Abstract
The past several years have seen an increased appreciation of the potential role of the endocrine system in the recovery process following surgery for congenital heart disease. Many of the hormonal changes following cardiac surgery are adaptive and necessary, whereas activation of proinflammatory cytokine and chemokine responses and some of the metabolic changes following surgery are likely mediators leading to detrimental outcomes. Additionally, other hormonal perturbations may contribute to adverse outcomes. This review examines the pain and the stress response, thyroid function and hyperglycemia following cardiopulmonary bypass (CPB), and the potential role of corticosteroids in the pediatric cardiac critical care unit.
Introduction
The past several years have seen an increased appreciation of the potential role of the endocrine system in the recovery process following surgery for congenital heart disease (CHD). Early characterization of the stress response was met with efforts to suppress this response in the belief that it contributed to postoperative morbidity and mortality. Continued research in this area has made it clear that many of the hormonal changes following cardiac surgery are adaptive and necessary, whereas activation of proinflammatory cytokine and chemokine responses and some of the metabolic changes following surgery are the likely mediators leading to detrimental outcomes. 1 –4 Additionally, other hormonal perturbations have been identified as possible contributors to adverse outcomes. The recognition that exposure to cardiopulmonary bypass (CPB) is associated with a decreased thyroid function 5,6 has met with efforts to normalize thyroid hormone levels pharmacologically, with some evidence that there may be some benefit in selected populations. 7,8 Similarly, the recognition that hyperglycemia is common after CPB and several studies showing an association of hyperglycemia with adverse clinical outcomes 9 –13 have met with efforts to establish tight glycemic control using intensive insulin therapy. 14 –16 Again, the clinical outcomes resulting from this therapy have been mixed. 17 Finally, suppression of adrenal function has also been suggested to be an important contributor to complications in critically ill patients. Like thyroid hormone and insulin, pharmacologic hormone replacement with hydrocortisone or other steroids is readily available and has been tried in several studies with variable results. 18,19
This review examines the pain and the stress response, thyroid function and hyperglycemia following CPB and the potential role of corticosteroids in the pediatric cardiac critical care unit.
Postoperative Analgesia/Sedation and Modulation of the Stress Response
Almost 20 years ago, Anand and Hickey showed that infants mounted substantial stress responses to cardiac surgery and postoperative pain and that continuous opioid infusion for postoperative analgesia reduced these responses and that reduced stress responses were associated with improved clinical outcomes. 2 Although reduced stress responses have been associated with lower morbidity and mortality in this and other studies, establishment of any cause-and-effect relationship has been elusive. Postoperative pain is one of the many stressors that can activate these endocrine responses in the surgical patient, but it is a factor that can be treated effectively. Attempts to treat postoperative pain will consequently modulate stress responses, although effective analgesia remains as the primary therapeutic goal whereas modulation of the surgical stress response is secondary.
This section discusses some of the current approaches used to treat postoperative pain following cardiac surgery and explores the minimal data on how these approaches may modulate the endocrine stress response.
Types of Pain
Different types of pain include physiological pain, inflammatory pain, neuropathic pain, and pain of visceral origin. These types of pain activate unique pain mechanisms, although considerable overlap occurs in their nociceptors, nerve fibers, dorsal root ganglion cells, or processing loci in spinal and supraspinal areas. Pain is a multilayered phenomenon, often associated with primary or secondary hyperalgesia (lowered pain thresholds in the area of injury or away from it) and allodynia (pain resulting from nonnoxious stimuli) in the postoperative period. 20
Neonates and infants are more sensitive to acute pain than older children or adults because they have lower thresholds for the cutaneous flexor reflex, 21 –23 delayed maturation of descending noxious inhibitory controls, 24,25 excitatory effects of the classical inhibitory neurotransmitters (γ-aminobutyric acid [GABA], glycine), 26,27 often leading to prolonged windup and hyperalgesia 23,28 and accentuated metabolic stress responses. 1,2,29 –33
Approaches to Pain
Nonpharmacologic approaches to achieve pain control, particularly in small infants, include nonnutritive sucking using a pacifier, nesting or swaddling, allowing routine skin-to-skin contact with the patient’s mother or father, using guided imagery in older children, respecting the patient’s sleep—wake cycles, and using sucrose therapy for skin-breaking procedures. 34 –37 In a systematic review examining the efficacy of sweet-tasting solutions for infants, 38,39 sucrose decreased crying, facial expressions of pain, heart rate changes, and other parameters in 13 of 14 randomized clinical trials. Although the efficacy of sucrose analgesia was challenged in a recent study, 40 methodological aspects of study design and data management cast doubt on these findings. There is no direct data regarding the effects of these nonpharmacologic approaches on the stress response.
