Steroids in the acutely ill: Evolving recommendations and practice

THREE MAIN MECHANISMS

In 2008, we described CIRCI as a syndrome of inadequate corticosteroid activity for the severity of the patient’s illness, that may occur with a decrease in adrenal steroid production (adrenal insufficiency) or from tissue resistance to glucocorticoids with or without adrenal insufficiency. To these mechanisms we now add alterations in cortisol metabolism.

During critical illness, production of adrenocorticotropic hormone (ACTH) is often low, while cortisol levels tend to be normal or, usually, high. The diurnal rhythm in cortisol levels (normally lower in the evening and higher in the morning) is also lost.

More than 90% of circulating cortisol is bound to corticosteroid-binding globulin, which tends to fall during critical illness. Only 5% to 10% of cortisol is free and biologically active.

Normal total cortisol levels are between 5 and 24 μg/dL.

Cortisol is a lipophilic hormone that enters cells passively and binds to specific glucocorticoid receptors in the cytoplasm, or to membrane sites. Cortisol is metabolized primarily in the liver by 5 alpha/beta-reductases and in the kidneys by the 11-beta-hydroxysteroid dehydrogenase (11-beta-HSD) type 2 enzyme. During severe stress states, the expression and activity of these enzyme systems is reduced, leading to decreased clearance of cortisol. Upon binding of glucocorticoid, the glucocorticoid receptor undergoes a conformational change, dissociates from the chaperone proteins, and enters the nucleus and mitochondria, where it binds to positive (transactivation) or negative (cis-repression) specific DNA regions termed glucocorticoid responsive elements to regulate transcription and translation of target genes in a cell- and gene-specific manner. Nuclear factor kappa B is the major transcription protein that glucocorticoids inhibit, and this inhibition is responsible for downregulating the actions of proinflammatory cytokines such as tumor necrosis factor-alpha, interleukin 1, and interleukin 6. Glucocorticoids also act through a nongenomic pathway.

CIRCI IS COMMON

The incidence of CIRCI varies widely, depending on the population studied and the diagnostic criteria used: 20% to 60% of patients with septic shock in a medical intensive care unit, 21% to 75% of patients with human immunodeficiency virus infection who are critically ill, and 15% to 50% of patients with traumatic brain injury.^4,5

The common clinical features are hemodynamic instability despite adequate fluid resuscitation, and fever and inflammation without an obvious source of infection that does not respond to empiric antimicrobial therapy. Other, nonspecific signs and symptoms can include altered mental status, gastrointestinal symptoms, hypoglycemia, eosinophilia, hyponatremia, and hyperkalemia. Hyponatremia and hyperkalemia are less common than in patients with primary adrenal insufficiency, in whom they may be more prominent.

Many conditions can be associated with CIRCI: sepsis, septic shock, acute respiratory distress syndrome (ARDS), major trauma, severe bacterial pneumonia, bacterial meningitis, and nonseptic shock states such in patients cardiopulmonary bypass, post cardiac arrest, cardiogenic shock, and burns.

NO SINGLE TEST CAN RELIABLY DIAGNOSE CIRCI

In 2008, the Corticosteroid Guideline Task Force recommended two ways of testing for CIRCI in the critically ill:

• The cosyntropin stimulation test (an increase in the total serum cortisol from baseline of less than 9 μg/dL 60 minutes after giving an intravenous (IV) 250-μg dose of cosyntropin, a synthetic formulation of natural ACTH)

• A random total plasma cortisol level (of less than 10 μg/dL).

However in 2017, we concluded that no single test can reliably diagnose the syndrome, and we could not reach a consensus (> 80% agreement) on whether the ACTH stimulation test is superior to random cortisol for the routine diagnosis of CIRCI.

If the ACTH stimulation test is used, we recommend that the high dose of cosyntropin (250 μg IV) be used for testing rather than the low dose (1 μg), because the latter has mediocre sensitivity in critically ill patients. This was a weak recommendation based on low-quality evidence.

We advised against measuring plasma free cortisol. We found no randomized trial that compared serum total vs free cortisol levels to diagnose CIRCI. A prospective study of 112 critically ill adults with treatment-insensitive hypotension, published after the 2008 recommendations, found a good correlation between serum concentrations of free and total cortisol before and after 250 μg ACTH stimulation testing.⁶ Measurement of free cortisol is cumbersome to perform and may not be widely available in hospital laboratories, and so for practical purposes we felt that measurement of total cortisol would be preferable for most clinicians.

We also suggested against the use of salivary cortisol levels, as it would not be cost-effective, practical, or feasible given that it is tested by enzyme immunoassay, which may not be routinely available at most centers.

Most importantly, we emphasized that if corticosteroids are clinically indicated in acutely ill or critically ill patients, there is no need to do a cosyntropin stimulation test.

