Abstract
Hypotension in the neonate is defined by low blood pressure associated with decreased systemic and organ blood flow. In a proportion of critically ill neonates, especially preterm neonates, hypotension is associated with adrenal insufficiency. Accordingly, corticosteroids have been increasingly used in the treatment of systemic hypotension in this patient population. Although recent information has led to a better understanding of the underlying pathophysiology, clinical presentation, diagnostic criteria, and treatment strategies of neonatal hypotension associated with adrenal insufficiency, more data are needed to understand the benefits and risks of the use of corticosteroids in the treatment of adrenal insufficiency and the associated hypotension.
Keywords
adrenal insufficiency, cardiovascular insufficiency, corticosteroids, dexamethasone, hydrocortisone, neonate, preterm infant, relative adrenal insufficiency, shock, systemic hypotension, term infant
- •
Relative adrenal insufficiency has become increasingly recognized in sick premature and term neonates, infants, children, and adults.
- •
Relative adrenal insufficiency often presents with cardiovascular insufficiency and increases the risk of morbidity and mortality in all populations.
- •
In general, relative adrenal insufficiency associated with illness and hypotension lasts less than a week in term infants and no longer than 2 weeks in very preterm infants.
- •
Corticosteroids increase blood pressure, but more studies are needed to determine the treatment criteria, dosage, duration, and long-term effects.
- •
Glucocorticoid therapy should be tailored for each patient and account for gestational age, condition, and response of the patient, limiting the exposure as much as possible.
Hypotension in the neonate is defined as low blood pressure and signs of decreased systemic and organ blood flow and has been shown to be associated with adrenal insufficiency in critically ill patients, including preterm and term neonates. Accordingly, corticosteroids have been increasingly used for the treatment of neonatal hypotension not responding to vasopressor-inotropes. Although there is a better understanding of the underlying pathophysiology, clinical presentation, diagnostic criteria, and treatment strategies for adrenal insufficiency-associated neonatal hypotension, data are still lacking on the long-term benefits and risks of steroid use in neonates with adrenal insufficiency-associated cardiovascular compromise.
Adrenal Insufficiency and Relative Adrenal Insufficiency
In 1885 Thomas Addison described manifestations of primary adrenal insufficiency as “general languor and debility, remarkable feebleness of the heart’s action, irritability of the stomach…occurring in connection with a diseased condition of the suprarenal capsules…” ( http://wehner.org/addison.htm ). Currently, more than 150 years after the original description of Addison’s disease, these manifestations of adrenal insufficiency extend to the recently described condition of “relative adrenal insufficiency.” Relative adrenal insufficiency is defined as a condition when there is inadequate corticosteroid activity compared with the level of illness in a critically ill patient and may arise from inadequate cortisol levels resulting from problems anywhere along the hypothalamic-pituitary-adrenal (HPA) axis or from tissue resistance to corticosteroids. Adrenal insufficiency in critically ill patients is also known as “functional adrenal insufficiency,” “transient adrenocortical insufficiency of prematurity,” or “critical illness-related corticosteroid insufficiency.” For the purposes of this chapter, the term “relative adrenal insufficiency” will be used interchangeably with these terms.
Relative adrenal insufficiency is not characterized by structural abnormality of the adrenal glands but by its transience, because the majority of patients who recover will have normal HPA axis function and corticosteroid activity. However, the diagnosis of relative adrenal insufficiency in a critically ill patient should not make one complacent about or underestimate the potentially life-threatening nature of this type of adrenal insufficiency. Higher mortality rates occur in those whose adrenal response to stress is blunted, as evidenced by trauma patients receiving etomidate during surgery. Etomidate, a sedative commonly used for intubation, suppresses cortisol production by inhibiting 11β-hydroxylase activity which leads to adrenal suppression. Since the 1980s, relative adrenal insufficiency has become increasingly recognized in sick premature and term neonates, infants, children, and adults.
In this chapter, we will review the incidence and pathophysiology of relative adrenal insufficiency, along with its clinical presentation, the diagnostic considerations, and treatment strategies the neonatologist needs to understand and consider.