Local and topical anesthetics can block pain at the point of entry into the pain system. Various preparations are available. 41 –45 A randomized trial using continuous spinal anesthesia with intrathecal bupivacaine (0.5 mL/kg of 0.25%) compared with an opioid-based general anesthesia during cardiac surgery was associated with significantly reduced catecholamine responses and diminished changes in blood lactate, but no differences occurred in their cortisol responses or hemodynamic changes during and after surgery. 46 Other trials of spinal or epidural anesthesia in noncardiac surgery have also noted significant reductions in hormonal stress responses but with little or no impact on clinical outcomes. 47
Intravenous opioids still remain the gold standard for postoperative pain relief in the cardiac surgical patient although increasing use of “fast-track” protocols for pediatric patients undergoing cardiac surgery may reduce the use of opioid drugs as sole analgesics or may lead to the use of lower doses, possibly delivering ineffective analgesia to these patients. 48 –50 A recent survey of Canadian neonatologists noted that the frequency of opioid use was 28% less likely in nonventilated than in ventilated infants, most clinicians used lower doses in the nonventilated infants, and 82% respondents voiced concerns for respiratory depression in nonventilated infants. 51
Morphine remains the most widely used analgesic for postoperative patients. 52,53 Morphine is metabolized in the liver to 2 active compounds, including morphine-6-glucuronide, which is a potent analgesic with a longer half-life than morphine and morphine-3-glucuronide, which acts as an opioid antagonist and may cause central nervous system (CNS) excitability, even leading to seizures. Neonates and small infants typically produce morphine-3-glucuronide, whereas older children and adults exclusively produce morphine-6-glucuronide. 53 –56 Accumulating data suggest that morphine may not alleviate acute pain, particularly in preterm or term neonates. Franck et al found no changes in plasma norepinephrine, vagal tone, or flexor withdrawal reflex following postoperative morphine use. 57 Simons and colleagues reported no analgesic effects with tracheal suctioning, 58 whereas Carbajal et al found no change in pain scores to heel sticks performed before or after loading dose of morphine (0.1 mg/kg) and that plasma morphine levels were not correlated with these pain scores. 59
Fentanyl, a synthetic opioid used frequently in the postoperative period, has an increased lipid solubility associated with a rapid onset of action (2-3 minutes) but a shorter duration of effect. 60,61 Delayed elimination of fentanyl occurs following prolonged infusions, particularly due to drug binding in lipid-rich tissues and in the third space fluid. The advantages of fentanyl in the cardiac surgical population result from its diminished hemodynamic effects compared to morphine. 62,63 A randomized trial in patients of age 6 months to 6 years compared the effects of intravenous, intrathecal, or combined intrathecal—intravenous fentanyl anesthesia for cardiac surgical operations. The combined use of intrathecal + intravenous fentanyl reduced the hemodynamic changes during surgery, reduced the perioperative changes in plasma cortisol and blood glucose, as well as the urinary cortisol excretion rates for 24 hours postoperatively. 64
Ketamine is also a useful drug in the cardiac surgical patient. It is routinely used as a general anesthetic agent, but recent years have seen a surge in its use for conscious sedation or procedural pain. Ketamine maintains hemodynamic stability and the respiratory drive, promotes bronchodilation, and is used popularly because of its hemodynamic effects and lack of respiratory depression.
