hypovitaminosis D remain unclear but could be a consequence of renal and intestinal losses, decreased concentrations vitamin D binding protein, and renal PTH resistance.

Another system of relevance in calcium and phosphorus homeostasis is the fibroblast growth factor-23 (FGF-23)/klotho axis. FGF-23 inhibits renal 1α-hydroxylase activity (vitamin D activation) and sodium/phosphorus cotransporter expression (promotes phosphaturia) and suppresses PTH secretion. Klotho is the coreceptor for FGF-23 and has multiple properties, including antisenescence, antioxidation, antiapoptosis, and cytoprotection. Disturbances in the FGF-23/klotho axis were recently documented in critically ill foals, and hospitalized foals with high FGF-23 or low klotho concentrations were 77

more likely to die.⁷⁷

Energy Axes (Endocrine Pancreas and Orexigenic System)

Insulin, glucagon, cortisol, and epinephrine are considered the main endocrine components of the energy axis; however, it is clear that this is a complex system in which many other hormones (growth hormone, ghrelin, leptin, adiponectin, somatostatin, and incretins) are also important.

Glucose derangements are common in critically ill people and animals. In people, hyperglycemia due to previously undiagnosed diabetes mellitus or “stress hyperglycemia” is most common and is associated with increased morbidity and mortality.¹ Hyperglycemia is also common in horses with GI disease, and persistent hyperglycemia is associated with decreased survival in this population as well.^79,80 In critically ill neonatal foals, both hypoglycemia and hyperglycemia are reported and are associated with a poorer prognosis.⁸¹ Hypoglycemia in septic neonatal foals is most likely related to decreased nursing due to illness and lack of substantial glycogen stores, but mechanisms resulting in hyperglycemia in critically ill people and animals appear to be more complication. Stresses of critical illness activate the HPA axis, sympathoadrenomedullary responses, and the GH axis, resulting in increased concentrations of the counter-regulatory hormones cortisol, catecholamines, GH, and glucagon. These hormones act directly and indirectly to increase blood glucose by increasing hepatic gluconeogenesis, inhibiting insulin release, and impairing glucose utilization in the periphery.¹ Proinflammatory cytokines like TNF-α and IL-1 are also frequently increased in critically ill patients and can induce peripheral insulin resistance in people and laboratory animals by interfering with insulin signaling.¹ Experimental endotoxemia in healthy horses has been shown to decrease insulin sensitiv­ity.^82,83 Insulin sensitivity has not been thoroughly investigated in critically ill horses and foals to date, but one study documented higher insulin concentrations in nonsurviving septic foals⁸⁴ and a recent report found evidence of insulin dysregulation and endocrine pancreatic dysfunction in nonsurviving horses with the systemic inflammatory response syndrome.⁸⁵

Glucose derangements have important clinical consequences in critical illness, as evidenced by the increases in morbidity and mortality discussed earlier. The consequences of hypo­glycemia are obvious, as CNS function depends on glucose availability. Clinical sequelae related to hyperglycemia can be equally important, as hyperglycemia results in volume depletion through osmotic diuresis and subsequent electrolyte and acid­base derangements and perfusion impairment.¹ Hyperglycemia can also result in tissue damage via increased production of reactive oxygen species and related lipid peroxidation, damaging tissues like cardiac muscle, vascular beds, and worsening ischemia-reperfusion injury in various tissues such as ischemic intestine.¹ Hyperglycemia also alters blood flow to certain tissues by inhibiting formation of the vasodilator nitric oxide and can cause hemostatic abnormalities through a variety of mechanisms.¹ Finally, immune function is impaired in hyper­glycemia by inhibiting phagocytic and chemotactic function in macrophages and neutrophils, respectively.¹

Because of the negative consequences associated with hyper­glycemia, the effects of tight glycemic control with exogenous insulin therapy have been investigated in a number of human studies. Initial studies found a substantial reduction in mortality in patients subject to tight glycemic control protocols that aimed to keep blood glucose concentrations between 80 and 110mg/dL in comparison with less stringent control (illness.⁹⁶

GH Axis

The GH axis plays an important role in supporting metabolism, gut mucosal function, wound healing, and immune function during critical illness.⁹⁷ Loss of the GH pulsatile pattern is common in critically ill humans, where high plasma GH concentrations are not followed by increases in IGF-I, indicating GH resistance. However, GH deficiency may occur in some patients; either mechanism can result in relative GH insuffi­ciency, which can increase catabolism and impair wound healing.^73,97 In critically ill people, recombinant GH supple­mentation has been used to decrease protein catabolism and initially appeared beneficial, though more recent studies have demonstrated associations with increased mortality, resulting in recommended cessation of the use of the compound in human ICUs.⁹⁷ GH actions are predominantly mediated via IGF-1, which is often decreased in critical illness.^73,97 GH has not been extensively studied in critically ill horses or foals, but a recent study found higher GH and lower IGF-1 concentrations in septic foals, suggesting that GH resistance may occur in this population of foals, as well as in people.⁹⁸

Recent work in people has also demonstrated interactions between ghrelin and the GH axis—ghrelin activates the GH receptor, and in addition to hunger, it also stimulates GH secretion.^73,97 One study demonstrated higher ghrelin concen­trations in septic compared with healthy foals; however, no association with survival was found.⁹⁸ High ghrelin and low IGF-1 (negative regulator for GH secretion) likely enhances GH secretion in critically ill foals. Ghrelin has been evalu­ated as a potential therapeutic agent to reduce inflammation, increase muscle mass, and treat cachexia and anorexia (anorexia nervosa) in other species. Ghrelin or related compounds may provide a means for provision of anabolic support to critically ill patients.

■