Adrenal Insufficiency and Relative Adrenal Insufficiency
In 1885 Thomas Addison described manifestations of primary adrenal insufficiency as “general languor and debility, remarkable feebleness of the heart’s action, irritability of the stomach…occurring in connection with a diseased condition of the suprarenal capsules…” ( http://wehner.org/addison.htm ). Currently, more than 150 years after the original description of Addison’s disease, these manifestations of adrenal insufficiency extend to the recently described condition of “relative adrenal insufficiency.” Relative adrenal insufficiency is defined as a condition when there is inadequate corticosteroid activity compared with the level of illness in a critically ill patient and may arise from inadequate cortisol levels resulting from problems anywhere along the hypothalamic-pituitary-adrenal (HPA) axis or from tissue resistance to corticosteroids. Adrenal insufficiency in critically ill patients is also known as “functional adrenal insufficiency,” “transient adrenocortical insufficiency of prematurity,” or “critical illness-related corticosteroid insufficiency.” For the purposes of this chapter, the term “relative adrenal insufficiency” will be used interchangeably with these terms.
Relative adrenal insufficiency is not characterized by structural abnormality of the adrenal glands but by its transience, because the majority of patients who recover will have normal HPA axis function and corticosteroid activity. However, the diagnosis of relative adrenal insufficiency in a critically ill patient should not make one complacent about or underestimate the potentially life-threatening nature of this type of adrenal insufficiency. Higher mortality rates occur in those whose adrenal response to stress is blunted, as evidenced by trauma patients receiving etomidate during surgery. Etomidate, a sedative commonly used for intubation, suppresses cortisol production by inhibiting 11β-hydroxylase activity which leads to adrenal suppression. Since the 1980s, relative adrenal insufficiency has become increasingly recognized in sick premature and term neonates, infants, children, and adults.
In this chapter, we will review the incidence and pathophysiology of relative adrenal insufficiency, along with its clinical presentation, the diagnostic considerations, and treatment strategies the neonatologist needs to understand and consider.
Incidence of Relative Adrenal Insufficiency
The overall incidence of relative adrenal insufficiency in critically ill adult patients is approximately 20% in general and approximately 60% in patients with severe sepsis and septic shock in particular. Reported proportions of critically ill pediatric patients with “inadequate” cortisol response vary widely, from as low as 2% often up to 87%. Overall, in the critically ill neonate the incidence is less well understood because there is no consensus on the diagnostic criteria of relative adrenal insufficiency of the ill newborn. But, if relative adrenal insufficiency were to be defined by the use of corticosteroid treatment for hypotension, one large multicenter study found that 17% of the 647 term and late preterm newborns with cardiovascular insufficiency received corticosteroids. In another study of 8019 hypotensive infants born between 2011 and 2012 in 43 neonatal intensive care units (NICUs), 18.8% received hydrocortisone. Moreover, in the extremely low birth weight (ELBW) subset of this population, 33.3% received hydrocortisone. However, these data do not reflect on the true incidence of the condition in the newborn population receiving care in the NICU.
Pathophysiology of Relative Adrenal Insufficiency in The Ill Patient
Postulated pathophysiologic mechanisms for relative adrenal insufficiency in ill patients include adrenergic receptor insensitivity due to receptor downregulation ( Fig. 30.1 ), proinflammatory cytokine-mediated suppression of the function of the pituitary and adrenal glands, inadequate HPA axis response to stress, limited adrenal reserve, gestational age-associated immaturity of the adrenal gland, corticosteroid tissue resistance, and limited adrenal perfusion ( Fig. 30.2 ).