Dexmedetomidine produces potent analgesia, anxiolysis, and moderate sedation via the stimulation of α2A- and α2C-adrenergic receptors, but indirect effects also occur via the activation of opioid and GABA receptors. 65,66 Dexmedetomidine showed significant neuroprotective effects in animal models of hypoxic–ischemic injury, with reduced brain matter loss and improved neurologic function. 67 Multiple studies using dexmedetomidine have consistently demonstrated improved neurologic outcomes, associated with significant reductions in apoptotic and excitotoxic cell death, mediated via the α2A-adrenergic receptors, 68 –70 although no clinical studies have systematically investigated these effects.
Thyroid Hormone as Adjunctive Therapy Following Pediatric Cardiac Surgery
Cardiopulmonary bypass induces marked and persistent depression of circulating thyroid hormone levels in both adults and children, 5,6 with inhibition of the thyroid hormone axis by proinflammatory cytokines a likely contributor. Small clinical series of children undergoing CPB have shown inverse relationships between postoperative cytokines including interleukin 6 (IL-6), macrophage inhibitory factor, and IL-8 and circulating thyroid hormones, either triiodothyronine (T3) or thyroxine (T4). Sites of cytokine inhibitory action include the hypothalamus, pituitary, thyroid gland, and peripheral conversion of T4 to T3. 71
Regardless of the mechanism, depression of thyroid hormone levels contributes to postoperative morbidity in adults undergoing CPB. Thus, thyroid hormone repletion in the form of T3 during and after CPB is an intuitive solution that has been the subject of several investigations. 6,72 Controlled randomized studies demonstrate that parental T3 repletion during coronary bypass surgery improves postoperative ventricular function 6,72,73 but the clinical response has been variable. Mullis-Jansson and coauthors showed T3 reduces the use of inotropic agents and mechanical devices and decreases the incidence of myocardial ischemia. 73 Klemperer et al could not find a T3 influence on these clinical outcome parameters 6 but did demonstrate a reduced rate of atrial fibrillation. 72
Infants and young children undergoing CPB demonstrate more profound and persistent decreases in thyroid hormone levels when compared to adults, 5,74 and postoperative hemodynamic abnormalities are associated with depression of thyroid hormone levels. 5,75 Accordingly, several investigators have conducted randomized clinical trials with relatively less number of participants (between 28 and 42 total participants per study) to explore the utility of thyroid hormone supplementation in children undergoing CPB, 7,76,77 or to evaluate pharmacokinetics. 78 The authors of those studies acknowledged some clinical design issues that hamper the interpretation of the results. These include the use of surrogate end points and inadequate power. 7,76 –78 Furthermore, their published reports often lacked details regarding adverse events. 7,76 –78 Bettendorf et al showed some elevation in the Therapeutic Interventional Scoring System (TISS) score and cardiac index in patients receiving T3 infusion. 76 Chowdhury and coauthors showed that treating patients with a defined nadir in perioperative T3 levels raised the TISS score and lowered inotropic score. 77 Mackie et al showed that T3 infusion improved composite clinical score and time to negative fluid balance in neonates undergoing aortic arch reconstruction. 7
The TRiiodothyronine supplementation in Infants and Children undergoing Cardiopulmonary bypass (TRICC) 73 study was a prospective, multicenter, double blind, randomized, placebo-controlled trial to evaluate the effects of T3 supplementation after CPB in children under the age of 2 years. Triiodothyronine (Triostat) was given as multiple bolus doses immediately before and after CPB. The primary efficacy analysis evaluated the effect of T3 supplementation on time from aortic cross clamp removal to extubation (transthoracic echocardiography [TTE]). The study enrolled 198 patients and 193 completed the trial and were included in the analyses, using a stratified diagnosis format. The study used Cox Proportional Hazards modeling analysis of the entire cohort and showed no change in TTE, but T3 supplementation was found to be safe. Adverse events including postoperative arrhythmia occurred at equivalent rates between groups. However, further analyses defined age-related differences in response to thyroid supplementation. The diagnostic distribution of the 2 age groups evaluated is shown in Figure 1 .
