Steroids After Severe Injury

Abstract

N THIS ISSUE OF JAMA ,R OQUILLY AND COLLEAGUES 1 report the results of the HYPOLYTE (Hydrocortisone Polytraumatise) study, a randomized controlled trial of early hydrocortisone administration to severely injured patients with a primary outcome of hospital-acquired pneumonia within 28 days of injury. All patients were given a short corticotrophin stimulation test, and those meeting their definition of corticosteroid insufficiency continued taking the blinded study treatment, while all other patients had study treatment discontinued. All patients were followed up for the clinical outcomes, and both an intentionto-treat and modified intention-to-treat analysis were conducted. Patients in the hydrocortisone group were found to have significant reduction in pneumonia risk (35.6% vs 51.3%) and a shorter duration of mechanical ventilation (ventilator-free days, 16 days vs 12 days). This study has many strengths including the randomized, placebocontrolled, blinded design, strict definition of the clinical outcome including bronchoalveolar lavage to confirm the diagnosis of pneumonia, and corticotrophin stimulation testing for all patients. Even though the authors demonstrated a significant reduction in the rate of hospital-acquired pneumonia and a reduction in ventilator days, this study, with a sample size of 149 patients, was not adequately powered to assess the effects of steroid on mortality in this patient population. There was no statistical difference in mortality between the 2 treatment groups, although the absolute mortality was higher in the hydrocortisone group (6 deaths, 8.2% vs 4 deaths, 5.3%). Previous studies of steroid administration suggest the need for caution in the interpretation of these results. The prior Corticosteroid Randomization After Significant Head Injury (CRASH) trial, 2 which randomized more than 10000 patients with traumatic brain injury to administration of methylprednisolone for 48 hours early after injury, demonstrated a higher mortality in the steroidtreated group (relative risk, 1.15; 95% confidence interval, 1.07-1.24), but found no difference in the rate of pneumonia between treatment groups. In a trial by Steinberg et al, 3 administration of methylprednisolone to patients with acute respiratory distress syndrome (ARDS) resulted in lower rates of suspected or probable pneumonia in the steroid-treated group but resulted in a significantly higher 60-day mortality for patients given steroids more than 14 days after the onset of ARDS. Although the dosage strategies and timing were different in these studies compared with those in the HYPOLYTE study, the overall evidence suggests that further study with a larger sample size is needed to better define the safety profile and risk of mortality in this patient population. There have been several reports of adrenal insufficiency after severe injury based on corticotrophin stimulation testing; however, comparison between studies is limited by variability in the definition of critical illness–related corticosteroid insufficiency. 4-7 In the HYPOLYTE study, the definition of corticosteroid insufficiency was based on a basal cortisol level of less than 15 µg/dL or a maximal increase of less than 9 µg/dL following a corticotrophin stimulation test. The consensus conference definition of corticosteroid insufficiency is based on patients who are hypotensive and have responded poorly to fluid resuscitation or vasopressor agents, primarily in the setting of sepsis. 8 The recommended threshold for diagnosis in this patient population is a random total cortisol less than 10 µg/dL or maximal increase of less than 9 µg/dL following a corticotrophin stimulation test. 8 The increasing use of etomidate for intubation in critically ill patients also challenges the true incidence of corticosteroid insufficiency because early assessment of adrenal function may be limited by drug-induced adrenal suppression. 5 Roquilly et al 1 acknowledge that further studies are needed to better define the true incidence of corticosteroid insufficiency in severely injured patients. Additional studies also are needed to better define the effectsofearlysteroidadministrationontheinflammatoryresponse of patients following severe injury. Traditional understanding of the immune-inflammatory response after