Prolonged exposure to inflammatory mediators is one of the proposed mechanisms for both vasopressor-resistant hypotension manifesting via receptor downregulation and relative adrenal insufficiency presenting with suppression of the function of the adrenal gland and/or HPA axis. Decreased vascular responsiveness to adrenergic agents is due to downregulation of adrenergic receptors. Downregulation of adrenergic receptors in clinical conditions occurs within hours due to prolonged exposure to intrinsic (stress response) or extrinsic (vasopressor therapy) catecholamines and inflammatory mediators such as nitric oxide (NO), tumor necrosis factor, and other inflammatory cytokines (interleukin [IL]-1, IL-2, IL-6, interferon gamma). Thus production of proinflammatory cytokines also contributes to decreased vascular reactivity to catecholamines and thus to the development of vasopressor-resistant hypotension.
Decreased vascular responsiveness to adrenergic stimulation has been described in critically ill patients who meet specific criteria for adrenal insufficiency. Annane et al. have reported worse vascular responsiveness to norepinephrine in adult patients with septic shock presenting with adrenal insufficiency compared with those without the condition. In the scenario in which severe illness has both inflammation and concomitant cortisol insufficiency-associated decreased cardiovascular responsiveness to catecholamines, corticosteroid therapy reestablishes vascular responsiveness and counteracts inflammation.
Newborns, especially those born prematurely, have unique susceptibilities to relative adrenal insufficiency which is, at least in part, due to the immaturity of their HPA axis and especially the cortical function of the adrenal gland. In addition, the changes in adrenal cortical hormone production, especially cortisol during the transition to extrauterine life and the gestational age-dependent changes in placental 11β-hydroxysteroid dehydrogenase type 2 activity and thus the effect of maternal corticosteroids on fetal corticosteroid production, pose special challenges for the newborn, especially the preterm neonate. Fetal adrenal glands begin to synthetize cortisol de novo only at approximately 22 to 24 weeks’ gestation, followed by a steady increase throughout the rest of the pregnancy. Corticotropin-releasing hormone (CRH) production by the placenta also increases throughout gestation, resulting in maternal and fetal serum CRH concentrations at term that are much higher than at any other time in life. Placental CRH stimulates cortisol production in the fetal adrenal gland. At birth, the very high placental CRH production ceases to have an effect on the newborn. The pituitary gland of the newborn which has been exposed to high concentrations of CRH during fetal life may become transiently insensitive to the lower concentrations of CRH produced by the hypothalamus of the neonate. Therefore it may not be able to increase adrenocorticotropic hormone (ACTH) production to appropriately stimulate the adrenal gland for several days after delivery. Healthy term newborns tolerate this period of relative HPA insufficiency. However, this situation may predispose the newborn to the development of relative adrenal insufficiency in critical illness. Relative adrenal insufficiency could then contribute to the severity of systemic hypotension and attenuate the response of the cardiovascular system to treatment.
In summary, the physiology and complex regulation of the fetal HPA axis and the sympathoadrenal system and their immaturity at birth, coupled with prolonged exposure to free radicals and the subsequent production of proinflammatory cytokines in critical illness, set the stage for the presentation of adrenal and cardiovascular insufficiency in the acutely ill premature or term infant.
Clinical Presentation of Relative Adrenal Insufficiency and Hypotension
Cardiovascular instability, a key clinical manifestation of relative adrenal insufficiency, may present with hypotension that is either responsive, dependent, or resistant to vasopressor/inotropes and, depending on myocardial function and the loading conditions of the heart, with high or low cardiac output. ∗
∗ References , , , , , , , , and .
Other clinical features of relative adrenal insufficiency may include hyponatremia, hyperkalemia, hypoglycemia, metabolic acidosis, and feeding intolerance ( Fig. 30.3 ). However, the electrolyte and/or metabolic abnormalities are less likely to occur in critically ill patients whose electrolyte and fluid management are tightly controlled.
In addition to cardiovascular instability and the potential electrolyte and metabolic abnormalities, relative adrenal insufficiency is characterized by random and/or stimulated cortisol levels that are inadequate for the degree of illness severity and by rapid clinical and hemodynamic improvement following corticosteroid therapy.