Participant numbers for each surgical diagnostic category. COA indicates coarctation of the aorta; TGA, transposition of the great arteries; HLV, hypoplastic left ventricle; TAPVD, total anomalous pulmonary venous drainage; VSD, ventricular septal defect; TOF, tetralogy of Fallot; CAVC, complete atrioventicular canal; SVC-PA, superior vena cava to pulmonary artery shunt (reprint by permission, Circulation). 8
Infants of age >5 months were extubated sooner than infants of age <5 months (hazard ratio = 1.76 for older compared to younger patients, P = .0054). For patients <5 months of age, randomization to Triostat resulted in significant shortening in TTE. The placebo group’s median TTE was 98 hours with 95% confidence interval (CI): 71 to 142. The Triostat group’s median TTE was 55 hours and 95% CI: 44 to 92. The hazard ratio among patients <5 months was 1.72 (P = .0216). For patients >5 months of age, randomization to Triostat resulted in small but significant delay in median TTE. The median TTE in hours was 16 (95% CI: 7-22) for placebo and 20 (95% CI: 16-45) for Triostat, and the hazard ratio was 0.60 (P = .0220). The Kaplan-Meier curves for these populations are shown in Figure 2 . Triiodothyronine supplementation also improved cardiac function in the younger cohort as assessed by echocardiography and inotropic score.
Kaplan-Meier curves for participants <5 months of age and participants >5 months of age (Statistics by Cox proportional hazards). Hazard ratio for the Triostat group among patients of age <5 months was 1.72 (P = .0216) and for patients of age >5 months was 0.60 (P = .0220). Below the graph are number at risk, number of extubations (percentage intubated; reprint by permission, Circulation). 8
The phenomenon, whereby T3 poses benefit for age <5 months but some detriment for age >5 months, suggests that the difference in response according to age has a physiological basis. Additionally, T3 levels in the Triostat group return to levels near those in the placebo group by 24 hours. One might therefore question how the temporary elevation impacts clinical parameters several hours or days after declination. Triiodothyronine exerts action through both nongenomic and genomic mechanisms. 8,74 The nongenomic mechanisms likely occur at the cell membrane and are immediate. 76 However, the genomic mechanisms involve T3 transport across membranes, binding to thyroid nuclear receptors, and modulation of transcription and protein translation. Triiodothyronine rapidly initiates transcription of some genes in infants, 77 but the ultimate result of this transactivation may not be apparent for several hours or days. 74 Delay in some T3-mediated actions may explain in part the lack of treatment benefit in the patients older than 5 months, particularly as most are extubated by 24 hours. The relatively minor but statistically significant prolongation of TTE in this age group with T3 supplementation might reflect dominance of the nongenomic action relative to genomic action since the outcome was measured shortly after receiving the loading doses.
What Is Unique About Strict Glycemic Control in the Cardiac Intensive Care Unit?
Hyperglycemia in critically ill patients has been the subject of much recent study. Work has focused on both the association of hyperglycemia with adverse outcomes and on the potential value of treatment with insulin. These studies have yielded conflicting results and the importance of hyperglycemia and the value of treatment have become subjects of intense debate. The belief that hyperglycemia leads to poor clinical outcomes and should be treated aggressively stems from data in both adults and children showing an increased incidence of adverse outcomes when intensive care unit (ICU) patients experience sustained hyperglycemia. 9 –13 A small number of these studies have focused on infants and children after cardiac surgery. In particular, 2 retrospective studies have found strong associations between the intensity and/or duration of hyperglycemia and adverse outcomes including renal failure, hepatic failure, adverse CNS events, infection, or death. 79,80 Length of ICU stay and duration of mechanical ventilation were also associated with hyperglycemia. When one looks at the potential benefits of tight glycemic control with insulin, the Leuven study 81 showed benefit in adult surgical ICU patients with ICU stays ≥3 days. The Leuven investigators also examined tight glycemic control in pediatric patients. A randomized trial in over 700 participants showed a shorter length of stay, a lower incidence of prolonged ICU stay, and lower mortality in the more intensively treated group. 16 Approximately 75% of the patients in this study had undergone cardiac surgery, and 25% of the participants receiving intensive insulin therapy had at least one episode of hypoglycemia.