Patients already on steroids (ie, ≥ 5 mg of prednisone or 20 mg of hydrocortisone per day for at least 3 weeks) are at risk of hypothalamic-pituitary-adrenal axis suppression. Cosyntropin stimulation testing is recommended only for those who are more likely to develop either permanent, secondary adrenal insufficiency or CIRCI, to confirm or deny need for permanent replacement therapy or to monitor their recovery during the final phases when tapering the corticosteroid. Dexamethasone does not interfere with the cortisol response or with the cortisol assay. However, if someone is taking hydrocortisone, it should be discontinued the evening before stimulation testing.

DIFFERENT GLUCOCORTICOID PREPARATIONS

The biologic rationale for glucorticoid use in acutely ill patients relates to its potent anti-inflammatory effects and to its effects on cardiovascular tone, including enhanced vasoconstrictor responde to exogenous catecholamines and inhibition of cyclo-oxygenase-2 and inducible nitric oxide synthase. Hydrocortisone tends to be preferred as replacement therapy, as it has a short duration of action which allows the hypo-thalamic-pituitary-adrenal axis to recover between doses. However, prednisone, dexamethasone, and methylprednisolone are the agents we generally use, particularly in those with severe inflammation including bacterial pneumonia and ARDS.

SIDE EFFECTS

Steroids are not benign drugs. They can be associated with several side effects, most notably hyperglycemia, but also hypernatremia, neuromuscular weakness, myopathy, superinfections, upper gastrointestinal bleeds, arrhythmias, and steroid-induced psychosis.

STEROIDS TO TREAT SEPTIC SHOCK

Corticosteroids have been used in septic shock for nearly 6 decades, and practice has changed as information has emerged.

1976—Schumer⁷ performs the first randomized controlled trial, using very large doses of steroids, ie, dexamethasone 3 mg/ kg or methylprednisolone in doses of nearly 2 g daily, in 172 patients with clinical septic shock and finds a reduction in mortality.

1980s—Bone et al⁸ and Sprung et al⁹ find that high-dose methylprednisolone (30 mg/kg daily) does not increase survival, and it induces superinfections in many patients. For the next 15 years, steroids are avoided in patients with septic shock.

2002—Annane et al¹⁰ find that the combination of hydrocortisone 50 mg IV every 6 hours plus flud-rocortisone 50 μg enterally daily is associated with a lower mortality rate, especially in those who do not respond to the cosyntropin stimulation test. However, patients who do respond derive no benefit. In the next few years, cosyntropin stimulation testing becomes popular, and nonresponders get hydrocortisone plus fludrocortisone.

2008—The Corticosteroid Therapy of Septic Shock (CORTICUS) trial¹¹ finds no benefit to the use of hydrocortisone 50 mg IV every 6 hours in terms of survival, but shock reversal is faster. Response to the cosyntropin stimulation test does not seem to matter. Many clinicians begin to abandon the stimulation test in patients deemed to already have an indication for corticosteroids, ie, refractory septic shock.

2016—The Hydrocortisone for Prevention of Septic Shock (HYPRESS) trial,¹² with 380 patients, finds that hydrocortisone 200 mg/day IV for 5 days does not prevent progression from sepsis to septic shock, indicating steroids should not be used for patients with sepsis who are not in refractory shock.

2018—The Adjunctive Corticosteroid Treatment in Critically Ill Patients With Septic Shock (ADRENAL) trial,¹³ in 3,800 patients, finds that hydrocortisone 200 mg by IV infusion for 7 days compared with placebo was not associated with survival benefit, but shock reversal is faster, duration of mechanical ventilation is shorter, and fewer blood transfusions are needed, with relatively few side effects, mainly hyperglycemia.

2018—The Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) trial,¹⁴ in 1,241 patients, uses the combination of hydrocortisone 200 mg IV in divided doses (50 mg IV every 6 hours) and fludrocortisone 50 μg enterally daily for 7 days. At 90 days, fewer patients in the steroid group had died, shock reversal was faster, and the steroid regimen was found to be safe; hyperglycemia was the most common adverse reaction. At 180 days, the survival benefit remained.

2019—A Cochrane meta-analysis¹⁵ finds that corticosteroid therapy “probably” reduces hospital mortality (by about 9%), and results in large reductions in length of stay in the intensive care unit (by about 1 day) and in the hospital (by about 1 and one-half days). Corticosteroid therapy was associated with increased risk of hyperglycemia, hypernatremia, and muscle weakness but no increased risk of superin-fection and little or no difference in gastrointestinal bleeding, neuropsychiatric, stroke, or cardiac events.

2020—Enter artificial intelligence. Pirracchio et al¹⁶ enter data from about 2,500 patients from 4 septic shock clinical trials into an artificial intelligence program to attempt to more precisely determine who should receive corticosteroids, based on multiple patient variables. This seems to be where the future lies. They found that strategies to treat all patients with corticosteroids or to treat no one were associated with a worse outcome. In contrast, an individual estimation-based treatment strategy always yielded a positive net benefit.