In preterm infants, blood pressure positively correlates with cortisol production rate. Thus patients with low random or basal serum cortisol values are more likely to have low blood pressure. In a randomized controlled trial (RCT) of hydrocortisone administration in hypotensive preterm infants, the overall baseline (control) median serum cortisol concentrations were 3.3 and 4.1 μg/dL in the treated and placebo groups, respectively. These values are considered low for the degree of illness severity in preterm infants when compared with the 50th percentile of the serum cortisol values in ill infants (7.2 μg/dL). In 54 surfactant-treated, 1-day-old preterm infants of 24 to 36 weeks’ gestation, serum cortisol values were lower in those with left ventricular output less than or equal to 180 mL/min/kg compared with those greater than 180 mL/min/kg.
The study by Yoder and colleagues in very premature baboons provides the most direct, albeit nonhuman, evidence that relative adrenal insufficiency, cardiovascular insufficiency, and prematurity are related. In earlier studies, this group of investigators documented that the majority of extremely premature baboons, delivered at approximately 67% of baboon gestation (∼26 weeks’ human gestation), required volume expansion and vasopressor-inotrope therapy to treat the subjects’ hypotension, oliguria, and acid-base imbalance. Many of the premature baboons (38%) also required hydrocortisone to successfully treat their hypotension despite receiving vasopressor-inotrope therapy. This group of researchers also demonstrated that decreased urinary free cortisol excretion in the first postnatal day correlated with decreased left ventricular function. Furthermore, hydrocortisone therapy (0.5 to 1.0 mg/kg/day for 1 to 2 days) corrected the hypotension and the left ventricular dysfunction, reduced vasopressor-inotrope requirement and mortality, and increased serum cortisol to levels comparable to those seen in baboons with no evidence of relative adrenal insufficiency or cardiovascular dysfunction.
In term infants, low cortisol values have also been found in association with hypotension. In some studies on late preterm and/or term neonates with refractory hypotension, low median cortisol values were in the range of 4.5 to 11.7 μg/dL.
Although recent evidence has shown that many infants will respond to corticosteroid treatment regardless of cortisol values, there are reports that show documented low cortisol values in neonates with vasopressor-resistant hypotension responsive to corticosteroid treatment. †
† References , , , , , , , and .
Other studies report a similar presentation of responsiveness to corticosteroids treatment without cortisol data. Attenuated cortisol response to adrenal stimulation with ACTH and extremely high random cortisol levels may also represent a clinical presentation in which vasopressor-resistant hypotension is the predominant factor of the cardiovascular collapse.Relative Adrenal Insufficiency in Preterm Infants
In 1989 Ward and Colasurdo first described ventilated, sick, extremely premature infants who presented with signs consistent with adrenal insufficiency or “Addisonian crisis.” These infants had signs of relative adrenal insufficiency, including hypotension, oliguria, hyponatremia, and cortisol values less than 15 μg/dL (414 nmol/L), and responded to hydrocortisone therapy. Since this time, subsequent investigations have further elucidated the clinical and laboratory presentation of relative adrenal insufficiency in ill, hypotensive premature infants. ‡
† References , , , , , , , , , , and .
Investigators have documented two intriguing observations regarding cortisol levels in well and sick premature infants. Many healthy premature infants with no signs of relative adrenal insufficiency have random cortisol levels that are not detectable or less than 5 μg/dL (138 nmol/L), a threshold considered to indicate adrenal insufficiency. §§ References , , , , , , , and .
On the other hand, there is a population of sick premature infants with serum cortisol levels similar to or lower than cortisol levels in well preterm or term infants. ¶¶ References , , , , , , , and .
The finding that sick premature infants do not have the expected increase in random cortisol levels commensurate with their illness acuity is supportive of the presence of relative adrenal insufficiency in this population.Further support for relative adrenal insufficiency in the preterm population comes from studies that have found that adrenal insufficiency results from immature cortisol synthesis in the adrenal gland. Ng et al. reported normal pituitary response to CRH but blunted adrenal response in sick premature infants with vasopressor-resistant hypotension and low serum cortisol values. These findings led Ng et al. to speculate that, in preterm neonates, the adrenal gland was responsible for the adrenal insufficiency. Findings from other investigators support this notion because sick premature infants have been found to have low random cortisol levels, elevated cortisol precursors, and blunted response to ACTH stimulation. ‖
‖ References , , , , , , , and .