The problem is that several other studies have yielded very different results. Ballweg et al found no association between hyperglycemia and neurologic injury in infants who had undergone corrective cardiac surgery. 82 Rossano et al found no link between glucose levels and other adverse outcomes after the arterial switch operation, 83 and most recently DeCampli et al found no association between glucose concentration and outcome in infants undergoing operation for CHD. 84 Polito and coauthors suggested that both hyperglycemia and hypoglycemia were associated with adverse outcomes after cardiac surgery in infants and children and that the ideal blood glucose concentration might be between 110 and 126 mg/dL. 85 Likewise, the value of glycemic control remains debatable, as the recently published Normoglycemia in Intensive Care Evaluation-Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial examined tight glycemic control in over 6000 adult ICU patients and was terminated early due to increased mortality in the intensively treated group. 17 Other large-scale studies including another study from Leuven in medical ICU patients have also failed to show clear benefit from treatment. 15 Additional trials of glycemic control in pediatric patients, including those recovering from cardiac surgery, are currently underway.
When one considers controversy regarding the consequences of hyperglycemia and the risks and benefits of treatment, it seems reasonable to ask the question is all hyperglycemia the same? Hyperglycemia is almost universal following pediatric heart surgery, yet most patients recover uneventfully. A study of both the degree and the duration of hyperglycemia in 772 patients undergoing repair or palliation of CHD using CPB over a 2-year period 86 showed that roughly 90% of patients had an elevated glucose level upon admission to the ICU. Moderate-to-severe hyperglycemia (≥180 mg/dL) was less common than mild hyperglycemia (elevated, but <180 mg/dL). The incidence of hyperglycemia steadily decreased over time, with roughly 50% of glucose measurements at 72 hours remaining elevated. There was an overall relationship between both the degree and the duration of hyperglycemia and adverse clinical outcomes. Despite this, when considering the patients with hyperglycemia, hospital mortality was only 2%, and the occurrence of individual adverse outcomes such as renal failure (3%), seizures (2%), or stroke (<1%) was low. Low blood pressure (4%) and low cardiac output (10%) were somewhat more common, but 68% of patients with hyperglycemia were discharged from the ICU in <3 days, with an average length of stay of only 2 days. Consideration of factors such as age and the presence of important residual cardiac lesions lessened the incremental contribution of hyperglycemia to adverse outcomes, again suggesting that the relationship between glucose and outcomes is not universal and direct.
A compelling case can be made that the clinical phenomenon of hyperglycemia might result from either a physiologic stress response or insulin resistance. It is specifically the insulin resistant state that is associated with ongoing inflammation, disrupted cellular energy metabolism, and adverse outcomes. 87,88 Hyperglycemia associated with a less profound stress response may be adaptive or even beneficial. 89 The inflammatory response to CPB is mediated by cytokines including IL-6, IL-8, IL-10, and tumor necrosis factor-α (TNF-α). 90 –94 In some situations, it is conceivable that the interaction of inflammatory and hyperglycemic stress responses to CPB can lead to a positive feedback loop wherein an insulin-resistant state develops. When insulin binds to its receptor, it results in autophosphorylation of multiple tyrosine residues in the receptor. Intracellular tyrosine kinase becomes activated, which phosphorylates insulin receptor substrate (IRS) proteins, leading to the activation of phosphatidylinositol 3 kinase (PI3K), a step that is essential in mediating translocation of the glucose transporter, GLUT-4. Inflammation causes phosphorylation of serine and threonine rather than tyrosine residues of the insulin receptor, which can render the insulin receptor less active. 95 Interleukin 6 can activate the suppressor of cytokine signaling (SOCS-3) resulting in the degradation of IRS, 95 thereby diminishing intracellular insulin signaling. An increase in hepatic TNF-α messenger RNA (mRNA) and protein can promote insulin resistance via serine phosphorylation rather than tyrosine phosphorylation of IRS-1/2, subsequently decreasing the association of IRS-1/2 with PI3K. 96 Vlasselaers et al recently demonstrated that tight glycemic control during and following cardiac surgery was associated with decreased IL-6, less need for blood pressure support, and less myocardial injury. 16