STEROID USE IN ACUTE LUNG INJURY AND ARDS

The ARDS Network study¹⁸ showed that corticosteroids should not be used in patients in the late phase of ARDS, ie, after 2 weeks on a ventilator. But what about early?

Meduri et al¹⁹ performed a randomized, double-blind placebo-controlled trial in 91 patients with severe early ARDS (within 72 hours), 66% of whom had sepsis. They randomized these patients in a 2:1 ratio to methylprednisolone infusion 1 mg/ kg/day vs placebo. They tapered the steroid slowly, keeping the same dose for 14 days, cutting it in half on day 15 to day 21, in half again from day 21 to 28, and then stopping. More importantly, they did surveillance with frequent bronchoscopies to rule out infection, and they avoided neuromuscular blockade so as not to accentuate neuro-muscular weakness, a side effect of steroids.

The primary outcome was a reduction in the Lung Injury Score or successful extubation by day 7, both of which were achieved. There were also significant reductions in the duration of mechanical ventilation, length of stay in the intensive care unit, and mortality.

The 2017 guidelines³ gave a weak recommendation for the use of methylprednisolone 1 mg/kg/day in patients with ARDS, specifically patients receiving at least 50% FiO₂ on mechanical ventilation with positive end-expiratory pressure of 10 cm H₂O or more who have a PaO₂/FiO₂ ratio under 200. The tradeoff is hyperglycemia. We felt there was no increased risk of neuromuscular weakness, gastrointestinal bleeding, or nosocomial infection.

Villar et al²⁰ performed a study in 277 patients who had moderate to severe ARDS, with PaO₂/FiO₂ ratios less than 200 despite lung-protective ventilation and other strategies to optimize them on mechanical ventilation; 77% of these patients had either pneumonia- or sepsis-associated ARDS. They randomized the patients to dexamethasone 20 mg/day for 5 days, and then 10 mg/day on days 6 to 10, and then stopped the steroid shortly after that in most patients. The study was conducted over 5 years. The trial was stopped by the Data and Safety Monitoring Board due to low enrollment rate after enrolling nearly 90% of the planned sample size. The investigators found significant benefits with dexamethasone in terms of ventilator-free days, all-cause mortality at day 60, and duration of mechanical ventilation.

STEROIDS IN SEVERE BACTERIAL PNEUMONIA

In 2017, we suggested using corticosteroids in hospitalized patients with severe community-acquired pneumonia who required mechanical ventilation or vasopressor therapy (conditional recommendation, moderate quality of evidence).²¹

This was largely based on a trial by Torres et al,²² published in 2015, in patients with severe community-acquired pneumonia and a high systemic inflammatory response, with C-reactive protein levels greater than 15 mg/L, requiring mechanical ventilation or vasopressor therapy. Corticosteroid therapy was associated with improvement in survival. Hyperglycemia was the most common adverse effect.

A Cochrane review²³ in 2017 looked at 17 randomized controlled trials involving 2,264 patients and showed corticosteroids to significantly reduce mortality and morbidity rates in patients with severe community-acquired pneumonia, but not in those with nonsevere community-acquired pneumonia.

COVID-19

The initial flulike symptoms of COVID-19 are usually associated with bioreplication, and by the time the patients come to the hospital with respiratory symptoms they usually are already at risk with a severe inflammatory response that can progress to ARDS and the need for invasive mechanical ventilation. These patients have elevations in proinflammatory markers and inflammatory cytokines such as C-reactive protein, IL-6, and IL-1.

In April 2020, we argued in favor of corticosteroid therapy in patients with severe COVID-19-associated ARDS based on available observational evidence at that time and the extremely high mortality associated with the disease, and in favor of not waiting for the results of the randomized controlled trials.²⁴

The Randomized Evaluation of COVID-19 Therapy (RECOVERY) trial²⁵ published its results a few months later. It was large, with 6,425 patients, of whom 2,104 received dexamethasone 6 mg IV daily for up to 10 days, and the rest usual care. Dexamethasone reduced mortality by one-third in those who required mechanical ventilation and by one-fifth in those receiving supplemental oxygen. However, there was no benefit in patients who were not receiving respiratory support, and there was even the suggestion of harm that could not be excluded.

A meta-analysis conducted by the World Health Organization Rapid Evidence Appraisal for COVID-19 Therapies Working Group²⁶ showed that the mortality rate at 28 days was significantly lower in patients with COVID-19 pneumonia and ARDS who got corticosteroid therapy, compared with those who did not receive corticosteroids (32.7% vs 41.5%, respectively).

DISCLOSURES

Dr. Pastores has disclosed serving as advisor or review panel participant for AbbVie Pharmaceuticals, as research investigator with bioMerieux, Eisai, and Revimmune, receiving royalty fees as textbook editor for McGraw-Hill.