Watterberg et al. reported that compared with term infants, sick premature infants had higher cortisol precursor concentrations (17α-OH pregnenolone, 17α-OH progesterone, and 11-deoxycortisol) and lower serum cortisol concentrations. In another, small, prospective study of infants born at less than 30 weeks’ gestation, Huysman et al. demonstrated that critically ill, ventilated infants, compared with less sick, nonventilated infants, had lower cortisol levels, elevated cortisol precursor levels of 17-hydroxyprogesterone, and insufficient cortisol response to ACTH (0.5 μg/kg) stimulation. Korte et al. also reported an abnormal adrenal response to cosyntropin (ACTH) in 51 ventilated, premature infants of less than 32 weeks’ gestation who had baseline cortisol levels of less than 5 μg/dL (138 nmol/L). Among the 51 infants, 64% and 37% had inadequate cortisol response (<9 μg/dL) to stimulation with 0.1 and 0.2 μg/kg cosyntropin, respectively.Evidence of Relative Adrenal Insufficiency in Sick Late Preterm and Term Infants
There is growing evidence that a significant number of ill term and late preterm infants also exhibit evidence of adrenal insufficiency. In 1972 Gutai et al. were one of the first groups of authors to describe suboptimal cortisol responses to illness in a small number of stressed term newborn infants. Indeed, there was no difference in the median random cortisol value between the 12 ill infants (5.2 μg/dL) and 28 healthy newborns (4.1 μg/dL), but all infants responded to 5 units of ACTH appropriately. A subsequent study by Thomas et al. found that 27% of ill term newborns studied had basal cortisol less than 2 μg/dL, and only 33% of these infants had an appropriate response to ACTH (>18 μg/dL). In the first study to investigate cardiovascular responses to dexamethasone in hypotensive term newborn infants, Tantivit et al. reported that five of the seven infants studied had cortisol values less than 10 μg/dL. All of these infants responded to dexamethasone administration with prompt hemodynamic stabilization. In 2000 Pittinger et al. described low cortisol values in sick infants with congenital diaphragmatic hernia and found that 79% had random cortisol levels of less than 7 μg/dL. Two of the four critically ill patients had an inappropriately low cortisol response to cosyntropin. On the other hand, Soliman et al. found an overall increase in basal circulating cortisol concentrations by twofold to threefold in neonates with sepsis and respiratory distress. Yet, more than 30% of these infants had cortisol values suggestive of relative adrenal insufficiency (<15 μg/dL). Patients with lower basal cortisol levels and peak cortisol responses to ACTH had higher mortality. Recent reports also suggest that, similar to preterm infants, critically ill late preterm and term infants with hypotension may have problems with adrenal synthesis of cortisol as higher cortisol precursor levels have been documented in those presenting with vasopressor-resistant hypotension. In a larger, prospective, observational study, 35 sick late preterm and term infants on mechanical ventilation had a median random cortisol level of 4.6 μg/dL, a value very low for the degree of illness. These infants also had relatively low ACTH values but demonstrated appropriate cortisol responses to ACTH stimulation (1 μg/kg), suggesting they had secondary adrenal insufficiency, a mechanism different from the adrenal insufficiency of the sick very low birth weight (VLBW) infant.
Duration of Relative Adrenal Insufficiency
Speculation about the duration of relative adrenal insufficiency in different patient populations is based on the duration of hormone replacement therapy and serum cortisol levels. In adults , Hoen et al. and Annane et al. reported that relative adrenal insufficiency and/or vasopressor dependency may be sustained for several days, weeks, or even beyond 1 month in patients with sepsis or hemorrhagic shock from trauma.