Insulin resistance increases hepatic glucose output but decreases glucose uptake and oxygenation peripherally. There is a decrease in glycogen synthesis and protein anabolism in muscle, and an increase in lipolysis and production of free fatty acids (FFAs) and glycerol. Free fatty acids stimulate further hepatic glucose output via enhanced gluconeogenesis and also stimulate metabolic processes that convert the excess glucose to fatty acids, promoting positive feedback. Furthermore, fatty acid transport into the mitochondria is inhibited, thus diverting FFA to produce triglycerides. Because glucose that is transported into the cell may be preferentially metabolized to lactate rather being oxidized, 97 and because insulin resistance with the loss of fatty acid oxidation in the mitochondria reduces adenosine triphosphate (ATP) synthesis, 98 insulin resistance is likely to significantly reduce the amount of ATP available for myocardial work. Data from the Leuven et al study suggests that the control of lipid metabolism might be the most important aspect of insulin treatment of hyperglycemia in critically ill patients. 99
Pros and Cons of Utilizing Hydrocortisone as a Vasoactive—Inotropic Agent
The clinical use of hydrocortisone to improve hemodynamic stability in the setting of escalating vasopressor doses and fluid requirements is long-standing and widespread. Despite the prevalent use of this approach to hemodynamic management, there remains little evidence of its efficacy for improving long-term outcomes. Risks associated with systemic corticosteroids include infection, hyperglycemia, and even neurocognitive impairment. A closer review of the pros and cons of hydrocortisone therapy is crucial to ensure that short-term gains in hemodynamics are not at the detriment of long-term outcomes.
Pro: Hydrocortisone Efficacy for Shock Reversal
The efficacy of hydrocortisone therapy in the post-congenital cardiac surgical period has not been rigorously tested. There have been a number of clinical trial investigations examining corticosteroid use in adult septic shock and in neonates, most prominently in extremely and very-low-birth-weight preterm infants. The Corticosteroid Therapy of Septic Shock (CORTICUS) trial 19 randomized 499 adults with septic shock to hydrocortisone or placebo. The hydrocortisone-treated patients experienced significantly earlier shock reversal than those who were administered placebo, but there was no difference in the 28-day mortality between groups. Markovitz et al 18 used the Pediatric Health Information System (PHIS) and Administrative Database of the Child Health Corporation of America to study 6693 children of age 0 to 17 years, who had been treated with the combination of systemic corticosteroids, mechanical ventilation, and vasopressor support during 2001 to 2002. Almost half (47.9%) of the children who met inclusion criteria had received at least 1 day or more of corticosteroids. Children treated with corticosteroids had longer hospital lengths of stay (48.3 vs 45.4; mean difference 2.8 days; 95% CI, 0.4-5.3; P = .02), greater days of mechanical ventilation (26.6 vs 22.4; mean difference 4.2; 95% CI, 2.4-6.0; P < .001), and longer vasopressor infusion times (10.6 vs 5.2; mean difference 5.4; 95% CI, 4.9-5.8; P < .001). The retrospective nature of this study makes it difficult to determine whether steroid utilization was detrimental, or whether sicker patients were more likely to be treated with steroids. Neither this study nor the CORTICUS trial showed an association between corticosteroid administration and reduced mortality. 18,19
With regard to the neonatal literature, there have been several trials of hydrocortisone for treatment of hypotension, 100 –103 and virtually all of which have shown hemodynamic benefit. A recent meta-analysis has found that hydrocortisone reduces vasopressor requirement and increases blood pressure in preterm neonates and that these results were “robust with a large tolerance for future null results.” 104
Adrenal insufficiency (AI) or critical illness-related corticosteroid insufficiency might also contribute to post-cardiotomy hemodynamic instability. 105 Despite several studies of the hypothalamic—pituitary—adrenal axis, there has been little progress determining the true incidence and impact of this endocrinopathy. Depending on the definition, cosyntropin test, and laboratory technique utilized, the incidence of AI in the pediatric septic literature ranges from 15% to 61%.