In preterm infants , Colasurdo et al. found that in nine sick infants of 26 weeks’ gestation with clinical symptoms of relative adrenal insufficiency and low cortisol levels (mean, 251 ± 102 nmol/L −9.1 ± 3.7 μg/dL), there was reversal of clinical signs within 2 days. Since then, based on serum cortisol data or response to corticosteroids, studies have reported inconsistent findings for the presumed duration of relative adrenal insufficiency. Ng et al. reported that inadequate cortisol response on postnatal day 7 in sick VLBW infants had resolved by day 14. In contrast, Guttentag et al. reported that no increase in either cortisol or ACTH serum levels occurred through the first 14 days in response to critical illness. If duration of relative adrenal insufficiency was based on the response to hormone replacement therapy, according to an earlier study by Gaissmaier and Pohlandt, a single dose of dexamethasone would be effective in reversing vasopressor-resistant hypotension in sick premature infants. In addition, Ng et al. reported that 79% of preterm infants less than 32 weeks’ gestation with refractory hypotension receiving hydrocortisone for 5 days were successfully weaned from vasopressor support by 72 hours of age compared with only 33% who were receiving placebo. Hochwald et al. also showed that a 2-day course of hydrocortisone compared with placebo significantly decreased dopamine exposure by 34 hours versus 67 hours ( P = .04). Similarly, Efird et al. reported that prophylactic hydrocortisone therapy for 5 days (vs. placebo) in ELBW infants reduced the use of vasopressors during the first two postnatal days. On the other hand, in very premature baboons with relative adrenal insufficiency and vasopressor-resistant hypotension at a gestational age approximately equivalent to human gestation of 26 weeks, there was no decrease in urinary free cortisol levels over the first 2 postnatal weeks (the entire duration of the study).
In term ill infants , Economou et al. found that, in 4 of 15 infants who were “low cortisol responders” relative to their illness, this state lasted until the fifth postnatal day. In a more recent study of critically ill newborn infants with refractory hypotension, Baker et al. used cortisol levels and the presence of hemodynamic instability to determine the duration of relative adrenal insufficiency and thereby the duration of hydrocortisone replacement therapy. Sixty-one term infants received 3.5 median days of hydrocortisone therapy, whereas the 37 extremely preterm infants enrolled in the study remained on hydrocortisone for a median of 15 days. In another study, 1-day-old term and late preterm ill infants with random low cortisol values of less than 15 μg/dL received longer courses of treatment with hydrocortisone for hypotension compared with those with higher cortisol values (5 vs. 2.5 median days). Kamath et al. found a high incidence (67%) of cortisol values of less than 15 μg/dL in hypotensive infants with congenital diaphragmatic hernia, all of whom received hydrocortisone for 10.9 ± 7.0 days. Thus it seems that the duration of relative adrenal insufficiency varies with age, disease severity, etiology, and response to treatment. In general, relative adrenal insufficiency seems to last less than a week in term infants and no longer than 2 weeks in very preterm infants irrespective of whether it is defined by cortisol values or by the short-term cardiovascular response to corticosteroid replacement therapy.
Late-onset Relative Adrenal Insufficiency-Associated Hypotension
Most reports of relative adrenal insufficiency are in critically ill newborn infants in the first postnatal days. There is some evidence that adrenal insufficiency can also occur later in life (4 to 7 days of age) in ill newborn infants with respiratory distress or sepsis. A survey in Japan found that approximately 4% of VLBW infants were receiving postnatal corticosteroid therapy after 7 days of age with symptoms suggestive of relative adrenal insufficiency, including hypotension and high cortisol precursors.
Relative Adrenal Insufficiency and Outcomes
In addition to hypotension, transient adrenal insufficiency in the newborn population is associated with other morbidities as well. Term infants born with meconium-stained amniotic fluid who presented with respiratory distress had lower ACTH and cortisol levels than did infants with meconium-stained amniotic fluid without respiratory distress. In another study, cortisol levels of less than 17 μg/dL were associated with more bronchopulmonary dysplasia and increased length of hospital stay. In addition, a number of studies found an association in sick, extremely premature infants between relative adrenal insufficiency, diagnosed by low cortisol values and/or elevated cortisol precursors, and increased pulmonary disease severity and subsequent chronic lung disease. Relative adrenal insufficiency has also been associated with the presence of patent ductus arteriosus. Of note is that these findings reveal only an association between low serum cortisol values and morbidity, and causation has not been documented.
Diagnosis of Adrenal Insufficiency
The reported incidence of relative adrenal insufficiency is obviously influenced by the criteria used for “adequate” and “inadequate” cortisol response or cortisol production . In critically ill adults the proposed criterion for diagnosing relative adrenal insufficiency is a change in total serum cortisol levels of less than 9 μg/dL following 250 μg of cosyntropin administration or a random total cortisol of less than 10 μg/dL (276 nmol/L). Of note is that a cortisol level of less than 10 μg/dL in ill adult patients has high predictive value (0.93) but low sensitivity (0.19) for relative adrenal insufficiency. In their most recent publication on clinical practice parameters for hemodynamic support of pediatric and neonatal septic shock, the American College of Critical Care Medicine propose to maintain “equipoise on the question of adjunctive steroid therapy and thus diagnosis of relative adrenal insufficiency for pediatric and newborn septic shock (outside of classic adrenal insufficiency) pending further trials.”
Devising diagnostic criteria for relative adrenal insufficiency in infants is particularly challenging because of the relatively little amount of available data. The diagnosis and thus the reported incidence of relative adrenal insufficiency are greatly influenced by the choice of the test. Neonatal investigators most often report isolated, random serum cortisol levels or stimulated serum cortisol levels at specific time points following CRH or ACTH stimulation. Using random cortisol values at the time of stress in VLBW infants, Korte et al. defined normal adrenal function as a random cortisol level of 15 μg/dL (414 nmol/L) and an inadequate serum cortisol as a serum cortisol level less than 5 μg/dL (<138 nmol/L). In critically ill term and late preterm infants, Fernandez et al. showed that those with random cortisol values less than 15 μg/dL had higher blood pressure within 24 hours after hydrocortisone administration. In addition, these infants demonstrated a decreased need for vasopressor-inotropes and had a lower heart rate after hydrocortisone administration compared with those with higher cortisol values. However, debate remains on the use of random cortisol values as a diagnostic tool because total plasma cortisol can have marked hourly variability in both adults and neonates. In addition, random cortisol values may not correlate with outcome, response to therapy, or severity of illness. However, a baseline random cortisol is often the most practical and easiest to obtain in the acute setting of the rapidly deteriorating newborn. In term and late preterm infants, a cutoff random cortisol value of 15 μg/dL is often used. For preterm infants, a recent review has suggested that a serum cortisol value of less than 200 nmol/L (7.25 μg/dL) or the less than 50th percentile for adjusted serum cortisol levels be used to diagnose relative (transient) adrenal insufficiency of prematurity. This value represents the 25th percentile of serum cortisol levels in well VLBW and the 50th percentile in vasopressor-dependent VLBW infants. In his review, PC Ng published a mathematical model which takes into account clinical variables that may change the percentile of the diagnostic serum cortisol.
To further test for relative adrenal insufficiency, some investigators used CRH alone or CRH followed by ACTH administration. However, investigators more often selected ACTH alone as the stimulating agent with dosages ranging from 0.1 to 62.5 μg/kg. #