A recent meta-analysis of steroid utilization in the adult cardiac surgery population identified 12 randomized clinical trials spanning the last 15 years, 4 of which focused specifically on hydrocortisone. While the analysis did find an association between steroid treatment and shorter hospital stay, there was no impact on mortality. 106 Only 3 studies, 2 retrospective and 1 prospective, have looked specifically at postoperative hydrocortisone use in the pediatric CHD population. The 2 retrospective studies observed an increase in blood pressure with a parallel decrease in heart rate and inotrope requirement, following hydrocortisone administration. White blood cell count and blood glucose concentrations rose but were not significantly different from prehydrocortisone levels. Longer-term outcomes such as days of mechanical ventilation, hospital length of stay, and mortality could not be analyzed due to the absence of a control group. 107 –109 The only prospective study to date randomized 20 neonates to hydrocortisone or placebo. Mean arterial blood pressures were similar 12 hours after surgery (hydrocortisone 59.7 ± 4.3 vs placebo 58.6 ± 5.6, P = .3), and there were no significant differences between groups for inotrope or mechanical ventilation duration. At 24 hours, inotrope score had decreased for both groups but was not significantly different between groups (hydrocortisone 7.6 ± 3.4 vs placebo 9.1 ± 3.0, P = .3). A potential clinical improvement in cardiac function and perfusion was observed in the hydrocortisone group as measured by higher left ventricle-shortening fraction (23.4 ± 13.2 vs 19 ± 17.5, P = .02) and lower lactic acid levels (2.3 ± 1.1 vs 3.3 ± 1.0, P = .049) 18 hours after surgery. 110
Con: Risks of Hydrocortisone Therapy
The risks associated with corticosteroids include infection, hyperglycemia, neuromuscular weakness, and potential compromise of neurologic development. Establishing risk factors specific to hydrocortisone therapy in the post-congenital surgical period are again limited by the lack of scientific evaluation for this strategy. The CORTICUS trial found an increased incidence of superinfections in the hydrocortisone-treated cohort (odds ratio [OR] 1.37, 95% CI, 1.05-1.79). 19 The meta-analysis of corticosteroid utilization surrounding adult cardiac surgery did not find an association between corticosteroid administration and the relative risk (RR) of infection (slope of regression line, 1.00, 95% CI, 0.99-1.01, P = .89). 106 Pasquali et al recently utilized the PHIS database to compare the outcomes of 46 730 children undergoing congenital cardiac surgery in 38 different centers over a 5-year period. Just over half of the children (54%) received corticosteroids the day preceding surgery or the day of surgery. While the study focused on pre- or intraoperative steroid administration, 35.4% of the steroid group, and 27.8% of the nonsteroid group, were administered at least 1 dose of corticosteroid during their postoperative period. This review was unable to find any outcome benefits in the corticosteroid-treated patients. However, there was considerable evidence suggesting increased morbidity risk in the steroid-treated group. Children who received perioperative steroids had longer hospital (OR 2.18, CI 1.62-2.74, P < .001) and ICU length of stays (OR 1.90, CI 1.56-2.23, P < .001). Postoperative infection rates were also significantly higher in the steroid-treated group (OR 1.27, CI 1.10-1.46, P = .001). 111 The single prospective trial comparison of post-cardiac surgery hydrocortisone use in children did not compare infection rates between the hydrocortisone and placebo groups. White blood cell count was not significantly different for hydrocortisone compared to placebo. 110
Hyperglycemia is a well-established side effect of high-dose corticosteroid administration. Patients in the hydrocortisone arm of the CORTICUS trial had a RR of 1.8 for hyperglycemia compared to controls. 19 In the adult cardiac surgery meta-analysis, hyperglycemia requiring insulin infusion was 28.2% with a RR of 1.49 (P < .01) for the hydrocortisone-treated groups. 106 The smaller pediatric studies looking specifically at postoperative hydrocortisone found no significant difference in blood glucose levels before and after hydrocortisone initiation or between the hydrocortisone and placebo groups. 107,108,110 The large pediatric cohort in the Pasquali et al study had higher frequency of postoperative insulin use in the children who received corticosteroids surrounding surgery (OR 2.45, CI 2.24-2.67, P < .001). 111
Neuromuscular weakness as a potential result of hydrocortisone therapy has not been well evaluated. The CORTICUS trial reported no loss of strength for patients receiving hydrocortisone compared to controls. No electrophysiologic testing was performed and assessment of weakness was subjectively determined. 19 Pediatric studies examining corticosteroid use for sepsis, chronic lung disease, CPB inflammatory modulation, and post-cardiotomy hypotension have not specifically addressed neuromuscular weakness.
There are no studies to date evaluating the impact of post-cardiac surgery hydrocortisone therapy on pediatric long-term outcomes. Steroids have been used in some centers during pediatric cardiac surgery to mediate the CPB-related systemic inflammatory response. 112 There is now convincing data that dexamethasone treatment in the neonatal period for prevention of chronic lung disease of prematurity is associated with poorer neurodevelopmental outcomes, 113 leading some to question whether the postoperative use of hydrocortisone or the pre-/intraoperative use of other steroids might also be detrimental. A retrospective analysis of preterm infants (<32 weeks' gestational age [GA]) from 2 centers with different steroid practices (Center #1-hydrocortisone, Center # 2-Dex) shows that hydrocortisone and dexamethasone were equally effective in reducing oxygen requirements. At 5 to 7 years' follow-up, neurologic outcome, psychomotor development, and school performance at center #1 were similar in both the hydrocortisone and control groups. At center #2, school performance was poorer (P < .02) and utilization of special education services was higher (P < .01) in the dexamethasone group compared to controls. 114 Comparisons in the neonatal rat model have shown that dexamethasone, but not hydrocortisone, alters hippocampal synaptic plasticity impacting memory retention and associative memory formation. 115 Based on consistent findings of poorer neurocognitive outcomes in children treated with dexamethasone during the neonatal period, the American Academy of Pediatrics has now recommended against the use of dexamethasone in very-low-birth-weight infants. 116 Of note, the Pasquali study observed that only 3.6% of the 25 113 children treated with corticosteroids surrounding their cardiac surgery were given hydrocortisone. The remainder was treated with methylprednisolone (69.8%) and dexamethasone (26.6%). 111
Certainly there are important pharmacologic differences between dexamethasone and hydrocortisone. Hydrocortisone has a shorter half-life and lower biologic activity than dexamethasone. Hydrocortisone’s preferential binding to mineralocorticoid receptors may narrow its therapeutic action targeting desired outcomes like improved vascular responsiveness and reduced capillary leak, without impacting glucocorticoid functions like glucose homeostasis. 116 However, outcome comparisons must be interpreted with caution, as dexamethasone follow-up has been prospective, whereas hydrocortisone outcome analyses are retrospective. The only prospective randomized clinical trial of hydrocortisone since 1972 followed a study cohort out to 18 to 22 months' adjusted age. Negative effects of hydrocortisone were not identified but the results were confounded by dexamethasone administration in a large proportion of the cohort during or after the hydrocortisone intervention. 117 The remaining studies are retrospective and found no difference in neurocognitive outcomes, motor development, or magnetic resonance imaging (MRI) brain lesions at 7 to 10 years of age when compared to controls who did not receive hydrocortisone during their neonatal period. 118 There have been no long-term follow-up studies of hydrocortisone or methylprednisolone use in pediatric sepsis or congenital cardiac surgery.
Conclusions
Understanding and modulation of the physiologic stress response, thyroid function, blood glucose, and the hypothalamic—pituitary—adrenal axis has become an increasing focus of pediatric cardiac critical care. To date, data suggest that attention to any or all of these issues has the potential to improve short-term outcomes, but as with many aspects of pediatric critical care, long-term data on safety and developmental impact are limited. Unwarranted intervention in any of these complex systems also has the potential for harm consequent to over- or undersedation and the need for prolonged mechanical ventilation, metabolic acceleration due to thyroid hormone, hypoglycemia, or the adverse effects of steroids on the developing brain. Future studies will ideally measure both short- and long-term outcomes and will help develop better understanding of the mechanisms that underlie endocrinopathies in the cardiac ICU so that these multisystem treatments can be directed toward those most at risk.
Footnotes
Abbreviations and Acronyms
Presented at the Eighth International Meeting of The Pediatric Cardiac Intensive Care Society, Miami, USA, December 8-11, 2010.
No undisclosed authors contributed to the manuscript.
The author(s) declared no conflicts of interest with respect to the authorship and/or publication of this article.
Dr. Portman received funding from the U.S. Food and Drug Administration Office of Orphan Product Development Program (R01 FD-R-1971-01).