# References , , , , , , , and .
Of note is that adrenal stimulation with supraphysiologic ACTH doses can induce a compromised adrenal gland to produce an “adequate”-appearing serum cortisol response and thus miss the diagnosis of relative adrenal insufficiency. Indeed, ACTH doses at 0.1 to 0.2 μg/kg and perhaps at 0.5 to 1.0 μg/kg are more likely to reveal “relative” adrenal insufficiency than higher ACTH doses. Yet, even within the dose range of 0.1 to 1.0 μg/kg, the proportion of sick premature infants with “inadequate” cortisol response (defined as a serum cortisol level of <9 μg/dL) varies greatly. In a study of ventilated VLBW infants of less than 32 weeks’ gestation and with random serum cortisol levels of less than 5 μg/dL (138 nmol/L), Korte et al. reported that 64% and 37% of infants had inadequate cortisol response (<9 μg/dL) to ACTH at doses of 0.1 and 0.2 μg/dL, respectively. In a trial of ventilated 3-week-old 25 weeks’ gestation preterm infants , Watterberg et al. reported 21% and 2% of infants had inadequate cortisol response to 0.1 and 1.0 μg/kg ACTH, respectively. In critically ill term infants , after giving 1 μg/kg of ACTH, Fernandez et al. found no infants with a cortisol value of less than 18 μg/dL and all patients increased their serum cortisol by greater than 9 μg/dL. In addition, there was no association between ACTH-stimulated cortisol values and severity of illness, need for vasopressor-inotropes, or days on mechanical ventilation. In nonmechanically ventilated infants with sepsis, Soliman et al. found that, when an ACTH stimulation dose of 250 μg/1.73 m 2 was used, no patients met the criteria of having adrenal insufficiency, whereas 13% of the patients were diagnosed with adrenal insufficiency when 1 μg/1.73 m 2 of ACTH was given. Importantly, patients with lower stimulated cortisol values did have higher mortality. In 2012 Hochwald et al. reported that, if an ACTH stimulation test using a dose of 1 μg/kg is performed in premature infants of less than 29 weeks’ gestation with varying degrees of illness in the first 8 postnatal hours, the test could predict those with adrenal insufficiency using hypotension requiring vasopressor-inotrope treatment as a marker. They determined that an ACTH-induced change in cortisol of less than 12% from baseline provided the highest sensitivity (75%) and specificity (93%) for detecting the development of hypotension in this population, On the other hand, random basal cortisol levels did not predict hypotension with an area under the receiver operating characteristic curve (ROC) of 48%.Free cortisol and not protein-bound cortisol is responsible for the physiologic effects at the cellular level, and it may prove to be a better diagnostic tool for adrenal insufficiency. However, the test to measure free cortisol is not widely available because there are no commercial tests on the market. Yoder et al. measured urinary free cortisol levels in 6-hourly block intervals in their premature baboon model to define the ontogeny of cortisol release over the initial hours and days of postnatal life. They found that urinary free cortisol levels were directly proportional to and highly correlated with plasma cortisol levels. Moreover, urinary free cortisol measurement avoids the potential fluctuations seen in serum cortisol levels and the need for frequent blood sampling. However, one needs to be able to effectively collect timed urine samples, which is not a very practical approach in the NICU. Vezina et al. described unbound cortisol concentrations in critically ill newborns with hypotension requiring vasopressor-inotrope administration. Similar data may serve to better tailor the dosing of hydrocortisone and predict its clinical effects in the future.
The lack of consensus on diagnostic evaluation and timing of evaluation is further fueled by the uncertainty about the definition of “adequate” and “inadequate” cortisol response or cortisol production across gestational and postnatal ages for different diagnostic evaluations. Indeed, recent studies have continued to complicate the consensus on the utility of adrenal function tests. For instance, Miletin et al. found no correlation between serum cortisol values and superior vena cava (SVC) flow or mean blood pressure. However, this finding might be explained by the lack of our ability to appropriately define the “normal range” for SVC flow and hypotension (see Chapter 3 ) and by the fact that additional hemodynamic parameters affect the normal range of SVC flow in an individual patient (see Chapters 1 and 11 ). At present, the validity, reproducibility, and clinical importance of identifying inadequate cortisol response to illness or ACTH at different stimulating doses (0.1, 0.2, 0.5, or 1.0 μg/kg ACTH) in terms of mortality, management, and long-term outcomes requires further study. To further complicate the diagnosis of adrenal insufficiency in the preterm and term neonate, measurements of serum cortisol and the response to ACTH stimulation may be influenced by the various procedures and the testing itself in acutely ill patients. Thus the validity and interpretation of adrenal function tests, especially in the midst of critical illness, is subject to an ongoing debate among adult, pediatric, and neonatal critical care physicians.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
